mirror/opencloud
1
0
Fork 0
mirror of https://github.com/opencloud-eu/opencloud.git synced 2026-01-04 11:19:39 -06:00
Code Issues Packages Projects Releases Wiki Activity

bump reva

Browse Source 
Signed-off-by: Jörn Friedrich Dreyer <jfd@butonic.de>
This commit is contained in:
Jörn Friedrich Dreyer
2024-05-30 11:27:56 +02:00
parent 933b1eb76c
commit e4b826d1ae
302 changed files with 63004 additions and 434 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
go.mod
View File

@@ -15,7 +15,7 @@ require (
github.com/cenkalti/backoff v2.2.1+incompatible
github.com/coreos/go-oidc/v3 v3.10.0
github.com/cs3org/go-cs3apis v0.0.0-20231023073225-7748710e0781
github.com/cs3org/reva/v2 v2.19.2-0.20240529081036-419196f2342e
github.com/cs3org/reva/v2 v2.19.2-0.20240530092407-7f72f379ea89
github.com/dhowden/tag v0.0.0-20230630033851-978a0926ee25
github.com/dutchcoders/go-clamd v0.0.0-20170520113014-b970184f4d9e
github.com/egirna/icap-client v0.1.1
@@ -290,8 +290,10 @@ require (
github.com/opencontainers/runtime-spec v1.1.0 // indirect
github.com/opentracing/opentracing-go v1.2.0 // indirect
github.com/oxtoacart/bpool v0.0.0-20190530202638-03653db5a59c // indirect
github.com/pablodz/inotifywaitgo v0.0.6 // indirect
github.com/patrickmn/go-cache v2.1.0+incompatible // indirect
github.com/pbnjay/memory v0.0.0-20210728143218-7b4eea64cf58 // indirect
github.com/pierrec/lz4/v4 v4.1.15 // indirect
github.com/pjbgf/sha1cd v0.3.0 // indirect
github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2 // indirect
github.com/pquerna/cachecontrol v0.2.0 // indirect
@@ -306,6 +308,7 @@ require (
github.com/rs/xid v1.5.0 // indirect
github.com/russellhaering/goxmldsig v1.4.0 // indirect
github.com/russross/blackfriday/v2 v2.1.0 // indirect
github.com/segmentio/kafka-go v0.4.47 // indirect
github.com/segmentio/ksuid v1.0.4 // indirect
github.com/sergi/go-diff v1.3.1 // indirect
github.com/sethvargo/go-password v0.2.0 // indirect

26
go.sum
View File

@@ -1025,8 +1025,8 @@ github.com/crewjam/saml v0.4.14 h1:g9FBNx62osKusnFzs3QTN5L9CVA/Egfgm+stJShzw/c=
github.com/crewjam/saml v0.4.14/go.mod h1:UVSZCf18jJkk6GpWNVqcyQJMD5HsRugBPf4I1nl2mME=
github.com/cs3org/go-cs3apis v0.0.0-20231023073225-7748710e0781 h1:BUdwkIlf8IS2FasrrPg8gGPHQPOrQ18MS1Oew2tmGtY=
github.com/cs3org/go-cs3apis v0.0.0-20231023073225-7748710e0781/go.mod h1:UXha4TguuB52H14EMoSsCqDj7k8a/t7g4gVP+bgY5LY=
github.com/cs3org/reva/v2 v2.19.2-0.20240529081036-419196f2342e h1:nooOncj0aaM6S8TuNSLY4d7KEb8xSzO9/LfqL0loV60=
github.com/cs3org/reva/v2 v2.19.2-0.20240529081036-419196f2342e/go.mod h1:BOlJApKFrWRiaOoBCRxCTG5bghTTMlYaEZrRxOzKaS8=
github.com/cs3org/reva/v2 v2.19.2-0.20240530092407-7f72f379ea89 h1:74khAslYAD8kXrBZVJOOCd3iXcp1gY00vgpqGw8lmh0=
github.com/cs3org/reva/v2 v2.19.2-0.20240530092407-7f72f379ea89/go.mod h1:lKqw0VuP1NcZbhj0e6tGoAGq3tgWO/pLafVJyDK0yVI=
github.com/cyberdelia/templates v0.0.0-20141128023046-ca7fffd4298c/go.mod h1:GyV+0YP4qX0UQ7r2MoYZ+AvYDp12OF5yg4q8rGnyNh4=
github.com/cyphar/filepath-securejoin v0.2.4 h1:Ugdm7cg7i6ZK6x3xDF1oEu1nfkyfH53EtKeQYTC3kyg=
github.com/cyphar/filepath-securejoin v0.2.4/go.mod h1:aPGpWjXOXUn2NCNjFvBE6aRxGGx79pTxQpKOJNYHHl4=
@@ -1230,6 +1230,8 @@ github.com/go-micro/plugins/v4/wrapper/monitoring/prometheus v1.2.0 h1:UWBUYtMXC
github.com/go-micro/plugins/v4/wrapper/monitoring/prometheus v1.2.0/go.mod h1:8BYxs/wEE4ZJayHZQffw4A8s9rcPumyoNms0hYoNocM=
github.com/go-micro/plugins/v4/wrapper/trace/opentelemetry v1.2.0 h1:e2hgtWMNqJ3DmbMt9ZxzmH/BkVAw9Xg23l6CHrXQfKw=
github.com/go-micro/plugins/v4/wrapper/trace/opentelemetry v1.2.0/go.mod h1:BBqL7ckGNb7rFfk3vU2Yj/CILVsz/WF19CkAyveQl8A=
github.com/go-ole/go-ole v1.2.6 h1:/Fpf6oFPoeFik9ty7siob0G6Ke8QvQEuVcuChpwXzpY=
github.com/go-ole/go-ole v1.2.6/go.mod h1:pprOEPIfldk/42T2oK7lQ4v4JSDwmV0As9GaiUsvbm0=
github.com/go-ozzo/ozzo-validation/v4 v4.3.0 h1:byhDUpfEwjsVQb1vBunvIjh2BHQ9ead57VkAEY4V+Es=
github.com/go-ozzo/ozzo-validation/v4 v4.3.0/go.mod h1:2NKgrcHl3z6cJs+3Oo940FPRiTzuqKbvfrL2RxCj6Ew=
github.com/go-pdf/fpdf v0.5.0/go.mod h1:HzcnA+A23uwogo0tp9yU+l3V+KXhiESpt1PMayhOh5M=
@@ -1808,6 +1810,8 @@ github.com/owncloud/libre-graph-api-go v1.0.5-0.20240425090020-dba6d1507c38 h1:L
github.com/owncloud/libre-graph-api-go v1.0.5-0.20240425090020-dba6d1507c38/go.mod h1:yXI+rmE8yYx+ZsGVrnCpprw/gZMcxjwntnX2y2+VKxY=
github.com/oxtoacart/bpool v0.0.0-20190530202638-03653db5a59c h1:rp5dCmg/yLR3mgFuSOe4oEnDDmGLROTvMragMUXpTQw=
github.com/oxtoacart/bpool v0.0.0-20190530202638-03653db5a59c/go.mod h1:X07ZCGwUbLaax7L0S3Tw4hpejzu63ZrrQiUe6W0hcy0=
github.com/pablodz/inotifywaitgo v0.0.6 h1:BTjQfnixXwG7oYmlIiyhWA6iyO9BtxatB3YgiibOTFc=
github.com/pablodz/inotifywaitgo v0.0.6/go.mod h1:OtzRCsYTJlIr+vAzlOtauTkfQ1c25ebFuXq8tbbf8cw=
github.com/pascaldekloe/goe v0.0.0-20180627143212-57f6aae5913c/go.mod h1:lzWF7FIEvWOWxwDKqyGYQf6ZUaNfKdP144TG7ZOy1lc=
github.com/pascaldekloe/goe v0.1.0 h1:cBOtyMzM9HTpWjXfbbunk26uA6nG3a8n06Wieeh0MwY=
github.com/pascaldekloe/goe v0.1.0/go.mod h1:lzWF7FIEvWOWxwDKqyGYQf6ZUaNfKdP144TG7ZOy1lc=
@@ -1821,6 +1825,7 @@ github.com/phpdave11/gofpdf v1.4.2/go.mod h1:zpO6xFn9yxo3YLyMvW8HcKWVdbNqgIfOOp2
github.com/phpdave11/gofpdi v1.0.12/go.mod h1:vBmVV0Do6hSBHC8uKUQ71JGW+ZGQq74llk/7bXwjDoI=
github.com/phpdave11/gofpdi v1.0.13/go.mod h1:vBmVV0Do6hSBHC8uKUQ71JGW+ZGQq74llk/7bXwjDoI=
github.com/pierrec/lz4 v2.0.5+incompatible/go.mod h1:pdkljMzZIN41W+lC3N2tnIh5sFi+IEE17M5jbnwPHcY=
github.com/pierrec/lz4/v4 v4.1.15 h1:MO0/ucJhngq7299dKLwIMtgTfbkoSPF6AoMYDd8Q4q0=
github.com/pierrec/lz4/v4 v4.1.15/go.mod h1:gZWDp/Ze/IJXGXf23ltt2EXimqmTUXEy0GFuRQyBid4=
github.com/pjbgf/sha1cd v0.3.0 h1:4D5XXmUUBUl/xQ6IjCkEAbqXskkq/4O7LmGn0AqMDs4=
github.com/pjbgf/sha1cd v0.3.0/go.mod h1:nZ1rrWOcGJ5uZgEEVL1VUM9iRQiZvWdbZjkKyFzPPsI=
@@ -1942,6 +1947,8 @@ github.com/sacloud/libsacloud v1.36.2/go.mod h1:P7YAOVmnIn3DKHqCZcUKYUXmSwGBm3yS
github.com/scaleway/scaleway-sdk-go v1.0.0-beta.7.0.20210127161313-bd30bebeac4f/go.mod h1:CJJ5VAbozOl0yEw7nHB9+7BXTJbIn6h7W+f6Gau5IP8=
github.com/sean-/seed v0.0.0-20170313163322-e2103e2c3529 h1:nn5Wsu0esKSJiIVhscUtVbo7ada43DJhG55ua/hjS5I=
github.com/sean-/seed v0.0.0-20170313163322-e2103e2c3529/go.mod h1:DxrIzT+xaE7yg65j358z/aeFdxmN0P9QXhEzd20vsDc=
github.com/segmentio/kafka-go v0.4.47 h1:IqziR4pA3vrZq7YdRxaT3w1/5fvIH5qpCwstUanQQB0=
github.com/segmentio/kafka-go v0.4.47/go.mod h1:HjF6XbOKh0Pjlkr5GVZxt6CsjjwnmhVOfURM5KMd8qg=
github.com/segmentio/ksuid v1.0.4 h1:sBo2BdShXjmcugAMwjugoGUdUV0pcxY5mW4xKRn3v4c=
github.com/segmentio/ksuid v1.0.4/go.mod h1:/XUiZBD3kVx5SmUOl55voK5yeAbBNNIed+2O73XgrPE=
github.com/sergi/go-diff v1.3.1 h1:xkr+Oxo4BOQKmkn/B9eMK0g5Kg/983T9DqqPHwYqD+8=
@@ -1951,6 +1958,8 @@ github.com/sethvargo/go-password v0.2.0 h1:BTDl4CC/gjf/axHMaDQtw507ogrXLci6XRiLc
github.com/sethvargo/go-password v0.2.0/go.mod h1:Ym4Mr9JXLBycr02MFuVQ/0JHidNetSgbzutTr3zsYXE=
github.com/shamaton/msgpack/v2 v2.2.0 h1:IP1m01pHwCrMa6ZccP9B3bqxEMKMSmMVAVKk54g3L/Y=
github.com/shamaton/msgpack/v2 v2.2.0/go.mod h1:6khjYnkx73f7VQU7wjcFS9DFjs+59naVWJv1TB7qdOI=
github.com/shirou/gopsutil v3.21.11+incompatible h1:+1+c1VGhc88SSonWP6foOcLhvnKlUeu/erjjvaPEYiI=
github.com/shirou/gopsutil v3.21.11+incompatible/go.mod h1:5b4v6he4MtMOwMlS0TUMTu2PcXUg8+E1lC7eC3UO/RA=
github.com/shurcooL/httpfs v0.0.0-20190707220628-8d4bc4ba7749 h1:bUGsEnyNbVPw06Bs80sCeARAlK8lhwqGyi6UT8ymuGk=
github.com/shurcooL/httpfs v0.0.0-20190707220628-8d4bc4ba7749/go.mod h1:ZY1cvUeJuFPAdZ/B6v7RHavJWZn2YPVFQ1OSXhCGOkg=
github.com/shurcooL/sanitized_anchor_name v1.0.0/go.mod h1:1NzhyTcUVG4SuEtjjoZeVRXNmyL/1OwPU0+IJeTBvfc=
@@ -2038,6 +2047,10 @@ github.com/tidwall/pretty v1.2.1 h1:qjsOFOWWQl+N3RsoF5/ssm1pHmJJwhjlSbZ51I6wMl4=
github.com/tidwall/pretty v1.2.1/go.mod h1:ITEVvHYasfjBbM0u2Pg8T2nJnzm8xPwvNhhsoaGGjNU=
github.com/tidwall/sjson v1.2.5 h1:kLy8mja+1c9jlljvWTlSazM7cKDRfJuR/bOJhcY5NcY=
github.com/tidwall/sjson v1.2.5/go.mod h1:Fvgq9kS/6ociJEDnK0Fk1cpYF4FIW6ZF7LAe+6jwd28=
github.com/tklauser/go-sysconf v0.3.14 h1:g5vzr9iPFFz24v2KZXs/pvpvh8/V9Fw6vQK5ZZb78yU=
github.com/tklauser/go-sysconf v0.3.14/go.mod h1:1ym4lWMLUOhuBOPGtRcJm7tEGX4SCYNEEEtghGG/8uY=
github.com/tklauser/numcpus v0.8.0 h1:Mx4Wwe/FjZLeQsK/6kt2EOepwwSl7SmJrK5bV/dXYgY=
github.com/tklauser/numcpus v0.8.0/go.mod h1:ZJZlAY+dmR4eut8epnzf0u/VwodKmryxR8txiloSqBE=
github.com/tmc/grpc-websocket-proxy v0.0.0-20190109142713-0ad062ec5ee5/go.mod h1:ncp9v5uamzpCO7NfCPTXjqaC+bZgJeR0sMTm6dMHP7U=
github.com/toorop/go-dkim v0.0.0-20201103131630-e1cd1a0a5208 h1:PM5hJF7HVfNWmCjMdEfbuOBNXSVF2cMFGgQTPdKCbwM=
github.com/toorop/go-dkim v0.0.0-20201103131630-e1cd1a0a5208/go.mod h1:BzWtXXrXzZUvMacR0oF/fbDDgUPO8L36tDMmRAf14ns=
@@ -2062,6 +2075,12 @@ github.com/wk8/go-ordered-map v1.0.0 h1:BV7z+2PaK8LTSd/mWgY12HyMAo5CEgkHqbkVq2th
github.com/wk8/go-ordered-map v1.0.0/go.mod h1:9ZIbRunKbuvfPKyBP1SIKLcXNlv74YCOZ3t3VTS6gRk=
github.com/xanzy/ssh-agent v0.3.3 h1:+/15pJfg/RsTxqYcX6fHqOXZwwMP+2VyYWJeWM2qQFM=
github.com/xanzy/ssh-agent v0.3.3/go.mod h1:6dzNDKs0J9rVPHPhaGCukekBHKqfl+L3KghI1Bc68Uw=
github.com/xdg-go/pbkdf2 v1.0.0 h1:Su7DPu48wXMwC3bs7MCNG+z4FhcyEuz5dlvchbq0B0c=
github.com/xdg-go/pbkdf2 v1.0.0/go.mod h1:jrpuAogTd400dnrH08LKmI/xc1MbPOebTwRqcT5RDeI=
github.com/xdg-go/scram v1.1.2 h1:FHX5I5B4i4hKRVRBCFRxq1iQRej7WO3hhBuJf+UUySY=
github.com/xdg-go/scram v1.1.2/go.mod h1:RT/sEzTbU5y00aCK8UOx6R7YryM0iF1N2MOmC3kKLN4=
github.com/xdg-go/stringprep v1.0.4 h1:XLI/Ng3O1Atzq0oBs3TWm+5ZVgkq2aqdlvP9JtoZ6c8=
github.com/xdg-go/stringprep v1.0.4/go.mod h1:mPGuuIYwz7CmR2bT9j4GbQqutWS1zV24gijq1dTyGkM=
github.com/xeipuuv/gojsonpointer v0.0.0-20180127040702-4e3ac2762d5f/go.mod h1:N2zxlSyiKSe5eX1tZViRH5QA0qijqEDrYZiPEAiq3wU=
github.com/xeipuuv/gojsonpointer v0.0.0-20190905194746-02993c407bfb h1:zGWFAtiMcyryUHoUjUJX0/lt1H2+i2Ka2n+D3DImSNo=
github.com/xeipuuv/gojsonpointer v0.0.0-20190905194746-02993c407bfb/go.mod h1:N2zxlSyiKSe5eX1tZViRH5QA0qijqEDrYZiPEAiq3wU=
@@ -2085,6 +2104,8 @@ github.com/yuin/goldmark v1.2.1/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9dec
github.com/yuin/goldmark v1.3.5/go.mod h1:mwnBkeHKe2W/ZEtQ+71ViKU8L12m81fl3OWwC1Zlc8k=
github.com/yuin/goldmark v1.4.1/go.mod h1:mwnBkeHKe2W/ZEtQ+71ViKU8L12m81fl3OWwC1Zlc8k=
github.com/yuin/goldmark v1.4.13/go.mod h1:6yULJ656Px+3vBD8DxQVa3kxgyrAnzto9xy5taEt/CY=
github.com/yusufpapurcu/wmi v1.2.4 h1:zFUKzehAFReQwLys1b/iSMl+JQGSCSjtVqQn9bBrPo0=
github.com/yusufpapurcu/wmi v1.2.4/go.mod h1:SBZ9tNy3G9/m5Oi98Zks0QjeHVDvuK0qfxQmPyzfmi0=
github.com/zeebo/assert v1.3.0/go.mod h1:Pq9JiuJQpG8JLJdtkwrJESF0Foym2/D9XMU5ciN/wJ0=
github.com/zeebo/xxh3 v1.0.2/go.mod h1:5NWz9Sef7zIDm2JHfFlcQvNekmcEl9ekUZQQKCYaDcA=
go-micro.dev/v4 v4.10.2 h1:GWQf1+FcAiMf1yca3P09RNjB31Xtk0C5HiKHSpq/2qA=
@@ -2327,6 +2348,7 @@ golang.org/x/net v0.9.0/go.mod h1:d48xBJpPfHeWQsugry2m+kC02ZBRGRgulfHnEXEuWns=
golang.org/x/net v0.10.0/go.mod h1:0qNGK6F8kojg2nk9dLZ2mShWaEBan6FAoqfSigmmuDg=
golang.org/x/net v0.12.0/go.mod h1:zEVYFnQC7m/vmpQFELhcD1EWkZlX69l4oqgmer6hfKA=
golang.org/x/net v0.14.0/go.mod h1:PpSgVXXLK0OxS0F31C1/tv6XNguvCrnXIDrFMspZIUI=
golang.org/x/net v0.17.0/go.mod h1:NxSsAGuq816PNPmqtQdLE42eU2Fs7NoRIZrHJAlaCOE=
golang.org/x/net v0.21.0/go.mod h1:bIjVDfnllIU7BJ2DNgfnXvpSvtn8VRwhlsaeUTyUS44=
golang.org/x/net v0.22.0/go.mod h1:JKghWKKOSdJwpW2GEx0Ja7fmaKnMsbu+MWVZTokSYmg=
golang.org/x/net v0.23.0/go.mod h1:JKghWKKOSdJwpW2GEx0Ja7fmaKnMsbu+MWVZTokSYmg=

162
vendor/github.com/cs3org/reva/v2/internal/grpc/services/storageprovider/storageprovider.go generated vendored
View File

@@ -101,20 +101,20 @@ func (c *config) init() {
}
}
type service struct {
type Service struct {
conf *config
storage storage.FS
Storage storage.FS
dataServerURL *url.URL
availableXS []*provider.ResourceChecksumPriority
}
func (s *service) Close() error {
return s.storage.Shutdown(context.Background())
func (s *Service) Close() error {
return s.Storage.Shutdown(context.Background())
}
func (s *service) UnprotectedEndpoints() []string { return []string{} }
func (s *Service) UnprotectedEndpoints() []string { return []string{} }
func (s *service) Register(ss *grpc.Server) {
func (s *Service) Register(ss *grpc.Server) {
provider.RegisterProviderAPIServer(ss, s)
}
@@ -199,9 +199,9 @@ func New(m map[string]interface{}, ss *grpc.Server) (rgrpc.Service, error) {
return nil, err
}
service := &service{
service := &Service{
conf: c,
storage: fs,
Storage: fs,
dataServerURL: u,
availableXS: xsTypes,
}
@@ -209,20 +209,20 @@ func New(m map[string]interface{}, ss *grpc.Server) (rgrpc.Service, error) {
return service, nil
}
func (s *service) SetArbitraryMetadata(ctx context.Context, req *provider.SetArbitraryMetadataRequest) (*provider.SetArbitraryMetadataResponse, error) {
func (s *Service) SetArbitraryMetadata(ctx context.Context, req *provider.SetArbitraryMetadataRequest) (*provider.SetArbitraryMetadataResponse, error) {
ctx = ctxpkg.ContextSetLockID(ctx, req.LockId)
err := s.storage.SetArbitraryMetadata(ctx, req.Ref, req.ArbitraryMetadata)
err := s.Storage.SetArbitraryMetadata(ctx, req.Ref, req.ArbitraryMetadata)
return &provider.SetArbitraryMetadataResponse{
Status: status.NewStatusFromErrType(ctx, "set arbitrary metadata", err),
}, nil
}
func (s *service) UnsetArbitraryMetadata(ctx context.Context, req *provider.UnsetArbitraryMetadataRequest) (*provider.UnsetArbitraryMetadataResponse, error) {
func (s *Service) UnsetArbitraryMetadata(ctx context.Context, req *provider.UnsetArbitraryMetadataRequest) (*provider.UnsetArbitraryMetadataResponse, error) {
ctx = ctxpkg.ContextSetLockID(ctx, req.LockId)
err := s.storage.UnsetArbitraryMetadata(ctx, req.Ref, req.ArbitraryMetadataKeys)
err := s.Storage.UnsetArbitraryMetadata(ctx, req.Ref, req.ArbitraryMetadataKeys)
return &provider.UnsetArbitraryMetadataResponse{
Status: status.NewStatusFromErrType(ctx, "unset arbitrary metadata", err),
@@ -230,13 +230,13 @@ func (s *service) UnsetArbitraryMetadata(ctx context.Context, req *provider.Unse
}
// SetLock puts a lock on the given reference
func (s *service) SetLock(ctx context.Context, req *provider.SetLockRequest) (*provider.SetLockResponse, error) {
func (s *Service) SetLock(ctx context.Context, req *provider.SetLockRequest) (*provider.SetLockResponse, error) {
if !canLockPublicShare(ctx) {
return &provider.SetLockResponse{
Status: status.NewPermissionDenied(ctx, nil, "no permission to lock the share"),
}, nil
}
err := s.storage.SetLock(ctx, req.Ref, req.Lock)
err := s.Storage.SetLock(ctx, req.Ref, req.Lock)
return &provider.SetLockResponse{
Status: status.NewStatusFromErrType(ctx, "set lock", err),
@@ -244,8 +244,8 @@ func (s *service) SetLock(ctx context.Context, req *provider.SetLockRequest) (*p
}
// GetLock returns an existing lock on the given reference
func (s *service) GetLock(ctx context.Context, req *provider.GetLockRequest) (*provider.GetLockResponse, error) {
lock, err := s.storage.GetLock(ctx, req.Ref)
func (s *Service) GetLock(ctx context.Context, req *provider.GetLockRequest) (*provider.GetLockResponse, error) {
lock, err := s.Storage.GetLock(ctx, req.Ref)
return &provider.GetLockResponse{
Status: status.NewStatusFromErrType(ctx, "get lock", err),
@@ -254,14 +254,14 @@ func (s *service) GetLock(ctx context.Context, req *provider.GetLockRequest) (*p
}
// RefreshLock refreshes an existing lock on the given reference
func (s *service) RefreshLock(ctx context.Context, req *provider.RefreshLockRequest) (*provider.RefreshLockResponse, error) {
func (s *Service) RefreshLock(ctx context.Context, req *provider.RefreshLockRequest) (*provider.RefreshLockResponse, error) {
if !canLockPublicShare(ctx) {
return &provider.RefreshLockResponse{
Status: status.NewPermissionDenied(ctx, nil, "no permission to refresh the share lock"),
}, nil
}
err := s.storage.RefreshLock(ctx, req.Ref, req.Lock, req.ExistingLockId)
err := s.Storage.RefreshLock(ctx, req.Ref, req.Lock, req.ExistingLockId)
return &provider.RefreshLockResponse{
Status: status.NewStatusFromErrType(ctx, "refresh lock", err),
@@ -269,21 +269,21 @@ func (s *service) RefreshLock(ctx context.Context, req *provider.RefreshLockRequ
}
// Unlock removes an existing lock from the given reference
func (s *service) Unlock(ctx context.Context, req *provider.UnlockRequest) (*provider.UnlockResponse, error) {
func (s *Service) Unlock(ctx context.Context, req *provider.UnlockRequest) (*provider.UnlockResponse, error) {
if !canLockPublicShare(ctx) {
return &provider.UnlockResponse{
Status: status.NewPermissionDenied(ctx, nil, "no permission to unlock the share"),
}, nil
}
err := s.storage.Unlock(ctx, req.Ref, req.Lock)
err := s.Storage.Unlock(ctx, req.Ref, req.Lock)
return &provider.UnlockResponse{
Status: status.NewStatusFromErrType(ctx, "unlock", err),
}, nil
}
func (s *service) InitiateFileDownload(ctx context.Context, req *provider.InitiateFileDownloadRequest) (*provider.InitiateFileDownloadResponse, error) {
func (s *Service) InitiateFileDownload(ctx context.Context, req *provider.InitiateFileDownloadRequest) (*provider.InitiateFileDownloadResponse, error) {
// TODO(labkode): maybe add some checks before download starts? eg. check permissions?
// TODO(labkode): maybe add short-lived token?
// We now simply point the client to the data server.
@@ -329,7 +329,7 @@ func validateIfUnmodifiedSince(ifUnmodifiedSince *typesv1beta1.Timestamp, info *
}
}
func (s *service) InitiateFileUpload(ctx context.Context, req *provider.InitiateFileUploadRequest) (*provider.InitiateFileUploadResponse, error) {
func (s *Service) InitiateFileUpload(ctx context.Context, req *provider.InitiateFileUploadRequest) (*provider.InitiateFileUploadResponse, error) {
// TODO(labkode): same considerations as download
log := appctx.GetLogger(ctx)
if req.Ref.GetPath() == "/" {
@@ -412,7 +412,7 @@ func (s *service) InitiateFileUpload(ctx context.Context, req *provider.Initiate
metadata["expires"] = strconv.Itoa(int(expirationTimestamp.Seconds))
}
uploadIDs, err := s.storage.InitiateUpload(ctx, req.Ref, uploadLength, metadata)
uploadIDs, err := s.Storage.InitiateUpload(ctx, req.Ref, uploadLength, metadata)
if err != nil {
var st *rpc.Status
switch err.(type) {
@@ -477,9 +477,9 @@ func (s *service) InitiateFileUpload(ctx context.Context, req *provider.Initiate
return res, nil
}
func (s *service) GetPath(ctx context.Context, req *provider.GetPathRequest) (*provider.GetPathResponse, error) {
func (s *Service) GetPath(ctx context.Context, req *provider.GetPathRequest) (*provider.GetPathResponse, error) {
// TODO(labkode): check that the storage ID is the same as the storage provider id.
fn, err := s.storage.GetPathByID(ctx, req.ResourceId)
fn, err := s.Storage.GetPathByID(ctx, req.ResourceId)
if err != nil {
return &provider.GetPathResponse{
Status: status.NewStatusFromErrType(ctx, "get path", err),
@@ -492,17 +492,17 @@ func (s *service) GetPath(ctx context.Context, req *provider.GetPathRequest) (*p
return res, nil
}
func (s *service) GetHome(ctx context.Context, req *provider.GetHomeRequest) (*provider.GetHomeResponse, error) {
func (s *Service) GetHome(ctx context.Context, req *provider.GetHomeRequest) (*provider.GetHomeResponse, error) {
return nil, errtypes.NotSupported("unused, use the gateway to look up the user home")
}
func (s *service) CreateHome(ctx context.Context, req *provider.CreateHomeRequest) (*provider.CreateHomeResponse, error) {
func (s *Service) CreateHome(ctx context.Context, req *provider.CreateHomeRequest) (*provider.CreateHomeResponse, error) {
return nil, errtypes.NotSupported("use CreateStorageSpace with type personal")
}
// CreateStorageSpace creates a storage space
func (s *service) CreateStorageSpace(ctx context.Context, req *provider.CreateStorageSpaceRequest) (*provider.CreateStorageSpaceResponse, error) {
resp, err := s.storage.CreateStorageSpace(ctx, req)
func (s *Service) CreateStorageSpace(ctx context.Context, req *provider.CreateStorageSpaceRequest) (*provider.CreateStorageSpaceResponse, error) {
resp, err := s.Storage.CreateStorageSpace(ctx, req)
if err != nil {
var st *rpc.Status
switch err.(type) {
@@ -513,7 +513,7 @@ func (s *service) CreateStorageSpace(ctx context.Context, req *provider.CreateSt
case errtypes.NotSupported:
// if trying to create a user home fall back to CreateHome
if u, ok := ctxpkg.ContextGetUser(ctx); ok && req.Type == "personal" && utils.UserEqual(req.GetOwner().Id, u.Id) {
if err := s.storage.CreateHome(ctx); err != nil {
if err := s.Storage.CreateHome(ctx); err != nil {
st = status.NewInternal(ctx, "error creating home")
} else {
st = status.NewOK(ctx)
@@ -544,7 +544,7 @@ func (s *service) CreateStorageSpace(ctx context.Context, req *provider.CreateSt
return resp, nil
}
func (s *service) ListStorageSpaces(ctx context.Context, req *provider.ListStorageSpacesRequest) (*provider.ListStorageSpacesResponse, error) {
func (s *Service) ListStorageSpaces(ctx context.Context, req *provider.ListStorageSpacesRequest) (*provider.ListStorageSpacesResponse, error) {
log := appctx.GetLogger(ctx)
// TODO this is just temporary. Update the API to include this flag.
@@ -555,7 +555,7 @@ func (s *service) ListStorageSpaces(ctx context.Context, req *provider.ListStora
}
}
spaces, err := s.storage.ListStorageSpaces(ctx, req.Filters, unrestricted)
spaces, err := s.Storage.ListStorageSpaces(ctx, req.Filters, unrestricted)
if err != nil {
var st *rpc.Status
switch err.(type) {
@@ -593,8 +593,8 @@ func (s *service) ListStorageSpaces(ctx context.Context, req *provider.ListStora
}, nil
}
func (s *service) UpdateStorageSpace(ctx context.Context, req *provider.UpdateStorageSpaceRequest) (*provider.UpdateStorageSpaceResponse, error) {
res, err := s.storage.UpdateStorageSpace(ctx, req)
func (s *Service) UpdateStorageSpace(ctx context.Context, req *provider.UpdateStorageSpaceRequest) (*provider.UpdateStorageSpaceResponse, error) {
res, err := s.Storage.UpdateStorageSpace(ctx, req)
if err != nil {
appctx.GetLogger(ctx).
Error().
@@ -607,14 +607,14 @@ func (s *service) UpdateStorageSpace(ctx context.Context, req *provider.UpdateSt
return res, nil
}
func (s *service) DeleteStorageSpace(ctx context.Context, req *provider.DeleteStorageSpaceRequest) (*provider.DeleteStorageSpaceResponse, error) {
func (s *Service) DeleteStorageSpace(ctx context.Context, req *provider.DeleteStorageSpaceRequest) (*provider.DeleteStorageSpaceResponse, error) {
// we need to get the space before so we can return critical information
// FIXME: why is this string parsing necessary?
idraw, _ := storagespace.ParseID(req.Id.GetOpaqueId())
idraw.OpaqueId = idraw.GetSpaceId()
id := &provider.StorageSpaceId{OpaqueId: storagespace.FormatResourceID(idraw)}
spaces, err := s.storage.ListStorageSpaces(ctx, []*provider.ListStorageSpacesRequest_Filter{{Type: provider.ListStorageSpacesRequest_Filter_TYPE_ID, Term: &provider.ListStorageSpacesRequest_Filter_Id{Id: id}}}, true)
spaces, err := s.Storage.ListStorageSpaces(ctx, []*provider.ListStorageSpacesRequest_Filter{{Type: provider.ListStorageSpacesRequest_Filter_TYPE_ID, Term: &provider.ListStorageSpacesRequest_Filter_Id{Id: id}}}, true)
if err != nil {
var st *rpc.Status
switch err.(type) {
@@ -636,7 +636,7 @@ func (s *service) DeleteStorageSpace(ctx context.Context, req *provider.DeleteSt
}, nil
}
if err := s.storage.DeleteStorageSpace(ctx, req); err != nil {
if err := s.Storage.DeleteStorageSpace(ctx, req); err != nil {
var st *rpc.Status
switch err.(type) {
case errtypes.IsNotFound:
@@ -670,7 +670,7 @@ func (s *service) DeleteStorageSpace(ctx context.Context, req *provider.DeleteSt
return res, nil
}
func (s *service) CreateContainer(ctx context.Context, req *provider.CreateContainerRequest) (*provider.CreateContainerResponse, error) {
func (s *Service) CreateContainer(ctx context.Context, req *provider.CreateContainerRequest) (*provider.CreateContainerResponse, error) {
// FIXME these should be part of the CreateContainerRequest object
if req.Opaque != nil {
if e, ok := req.Opaque.Map["lockid"]; ok && e.Decoder == "plain" {
@@ -678,14 +678,14 @@ func (s *service) CreateContainer(ctx context.Context, req *provider.CreateConta
}
}
err := s.storage.CreateDir(ctx, req.Ref)
err := s.Storage.CreateDir(ctx, req.Ref)
return &provider.CreateContainerResponse{
Status: status.NewStatusFromErrType(ctx, "create container", err),
}, nil
}
func (s *service) TouchFile(ctx context.Context, req *provider.TouchFileRequest) (*provider.TouchFileResponse, error) {
func (s *Service) TouchFile(ctx context.Context, req *provider.TouchFileRequest) (*provider.TouchFileResponse, error) {
// FIXME these should be part of the TouchFileRequest object
var mtime string
if req.Opaque != nil {
@@ -695,14 +695,14 @@ func (s *service) TouchFile(ctx context.Context, req *provider.TouchFileRequest)
mtime = utils.ReadPlainFromOpaque(req.Opaque, "X-OC-Mtime")
}
err := s.storage.TouchFile(ctx, req.Ref, utils.ExistsInOpaque(req.Opaque, "markprocessing"), mtime)
err := s.Storage.TouchFile(ctx, req.Ref, utils.ExistsInOpaque(req.Opaque, "markprocessing"), mtime)
return &provider.TouchFileResponse{
Status: status.NewStatusFromErrType(ctx, "touch file", err),
}, nil
}
func (s *service) Delete(ctx context.Context, req *provider.DeleteRequest) (*provider.DeleteResponse, error) {
func (s *Service) Delete(ctx context.Context, req *provider.DeleteRequest) (*provider.DeleteResponse, error) {
if req.Ref.GetPath() == "/" {
return &provider.DeleteResponse{
Status: status.NewInternal(ctx, "can't delete mount path"),
@@ -720,7 +720,7 @@ func (s *service) Delete(ctx context.Context, req *provider.DeleteRequest) (*pro
}
}
md, err := s.storage.GetMD(ctx, req.Ref, []string{}, []string{"id", "status"})
md, err := s.Storage.GetMD(ctx, req.Ref, []string{}, []string{"id", "status"})
if err != nil {
return &provider.DeleteResponse{
Status: status.NewStatusFromErrType(ctx, "can't stat resource to delete", err),
@@ -741,7 +741,7 @@ func (s *service) Delete(ctx context.Context, req *provider.DeleteRequest) (*pro
}, nil
}
err = s.storage.Delete(ctx, req.Ref)
err = s.Storage.Delete(ctx, req.Ref)
return &provider.DeleteResponse{
Status: status.NewStatusFromErrType(ctx, "delete", err),
@@ -753,17 +753,17 @@ func (s *service) Delete(ctx context.Context, req *provider.DeleteRequest) (*pro
}, nil
}
func (s *service) Move(ctx context.Context, req *provider.MoveRequest) (*provider.MoveResponse, error) {
func (s *Service) Move(ctx context.Context, req *provider.MoveRequest) (*provider.MoveResponse, error) {
ctx = ctxpkg.ContextSetLockID(ctx, req.LockId)
err := s.storage.Move(ctx, req.Source, req.Destination)
err := s.Storage.Move(ctx, req.Source, req.Destination)
return &provider.MoveResponse{
Status: status.NewStatusFromErrType(ctx, "move", err),
}, nil
}
func (s *service) Stat(ctx context.Context, req *provider.StatRequest) (*provider.StatResponse, error) {
func (s *Service) Stat(ctx context.Context, req *provider.StatRequest) (*provider.StatResponse, error) {
ctx, span := appctx.GetTracerProvider(ctx).Tracer(tracerName).Start(ctx, "stat")
defer span.End()
@@ -772,7 +772,7 @@ func (s *service) Stat(ctx context.Context, req *provider.StatRequest) (*provide
Value: attribute.StringValue(req.GetRef().String()),
})
md, err := s.storage.GetMD(ctx, req.GetRef(), req.GetArbitraryMetadataKeys(), req.GetFieldMask().GetPaths())
md, err := s.Storage.GetMD(ctx, req.GetRef(), req.GetArbitraryMetadataKeys(), req.GetFieldMask().GetPaths())
if err != nil {
return &provider.StatResponse{
Status: status.NewStatusFromErrType(ctx, "stat", err),
@@ -789,11 +789,11 @@ func (s *service) Stat(ctx context.Context, req *provider.StatRequest) (*provide
}, nil
}
func (s *service) ListContainerStream(req *provider.ListContainerStreamRequest, ss provider.ProviderAPI_ListContainerStreamServer) error {
func (s *Service) ListContainerStream(req *provider.ListContainerStreamRequest, ss provider.ProviderAPI_ListContainerStreamServer) error {
ctx := ss.Context()
log := appctx.GetLogger(ctx)
mds, err := s.storage.ListFolder(ctx, req.GetRef(), req.GetArbitraryMetadataKeys(), req.GetFieldMask().GetPaths())
mds, err := s.Storage.ListFolder(ctx, req.GetRef(), req.GetArbitraryMetadataKeys(), req.GetFieldMask().GetPaths())
if err != nil {
var st *rpc.Status
switch err.(type) {
@@ -836,8 +836,8 @@ func (s *service) ListContainerStream(req *provider.ListContainerStreamRequest,
return nil
}
func (s *service) ListContainer(ctx context.Context, req *provider.ListContainerRequest) (*provider.ListContainerResponse, error) {
mds, err := s.storage.ListFolder(ctx, req.GetRef(), req.GetArbitraryMetadataKeys(), req.GetFieldMask().GetPaths())
func (s *Service) ListContainer(ctx context.Context, req *provider.ListContainerRequest) (*provider.ListContainerResponse, error) {
mds, err := s.Storage.ListFolder(ctx, req.GetRef(), req.GetArbitraryMetadataKeys(), req.GetFieldMask().GetPaths())
res := &provider.ListContainerResponse{
Status: status.NewStatusFromErrType(ctx, "list container", err),
Infos: mds,
@@ -854,8 +854,8 @@ func (s *service) ListContainer(ctx context.Context, req *provider.ListContainer
return res, nil
}
func (s *service) ListFileVersions(ctx context.Context, req *provider.ListFileVersionsRequest) (*provider.ListFileVersionsResponse, error) {
revs, err := s.storage.ListRevisions(ctx, req.Ref)
func (s *Service) ListFileVersions(ctx context.Context, req *provider.ListFileVersionsRequest) (*provider.ListFileVersionsResponse, error) {
revs, err := s.Storage.ListRevisions(ctx, req.Ref)
sort.Sort(descendingMtime(revs))
@@ -865,22 +865,22 @@ func (s *service) ListFileVersions(ctx context.Context, req *provider.ListFileVe
}, nil
}
func (s *service) RestoreFileVersion(ctx context.Context, req *provider.RestoreFileVersionRequest) (*provider.RestoreFileVersionResponse, error) {
func (s *Service) RestoreFileVersion(ctx context.Context, req *provider.RestoreFileVersionRequest) (*provider.RestoreFileVersionResponse, error) {
ctx = ctxpkg.ContextSetLockID(ctx, req.LockId)
err := s.storage.RestoreRevision(ctx, req.Ref, req.Key)
err := s.Storage.RestoreRevision(ctx, req.Ref, req.Key)
return &provider.RestoreFileVersionResponse{
Status: status.NewStatusFromErrType(ctx, "restore file version", err),
}, nil
}
func (s *service) ListRecycleStream(req *provider.ListRecycleStreamRequest, ss provider.ProviderAPI_ListRecycleStreamServer) error {
func (s *Service) ListRecycleStream(req *provider.ListRecycleStreamRequest, ss provider.ProviderAPI_ListRecycleStreamServer) error {
ctx := ss.Context()
log := appctx.GetLogger(ctx)
key, itemPath := router.ShiftPath(req.Key)
items, err := s.storage.ListRecycle(ctx, req.Ref, key, itemPath)
items, err := s.Storage.ListRecycle(ctx, req.Ref, key, itemPath)
if err != nil {
var st *rpc.Status
switch err.(type) {
@@ -922,9 +922,9 @@ func (s *service) ListRecycleStream(req *provider.ListRecycleStreamRequest, ss p
return nil
}
func (s *service) ListRecycle(ctx context.Context, req *provider.ListRecycleRequest) (*provider.ListRecycleResponse, error) {
func (s *Service) ListRecycle(ctx context.Context, req *provider.ListRecycleRequest) (*provider.ListRecycleResponse, error) {
key, itemPath := router.ShiftPath(req.Key)
items, err := s.storage.ListRecycle(ctx, req.Ref, key, itemPath)
items, err := s.Storage.ListRecycle(ctx, req.Ref, key, itemPath)
if err != nil {
var st *rpc.Status
switch err.(type) {
@@ -957,12 +957,12 @@ func (s *service) ListRecycle(ctx context.Context, req *provider.ListRecycleRequ
return res, nil
}
func (s *service) RestoreRecycleItem(ctx context.Context, req *provider.RestoreRecycleItemRequest) (*provider.RestoreRecycleItemResponse, error) {
func (s *Service) RestoreRecycleItem(ctx context.Context, req *provider.RestoreRecycleItemRequest) (*provider.RestoreRecycleItemResponse, error) {
ctx = ctxpkg.ContextSetLockID(ctx, req.LockId)
// TODO(labkode): CRITICAL: fill recycle info with storage provider.
key, itemPath := router.ShiftPath(req.Key)
err := s.storage.RestoreRecycleItem(ctx, req.Ref, key, itemPath, req.RestoreRef)
err := s.Storage.RestoreRecycleItem(ctx, req.Ref, key, itemPath, req.RestoreRef)
res := &provider.RestoreRecycleItemResponse{
Status: status.NewStatusFromErrType(ctx, "restore recycle item", err),
@@ -970,7 +970,7 @@ func (s *service) RestoreRecycleItem(ctx context.Context, req *provider.RestoreR
return res, nil
}
func (s *service) PurgeRecycle(ctx context.Context, req *provider.PurgeRecycleRequest) (*provider.PurgeRecycleResponse, error) {
func (s *Service) PurgeRecycle(ctx context.Context, req *provider.PurgeRecycleRequest) (*provider.PurgeRecycleResponse, error) {
// FIXME these should be part of the PurgeRecycleRequest object
if req.Opaque != nil {
if e, ok := req.Opaque.Map["lockid"]; ok && e.Decoder == "plain" {
@@ -981,7 +981,7 @@ func (s *service) PurgeRecycle(ctx context.Context, req *provider.PurgeRecycleRe
// if a key was sent as opaque id purge only that item
key, itemPath := router.ShiftPath(req.Key)
if key != "" {
if err := s.storage.PurgeRecycleItem(ctx, req.Ref, key, itemPath); err != nil {
if err := s.Storage.PurgeRecycleItem(ctx, req.Ref, key, itemPath); err != nil {
st := status.NewStatusFromErrType(ctx, "error purging recycle item", err)
appctx.GetLogger(ctx).
Error().
@@ -994,7 +994,7 @@ func (s *service) PurgeRecycle(ctx context.Context, req *provider.PurgeRecycleRe
Status: st,
}, nil
}
} else if err := s.storage.EmptyRecycle(ctx, req.Ref); err != nil {
} else if err := s.Storage.EmptyRecycle(ctx, req.Ref); err != nil {
// otherwise try emptying the whole recycle bin
st := status.NewStatusFromErrType(ctx, "error emptying recycle", err)
appctx.GetLogger(ctx).
@@ -1015,8 +1015,8 @@ func (s *service) PurgeRecycle(ctx context.Context, req *provider.PurgeRecycleRe
return res, nil
}
func (s *service) ListGrants(ctx context.Context, req *provider.ListGrantsRequest) (*provider.ListGrantsResponse, error) {
grants, err := s.storage.ListGrants(ctx, req.Ref)
func (s *Service) ListGrants(ctx context.Context, req *provider.ListGrantsRequest) (*provider.ListGrantsResponse, error) {
grants, err := s.Storage.ListGrants(ctx, req.Ref)
if err != nil {
var st *rpc.Status
switch err.(type) {
@@ -1045,7 +1045,7 @@ func (s *service) ListGrants(ctx context.Context, req *provider.ListGrantsReques
return res, nil
}
func (s *service) DenyGrant(ctx context.Context, req *provider.DenyGrantRequest) (*provider.DenyGrantResponse, error) {
func (s *Service) DenyGrant(ctx context.Context, req *provider.DenyGrantRequest) (*provider.DenyGrantResponse, error) {
// check grantee type is valid
if req.Grantee.Type == provider.GranteeType_GRANTEE_TYPE_INVALID {
return &provider.DenyGrantResponse{
@@ -1053,7 +1053,7 @@ func (s *service) DenyGrant(ctx context.Context, req *provider.DenyGrantRequest)
}, nil
}
err := s.storage.DenyGrant(ctx, req.Ref, req.Grantee)
err := s.Storage.DenyGrant(ctx, req.Ref, req.Grantee)
if err != nil {
var st *rpc.Status
switch err.(type) {
@@ -1086,7 +1086,7 @@ func (s *service) DenyGrant(ctx context.Context, req *provider.DenyGrantRequest)
return res, nil
}
func (s *service) AddGrant(ctx context.Context, req *provider.AddGrantRequest) (*provider.AddGrantResponse, error) {
func (s *Service) AddGrant(ctx context.Context, req *provider.AddGrantRequest) (*provider.AddGrantResponse, error) {
ctx = ctxpkg.ContextSetLockID(ctx, req.LockId)
// TODO: update CS3 APIs
@@ -1109,14 +1109,14 @@ func (s *service) AddGrant(ctx context.Context, req *provider.AddGrantRequest) (
}, nil
}
err := s.storage.AddGrant(ctx, req.Ref, req.Grant)
err := s.Storage.AddGrant(ctx, req.Ref, req.Grant)
return &provider.AddGrantResponse{
Status: status.NewStatusFromErrType(ctx, "add grant", err),
}, nil
}
func (s *service) UpdateGrant(ctx context.Context, req *provider.UpdateGrantRequest) (*provider.UpdateGrantResponse, error) {
func (s *Service) UpdateGrant(ctx context.Context, req *provider.UpdateGrantRequest) (*provider.UpdateGrantResponse, error) {
// FIXME these should be part of the UpdateGrantRequest object
if req.Opaque != nil {
if e, ok := req.Opaque.Map["lockid"]; ok && e.Decoder == "plain" {
@@ -1144,14 +1144,14 @@ func (s *service) UpdateGrant(ctx context.Context, req *provider.UpdateGrantRequ
}, nil
}
err := s.storage.UpdateGrant(ctx, req.Ref, req.Grant)
err := s.Storage.UpdateGrant(ctx, req.Ref, req.Grant)
return &provider.UpdateGrantResponse{
Status: status.NewStatusFromErrType(ctx, "update grant", err),
}, nil
}
func (s *service) RemoveGrant(ctx context.Context, req *provider.RemoveGrantRequest) (*provider.RemoveGrantResponse, error) {
func (s *Service) RemoveGrant(ctx context.Context, req *provider.RemoveGrantRequest) (*provider.RemoveGrantResponse, error) {
ctx = ctxpkg.ContextSetLockID(ctx, req.LockId)
// check targetType is valid
@@ -1168,14 +1168,14 @@ func (s *service) RemoveGrant(ctx context.Context, req *provider.RemoveGrantRequ
ctx = context.WithValue(ctx, utils.SpaceGrant, struct{}{})
}
err := s.storage.RemoveGrant(ctx, req.Ref, req.Grant)
err := s.Storage.RemoveGrant(ctx, req.Ref, req.Grant)
return &provider.RemoveGrantResponse{
Status: status.NewStatusFromErrType(ctx, "remove grant", err),
}, nil
}
func (s *service) CreateReference(ctx context.Context, req *provider.CreateReferenceRequest) (*provider.CreateReferenceResponse, error) {
func (s *Service) CreateReference(ctx context.Context, req *provider.CreateReferenceRequest) (*provider.CreateReferenceResponse, error) {
log := appctx.GetLogger(ctx)
// parse uri is valid
@@ -1187,7 +1187,7 @@ func (s *service) CreateReference(ctx context.Context, req *provider.CreateRefer
}, nil
}
if err := s.storage.CreateReference(ctx, req.Ref.GetPath(), u); err != nil {
if err := s.Storage.CreateReference(ctx, req.Ref.GetPath(), u); err != nil {
var st *rpc.Status
switch err.(type) {
case errtypes.IsNotFound:
@@ -1212,14 +1212,14 @@ func (s *service) CreateReference(ctx context.Context, req *provider.CreateRefer
}, nil
}
func (s *service) CreateSymlink(ctx context.Context, req *provider.CreateSymlinkRequest) (*provider.CreateSymlinkResponse, error) {
func (s *Service) CreateSymlink(ctx context.Context, req *provider.CreateSymlinkRequest) (*provider.CreateSymlinkResponse, error) {
return &provider.CreateSymlinkResponse{
Status: status.NewUnimplemented(ctx, errtypes.NotSupported("CreateSymlink not implemented"), "CreateSymlink not implemented"),
}, nil
}
func (s *service) GetQuota(ctx context.Context, req *provider.GetQuotaRequest) (*provider.GetQuotaResponse, error) {
total, used, remaining, err := s.storage.GetQuota(ctx, req.Ref)
func (s *Service) GetQuota(ctx context.Context, req *provider.GetQuotaRequest) (*provider.GetQuotaResponse, error) {
total, used, remaining, err := s.Storage.GetQuota(ctx, req.Ref)
if err != nil {
var st *rpc.Status
switch err.(type) {
@@ -1257,7 +1257,7 @@ func (s *service) GetQuota(ctx context.Context, req *provider.GetQuotaRequest) (
return res, nil
}
func (s *service) addMissingStorageProviderID(resourceID *provider.ResourceId, spaceID *provider.StorageSpaceId) {
func (s *Service) addMissingStorageProviderID(resourceID *provider.ResourceId, spaceID *provider.StorageSpaceId) {
// The storage driver might set the mount ID by itself, in which case skip this step
if resourceID != nil && resourceID.GetStorageId() == "" {
resourceID.StorageId = s.conf.MountID

49
vendor/github.com/cs3org/reva/v2/internal/http/services/appprovider/appprovider.go generated vendored
View File

@@ -56,11 +56,12 @@ func init() {
// Config holds the config options for the HTTP appprovider service
type Config struct {
Prefix string `mapstructure:"prefix"`
GatewaySvc string `mapstructure:"gatewaysvc"`
Insecure bool `mapstructure:"insecure"`
WebBaseURI string `mapstructure:"webbaseuri"`
Web Web `mapstructure:"web"`
Prefix string `mapstructure:"prefix"`
GatewaySvc string `mapstructure:"gatewaysvc"`
Insecure bool `mapstructure:"insecure"`
WebBaseURI string `mapstructure:"webbaseuri"`
Web Web `mapstructure:"web"`
SecureViewApp string `mapstructure:"secure_view_app"`
}
// Web holds the config options for the URL parameters for Web
@@ -342,6 +343,16 @@ func (s *svc) handleList(w http.ResponseWriter, r *http.Request) {
}
res := filterAppsByUserAgent(listRes.MimeTypes, r.UserAgent())
// if app name or address matches the configured secure view app add that flag to the response
for _, mt := range res {
for _, app := range mt.AppProviders {
if app.Name == s.conf.SecureViewApp {
app.SecureView = true
}
}
}
js, err := json.Marshal(map[string]interface{}{"mime-types": res})
if err != nil {
writeError(w, r, appErrorServerError, "error marshalling JSON response", err)
@@ -545,22 +556,38 @@ func newOpenInWebResponse(baseURI string, params, staticParams map[string]string
return openInWebResponse{URI: uri.String()}, nil
}
func filterAppsByUserAgent(mimeTypes []*appregistry.MimeTypeInfo, userAgent string) []*appregistry.MimeTypeInfo {
// MimeTypeInfo wraps the appregistry.MimeTypeInfo to change the app providers to ProviderInfos with a secure view flag
type MimeTypeInfo struct {
appregistry.MimeTypeInfo
AppProviders []*ProviderInfo `json:"app_providers"`
}
// ProviderInfo wraps the appregistry.ProviderInfo to add a secure view flag
type ProviderInfo struct {
appregistry.ProviderInfo
// TODO make this part of the CS3 provider info
SecureView bool `json:"secure_view"`
}
// filterAppsByUserAgent rewrites the mime type info to only include apps that can be called by the user agent
// it also wraps the provider info to be able to add a secure view flag
func filterAppsByUserAgent(mimeTypes []*appregistry.MimeTypeInfo, userAgent string) []*MimeTypeInfo {
ua := ua.Parse(userAgent)
res := []*appregistry.MimeTypeInfo{}
res := []*MimeTypeInfo{}
for _, m := range mimeTypes {
apps := []*appregistry.ProviderInfo{}
apps := []*ProviderInfo{}
for _, p := range m.AppProviders {
p.Address = "" // address is internal only and not needed in the client
// apps are called by name, so if it has no name it cannot be called and should not be advertised
// also filter Desktop-only apps if ua is not Desktop
if p.Name != "" && (ua.Desktop || !p.DesktopOnly) {
apps = append(apps, p)
apps = append(apps, &ProviderInfo{ProviderInfo: *p})
}
}
if len(apps) > 0 {
m.AppProviders = apps
res = append(res, m)
mt := &MimeTypeInfo{MimeTypeInfo: *m}
mt.AppProviders = apps
res = append(res, mt)
}
}
return res

8
vendor/github.com/cs3org/reva/v2/pkg/rhttp/datatx/manager/tus/tus.go generated vendored
View File

@@ -82,7 +82,7 @@ func New(m map[string]interface{}, publisher events.Publisher) (datatx.DataTX, e
}
func (m *manager) Handler(fs storage.FS) (http.Handler, error) {
composable, ok := fs.(composable)
composable, ok := fs.(storage.ComposableFS)
if !ok {
return nil, errtypes.NotSupported("file system does not support the tus protocol")
}
@@ -193,12 +193,6 @@ func (m *manager) Handler(fs storage.FS) (http.Handler, error) {
return h, nil
}
// Composable is the interface that a struct needs to implement
// to be composable, so that it can support the TUS methods
type composable interface {
UseIn(composer *tusd.StoreComposer)
}
func setHeaders(fs storage.FS, w http.ResponseWriter, r *http.Request) {
ctx := r.Context()
id := path.Base(r.URL.Path)

1
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/loader/loader.go generated vendored
View File

@@ -31,7 +31,6 @@ import (
_ "github.com/cs3org/reva/v2/pkg/storage/fs/nextcloud"
_ "github.com/cs3org/reva/v2/pkg/storage/fs/ocis"
_ "github.com/cs3org/reva/v2/pkg/storage/fs/owncloudsql"
_ "github.com/cs3org/reva/v2/pkg/storage/fs/posix"
_ "github.com/cs3org/reva/v2/pkg/storage/fs/s3"
_ "github.com/cs3org/reva/v2/pkg/storage/fs/s3ng"
// Add your own here

25
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/loader/loader_linux.go generated vendored Normal file
View File

@@ -0,0 +1,25 @@
// Copyright 2018-2024 CERN
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// In applying this license, CERN does not waive the privileges and immunities
// granted to it by virtue of its status as an Intergovernmental Organization
// or submit itself to any jurisdiction.
package loader
import (
// Load core storage filesystem backends.
_ "github.com/cs3org/reva/v2/pkg/storage/fs/posix"
// Add your own here
)

56
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/posix/blobstore/blobstore.go generated vendored
View File

@@ -20,14 +20,10 @@ package blobstore
import (
"bufio"
"fmt"
"io"
"os"
"path/filepath"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/lookup"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/node"
"github.com/cs3org/reva/v2/pkg/utils"
"github.com/pkg/errors"
)
@@ -38,11 +34,6 @@ type Blobstore struct {
// New returns a new Blobstore
func New(root string) (*Blobstore, error) {
err := os.MkdirAll(root, 0700)
if err != nil {
return nil, err
}
return &Blobstore{
root: root,
}, nil
@@ -50,35 +41,21 @@ func New(root string) (*Blobstore, error) {
// Upload stores some data in the blobstore under the given key
func (bs *Blobstore) Upload(node *node.Node, source string) error {
dest, err := bs.path(node)
if err != nil {
return err
}
// ensure parent path exists
if err := os.MkdirAll(filepath.Dir(dest), 0700); err != nil {
return errors.Wrap(err, "Decomposedfs: oCIS blobstore: error creating parent folders for blob")
}
if err := os.Rename(source, dest); err == nil {
return nil
}
// Rename failed, file needs to be copied.
file, err := os.Open(source)
if err != nil {
return errors.Wrap(err, "Decomposedfs: oCIS blobstore: Can not open source file to upload")
}
defer file.Close()
f, err := os.OpenFile(dest, os.O_CREATE|os.O_WRONLY, 0700)
f, err := os.OpenFile(node.InternalPath(), os.O_CREATE|os.O_WRONLY, 0700)
if err != nil {
return errors.Wrapf(err, "could not open blob '%s' for writing", dest)
return errors.Wrapf(err, "could not open blob '%s' for writing", node.InternalPath())
}
w := bufio.NewWriter(f)
_, err = w.ReadFrom(file)
if err != nil {
return errors.Wrapf(err, "could not write blob '%s'", dest)
return errors.Wrapf(err, "could not write blob '%s'", node.InternalPath())
}
return w.Flush()
@@ -86,37 +63,14 @@ func (bs *Blobstore) Upload(node *node.Node, source string) error {
// Download retrieves a blob from the blobstore for reading
func (bs *Blobstore) Download(node *node.Node) (io.ReadCloser, error) {
dest, err := bs.path(node)
file, err := os.Open(node.InternalPath())
if err != nil {
return nil, err
}
file, err := os.Open(dest)
if err != nil {
return nil, errors.Wrapf(err, "could not read blob '%s'", dest)
return nil, errors.Wrapf(err, "could not read blob '%s'", node.InternalPath())
}
return file, nil
}
// Delete deletes a blob from the blobstore
func (bs *Blobstore) Delete(node *node.Node) error {
dest, err := bs.path(node)
if err != nil {
return err
}
if err := utils.RemoveItem(dest); err != nil {
return errors.Wrapf(err, "could not delete blob '%s'", dest)
}
return nil
}
func (bs *Blobstore) path(node *node.Node) (string, error) {
if node.BlobID == "" {
return "", fmt.Errorf("blobstore: BlobID is empty")
}
return filepath.Join(
bs.root,
filepath.Clean(filepath.Join(
"/", "spaces", lookup.Pathify(node.SpaceID, 1, 2), "blobs", lookup.Pathify(node.BlobID, 4, 2)),
),
), nil
}

169
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/posix/lookup/lookup.go generated vendored
View File

@@ -24,16 +24,20 @@ import (
"os"
"path/filepath"
"strings"
"syscall"
user "github.com/cs3org/go-cs3apis/cs3/identity/user/v1beta1"
provider "github.com/cs3org/go-cs3apis/cs3/storage/provider/v1beta1"
"github.com/cs3org/reva/v2/pkg/appctx"
"github.com/cs3org/reva/v2/pkg/errtypes"
"github.com/cs3org/reva/v2/pkg/storage/fs/posix/options"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata/prefixes"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/node"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/options"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/usermapper"
"github.com/cs3org/reva/v2/pkg/storage/utils/templates"
"github.com/cs3org/reva/v2/pkg/storagespace"
"github.com/google/uuid"
"github.com/pkg/errors"
"github.com/rogpeppe/go-internal/lockedfile"
"go.opentelemetry.io/otel"
@@ -43,24 +47,127 @@ import (
var tracer trace.Tracer
var _spaceTypePersonal = "personal"
var _spaceTypeProject = "project"
func init() {
tracer = otel.Tracer("github.com/cs3org/reva/pkg/storage/utils/decomposedfs/lookup")
}
// IDCache is a cache for node ids
type IDCache interface {
Get(ctx context.Context, spaceID, nodeID string) (string, bool)
Set(ctx context.Context, spaceID, nodeID, val string) error
}
// Lookup implements transformations from filepath to node and back
type Lookup struct {
Options *options.Options
IDCache IDCache
metadataBackend metadata.Backend
userMapper usermapper.Mapper
}
// New returns a new Lookup instance
func New(b metadata.Backend, o *options.Options) *Lookup {
return &Lookup{
func New(b metadata.Backend, um usermapper.Mapper, o *options.Options) *Lookup {
lu := &Lookup{
Options: o,
metadataBackend: b,
IDCache: NewStoreIDCache(&o.Options),
userMapper: um,
}
go func() {
_ = lu.WarmupIDCache(o.Root)
}()
return lu
}
// CacheID caches the id for the given space and node id
func (lu *Lookup) CacheID(ctx context.Context, spaceID, nodeID, val string) error {
return lu.IDCache.Set(ctx, spaceID, nodeID, val)
}
// GetCachedID returns the cached id for the given space and node id
func (lu *Lookup) GetCachedID(ctx context.Context, spaceID, nodeID string) (string, bool) {
return lu.IDCache.Get(ctx, spaceID, nodeID)
}
// WarmupIDCache warms up the id cache
func (lu *Lookup) WarmupIDCache(root string) error {
spaceID := []byte("")
var gid int
return filepath.Walk(root, func(path string, info os.FileInfo, err error) error {
if err != nil {
return err
}
attribs, err := lu.metadataBackend.All(context.Background(), path)
if err == nil {
nodeSpaceID, ok := attribs[prefixes.SpaceIDAttr]
if ok {
spaceID = nodeSpaceID
// set the uid and gid for the space
fi, err := os.Stat(path)
if err != nil {
return err
}
sys := fi.Sys().(*syscall.Stat_t)
gid = int(sys.Gid)
_, err = lu.userMapper.ScopeUserByIds(-1, gid)
if err != nil {
return err
}
}
if len(spaceID) == 0 {
// try to find space
spaceCandidate := path
for strings.HasPrefix(spaceCandidate, lu.Options.Root) {
spaceID, err = lu.MetadataBackend().Get(context.Background(), spaceCandidate, prefixes.SpaceIDAttr)
if err == nil {
if lu.Options.UseSpaceGroups {
// set the uid and gid for the space
fi, err := os.Stat(spaceCandidate)
if err != nil {
return err
}
sys := fi.Sys().(*syscall.Stat_t)
gid := int(sys.Gid)
_, err = lu.userMapper.ScopeUserByIds(-1, gid)
if err != nil {
return err
}
}
break
}
spaceCandidate = filepath.Dir(spaceCandidate)
}
}
id, ok := attribs[prefixes.IDAttr]
if ok && len(spaceID) > 0 {
_ = lu.IDCache.Set(context.Background(), string(spaceID), string(id), path)
}
}
return nil
})
}
// NodeFromPath returns the node for the given path
func (lu *Lookup) NodeIDFromParentAndName(ctx context.Context, parent *node.Node, name string) (string, error) {
id, err := lu.metadataBackend.Get(ctx, filepath.Join(parent.InternalPath(), name), prefixes.IDAttr)
if err != nil {
if metadata.IsNotExist(err) {
return "", errtypes.NotFound(name)
}
return "", err
}
return string(id), nil
}
// MetadataBackend returns the metadata backend
@@ -118,17 +225,6 @@ func (lu *Lookup) TypeFromPath(ctx context.Context, path string) provider.Resour
return t
}
func (lu *Lookup) NodeIDFromParentAndName(ctx context.Context, parent *node.Node, name string) (string, error) {
id, err := lu.metadataBackend.Get(ctx, filepath.Join(parent.InternalPath(), name), prefixes.IDAttr)
if err != nil {
if metadata.IsNotExist(err) {
return "", errtypes.NotFound(name)
}
return "", err
}
return string(id), nil
}
// NodeFromResource takes in a request path or request id and converts it to a Node
func (lu *Lookup) NodeFromResource(ctx context.Context, ref *provider.Reference) (*node.Node, error) {
ctx, span := tracer.Start(ctx, "NodeFromResource")
@@ -272,11 +368,9 @@ func (lu *Lookup) InternalRoot() string {
// InternalPath returns the internal path for a given ID
func (lu *Lookup) InternalPath(spaceID, nodeID string) string {
return filepath.Join(lu.Options.Root, "spaces", Pathify(spaceID, 1, 2), "nodes", Pathify(nodeID, 4, 2))
}
path, _ := lu.IDCache.Get(context.Background(), spaceID, nodeID)
func (lu *Lookup) SpacePath(spaceID string) string {
return filepath.Join(lu.Options.Root, spaceID)
return path
}
// // ReferenceFromAttr returns a CS3 reference from xattr of a node.
@@ -358,30 +452,25 @@ func (lu *Lookup) CopyMetadataWithSourceLock(ctx context.Context, sourcePath, ta
// GenerateSpaceID generates a space id for the given space type and owner
func (lu *Lookup) GenerateSpaceID(spaceType string, owner *user.User) (string, error) {
switch spaceType {
case _spaceTypeProject:
return uuid.New().String(), nil
case _spaceTypePersonal:
return templates.WithUser(owner, lu.Options.UserLayout), nil
path := templates.WithUser(owner, lu.Options.UserLayout)
spaceID, err := lu.metadataBackend.Get(context.Background(), filepath.Join(lu.Options.Root, path), prefixes.IDAttr)
if err != nil {
if metadata.IsNotExist(err) || metadata.IsAttrUnset(err) {
return uuid.New().String(), nil
} else {
return "", err
}
}
resID, err := storagespace.ParseID(string(spaceID))
if err != nil {
return "", err
}
return resID.SpaceId, nil
default:
return "", fmt.Errorf("unsupported space type: %s", spaceType)
}
}
// DetectBackendOnDisk returns the name of the metadata backend being used on disk
func DetectBackendOnDisk(root string) string {
matches, _ := filepath.Glob(filepath.Join(root, "spaces", "*", "*"))
if len(matches) > 0 {
base := matches[len(matches)-1]
spaceid := strings.ReplaceAll(
strings.TrimPrefix(base, filepath.Join(root, "spaces")),
"/", "")
spaceRoot := Pathify(spaceid, 4, 2)
_, err := os.Stat(filepath.Join(base, "nodes", spaceRoot+".mpk"))
if err == nil {
return "mpk"
}
_, err = os.Stat(filepath.Join(base, "nodes", spaceRoot+".ini"))
if err == nil {
return "ini"
}
}
return "xattrs"
}

69
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/posix/lookup/store_idcache.go generated vendored Normal file
View File

@@ -0,0 +1,69 @@
// Copyright 2018-2021 CERN
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// In applying this license, CERN does not waive the privileges and immunities
// granted to it by virtue of its status as an Intergovernmental Organization
// or submit itself to any jurisdiction.
package lookup
import (
"context"
microstore "go-micro.dev/v4/store"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/options"
"github.com/cs3org/reva/v2/pkg/store"
)
type StoreIDCache struct {
cache microstore.Store
}
// NewMemoryIDCache returns a new MemoryIDCache
func NewStoreIDCache(o *options.Options) *StoreIDCache {
return &StoreIDCache{
cache: store.Create(
store.Store(o.IDCache.Store),
store.TTL(o.IDCache.TTL),
store.Size(o.IDCache.Size),
microstore.Nodes(o.IDCache.Nodes...),
microstore.Database(o.IDCache.Database),
microstore.Table(o.IDCache.Table),
store.DisablePersistence(o.IDCache.DisablePersistence),
store.Authentication(o.IDCache.AuthUsername, o.IDCache.AuthPassword),
),
}
}
// Add adds a new entry to the cache
func (c *StoreIDCache) Set(_ context.Context, spaceID, nodeID, val string) error {
return c.cache.Write(&microstore.Record{
Key: cacheKey(spaceID, nodeID),
Value: []byte(val),
})
}
// Get returns the value for a given key
func (c *StoreIDCache) Get(_ context.Context, spaceID, nodeID string) (string, bool) {
records, err := c.cache.Read(cacheKey(spaceID, nodeID))
if err != nil {
return "", false
}
return string(records[0].Value), true
}
func cacheKey(spaceid, nodeID string) string {
return spaceid + "!" + nodeID
}

52
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/posix/options/options.go generated vendored Normal file
View File

@@ -0,0 +1,52 @@
// Copyright 2018-2024 CERN
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// In applying this license, CERN does not waive the privileges and immunities
// granted to it by virtue of its status as an Intergovernmental Organization
// or submit itself to any jurisdiction.
package options
import (
decomposedoptions "github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/options"
"github.com/mitchellh/mapstructure"
"github.com/pkg/errors"
)
type Options struct {
decomposedoptions.Options
UseSpaceGroups bool `mapstructure:"use_space_groups"`
WatchType string `mapstructure:"watch_type"`
WatchPath string `mapstructure:"watch_path"`
WatchFolderKafkaBrokers string `mapstructure:"watch_folder_kafka_brokers"`
}
// New returns a new Options instance for the given configuration
func New(m map[string]interface{}) (*Options, error) {
o := &Options{}
if err := mapstructure.Decode(m, o); err != nil {
err = errors.Wrap(err, "error decoding conf")
return nil, err
}
do, err := decomposedoptions.New(m)
if err != nil {
return nil, err
}
o.Options = *do
return o, nil
}

126
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/posix/posix.go generated vendored
View File

@@ -1,3 +1,6 @@
//go:build linux
// +build linux
// Copyright 2018-2021 CERN
//
// Licensed under the Apache License, Version 2.0 (the "License");
@@ -19,9 +22,12 @@
package posix
import (
"context"
"fmt"
"path"
"os"
"syscall"
tusd "github.com/tus/tusd/pkg/handler"
microstore "go-micro.dev/v4/store"
"github.com/cs3org/reva/v2/pkg/events"
@@ -29,21 +35,31 @@ import (
"github.com/cs3org/reva/v2/pkg/storage"
"github.com/cs3org/reva/v2/pkg/storage/fs/posix/blobstore"
"github.com/cs3org/reva/v2/pkg/storage/fs/posix/lookup"
"github.com/cs3org/reva/v2/pkg/storage/fs/posix/options"
"github.com/cs3org/reva/v2/pkg/storage/fs/posix/tree"
"github.com/cs3org/reva/v2/pkg/storage/fs/registry"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/aspects"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/node"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/options"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/permissions"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/upload"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/usermapper"
"github.com/cs3org/reva/v2/pkg/storage/utils/middleware"
"github.com/cs3org/reva/v2/pkg/store"
"github.com/pkg/errors"
)
func init() {
registry.Register("posix", New)
}
type posixFS struct {
storage.FS
um usermapper.Mapper
}
// New returns an implementation to of the storage.FS interface that talk to
// a local filesystem.
func New(m map[string]interface{}, stream events.Stream) (storage.FS, error) {
@@ -52,22 +68,24 @@ func New(m map[string]interface{}, stream events.Stream) (storage.FS, error) {
return nil, err
}
bs, err := blobstore.New(path.Join(o.Root))
bs, err := blobstore.New(o.Root)
if err != nil {
return nil, err
}
um := usermapper.NewUnixMapper()
var lu *lookup.Lookup
switch o.MetadataBackend {
case "xattrs":
lu = lookup.New(metadata.XattrsBackend{}, o)
lu = lookup.New(metadata.XattrsBackend{}, um, o)
case "messagepack":
lu = lookup.New(metadata.NewMessagePackBackend(o.Root, o.FileMetadataCache), o)
lu = lookup.New(metadata.NewMessagePackBackend(o.Root, o.FileMetadataCache), um, o)
default:
return nil, fmt.Errorf("unknown metadata backend %s, only 'messagepack' or 'xattrs' (default) supported", o.MetadataBackend)
}
tp := tree.New(lu, bs, o, store.Create(
tp, err := tree.New(lu, bs, um, o, store.Create(
store.Store(o.IDCache.Store),
store.TTL(o.IDCache.TTL),
store.Size(o.IDCache.Size),
@@ -77,6 +95,9 @@ func New(m map[string]interface{}, stream events.Stream) (storage.FS, error) {
store.DisablePersistence(o.IDCache.DisablePersistence),
store.Authentication(o.IDCache.AuthUsername, o.IDCache.AuthPassword),
))
if err != nil {
return nil, err
}
permissionsSelector, err := pool.PermissionsSelector(o.PermissionsSVC, pool.WithTLSMode(o.PermTLSMode))
if err != nil {
@@ -86,15 +107,98 @@ func New(m map[string]interface{}, stream events.Stream) (storage.FS, error) {
p := permissions.NewPermissions(node.NewPermissions(lu), permissionsSelector)
aspects := aspects.Aspects{
Lookup: lu,
Tree: tp,
Permissions: p,
EventStream: stream,
Lookup: lu,
Tree: tp,
Permissions: p,
EventStream: stream,
UserMapper: um,
DisableVersioning: true,
}
fs, err := decomposedfs.New(o, aspects)
dfs, err := decomposedfs.New(&o.Options, aspects)
if err != nil {
return nil, err
}
hooks := []middleware.Hook{}
if o.UseSpaceGroups {
resolveSpaceHook := func(methodName string, ctx context.Context, spaceID string) (context.Context, middleware.UnHook, error) {
if spaceID == "" {
return ctx, nil, nil
}
spaceRoot := lu.InternalPath(spaceID, spaceID)
fi, err := os.Stat(spaceRoot)
if err != nil {
return ctx, nil, err
}
ctx = context.WithValue(ctx, decomposedfs.CtxKeySpaceGID, fi.Sys().(*syscall.Stat_t).Gid)
return ctx, nil, err
}
scopeSpaceGroupHook := func(methodName string, ctx context.Context, spaceID string) (context.Context, middleware.UnHook, error) {
spaceGID, ok := ctx.Value(decomposedfs.CtxKeySpaceGID).(uint32)
if !ok {
return ctx, nil, nil
}
unscope, err := um.ScopeUserByIds(-1, int(spaceGID))
if err != nil {
return ctx, nil, errors.Wrap(err, "failed to scope user")
}
return ctx, unscope, nil
}
hooks = append(hooks, resolveSpaceHook, scopeSpaceGroupHook)
}
mw := middleware.NewFS(dfs, hooks...)
fs := &posixFS{
FS: mw,
um: um,
}
return fs, nil
}
// ListUploadSessions returns the upload sessions matching the given filter
func (fs *posixFS) ListUploadSessions(ctx context.Context, filter storage.UploadSessionFilter) ([]storage.UploadSession, error) {
return fs.FS.(storage.UploadSessionLister).ListUploadSessions(ctx, filter)
}
// UseIn tells the tus upload middleware which extensions it supports.
func (fs *posixFS) UseIn(composer *tusd.StoreComposer) {
fs.FS.(storage.ComposableFS).UseIn(composer)
}
// NewUpload returns a new tus Upload instance
func (fs *posixFS) NewUpload(ctx context.Context, info tusd.FileInfo) (upload tusd.Upload, err error) {
return fs.FS.(tusd.DataStore).NewUpload(ctx, info)
}
// NewUpload returns a new tus Upload instance
func (fs *posixFS) GetUpload(ctx context.Context, id string) (upload tusd.Upload, err error) {
return fs.FS.(tusd.DataStore).GetUpload(ctx, id)
}
// AsTerminatableUpload returns a TerminatableUpload
// To implement the termination extension as specified in https://tus.io/protocols/resumable-upload.html#termination
// the storage needs to implement AsTerminatableUpload
func (fs *posixFS) AsTerminatableUpload(up tusd.Upload) tusd.TerminatableUpload {
return up.(*upload.OcisSession)
}
// AsLengthDeclarableUpload returns a LengthDeclarableUpload
// To implement the creation-defer-length extension as specified in https://tus.io/protocols/resumable-upload.html#creation
// the storage needs to implement AsLengthDeclarableUpload
func (fs *posixFS) AsLengthDeclarableUpload(up tusd.Upload) tusd.LengthDeclarableUpload {
return up.(*upload.OcisSession)
}
// AsConcatableUpload returns a ConcatableUpload
// To implement the concatenation extension as specified in https://tus.io/protocols/resumable-upload.html#concatenation
// the storage needs to implement AsConcatableUpload
func (fs *posixFS) AsConcatableUpload(up tusd.Upload) tusd.ConcatableUpload {
return up.(*upload.OcisSession)
}

277
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/posix/tree/assimilation.go generated vendored Normal file
View File

@@ -0,0 +1,277 @@
// Copyright 2018-2021 CERN
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// In applying this license, CERN does not waive the privileges and immunities
// granted to it by virtue of its status as an Intergovernmental Organization
// or submit itself to any jurisdiction.
package tree
import (
"context"
"fmt"
"io/fs"
"os"
"path/filepath"
"strings"
"sync"
"syscall"
"time"
provider "github.com/cs3org/go-cs3apis/cs3/storage/provider/v1beta1"
"github.com/cs3org/reva/v2/pkg/storage/fs/posix/lookup"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata/prefixes"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/node"
"github.com/google/uuid"
"github.com/pkg/errors"
"github.com/rs/zerolog/log"
)
type ScanDebouncer struct {
after time.Duration
f func(item scanItem)
pending map[string]*time.Timer
inProgress sync.Map
mutex sync.Mutex
}
// NewScanDebouncer returns a new SpaceDebouncer instance
func NewScanDebouncer(d time.Duration, f func(item scanItem)) *ScanDebouncer {
return &ScanDebouncer{
after: d,
f: f,
pending: map[string]*time.Timer{},
inProgress: sync.Map{},
}
}
// Debounce restars the debounce timer for the given space
func (d *ScanDebouncer) Debounce(item scanItem) {
d.mutex.Lock()
defer d.mutex.Unlock()
path := item.Path
force := item.ForceRescan
if t := d.pending[item.Path]; t != nil {
force = force || item.ForceRescan
t.Stop()
}
d.pending[item.Path] = time.AfterFunc(d.after, func() {
if _, ok := d.inProgress.Load(path); ok {
// Reschedule this run for when the previous run has finished
d.mutex.Lock()
d.pending[path].Reset(d.after)
d.mutex.Unlock()
return
}
d.inProgress.Store(path, true)
defer d.inProgress.Delete(path)
d.f(scanItem{
Path: path,
ForceRescan: force,
})
})
}
func (t *Tree) workScanQueue() {
for i := 0; i < t.options.MaxConcurrency; i++ {
go func() {
for {
item := <-t.scanQueue
err := t.assimilate(item)
if err != nil {
log.Error().Err(err).Str("path", item.Path).Msg("failed to assimilate item")
continue
}
}
}()
}
}
// Scan scans the given path and updates the id chache
func (t *Tree) Scan(path string, forceRescan bool) error {
t.scanDebouncer.Debounce(scanItem{
Path: path,
ForceRescan: forceRescan,
})
return nil
}
func (t *Tree) assimilate(item scanItem) error {
var err error
// find the space id, scope by the according user
spaceID := []byte("")
spaceCandidate := item.Path
for strings.HasPrefix(spaceCandidate, t.options.Root) {
spaceID, err = t.lookup.MetadataBackend().Get(context.Background(), spaceCandidate, prefixes.SpaceIDAttr)
if err == nil {
if t.options.UseSpaceGroups {
// set the uid and gid for the space
fi, err := os.Stat(spaceCandidate)
if err != nil {
return err
}
sys := fi.Sys().(*syscall.Stat_t)
gid := int(sys.Gid)
_, err = t.userMapper.ScopeUserByIds(-1, gid)
if err != nil {
return err
}
}
break
}
spaceCandidate = filepath.Dir(spaceCandidate)
}
if len(spaceID) == 0 {
return fmt.Errorf("did not find space id for path")
}
var id []byte
if !item.ForceRescan {
// already assimilated?
id, err := t.lookup.MetadataBackend().Get(context.Background(), item.Path, prefixes.IDAttr)
if err == nil {
_ = t.lookup.(*lookup.Lookup).CacheID(context.Background(), string(spaceID), string(id), item.Path)
return nil
}
}
// lock the file for assimilation
unlock, err := t.lookup.MetadataBackend().Lock(item.Path)
if err != nil {
return errors.Wrap(err, "failed to lock item for assimilation")
}
defer func() {
_ = unlock()
}()
// check for the id attribute again after grabbing the lock, maybe the file was assimilated/created by us in the meantime
id, err = t.lookup.MetadataBackend().Get(context.Background(), item.Path, prefixes.IDAttr)
if err == nil {
_ = t.lookup.(*lookup.Lookup).CacheID(context.Background(), string(spaceID), string(id), item.Path)
if item.ForceRescan {
_, err = t.updateFile(item.Path, string(id), string(spaceID))
if err != nil {
return err
}
}
} else {
// assimilate new file
newId := uuid.New().String()
_, err = t.updateFile(item.Path, newId, string(spaceID))
if err != nil {
return err
}
}
return nil
}
func (t *Tree) updateFile(path, id, spaceID string) (fs.FileInfo, error) {
retries := 1
parentID := ""
assimilate:
if id != spaceID {
// read parent
parentAttribs, err := t.lookup.MetadataBackend().All(context.Background(), filepath.Dir(path))
if err != nil {
return nil, fmt.Errorf("failed to read parent item attributes")
}
if len(parentAttribs) == 0 || len(parentAttribs[prefixes.IDAttr]) == 0 {
if retries == 0 {
return nil, fmt.Errorf("got empty parent attribs even after assimilating")
}
// assimilate parent first
err = t.assimilate(scanItem{Path: filepath.Dir(path), ForceRescan: false})
if err != nil {
return nil, err
}
// retry
retries--
goto assimilate
}
parentID = string(parentAttribs[prefixes.IDAttr])
}
// assimilate file
fi, err := os.Stat(path)
if err != nil {
return nil, errors.Wrap(err, "failed to stat item")
}
previousAttribs, err := t.lookup.MetadataBackend().All(context.Background(), path)
if err != nil && !metadata.IsAttrUnset(err) {
return nil, errors.Wrap(err, "failed to get item attribs")
}
attributes := node.Attributes{
prefixes.IDAttr: []byte(id),
prefixes.NameAttr: []byte(filepath.Base(path)),
prefixes.MTimeAttr: []byte(fi.ModTime().Format(time.RFC3339)),
}
if len(parentID) > 0 {
attributes[prefixes.ParentidAttr] = []byte(parentID)
}
sha1h, md5h, adler32h, err := node.CalculateChecksums(context.Background(), path)
if err == nil {
attributes[prefixes.ChecksumPrefix+"sha1"] = sha1h.Sum(nil)
attributes[prefixes.ChecksumPrefix+"md5"] = md5h.Sum(nil)
attributes[prefixes.ChecksumPrefix+"adler32"] = adler32h.Sum(nil)
}
if fi.IsDir() {
attributes.SetInt64(prefixes.TypeAttr, int64(provider.ResourceType_RESOURCE_TYPE_CONTAINER))
attributes.SetInt64(prefixes.TreesizeAttr, 0)
if previousAttribs != nil && previousAttribs[prefixes.TreesizeAttr] != nil {
attributes[prefixes.TreesizeAttr] = previousAttribs[prefixes.TreesizeAttr]
}
attributes[prefixes.PropagationAttr] = []byte("1")
} else {
attributes.SetInt64(prefixes.TypeAttr, int64(provider.ResourceType_RESOURCE_TYPE_FILE))
attributes.SetString(prefixes.BlobIDAttr, id)
attributes.SetInt64(prefixes.BlobsizeAttr, fi.Size())
// propagate the change
sizeDiff := fi.Size()
if previousAttribs != nil && previousAttribs[prefixes.BlobsizeAttr] != nil {
oldSize, err := attributes.Int64(prefixes.BlobsizeAttr)
if err == nil {
sizeDiff -= oldSize
}
}
n := node.New(spaceID, id, parentID, filepath.Base(path), fi.Size(), "", provider.ResourceType_RESOURCE_TYPE_FILE, nil, t.lookup)
n.SpaceRoot = &node.Node{SpaceID: spaceID, ID: spaceID}
err = t.Propagate(context.Background(), n, sizeDiff)
if err != nil {
return nil, errors.Wrap(err, "failed to propagate")
}
}
err = t.lookup.MetadataBackend().SetMultiple(context.Background(), path, attributes, false)
if err != nil {
return nil, errors.Wrap(err, "failed to set attributes")
}
_ = t.lookup.(*lookup.Lookup).CacheID(context.Background(), spaceID, id, path)
return fi, nil
}

85
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/posix/tree/gpfsfilauditloggingwatcher.go generated vendored Normal file
View File

@@ -0,0 +1,85 @@
package tree
import (
"bufio"
"encoding/json"
"io"
"os"
"strconv"
"time"
"github.com/cs3org/reva/v2/pkg/storage/fs/posix/lookup"
)
type GpfsFileAuditLoggingWatcher struct {
tree *Tree
}
type lwe struct {
Event string
Path string
BytesWritten string
}
func NewGpfsFileAuditLoggingWatcher(tree *Tree, auditLogFile string) (*GpfsFileAuditLoggingWatcher, error) {
w := &GpfsFileAuditLoggingWatcher{
tree: tree,
}
_, err := os.Stat(auditLogFile)
if err != nil {
return nil, err
}
return w, nil
}
func (w *GpfsFileAuditLoggingWatcher) Watch(path string) {
start:
file, err := os.Open(path)
if err != nil {
// try again later
time.Sleep(5 * time.Second)
goto start
}
defer file.Close()
// Seek to the end of the file
_, err = file.Seek(0, io.SeekEnd)
if err != nil {
time.Sleep(5 * time.Second)
goto start
}
reader := bufio.NewReader(file)
ev := &lwe{}
for {
line, err := reader.ReadString('\n')
switch err {
case nil:
err := json.Unmarshal([]byte(line), ev)
if err != nil {
continue
}
switch ev.Event {
case "CREATE":
go func() { _ = w.tree.Scan(ev.Path, false) }()
case "CLOSE":
bytesWritten, err := strconv.Atoi(ev.BytesWritten)
if err == nil && bytesWritten > 0 {
go func() { _ = w.tree.Scan(ev.Path, true) }()
}
case "RENAME":
go func() {
_ = w.tree.Scan(ev.Path, true)
_ = w.tree.lookup.(*lookup.Lookup).WarmupIDCache(ev.Path)
}()
}
case io.EOF:
time.Sleep(1 * time.Second)
default:
time.Sleep(5 * time.Second)
goto start
}
}
}

67
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/posix/tree/gpfswatchfolderwatcher.go generated vendored Normal file
View File

@@ -0,0 +1,67 @@
package tree
import (
"context"
"encoding/json"
"log"
"strconv"
"strings"
"github.com/cs3org/reva/v2/pkg/storage/fs/posix/lookup"
kafka "github.com/segmentio/kafka-go"
)
type GpfsWatchFolderWatcher struct {
tree *Tree
brokers []string
}
func NewGpfsWatchFolderWatcher(tree *Tree, kafkaBrokers []string) (*GpfsWatchFolderWatcher, error) {
return &GpfsWatchFolderWatcher{
tree: tree,
brokers: kafkaBrokers,
}, nil
}
func (w *GpfsWatchFolderWatcher) Watch(topic string) {
r := kafka.NewReader(kafka.ReaderConfig{
Brokers: w.brokers,
GroupID: "ocis-posixfs",
Topic: topic,
})
lwev := &lwe{}
for {
m, err := r.ReadMessage(context.Background())
if err != nil {
break
}
err = json.Unmarshal(m.Value, lwev)
if err != nil {
continue
}
if strings.HasSuffix(lwev.Path, ".flock") || strings.HasSuffix(lwev.Path, ".mlock") {
continue
}
switch {
case strings.Contains(lwev.Event, "IN_CREATE"):
go func() { _ = w.tree.Scan(lwev.Path, false) }()
case strings.Contains(lwev.Event, "IN_CLOSE_WRITE"):
bytesWritten, err := strconv.Atoi(lwev.BytesWritten)
if err == nil && bytesWritten > 0 {
go func() { _ = w.tree.Scan(lwev.Path, true) }()
}
case strings.Contains(lwev.Event, "IN_MOVED_TO"):
go func() {
_ = w.tree.Scan(lwev.Path, true)
_ = w.tree.lookup.(*lookup.Lookup).WarmupIDCache(lwev.Path)
}()
}
}
if err := r.Close(); err != nil {
log.Fatal("failed to close reader:", err)
}
}

73
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/posix/tree/inotifywatcher.go generated vendored Normal file
View File

@@ -0,0 +1,73 @@
package tree
import (
"fmt"
"strings"
"github.com/cs3org/reva/v2/pkg/storage/fs/posix/lookup"
"github.com/pablodz/inotifywaitgo/inotifywaitgo"
)
type InotifyWatcher struct {
tree *Tree
}
func NewInotifyWatcher(tree *Tree) *InotifyWatcher {
return &InotifyWatcher{
tree: tree,
}
}
func (iw *InotifyWatcher) Watch(path string) {
events := make(chan inotifywaitgo.FileEvent)
errors := make(chan error)
go inotifywaitgo.WatchPath(&inotifywaitgo.Settings{
Dir: path,
FileEvents: events,
ErrorChan: errors,
KillOthers: true,
Options: &inotifywaitgo.Options{
Recursive: true,
Events: []inotifywaitgo.EVENT{
inotifywaitgo.CREATE,
inotifywaitgo.MOVED_TO,
inotifywaitgo.CLOSE_WRITE,
},
Monitor: true,
},
Verbose: false,
})
for {
select {
case event := <-events:
for _, e := range event.Events {
if strings.HasSuffix(event.Filename, ".flock") || strings.HasSuffix(event.Filename, ".mlock") {
continue
}
switch e {
case inotifywaitgo.CREATE:
go func() { _ = iw.tree.Scan(event.Filename, false) }()
case inotifywaitgo.MOVED_TO:
go func() {
_ = iw.tree.Scan(event.Filename, true)
_ = iw.tree.lookup.(*lookup.Lookup).WarmupIDCache(event.Filename)
}()
case inotifywaitgo.CLOSE_WRITE:
go func() { _ = iw.tree.Scan(event.Filename, true) }()
}
}
case err := <-errors:
switch err.Error() {
case inotifywaitgo.NOT_INSTALLED:
panic("Error: inotifywait is not installed")
case inotifywaitgo.INVALID_EVENT:
// ignore
default:
fmt.Printf("Error: %s\n", err)
}
}
}
}

335
vendor/github.com/cs3org/reva/v2/pkg/storage/fs/posix/tree/tree.go generated vendored
View File

@@ -30,23 +30,28 @@ import (
"strings"
"time"
provider "github.com/cs3org/go-cs3apis/cs3/storage/provider/v1beta1"
"github.com/cs3org/reva/v2/pkg/appctx"
"github.com/cs3org/reva/v2/pkg/errtypes"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/lookup"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata/prefixes"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/node"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/options"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/tree/propagator"
"github.com/cs3org/reva/v2/pkg/utils"
"github.com/google/uuid"
"github.com/pkg/errors"
"github.com/rs/zerolog"
"github.com/rs/zerolog/log"
"go-micro.dev/v4/store"
"go.opentelemetry.io/otel"
"go.opentelemetry.io/otel/trace"
"golang.org/x/sync/errgroup"
provider "github.com/cs3org/go-cs3apis/cs3/storage/provider/v1beta1"
"github.com/cs3org/reva/v2/pkg/appctx"
"github.com/cs3org/reva/v2/pkg/errtypes"
"github.com/cs3org/reva/v2/pkg/logger"
"github.com/cs3org/reva/v2/pkg/storage/fs/posix/lookup"
"github.com/cs3org/reva/v2/pkg/storage/fs/posix/options"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata/prefixes"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/node"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/tree/propagator"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/usermapper"
"github.com/cs3org/reva/v2/pkg/utils"
)
var tracer trace.Tracer
@@ -62,6 +67,15 @@ type Blobstore interface {
Delete(node *node.Node) error
}
type Watcher interface {
Watch(path string)
}
type scanItem struct {
Path string
ForceRescan bool
}
// Tree manages a hierarchical tree
type Tree struct {
lookup node.PathLookup
@@ -70,26 +84,76 @@ type Tree struct {
options *options.Options
idCache store.Store
userMapper usermapper.Mapper
idCache store.Store
watcher Watcher
scanQueue chan scanItem
scanDebouncer *ScanDebouncer
log *zerolog.Logger
}
// PermissionCheckFunc defined a function used to check resource permissions
type PermissionCheckFunc func(rp *provider.ResourcePermissions) bool
// New returns a new instance of Tree
func New(lu node.PathLookup, bs Blobstore, o *options.Options, cache store.Store) *Tree {
return &Tree{
func New(lu node.PathLookup, bs Blobstore, um usermapper.Mapper, o *options.Options, cache store.Store) (*Tree, error) {
log := logger.New()
scanQueue := make(chan scanItem)
t := &Tree{
lookup: lu,
blobstore: bs,
userMapper: um,
options: o,
idCache: cache,
propagator: propagator.New(lu, o),
propagator: propagator.New(lu, &o.Options),
scanQueue: scanQueue,
scanDebouncer: NewScanDebouncer(500*time.Millisecond, func(item scanItem) {
scanQueue <- item
}),
log: log,
}
watchPath := o.WatchPath
var err error
switch o.WatchType {
case "gpfswatchfolder":
t.watcher, err = NewGpfsWatchFolderWatcher(t, strings.Split(o.WatchFolderKafkaBrokers, ","))
if err != nil {
return nil, err
}
case "gpfsfileauditlogging":
t.watcher, err = NewGpfsFileAuditLoggingWatcher(t, o.WatchPath)
if err != nil {
return nil, err
}
default:
t.watcher = NewInotifyWatcher(t)
watchPath = o.Root
}
// Start watching for fs events and put them into the queue
go t.watcher.Watch(watchPath)
// Handle queued fs events
go t.workScanQueue()
return t, nil
}
// Setup prepares the tree structure
func (t *Tree) Setup() error {
return os.MkdirAll(t.options.Root, 0700)
err := os.MkdirAll(t.options.Root, 0700)
if err != nil {
return err
}
err = os.MkdirAll(t.options.UploadDirectory, 0700)
if err != nil {
return err
}
return nil
}
// GetMD returns the metadata of a node in the tree
@@ -115,35 +179,48 @@ func (t *Tree) TouchFile(ctx context.Context, n *node.Node, markprocessing bool,
return errtypes.AlreadyExists(n.ID)
}
parentPath := n.ParentPath()
nodePath := filepath.Join(parentPath, n.Name)
// lock the meta file
unlock, err := t.lookup.MetadataBackend().Lock(nodePath)
if err != nil {
return err
}
defer func() {
_ = unlock()
}()
if n.ID == "" {
n.ID = uuid.New().String()
}
n.SetType(provider.ResourceType_RESOURCE_TYPE_FILE)
nodePath := n.InternalPath()
// Set id in cache
_ = t.lookup.(*lookup.Lookup).CacheID(context.Background(), n.SpaceID, n.ID, nodePath)
if err := os.MkdirAll(filepath.Dir(nodePath), 0700); err != nil {
return errors.Wrap(err, "Decomposedfs: error creating node")
}
_, err := os.Create(nodePath)
_, err = os.Create(nodePath)
if err != nil {
return errors.Wrap(err, "Decomposedfs: error creating node")
}
attributes := n.NodeMetadata(ctx)
attributes[prefixes.IDAttr] = []byte(n.ID)
if markprocessing {
attributes[prefixes.StatusPrefix] = []byte(node.ProcessingStatus)
}
nodeMTime := time.Now()
if mtime != "" {
if err := n.SetMtimeString(ctx, mtime); err != nil {
return errors.Wrap(err, "Decomposedfs: could not set mtime")
}
} else {
now := time.Now()
if err := n.SetMtime(ctx, &now); err != nil {
return errors.Wrap(err, "Decomposedfs: could not set mtime")
nodeMTime, err = utils.MTimeToTime(mtime)
if err != nil {
return err
}
}
err = n.SetXattrsWithContext(ctx, attributes, true)
attributes[prefixes.MTimeAttr] = []byte(nodeMTime.UTC().Format(time.RFC3339Nano))
err = n.SetXattrsWithContext(ctx, attributes, false)
if err != nil {
return err
}
@@ -192,47 +269,9 @@ func (t *Tree) Move(ctx context.Context, oldNode *node.Node, newNode *node.Node)
}
}
// remove cache entry in any case to avoid inconsistencies
defer func() { _ = t.idCache.Delete(filepath.Join(oldNode.ParentPath(), oldNode.Name)) }()
// Always target the old node ID for xattr updates.
// The new node id is empty if the target does not exist
// and we need to overwrite the new one when overwriting an existing path.
// are we just renaming (parent stays the same)?
if oldNode.ParentID == newNode.ParentID {
// parentPath := t.lookup.InternalPath(oldNode.SpaceID, oldNode.ParentID)
parentPath := oldNode.ParentPath()
// rename child
err = os.Rename(
filepath.Join(parentPath, oldNode.Name),
filepath.Join(parentPath, newNode.Name),
)
if err != nil {
return errors.Wrap(err, "Decomposedfs: could not rename child")
}
// update name attribute
if err := oldNode.SetXattrString(ctx, prefixes.NameAttr, newNode.Name); err != nil {
return errors.Wrap(err, "Decomposedfs: could not set name attribute")
}
return t.Propagate(ctx, newNode, 0)
}
// we are moving the node to a new parent, any target has been removed
// bring old node to the new parent
// rename child
err = os.Rename(
filepath.Join(oldNode.ParentPath(), oldNode.Name),
filepath.Join(newNode.ParentPath(), newNode.Name),
)
if err != nil {
return errors.Wrap(err, "Decomposedfs: could not move child")
}
// update target parentid and name
attribs := node.Attributes{}
attribs.SetString(prefixes.ParentidAttr, newNode.ParentID)
@@ -253,6 +292,28 @@ func (t *Tree) Move(ctx context.Context, oldNode *node.Node, newNode *node.Node)
sizeDiff = oldNode.Blobsize
}
// rename node
err = os.Rename(
filepath.Join(oldNode.ParentPath(), oldNode.Name),
filepath.Join(newNode.ParentPath(), newNode.Name),
)
if err != nil {
return errors.Wrap(err, "Decomposedfs: could not move child")
}
// update the id cache
if newNode.ID == "" {
newNode.ID = oldNode.ID
}
_ = t.lookup.(*lookup.Lookup).CacheID(ctx, newNode.SpaceID, newNode.ID, filepath.Join(newNode.ParentPath(), newNode.Name))
// update id cache for the moved subtree
if oldNode.IsDir(ctx) {
err = t.lookup.(*lookup.Lookup).WarmupIDCache(filepath.Join(newNode.ParentPath(), newNode.Name))
if err != nil {
return err
}
}
// TODO inefficient because we might update several nodes twice, only propagate unchanged nodes?
// collect in a list, then only stat each node once
// also do this in a go routine ... webdav should check the etag async
@@ -268,18 +329,6 @@ func (t *Tree) Move(ctx context.Context, oldNode *node.Node, newNode *node.Node)
return nil
}
func readChildNodeFromLink(ctx context.Context, path string) (string, error) {
_, span := tracer.Start(ctx, "readChildNodeFromLink")
defer span.End()
link, err := os.Readlink(path)
if err != nil {
return "", err
}
nodeID := strings.TrimLeft(link, "/.")
nodeID = strings.ReplaceAll(nodeID, "/", "")
return nodeID, nil
}
// ListFolder lists the content of a folder node
func (t *Tree) ListFolder(ctx context.Context, n *node.Node) ([]*node.Node, error) {
ctx, span := tracer.Start(ctx, "ListFolder")
@@ -329,22 +378,27 @@ func (t *Tree) ListFolder(ctx context.Context, n *node.Node) ([]*node.Node, erro
// Spawn workers that'll concurrently work the queue
for i := 0; i < numWorkers; i++ {
g.Go(func() error {
var err error
// switch user if necessary
spaceGID, ok := ctx.Value(decomposedfs.CtxKeySpaceGID).(uint32)
if ok {
unscope, err := t.userMapper.ScopeUserByIds(-1, int(spaceGID))
if err != nil {
return errors.Wrap(err, "failed to scope user")
}
defer func() { _ = unscope() }()
}
for name := range work {
path := filepath.Join(dir, name)
nodeID := getNodeIDFromCache(ctx, path, t.idCache)
if nodeID == "" {
nodeID, err = readChildNodeFromLink(ctx, path)
if err != nil {
return err
}
err = storeNodeIDInCache(ctx, path, nodeID, t.idCache)
if err != nil {
return err
nodeID, err := t.lookup.MetadataBackend().Get(ctx, path, prefixes.IDAttr)
if err != nil {
if metadata.IsAttrUnset(err) {
continue
}
return err
}
child, err := node.ReadNode(ctx, t.lookup, n.SpaceID, nodeID, false, n.SpaceRoot, true)
child, err := node.ReadNode(ctx, t.lookup, n.SpaceID, string(nodeID), false, n.SpaceRoot, true)
if err != nil {
return err
}
@@ -384,7 +438,12 @@ func (t *Tree) ListFolder(ctx context.Context, n *node.Node) ([]*node.Node, erro
// Delete deletes a node in the tree by moving it to the trash
func (t *Tree) Delete(ctx context.Context, n *node.Node) (err error) {
path := filepath.Join(n.ParentPath(), n.Name)
path := n.InternalPath()
if !strings.HasPrefix(path, t.options.Root) {
return errtypes.InternalError("invalid internal path")
}
// remove entry from cache immediately to avoid inconsistencies
defer func() { _ = t.idCache.Delete(path) }()
@@ -392,19 +451,7 @@ func (t *Tree) Delete(ctx context.Context, n *node.Node) (err error) {
if deletingSharedResource != nil && deletingSharedResource.(bool) {
src := filepath.Join(n.ParentPath(), n.Name)
return os.Remove(src)
}
// get the original path
origin, err := t.lookup.Path(ctx, n, node.NoCheck)
if err != nil {
return
}
// set origin location in metadata
nodePath := n.InternalPath()
if err := n.SetXattrString(ctx, prefixes.TrashOriginAttr, origin); err != nil {
return err
return os.RemoveAll(src)
}
var sizeDiff int64
@@ -418,53 +465,11 @@ func (t *Tree) Delete(ctx context.Context, n *node.Node) (err error) {
sizeDiff = -n.Blobsize
}
deletionTime := time.Now().UTC().Format(time.RFC3339Nano)
// Prepare the trash
trashLink := filepath.Join(t.options.Root, "spaces", lookup.Pathify(n.SpaceRoot.ID, 1, 2), "trash", lookup.Pathify(n.ID, 4, 2))
if err := os.MkdirAll(filepath.Dir(trashLink), 0700); err != nil {
// Roll back changes
_ = n.RemoveXattr(ctx, prefixes.TrashOriginAttr, true)
return err
}
// FIXME can we just move the node into the trash dir? instead of adding another symlink and appending a trash timestamp?
// can we just use the mtime as the trash time?
// TODO store a trashed by userid
// first make node appear in the space trash
// parent id and name are stored as extended attributes in the node itself
err = os.Symlink("../../../../../nodes/"+lookup.Pathify(n.ID, 4, 2)+node.TrashIDDelimiter+deletionTime, trashLink)
if err != nil {
// Roll back changes
_ = n.RemoveXattr(ctx, prefixes.TrashOriginAttr, true)
return
}
// at this point we have a symlink pointing to a non existing destination, which is fine
// rename the trashed node so it is not picked up when traversing up the tree and matches the symlink
trashPath := nodePath + node.TrashIDDelimiter + deletionTime
err = os.Rename(nodePath, trashPath)
if err != nil {
// To roll back changes
// TODO remove symlink
// Roll back changes
_ = n.RemoveXattr(ctx, prefixes.TrashOriginAttr, true)
return
}
err = t.lookup.MetadataBackend().Rename(nodePath, trashPath)
if err != nil {
_ = n.RemoveXattr(ctx, prefixes.TrashOriginAttr, true)
_ = os.Rename(trashPath, nodePath)
return
}
// Remove lock file if it exists
_ = os.Remove(n.LockFilePath())
// finally remove the entry from the parent dir
if err = os.Remove(path); err != nil {
if err = os.RemoveAll(path); err != nil {
// To roll back changes
// TODO revert the rename
// TODO remove symlink
@@ -729,18 +734,37 @@ func (t *Tree) InitNewNode(ctx context.Context, n *node.Node, fsize uint64) (met
func (t *Tree) createDirNode(ctx context.Context, n *node.Node) (err error) {
ctx, span := tracer.Start(ctx, "createDirNode")
defer span.End()
idcache := t.lookup.(*lookup.Lookup).IDCache
// create a directory node
nodePath := n.InternalPath()
if err := os.MkdirAll(nodePath, 0700); err != nil {
parentPath, ok := idcache.Get(ctx, n.SpaceID, n.ParentID)
if !ok {
return errtypes.NotFound(n.ParentID)
}
path := filepath.Join(parentPath, n.Name)
// lock the meta file
unlock, err := t.lookup.MetadataBackend().Lock(path)
if err != nil {
return err
}
defer func() {
_ = unlock()
}()
if err := os.MkdirAll(path, 0700); err != nil {
return errors.Wrap(err, "Decomposedfs: error creating node")
}
_ = idcache.Set(ctx, n.SpaceID, n.ID, path)
attributes := n.NodeMetadata(ctx)
attributes[prefixes.IDAttr] = []byte(n.ID)
attributes[prefixes.TreesizeAttr] = []byte("0") // initialize as empty, TODO why bother? if it is not set we could treat it as 0?
if t.options.TreeTimeAccounting || t.options.TreeSizeAccounting {
attributes[prefixes.PropagationAttr] = []byte("1") // mark the node for propagation
}
return n.SetXattrsWithContext(ctx, attributes, true)
return n.SetXattrsWithContext(ctx, attributes, false)
}
var nodeIDRegep = regexp.MustCompile(`.*/nodes/([^.]*).*`)
@@ -814,22 +838,3 @@ func (t *Tree) readRecycleItem(ctx context.Context, spaceID, key, path string) (
return
}
func getNodeIDFromCache(ctx context.Context, path string, cache store.Store) string {
_, span := tracer.Start(ctx, "getNodeIDFromCache")
defer span.End()
recs, err := cache.Read(path)
if err == nil && len(recs) > 0 {
return string(recs[0].Value)
}
return ""
}
func storeNodeIDInCache(ctx context.Context, path string, nodeID string, cache store.Store) error {
_, span := tracer.Start(ctx, "storeNodeIDInCache")
defer span.End()
return cache.Write(&store.Record{
Key: path,
Value: []byte(nodeID),
})
}

11
vendor/github.com/cs3org/reva/v2/pkg/storage/storage.go generated vendored
View File

@@ -23,6 +23,8 @@ import (
"io"
"net/url"
tusd "github.com/tus/tusd/pkg/handler"
provider "github.com/cs3org/go-cs3apis/cs3/storage/provider/v1beta1"
registry "github.com/cs3org/go-cs3apis/cs3/storage/registry/v1beta1"
)
@@ -71,6 +73,15 @@ type FS interface {
DeleteStorageSpace(ctx context.Context, req *provider.DeleteStorageSpaceRequest) error
}
// UnscopeFunc is a function that unscopes a user
type UnscopeFunc func()
// Composable is the interface that a struct needs to implement
// to be composable, so that it can support the TUS methods
type ComposableFS interface {
UseIn(composer *tusd.StoreComposer)
}
// Registry is the interface that storage registries implement
// for discovering storage providers
type Registry interface {

2
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/aspects/aspects.go generated vendored
View File

@@ -22,6 +22,7 @@ import (
"github.com/cs3org/reva/v2/pkg/events"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/node"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/permissions"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/usermapper"
)
// Aspects holds dependencies for handling aspects of the decomposedfs
@@ -31,4 +32,5 @@ type Aspects struct {
Permissions permissions.Permissions
EventStream events.Stream
DisableVersioning bool
UserMapper usermapper.Mapper
}

18
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/decomposedfs.go generated vendored
View File

@@ -51,6 +51,7 @@ import (
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/spaceidindex"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/tree"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/upload"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/usermapper"
"github.com/cs3org/reva/v2/pkg/storage/utils/filelocks"
"github.com/cs3org/reva/v2/pkg/storage/utils/templates"
"github.com/cs3org/reva/v2/pkg/storagespace"
@@ -65,6 +66,12 @@ import (
"golang.org/x/sync/errgroup"
)
type CtxKey int
const (
CtxKeySpaceGID CtxKey = iota
)
var (
tracer trace.Tracer
@@ -105,6 +112,7 @@ type Decomposedfs struct {
tp node.Tree
o *options.Options
p permissions.Permissions
um usermapper.Mapper
chunkHandler *chunking.ChunkHandler
stream events.Stream
sessionStore SessionStore
@@ -200,19 +208,25 @@ func New(o *options.Options, aspects aspects.Aspects) (storage.FS, error) {
return nil, err
}
// set a null usermapper if we don't have one
if aspects.UserMapper == nil {
aspects.UserMapper = &usermapper.NullMapper{}
}
fs := &Decomposedfs{
tp: aspects.Tree,
lu: aspects.Lookup,
o: o,
p: aspects.Permissions,
um: aspects.UserMapper,
chunkHandler: chunking.NewChunkHandler(filepath.Join(o.Root, "uploads")),
stream: aspects.EventStream,
UserCache: ttlcache.NewCache(),
userSpaceIndex: userSpaceIndex,
groupSpaceIndex: groupSpaceIndex,
spaceTypeIndex: spaceTypeIndex,
sessionStore: upload.NewSessionStore(aspects.Lookup, aspects.Tree, o.Root, aspects.EventStream, o.AsyncFileUploads, o.Tokens, aspects.DisableVersioning),
}
fs.sessionStore = upload.NewSessionStore(fs, aspects, o.Root, o.AsyncFileUploads, o.Tokens)
if o.AsyncFileUploads {
if fs.stream == nil {
@@ -884,7 +898,7 @@ func (fs *Decomposedfs) GetMD(ctx context.Context, ref *provider.Reference, mdKe
}
}
if addSpace {
if md.Space, err = fs.storageSpaceFromNode(ctx, node, true); err != nil {
if md.Space, err = fs.StorageSpaceFromNode(ctx, node, true); err != nil {
return nil, err
}
}

23
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/grants.go generated vendored
View File

@@ -20,7 +20,6 @@ package decomposedfs
import (
"context"
"os"
"path/filepath"
"strings"
@@ -29,11 +28,11 @@ import (
ctxpkg "github.com/cs3org/reva/v2/pkg/ctx"
"github.com/cs3org/reva/v2/pkg/errtypes"
"github.com/cs3org/reva/v2/pkg/storage/utils/ace"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata/prefixes"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/node"
"github.com/cs3org/reva/v2/pkg/storagespace"
"github.com/cs3org/reva/v2/pkg/utils"
"github.com/rogpeppe/go-internal/lockedfile"
)
// DenyGrant denies access to a resource.
@@ -80,7 +79,9 @@ func (fs *Decomposedfs) AddGrant(ctx context.Context, ref *provider.Reference, g
if err != nil {
return err
}
defer unlockFunc()
defer func() {
_ = unlockFunc()
}()
if grant != nil {
return errtypes.AlreadyExists(filepath.Join(grantNode.ParentID, grantNode.Name))
@@ -176,7 +177,9 @@ func (fs *Decomposedfs) RemoveGrant(ctx context.Context, ref *provider.Reference
if err != nil {
return err
}
defer unlockFunc()
defer func() {
_ = unlockFunc()
}()
if grant == nil {
return errtypes.NotFound("grant not found")
@@ -237,7 +240,9 @@ func (fs *Decomposedfs) UpdateGrant(ctx context.Context, ref *provider.Reference
if err != nil {
return err
}
defer unlockFunc()
defer func() {
_ = unlockFunc()
}()
if grant == nil {
// grant not found
@@ -264,9 +269,7 @@ func (fs *Decomposedfs) UpdateGrant(ctx context.Context, ref *provider.Reference
}
// checks if the given grant exists and returns it. Nil grant means it doesn't exist
func (fs *Decomposedfs) loadGrant(ctx context.Context, ref *provider.Reference, g *provider.Grant) (*node.Node, func(), *provider.Grant, error) {
var unlockFunc func()
func (fs *Decomposedfs) loadGrant(ctx context.Context, ref *provider.Reference, g *provider.Grant) (*node.Node, metadata.UnlockFunc, *provider.Grant, error) {
n, err := fs.lu.NodeFromResource(ctx, ref)
if err != nil {
return nil, nil, nil, err
@@ -275,11 +278,11 @@ func (fs *Decomposedfs) loadGrant(ctx context.Context, ref *provider.Reference,
return nil, nil, nil, errtypes.NotFound(filepath.Join(n.ParentID, n.Name))
}
f, err := lockedfile.OpenFile(fs.lu.MetadataBackend().LockfilePath(n.InternalPath()), os.O_RDWR|os.O_CREATE, 0600)
// lock the metadata file
unlockFunc, err := fs.lu.MetadataBackend().Lock(n.InternalPath())
if err != nil {
return nil, nil, nil, err
}
unlockFunc = func() { f.Close() }
grants, err := n.ListGrants(ctx)
if err != nil {

9
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata/errors.go generated vendored
View File

@@ -23,6 +23,7 @@ import (
"os"
"syscall"
"github.com/cs3org/reva/v2/pkg/errtypes"
"github.com/pkg/errors"
"github.com/pkg/xattr"
)
@@ -30,6 +31,9 @@ import (
// IsNotExist checks if there is a os not exists error buried inside the xattr error,
// as we cannot just use os.IsNotExist().
func IsNotExist(err error) bool {
if _, ok := err.(errtypes.IsNotFound); ok {
return true
}
if os.IsNotExist(errors.Cause(err)) {
return true
}
@@ -59,5 +63,10 @@ func IsNotDir(err error) bool {
return serr == syscall.ENOTDIR
}
}
if xerr, ok := errors.Cause(err).(*xattr.Error); ok {
if serr, ok2 := xerr.Err.(syscall.Errno); ok2 {
return serr == syscall.ENOTDIR
}
}
return false
}

52
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/spaces.go generated vendored
View File

@@ -76,6 +76,24 @@ func (fs *Decomposedfs) CreateStorageSpace(ctx context.Context, req *provider.Cr
spaceID = reqSpaceID
}
// Check if space already exists
rootPath := ""
switch req.Type {
case _spaceTypePersonal:
if fs.o.PersonalSpacePathTemplate != "" {
rootPath = filepath.Join(fs.o.Root, templates.WithUser(u, fs.o.PersonalSpacePathTemplate))
}
default:
if fs.o.GeneralSpacePathTemplate != "" {
rootPath = filepath.Join(fs.o.Root, templates.WithSpacePropertiesAndUser(u, req.Type, req.Name, spaceID, fs.o.GeneralSpacePathTemplate))
}
}
if rootPath != "" {
if _, err := os.Stat(rootPath); err == nil {
return nil, errtypes.AlreadyExists("decomposedfs: spaces: space already exists")
}
}
description := utils.ReadPlainFromOpaque(req.Opaque, "description")
alias := utils.ReadPlainFromOpaque(req.Opaque, "spaceAlias")
if alias == "" {
@@ -97,20 +115,18 @@ func (fs *Decomposedfs) CreateStorageSpace(ctx context.Context, req *provider.Cr
// create a directory node
root.SetType(provider.ResourceType_RESOURCE_TYPE_CONTAINER)
rootPath := root.InternalPath()
switch req.Type {
case _spaceTypePersonal:
if fs.o.PersonalSpacePathTemplate != "" {
rootPath = filepath.Join(fs.o.Root, templates.WithUser(u, fs.o.PersonalSpacePathTemplate))
}
default:
if fs.o.GeneralSpacePathTemplate != "" {
rootPath = filepath.Join(fs.o.Root, templates.WithSpacePropertiesAndUser(u, req.Type, req.Name, spaceID, fs.o.GeneralSpacePathTemplate))
}
if rootPath == "" {
rootPath = root.InternalPath()
}
if err := os.MkdirAll(rootPath, 0700); err != nil {
return nil, errors.Wrap(err, "Decomposedfs: error creating node")
// set 755 permissions for the base dir
if err := os.MkdirAll(filepath.Dir(rootPath), 0755); err != nil {
return nil, errors.Wrap(err, fmt.Sprintf("Decomposedfs: error creating spaces base dir %s", filepath.Dir(rootPath)))
}
// 770 permissions for the space
if err := os.MkdirAll(rootPath, 0770); err != nil {
return nil, errors.Wrap(err, fmt.Sprintf("Decomposedfs: error creating space %s", rootPath))
}
// Store id in cache
@@ -199,7 +215,7 @@ func (fs *Decomposedfs) CreateStorageSpace(ctx context.Context, req *provider.Cr
}
}
space, err := fs.storageSpaceFromNode(ctx, root, true)
space, err := fs.StorageSpaceFromNode(ctx, root, true)
if err != nil {
return nil, err
}
@@ -289,7 +305,7 @@ func (fs *Decomposedfs) ListStorageSpaces(ctx context.Context, filter []*provide
// return empty list
return spaces, nil
}
space, err := fs.storageSpaceFromNode(ctx, n, checkNodePermissions)
space, err := fs.StorageSpaceFromNode(ctx, n, checkNodePermissions)
if err != nil {
return nil, err
}
@@ -432,7 +448,7 @@ func (fs *Decomposedfs) ListStorageSpaces(ctx context.Context, filter []*provide
continue
}
space, err := fs.storageSpaceFromNode(ctx, n, checkNodePermissions)
space, err := fs.StorageSpaceFromNode(ctx, n, checkNodePermissions)
if err != nil {
switch err.(type) {
case errtypes.IsPermissionDenied:
@@ -485,7 +501,7 @@ func (fs *Decomposedfs) ListStorageSpaces(ctx context.Context, filter []*provide
return nil, err
}
if n.Exists {
space, err := fs.storageSpaceFromNode(ctx, n, checkNodePermissions)
space, err := fs.StorageSpaceFromNode(ctx, n, checkNodePermissions)
if err != nil {
return nil, err
}
@@ -670,7 +686,7 @@ func (fs *Decomposedfs) UpdateStorageSpace(ctx context.Context, req *provider.Up
}
// send back the updated data from the storage
updatedSpace, err := fs.storageSpaceFromNode(ctx, spaceNode, false)
updatedSpace, err := fs.StorageSpaceFromNode(ctx, spaceNode, false)
if err != nil {
return nil, err
}
@@ -779,7 +795,7 @@ func (fs *Decomposedfs) linkStorageSpaceType(ctx context.Context, spaceType, spa
return fs.spaceTypeIndex.Add(spaceType, spaceID, target)
}
func (fs *Decomposedfs) storageSpaceFromNode(ctx context.Context, n *node.Node, checkPermissions bool) (*provider.StorageSpace, error) {
func (fs *Decomposedfs) StorageSpaceFromNode(ctx context.Context, n *node.Node, checkPermissions bool) (*provider.StorageSpace, error) {
user := ctxpkg.ContextMustGetUser(ctx)
if checkPermissions {
rp, err := fs.p.AssemblePermissions(ctx, n)

12
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/tree/tree.go generated vendored
View File

@@ -601,9 +601,13 @@ func (t *Tree) RestoreRecycleItemFunc(ctx context.Context, spaceid, key, trashPa
return errors.Wrap(err, "Decomposedfs: could not resolve trash root")
}
deletePath = filepath.Join(resolvedTrashRoot, trashPath)
}
if err = os.Remove(deletePath); err != nil {
logger.Error().Err(err).Str("trashItem", trashItem).Msg("error deleting trash item")
if err = os.Remove(deletePath); err != nil {
logger.Error().Err(err).Str("trashItem", trashItem).Str("deletePath", deletePath).Str("trashPath", trashPath).Msg("error deleting trash item")
}
} else {
if err = utils.RemoveItem(deletePath); err != nil {
logger.Error().Err(err).Str("trashItem", trashItem).Str("deletePath", deletePath).Str("trashPath", trashPath).Msg("error recursively deleting trash item")
}
}
var sizeDiff int64
@@ -651,7 +655,7 @@ func (t *Tree) PurgeRecycleItemFunc(ctx context.Context, spaceid, key string, pa
}
deletePath = filepath.Join(resolvedTrashRoot, path)
}
if err = os.Remove(deletePath); err != nil {
if err = utils.RemoveItem(deletePath); err != nil {
logger.Error().Err(err).Str("deletePath", deletePath).Msg("error deleting trash item")
return err
}

36
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/upload.go generated vendored
View File

@@ -24,6 +24,7 @@ import (
"os"
"path/filepath"
"strings"
"syscall"
"time"
"github.com/google/uuid"
@@ -180,6 +181,13 @@ func (fs *Decomposedfs) InitiateUpload(ctx context.Context, ref *provider.Refere
session.SetStorageValue("SpaceRoot", n.SpaceRoot.ID) // TODO SpaceRoot -> SpaceID
session.SetStorageValue("SpaceOwnerOrManager", n.SpaceOwnerOrManager(ctx).GetOpaqueId()) // TODO needed for what?
// remember the gid of the space
fi, err := os.Stat(n.SpaceRoot.InternalPath())
if err != nil {
return nil, err
}
session.SetStorageValue("SpaceGid", fmt.Sprintf("%d", (fi.Sys().(*syscall.Stat_t).Gid)))
iid, _ := ctxpkg.ContextGetInitiator(ctx)
session.SetMetadata("initiatorid", iid)
@@ -298,18 +306,20 @@ func (fs *Decomposedfs) InitiateUpload(ctx context.Context, ref *provider.Refere
session.SetStorageValue("LogLevel", log.GetLevel().String())
log.Debug().Interface("session", session).Msg("Decomposedfs: built session info")
// Create binary file in the upload folder with no content
// It will be used when determining the current offset of an upload
err = session.TouchBin()
err = fs.um.RunInBaseScope(func() error {
// Create binary file in the upload folder with no content
// It will be used when determining the current offset of an upload
err := session.TouchBin()
if err != nil {
return err
}
return session.Persist(ctx)
})
if err != nil {
return nil, err
}
err = session.Persist(ctx)
if err != nil {
return nil, err
}
metrics.UploadSessionsInitiated.Inc()
if uploadLength == 0 {
@@ -345,7 +355,13 @@ func (fs *Decomposedfs) NewUpload(ctx context.Context, info tusd.FileInfo) (tusd
// GetUpload returns the Upload for the given upload id
func (fs *Decomposedfs) GetUpload(ctx context.Context, id string) (tusd.Upload, error) {
return fs.sessionStore.Get(ctx, id)
var ul tusd.Upload
var err error
_ = fs.um.RunInBaseScope(func() error {
ul, err = fs.sessionStore.Get(ctx, id)
return nil
})
return ul, err
}
// ListUploadSessions returns the upload sessions for the given filter

39
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/upload/store.go generated vendored
View File

@@ -34,10 +34,13 @@ import (
"github.com/cs3org/reva/v2/pkg/appctx"
"github.com/cs3org/reva/v2/pkg/errtypes"
"github.com/cs3org/reva/v2/pkg/events"
"github.com/cs3org/reva/v2/pkg/storage"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/aspects"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/metadata/prefixes"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/node"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/options"
"github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/usermapper"
"github.com/google/uuid"
"github.com/pkg/errors"
"github.com/rogpeppe/go-internal/lockedfile"
@@ -53,8 +56,10 @@ type PermissionsChecker interface {
// OcisStore manages upload sessions
type OcisStore struct {
fs storage.FS
lu node.PathLookup
tp node.Tree
um usermapper.Mapper
root string
pub events.Publisher
async bool
@@ -63,15 +68,17 @@ type OcisStore struct {
}
// NewSessionStore returns a new OcisStore
func NewSessionStore(lu node.PathLookup, tp node.Tree, root string, pub events.Publisher, async bool, tknopts options.TokenOptions, disableVersioning bool) *OcisStore {
func NewSessionStore(fs storage.FS, aspects aspects.Aspects, root string, async bool, tknopts options.TokenOptions) *OcisStore {
return &OcisStore{
lu: lu,
tp: tp,
fs: fs,
lu: aspects.Lookup,
tp: aspects.Tree,
root: root,
pub: pub,
pub: aspects.EventStream,
async: async,
tknopts: tknopts,
disableVersioning: disableVersioning,
disableVersioning: aspects.DisableVersioning,
um: aspects.UserMapper,
}
}
@@ -230,7 +237,7 @@ func (store OcisStore) CreateNodeForUpload(session *OcisSession, initAttrs node.
unlock, err = store.tp.InitNewNode(ctx, n, uint64(session.Size()))
if err != nil {
appctx.GetLogger(ctx).Error().Err(err).Msg("failed to init new node")
appctx.GetLogger(ctx).Error().Str("path", n.InternalPath()).Err(err).Msg("failed to init new node")
}
session.info.MetaData["sizeDiff"] = strconv.FormatInt(session.Size(), 10)
}
@@ -270,7 +277,10 @@ func (store OcisStore) CreateNodeForUpload(session *OcisSession, initAttrs node.
return nil, errors.Wrap(err, "Decomposedfs: could not write metadata")
}
if err := session.Persist(ctx); err != nil {
err = store.um.RunInBaseScope(func() error {
return session.Persist(ctx)
})
if err != nil {
return nil, err
}
@@ -340,9 +350,8 @@ func (store OcisStore) updateExistingNode(ctx context.Context, session *OcisSess
}
}
versionPath := n.InternalPath()
if !store.disableVersioning {
versionPath = session.store.lu.InternalPath(spaceID, n.ID+node.RevisionIDDelimiter+oldNodeMtime.UTC().Format(time.RFC3339Nano))
versionPath := session.store.lu.InternalPath(spaceID, n.ID+node.RevisionIDDelimiter+oldNodeMtime.UTC().Format(time.RFC3339Nano))
// create version node
if _, err := os.Create(versionPath); err != nil {
@@ -359,14 +368,14 @@ func (store OcisStore) updateExistingNode(ctx context.Context, session *OcisSess
}, f, true); err != nil {
return unlock, err
}
session.info.MetaData["versionsPath"] = versionPath
// keep mtime from previous version
if err := os.Chtimes(session.info.MetaData["versionsPath"], oldNodeMtime, oldNodeMtime); err != nil {
return unlock, errtypes.InternalError(fmt.Sprintf("failed to change mtime of version node: %s", err))
}
}
session.info.MetaData["sizeDiff"] = strconv.FormatInt((int64(fsize) - old.Blobsize), 10)
session.info.MetaData["versionsPath"] = versionPath
// keep mtime from previous version
if err := os.Chtimes(session.info.MetaData["versionsPath"], oldNodeMtime, oldNodeMtime); err != nil {
return unlock, errtypes.InternalError(fmt.Sprintf("failed to change mtime of version node: %s", err))
}
session.info.MetaData["sizeDiff"] = strconv.FormatInt((int64(fsize) - old.Blobsize), 10)
return unlock, nil
}

28
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/upload/upload.go generated vendored
View File

@@ -26,6 +26,7 @@ import (
"io"
"io/fs"
"os"
"strconv"
"strings"
"time"
@@ -143,6 +144,22 @@ func (session *OcisSession) FinishUpload(ctx context.Context) error {
prefixes.ChecksumPrefix + "adler32": adler32h.Sum(nil),
}
// At this point we scope by the space to create the final file in the final location
if session.store.um != nil && session.info.Storage["SpaceGid"] != "" {
gid, err := strconv.Atoi(session.info.Storage["SpaceGid"])
if err != nil {
return errors.Wrap(err, "failed to parse space gid")
}
unscope, err := session.store.um.ScopeUserByIds(-1, gid)
if err != nil {
return errors.Wrap(err, "failed to scope user")
}
if unscope != nil {
defer func() { _ = unscope() }()
}
}
n, err := session.store.CreateNodeForUpload(session, attrs)
if err != nil {
session.store.Cleanup(ctx, session, true, false, false)
@@ -198,7 +215,9 @@ func (session *OcisSession) Terminate(_ context.Context) error {
func (session *OcisSession) DeclareLength(ctx context.Context, length int64) error {
session.info.Size = length
session.info.SizeIsDeferred = false
return session.Persist(session.Context(ctx))
return session.store.um.RunInBaseScope(func() error {
return session.Persist(session.Context(ctx))
})
}
// ConcatUploads concatenates multiple uploads
@@ -231,7 +250,8 @@ func (session *OcisSession) Finalize() (err error) {
ctx, span := tracer.Start(session.Context(context.Background()), "Finalize")
defer span.End()
revisionNode := &node.Node{SpaceID: session.SpaceID(), BlobID: session.ID(), Blobsize: session.Size()}
revisionNode := node.New(session.SpaceID(), session.NodeID(), "", "", session.Size(), session.ID(),
provider.ResourceType_RESOURCE_TYPE_FILE, session.SpaceOwner(), session.store.lu)
// upload the data to the blobstore
_, subspan := tracer.Start(ctx, "WriteBlob")
@@ -270,9 +290,9 @@ func (session *OcisSession) Cleanup(revertNodeMetadata, cleanBin, cleanInfo bool
if revertNodeMetadata {
n, err := session.Node(ctx)
if err != nil {
appctx.GetLogger(ctx).Error().Err(err).Str("spaceid", n.SpaceID).Str("nodeid", n.ID).Str("uploadid", session.ID()).Msg("reading node for session failed")
appctx.GetLogger(ctx).Error().Err(err).Str("sessionid", session.ID()).Msg("reading node for session failed")
}
if session.NodeExists() {
if session.NodeExists() && session.info.MetaData["versionsPath"] != "" {
p := session.info.MetaData["versionsPath"]
if err := session.store.lu.CopyMetadata(ctx, p, n.InternalPath(), func(attributeName string, value []byte) (newValue []byte, copy bool) {
return value, strings.HasPrefix(attributeName, prefixes.ChecksumPrefix) ||

56
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/usermapper/usermapper.go generated vendored Normal file
View File

@@ -0,0 +1,56 @@
// Copyright 2018-2021 CERN
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// In applying this license, CERN does not waive the privileges and immunities
// granted to it by virtue of its status as an Intergovernmental Organization
// or submit itself to any jurisdiction.
package usermapper
import (
"context"
)
// Mapper is the interface that wraps the basic mapping methods
type Mapper interface {
RunInBaseScope(f func() error) error
ScopeBase() (func() error, error)
ScopeUser(ctx context.Context) (func() error, error)
ScopeUserByIds(uid, gid int) (func() error, error)
}
// UnscopeFunc is a function that unscopes the current user
type UnscopeFunc func() error
// NullMapper is a user mapper that does nothing
type NullMapper struct{}
// RunInBaseScope runs the given function in the scope of the base user
func (nm *NullMapper) RunInBaseScope(f func() error) error {
return f()
}
// ScopeBase returns to the base uid and gid returning a function that can be used to restore the previous scope
func (nm *NullMapper) ScopeBase() (func() error, error) {
return func() error { return nil }, nil
}
// ScopeUser returns to the base uid and gid returning a function that can be used to restore the previous scope
func (nm *NullMapper) ScopeUser(ctx context.Context) (func() error, error) {
return func() error { return nil }, nil
}
func (nm *NullMapper) ScopeUserByIds(uid, gid int) (func() error, error) {
return func() error { return nil }, nil
}

131
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/decomposedfs/usermapper/usermapper_linux.go generated vendored Normal file
View File

@@ -0,0 +1,131 @@
// Copyright 2018-2021 CERN
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
// In applying this license, CERN does not waive the privileges and immunities
// granted to it by virtue of its status as an Intergovernmental Organization
// or submit itself to any jurisdiction.
package usermapper
import (
"context"
"fmt"
"os/user"
"runtime"
"strconv"
"golang.org/x/sys/unix"
revactx "github.com/cs3org/reva/v2/pkg/ctx"
)
// UnixMapper is a user mapper that maps users to unix uids and gids
type UnixMapper struct {
baseUid int
baseGid int
}
// New returns a new user mapper
func NewUnixMapper() *UnixMapper {
baseUid, _ := unix.SetfsuidRetUid(-1)
baseGid, _ := unix.SetfsgidRetGid(-1)
return &UnixMapper{
baseUid: baseUid,
baseGid: baseGid,
}
}
// RunInUserScope runs the given function in the scope of the base user
func (um *UnixMapper) RunInBaseScope(f func() error) error {
unscope, err := um.ScopeBase()
if err != nil {
return err
}
defer func() { _ = unscope() }()
return f()
}
// ScopeBase returns to the base uid and gid returning a function that can be used to restore the previous scope
func (um *UnixMapper) ScopeBase() (func() error, error) {
return um.ScopeUserByIds(-1, um.baseGid)
}
// ScopeUser returns to the base uid and gid returning a function that can be used to restore the previous scope
func (um *UnixMapper) ScopeUser(ctx context.Context) (func() error, error) {
u := revactx.ContextMustGetUser(ctx)
uid, gid, err := um.mapUser(u.Username)
if err != nil {
return nil, err
}
return um.ScopeUserByIds(uid, gid)
}
// ScopeUserByIds scopes the current user to the given uid and gid returning a function that can be used to restore the previous scope
func (um *UnixMapper) ScopeUserByIds(uid, gid int) (func() error, error) {
runtime.LockOSThread() // Lock this Goroutine to the current OS thread
var err error
var prevUid int
var prevGid int
if uid >= 0 {
prevUid, err = unix.SetfsuidRetUid(uid)
if err != nil {
return nil, err
}
if testUid, _ := unix.SetfsuidRetUid(-1); testUid != uid {
return nil, fmt.Errorf("failed to setfsuid to %d", uid)
}
}
if gid >= 0 {
prevGid, err = unix.SetfsgidRetGid(gid)
if err != nil {
return nil, err
}
if testGid, _ := unix.SetfsgidRetGid(-1); testGid != gid {
return nil, fmt.Errorf("failed to setfsgid to %d", gid)
}
}
return func() error {
if uid >= 0 {
_ = unix.Setfsuid(prevUid)
}
if gid >= 0 {
_ = unix.Setfsgid(prevGid)
}
runtime.UnlockOSThread()
return nil
}, nil
}
func (u *UnixMapper) mapUser(username string) (int, int, error) {
userDetails, err := user.Lookup(username)
if err != nil {
return 0, 0, err
}
uid, err := strconv.Atoi(userDetails.Uid)
if err != nil {
return 0, 0, err
}
gid, err := strconv.Atoi(userDetails.Gid)
if err != nil {
return 0, 0, err
}
return uid, gid, nil
}

1090
vendor/github.com/cs3org/reva/v2/pkg/storage/utils/middleware/middleware.go generated vendored Normal file
View File

File diff suppressed because it is too large Load Diff

2
vendor/github.com/klauspost/compress/.gitattributes generated vendored Normal file
View File

@@ -0,0 +1,2 @@
* -text
*.bin -text -diff

32
vendor/github.com/klauspost/compress/.gitignore generated vendored Normal file
View File

@@ -0,0 +1,32 @@
# Compiled Object files, Static and Dynamic libs (Shared Objects)
*.o
*.a
*.so
# Folders
_obj
_test
# Architecture specific extensions/prefixes
*.[568vq]
[568vq].out
*.cgo1.go
*.cgo2.c
_cgo_defun.c
_cgo_gotypes.go
_cgo_export.*
_testmain.go
*.exe
*.test
*.prof
/s2/cmd/_s2sx/sfx-exe
# Linux perf files
perf.data
perf.data.old
# gdb history
.gdb_history

123
vendor/github.com/klauspost/compress/.goreleaser.yml generated vendored Normal file
View File

@@ -0,0 +1,123 @@
# This is an example goreleaser.yaml file with some sane defaults.
# Make sure to check the documentation at http://goreleaser.com
before:
hooks:
- ./gen.sh
builds:
-
id: "s2c"
binary: s2c
main: ./s2/cmd/s2c/main.go
flags:
- -trimpath
env:
- CGO_ENABLED=0
goos:
- aix
- linux
- freebsd
- netbsd
- windows
- darwin
goarch:
- 386
- amd64
- arm
- arm64
- ppc64
- ppc64le
- mips64
- mips64le
goarm:
- 7
-
id: "s2d"
binary: s2d
main: ./s2/cmd/s2d/main.go
flags:
- -trimpath
env:
- CGO_ENABLED=0
goos:
- aix
- linux
- freebsd
- netbsd
- windows
- darwin
goarch:
- 386
- amd64
- arm
- arm64
- ppc64
- ppc64le
- mips64
- mips64le
goarm:
- 7
-
id: "s2sx"
binary: s2sx
main: ./s2/cmd/_s2sx/main.go
flags:
- -modfile=s2sx.mod
- -trimpath
env:
- CGO_ENABLED=0
goos:
- aix
- linux
- freebsd
- netbsd
- windows
- darwin
goarch:
- 386
- amd64
- arm
- arm64
- ppc64
- ppc64le
- mips64
- mips64le
goarm:
- 7
archives:
-
id: s2-binaries
name_template: "s2-{{ .Os }}_{{ .Arch }}{{ if .Arm }}v{{ .Arm }}{{ end }}"
format_overrides:
- goos: windows
format: zip
files:
- unpack/*
- s2/LICENSE
- s2/README.md
checksum:
name_template: 'checksums.txt'
snapshot:
name_template: "{{ .Tag }}-next"
changelog:
sort: asc
filters:
exclude:
- '^doc:'
- '^docs:'
- '^test:'
- '^tests:'
- '^Update\sREADME.md'
nfpms:
-
file_name_template: "s2_package__{{ .Os }}_{{ .Arch }}{{ if .Arm }}v{{ .Arm }}{{ end }}"
vendor: Klaus Post
homepage: https://github.com/klauspost/compress
maintainer: Klaus Post <klauspost@gmail.com>
description: S2 Compression Tool
license: BSD 3-Clause
formats:
- deb
- rpm

700
vendor/github.com/klauspost/compress/README.md generated vendored Normal file
View File

@@ -0,0 +1,700 @@
# compress
This package provides various compression algorithms.
* [zstandard](https://github.com/klauspost/compress/tree/master/zstd#zstd) compression and decompression in pure Go.
* [S2](https://github.com/klauspost/compress/tree/master/s2#s2-compression) is a high performance replacement for Snappy.
* Optimized [deflate](https://godoc.org/github.com/klauspost/compress/flate) packages which can be used as a dropin replacement for [gzip](https://godoc.org/github.com/klauspost/compress/gzip), [zip](https://godoc.org/github.com/klauspost/compress/zip) and [zlib](https://godoc.org/github.com/klauspost/compress/zlib).
* [snappy](https://github.com/klauspost/compress/tree/master/snappy) is a drop-in replacement for `github.com/golang/snappy` offering better compression and concurrent streams.
* [huff0](https://github.com/klauspost/compress/tree/master/huff0) and [FSE](https://github.com/klauspost/compress/tree/master/fse) implementations for raw entropy encoding.
* [gzhttp](https://github.com/klauspost/compress/tree/master/gzhttp) Provides client and server wrappers for handling gzipped requests efficiently.
* [pgzip](https://github.com/klauspost/pgzip) is a separate package that provides a very fast parallel gzip implementation.
[![Go Reference](https://pkg.go.dev/badge/klauspost/compress.svg)](https://pkg.go.dev/github.com/klauspost/compress?tab=subdirectories)
[![Go](https://github.com/klauspost/compress/actions/workflows/go.yml/badge.svg)](https://github.com/klauspost/compress/actions/workflows/go.yml)
[![Sourcegraph Badge](https://sourcegraph.com/github.com/klauspost/compress/-/badge.svg)](https://sourcegraph.com/github.com/klauspost/compress?badge)
# changelog
* Feb 5th, 2024 - [1.17.6](https://github.com/klauspost/compress/releases/tag/v1.17.6)
* zstd: Fix incorrect repeat coding in best mode https://github.com/klauspost/compress/pull/923
* s2: Fix DecodeConcurrent deadlock on errors https://github.com/klauspost/compress/pull/925
* Jan 26th, 2024 - [v1.17.5](https://github.com/klauspost/compress/releases/tag/v1.17.5)
* flate: Fix reset with dictionary on custom window encodes https://github.com/klauspost/compress/pull/912
* zstd: Add Frame header encoding and stripping https://github.com/klauspost/compress/pull/908
* zstd: Limit better/best default window to 8MB https://github.com/klauspost/compress/pull/913
* zstd: Speed improvements by @greatroar in https://github.com/klauspost/compress/pull/896 https://github.com/klauspost/compress/pull/910
* s2: Fix callbacks for skippable blocks and disallow 0xfe (Padding) by @Jille in https://github.com/klauspost/compress/pull/916 https://github.com/klauspost/compress/pull/917
https://github.com/klauspost/compress/pull/919 https://github.com/klauspost/compress/pull/918
* Dec 1st, 2023 - [v1.17.4](https://github.com/klauspost/compress/releases/tag/v1.17.4)
* huff0: Speed up symbol counting by @greatroar in https://github.com/klauspost/compress/pull/887
* huff0: Remove byteReader by @greatroar in https://github.com/klauspost/compress/pull/886
* gzhttp: Allow overriding decompression on transport https://github.com/klauspost/compress/pull/892
* gzhttp: Clamp compression level https://github.com/klauspost/compress/pull/890
* gzip: Error out if reserved bits are set https://github.com/klauspost/compress/pull/891
* Nov 15th, 2023 - [v1.17.3](https://github.com/klauspost/compress/releases/tag/v1.17.3)
* fse: Fix max header size https://github.com/klauspost/compress/pull/881
* zstd: Improve better/best compression https://github.com/klauspost/compress/pull/877
* gzhttp: Fix missing content type on Close https://github.com/klauspost/compress/pull/883
* Oct 22nd, 2023 - [v1.17.2](https://github.com/klauspost/compress/releases/tag/v1.17.2)
* zstd: Fix rare *CORRUPTION* output in "best" mode. See https://github.com/klauspost/compress/pull/876
* Oct 14th, 2023 - [v1.17.1](https://github.com/klauspost/compress/releases/tag/v1.17.1)
* s2: Fix S2 "best" dictionary wrong encoding by @klauspost in https://github.com/klauspost/compress/pull/871
* flate: Reduce allocations in decompressor and minor code improvements by @fakefloordiv in https://github.com/klauspost/compress/pull/869
* s2: Fix EstimateBlockSize on 6&7 length input by @klauspost in https://github.com/klauspost/compress/pull/867
* Sept 19th, 2023 - [v1.17.0](https://github.com/klauspost/compress/releases/tag/v1.17.0)
* Add experimental dictionary builder https://github.com/klauspost/compress/pull/853
* Add xerial snappy read/writer https://github.com/klauspost/compress/pull/838
* flate: Add limited window compression https://github.com/klauspost/compress/pull/843
* s2: Do 2 overlapping match checks https://github.com/klauspost/compress/pull/839
* flate: Add amd64 assembly matchlen https://github.com/klauspost/compress/pull/837
* gzip: Copy bufio.Reader on Reset by @thatguystone in https://github.com/klauspost/compress/pull/860
<details>
<summary>See changes to v1.16.x</summary>
* July 1st, 2023 - [v1.16.7](https://github.com/klauspost/compress/releases/tag/v1.16.7)
* zstd: Fix default level first dictionary encode https://github.com/klauspost/compress/pull/829
* s2: add GetBufferCapacity() method by @GiedriusS in https://github.com/klauspost/compress/pull/832
* June 13, 2023 - [v1.16.6](https://github.com/klauspost/compress/releases/tag/v1.16.6)
* zstd: correctly ignore WithEncoderPadding(1) by @ianlancetaylor in https://github.com/klauspost/compress/pull/806
* zstd: Add amd64 match length assembly https://github.com/klauspost/compress/pull/824
* gzhttp: Handle informational headers by @rtribotte in https://github.com/klauspost/compress/pull/815
* s2: Improve Better compression slightly https://github.com/klauspost/compress/pull/663
* Apr 16, 2023 - [v1.16.5](https://github.com/klauspost/compress/releases/tag/v1.16.5)
* zstd: readByte needs to use io.ReadFull by @jnoxon in https://github.com/klauspost/compress/pull/802
* gzip: Fix WriterTo after initial read https://github.com/klauspost/compress/pull/804
* Apr 5, 2023 - [v1.16.4](https://github.com/klauspost/compress/releases/tag/v1.16.4)
* zstd: Improve zstd best efficiency by @greatroar and @klauspost in https://github.com/klauspost/compress/pull/784
* zstd: Respect WithAllLitEntropyCompression https://github.com/klauspost/compress/pull/792
* zstd: Fix amd64 not always detecting corrupt data https://github.com/klauspost/compress/pull/785
* zstd: Various minor improvements by @greatroar in https://github.com/klauspost/compress/pull/788 https://github.com/klauspost/compress/pull/794 https://github.com/klauspost/compress/pull/795
* s2: Fix huge block overflow https://github.com/klauspost/compress/pull/779
* s2: Allow CustomEncoder fallback https://github.com/klauspost/compress/pull/780
* gzhttp: Suppport ResponseWriter Unwrap() in gzhttp handler by @jgimenez in https://github.com/klauspost/compress/pull/799
* Mar 13, 2023 - [v1.16.1](https://github.com/klauspost/compress/releases/tag/v1.16.1)
* zstd: Speed up + improve best encoder by @greatroar in https://github.com/klauspost/compress/pull/776
* gzhttp: Add optional [BREACH mitigation](https://github.com/klauspost/compress/tree/master/gzhttp#breach-mitigation). https://github.com/klauspost/compress/pull/762 https://github.com/klauspost/compress/pull/768 https://github.com/klauspost/compress/pull/769 https://github.com/klauspost/compress/pull/770 https://github.com/klauspost/compress/pull/767
* s2: Add Intel LZ4s converter https://github.com/klauspost/compress/pull/766
* zstd: Minor bug fixes https://github.com/klauspost/compress/pull/771 https://github.com/klauspost/compress/pull/772 https://github.com/klauspost/compress/pull/773
* huff0: Speed up compress1xDo by @greatroar in https://github.com/klauspost/compress/pull/774
* Feb 26, 2023 - [v1.16.0](https://github.com/klauspost/compress/releases/tag/v1.16.0)
* s2: Add [Dictionary](https://github.com/klauspost/compress/tree/master/s2#dictionaries) support. https://github.com/klauspost/compress/pull/685
* s2: Add Compression Size Estimate. https://github.com/klauspost/compress/pull/752
* s2: Add support for custom stream encoder. https://github.com/klauspost/compress/pull/755
* s2: Add LZ4 block converter. https://github.com/klauspost/compress/pull/748
* s2: Support io.ReaderAt in ReadSeeker. https://github.com/klauspost/compress/pull/747
* s2c/s2sx: Use concurrent decoding. https://github.com/klauspost/compress/pull/746
</details>
<details>
<summary>See changes to v1.15.x</summary>
* Jan 21st, 2023 (v1.15.15)
* deflate: Improve level 7-9 by @klauspost in https://github.com/klauspost/compress/pull/739
* zstd: Add delta encoding support by @greatroar in https://github.com/klauspost/compress/pull/728
* zstd: Various speed improvements by @greatroar https://github.com/klauspost/compress/pull/741 https://github.com/klauspost/compress/pull/734 https://github.com/klauspost/compress/pull/736 https://github.com/klauspost/compress/pull/744 https://github.com/klauspost/compress/pull/743 https://github.com/klauspost/compress/pull/745
* gzhttp: Add SuffixETag() and DropETag() options to prevent ETag collisions on compressed responses by @willbicks in https://github.com/klauspost/compress/pull/740
* Jan 3rd, 2023 (v1.15.14)
* flate: Improve speed in big stateless blocks https://github.com/klauspost/compress/pull/718
* zstd: Minor speed tweaks by @greatroar in https://github.com/klauspost/compress/pull/716 https://github.com/klauspost/compress/pull/720
* export NoGzipResponseWriter for custom ResponseWriter wrappers by @harshavardhana in https://github.com/klauspost/compress/pull/722
* s2: Add example for indexing and existing stream https://github.com/klauspost/compress/pull/723
* Dec 11, 2022 (v1.15.13)
* zstd: Add [MaxEncodedSize](https://pkg.go.dev/github.com/klauspost/compress@v1.15.13/zstd#Encoder.MaxEncodedSize) to encoder https://github.com/klauspost/compress/pull/691
* zstd: Various tweaks and improvements https://github.com/klauspost/compress/pull/693 https://github.com/klauspost/compress/pull/695 https://github.com/klauspost/compress/pull/696 https://github.com/klauspost/compress/pull/701 https://github.com/klauspost/compress/pull/702 https://github.com/klauspost/compress/pull/703 https://github.com/klauspost/compress/pull/704 https://github.com/klauspost/compress/pull/705 https://github.com/klauspost/compress/pull/706 https://github.com/klauspost/compress/pull/707 https://github.com/klauspost/compress/pull/708
* Oct 26, 2022 (v1.15.12)
* zstd: Tweak decoder allocs. https://github.com/klauspost/compress/pull/680
* gzhttp: Always delete `HeaderNoCompression` https://github.com/klauspost/compress/pull/683
* Sept 26, 2022 (v1.15.11)
* flate: Improve level 1-3 compression https://github.com/klauspost/compress/pull/678
* zstd: Improve "best" compression by @nightwolfz in https://github.com/klauspost/compress/pull/677
* zstd: Fix+reduce decompression allocations https://github.com/klauspost/compress/pull/668
* zstd: Fix non-effective noescape tag https://github.com/klauspost/compress/pull/667
* Sept 16, 2022 (v1.15.10)
* zstd: Add [WithDecodeAllCapLimit](https://pkg.go.dev/github.com/klauspost/compress@v1.15.10/zstd#WithDecodeAllCapLimit) https://github.com/klauspost/compress/pull/649
* Add Go 1.19 - deprecate Go 1.16 https://github.com/klauspost/compress/pull/651
* flate: Improve level 5+6 compression https://github.com/klauspost/compress/pull/656
* zstd: Improve "better" compresssion https://github.com/klauspost/compress/pull/657
* s2: Improve "best" compression https://github.com/klauspost/compress/pull/658
* s2: Improve "better" compression. https://github.com/klauspost/compress/pull/635
* s2: Slightly faster non-assembly decompression https://github.com/klauspost/compress/pull/646
* Use arrays for constant size copies https://github.com/klauspost/compress/pull/659
* July 21, 2022 (v1.15.9)
* zstd: Fix decoder crash on amd64 (no BMI) on invalid input https://github.com/klauspost/compress/pull/645
* zstd: Disable decoder extended memory copies (amd64) due to possible crashes https://github.com/klauspost/compress/pull/644
* zstd: Allow single segments up to "max decoded size" by @klauspost in https://github.com/klauspost/compress/pull/643
* July 13, 2022 (v1.15.8)
* gzip: fix stack exhaustion bug in Reader.Read https://github.com/klauspost/compress/pull/641
* s2: Add Index header trim/restore https://github.com/klauspost/compress/pull/638
* zstd: Optimize seqdeq amd64 asm by @greatroar in https://github.com/klauspost/compress/pull/636
* zstd: Improve decoder memcopy https://github.com/klauspost/compress/pull/637
* huff0: Pass a single bitReader pointer to asm by @greatroar in https://github.com/klauspost/compress/pull/634
* zstd: Branchless getBits for amd64 w/o BMI2 by @greatroar in https://github.com/klauspost/compress/pull/640
* gzhttp: Remove header before writing https://github.com/klauspost/compress/pull/639
* June 29, 2022 (v1.15.7)
* s2: Fix absolute forward seeks https://github.com/klauspost/compress/pull/633
* zip: Merge upstream https://github.com/klauspost/compress/pull/631
* zip: Re-add zip64 fix https://github.com/klauspost/compress/pull/624
* zstd: translate fseDecoder.buildDtable into asm by @WojciechMula in https://github.com/klauspost/compress/pull/598
* flate: Faster histograms https://github.com/klauspost/compress/pull/620
* deflate: Use compound hcode https://github.com/klauspost/compress/pull/622
* June 3, 2022 (v1.15.6)
* s2: Improve coding for long, close matches https://github.com/klauspost/compress/pull/613
* s2c: Add Snappy/S2 stream recompression https://github.com/klauspost/compress/pull/611
* zstd: Always use configured block size https://github.com/klauspost/compress/pull/605
* zstd: Fix incorrect hash table placement for dict encoding in default https://github.com/klauspost/compress/pull/606
* zstd: Apply default config to ZipDecompressor without options https://github.com/klauspost/compress/pull/608
* gzhttp: Exclude more common archive formats https://github.com/klauspost/compress/pull/612
* s2: Add ReaderIgnoreCRC https://github.com/klauspost/compress/pull/609
* s2: Remove sanity load on index creation https://github.com/klauspost/compress/pull/607
* snappy: Use dedicated function for scoring https://github.com/klauspost/compress/pull/614
* s2c+s2d: Use official snappy framed extension https://github.com/klauspost/compress/pull/610
* May 25, 2022 (v1.15.5)
* s2: Add concurrent stream decompression https://github.com/klauspost/compress/pull/602
* s2: Fix final emit oob read crash on amd64 https://github.com/klauspost/compress/pull/601
* huff0: asm implementation of Decompress1X by @WojciechMula https://github.com/klauspost/compress/pull/596
* zstd: Use 1 less goroutine for stream decoding https://github.com/klauspost/compress/pull/588
* zstd: Copy literal in 16 byte blocks when possible https://github.com/klauspost/compress/pull/592
* zstd: Speed up when WithDecoderLowmem(false) https://github.com/klauspost/compress/pull/599
* zstd: faster next state update in BMI2 version of decode by @WojciechMula in https://github.com/klauspost/compress/pull/593
* huff0: Do not check max size when reading table. https://github.com/klauspost/compress/pull/586
* flate: Inplace hashing for level 7-9 by @klauspost in https://github.com/klauspost/compress/pull/590
* May 11, 2022 (v1.15.4)
* huff0: decompress directly into output by @WojciechMula in [#577](https://github.com/klauspost/compress/pull/577)
* inflate: Keep dict on stack [#581](https://github.com/klauspost/compress/pull/581)
* zstd: Faster decoding memcopy in asm [#583](https://github.com/klauspost/compress/pull/583)
* zstd: Fix ignored crc [#580](https://github.com/klauspost/compress/pull/580)
* May 5, 2022 (v1.15.3)
* zstd: Allow to ignore checksum checking by @WojciechMula [#572](https://github.com/klauspost/compress/pull/572)
* s2: Fix incorrect seek for io.SeekEnd in [#575](https://github.com/klauspost/compress/pull/575)
* Apr 26, 2022 (v1.15.2)
* zstd: Add x86-64 assembly for decompression on streams and blocks. Contributed by [@WojciechMula](https://github.com/WojciechMula). Typically 2x faster. [#528](https://github.com/klauspost/compress/pull/528) [#531](https://github.com/klauspost/compress/pull/531) [#545](https://github.com/klauspost/compress/pull/545) [#537](https://github.com/klauspost/compress/pull/537)
* zstd: Add options to ZipDecompressor and fixes [#539](https://github.com/klauspost/compress/pull/539)
* s2: Use sorted search for index [#555](https://github.com/klauspost/compress/pull/555)
* Minimum version is Go 1.16, added CI test on 1.18.
* Mar 11, 2022 (v1.15.1)
* huff0: Add x86 assembly of Decode4X by @WojciechMula in [#512](https://github.com/klauspost/compress/pull/512)
* zstd: Reuse zip decoders in [#514](https://github.com/klauspost/compress/pull/514)
* zstd: Detect extra block data and report as corrupted in [#520](https://github.com/klauspost/compress/pull/520)
* zstd: Handle zero sized frame content size stricter in [#521](https://github.com/klauspost/compress/pull/521)
* zstd: Add stricter block size checks in [#523](https://github.com/klauspost/compress/pull/523)
* Mar 3, 2022 (v1.15.0)
* zstd: Refactor decoder by @klauspost in [#498](https://github.com/klauspost/compress/pull/498)
* zstd: Add stream encoding without goroutines by @klauspost in [#505](https://github.com/klauspost/compress/pull/505)
* huff0: Prevent single blocks exceeding 16 bits by @klauspost in[#507](https://github.com/klauspost/compress/pull/507)
* flate: Inline literal emission by @klauspost in [#509](https://github.com/klauspost/compress/pull/509)
* gzhttp: Add zstd to transport by @klauspost in [#400](https://github.com/klauspost/compress/pull/400)
* gzhttp: Make content-type optional by @klauspost in [#510](https://github.com/klauspost/compress/pull/510)
Both compression and decompression now supports "synchronous" stream operations. This means that whenever "concurrency" is set to 1, they will operate without spawning goroutines.
Stream decompression is now faster on asynchronous, since the goroutine allocation much more effectively splits the workload. On typical streams this will typically use 2 cores fully for decompression. When a stream has finished decoding no goroutines will be left over, so decoders can now safely be pooled and still be garbage collected.
While the release has been extensively tested, it is recommended to testing when upgrading.
</details>
<details>
<summary>See changes to v1.14.x</summary>
* Feb 22, 2022 (v1.14.4)
* flate: Fix rare huffman only (-2) corruption. [#503](https://github.com/klauspost/compress/pull/503)
* zip: Update deprecated CreateHeaderRaw to correctly call CreateRaw by @saracen in [#502](https://github.com/klauspost/compress/pull/502)
* zip: don't read data descriptor early by @saracen in [#501](https://github.com/klauspost/compress/pull/501) #501
* huff0: Use static decompression buffer up to 30% faster by @klauspost in [#499](https://github.com/klauspost/compress/pull/499) [#500](https://github.com/klauspost/compress/pull/500)
* Feb 17, 2022 (v1.14.3)
* flate: Improve fastest levels compression speed ~10% more throughput. [#482](https://github.com/klauspost/compress/pull/482) [#489](https://github.com/klauspost/compress/pull/489) [#490](https://github.com/klauspost/compress/pull/490) [#491](https://github.com/klauspost/compress/pull/491) [#494](https://github.com/klauspost/compress/pull/494) [#478](https://github.com/klauspost/compress/pull/478)
* flate: Faster decompression speed, ~5-10%. [#483](https://github.com/klauspost/compress/pull/483)
* s2: Faster compression with Go v1.18 and amd64 microarch level 3+. [#484](https://github.com/klauspost/compress/pull/484) [#486](https://github.com/klauspost/compress/pull/486)
* Jan 25, 2022 (v1.14.2)
* zstd: improve header decoder by @dsnet [#476](https://github.com/klauspost/compress/pull/476)
* zstd: Add bigger default blocks [#469](https://github.com/klauspost/compress/pull/469)
* zstd: Remove unused decompression buffer [#470](https://github.com/klauspost/compress/pull/470)
* zstd: Fix logically dead code by @ningmingxiao [#472](https://github.com/klauspost/compress/pull/472)
* flate: Improve level 7-9 [#471](https://github.com/klauspost/compress/pull/471) [#473](https://github.com/klauspost/compress/pull/473)
* zstd: Add noasm tag for xxhash [#475](https://github.com/klauspost/compress/pull/475)
* Jan 11, 2022 (v1.14.1)
* s2: Add stream index in [#462](https://github.com/klauspost/compress/pull/462)
* flate: Speed and efficiency improvements in [#439](https://github.com/klauspost/compress/pull/439) [#461](https://github.com/klauspost/compress/pull/461) [#455](https://github.com/klauspost/compress/pull/455) [#452](https://github.com/klauspost/compress/pull/452) [#458](https://github.com/klauspost/compress/pull/458)
* zstd: Performance improvement in [#420]( https://github.com/klauspost/compress/pull/420) [#456](https://github.com/klauspost/compress/pull/456) [#437](https://github.com/klauspost/compress/pull/437) [#467](https://github.com/klauspost/compress/pull/467) [#468](https://github.com/klauspost/compress/pull/468)
* zstd: add arm64 xxhash assembly in [#464](https://github.com/klauspost/compress/pull/464)
* Add garbled for binaries for s2 in [#445](https://github.com/klauspost/compress/pull/445)
</details>
<details>
<summary>See changes to v1.13.x</summary>
* Aug 30, 2021 (v1.13.5)
* gz/zlib/flate: Alias stdlib errors [#425](https://github.com/klauspost/compress/pull/425)
* s2: Add block support to commandline tools [#413](https://github.com/klauspost/compress/pull/413)
* zstd: pooledZipWriter should return Writers to the same pool [#426](https://github.com/klauspost/compress/pull/426)
* Removed golang/snappy as external dependency for tests [#421](https://github.com/klauspost/compress/pull/421)
* Aug 12, 2021 (v1.13.4)
* Add [snappy replacement package](https://github.com/klauspost/compress/tree/master/snappy).
* zstd: Fix incorrect encoding in "best" mode [#415](https://github.com/klauspost/compress/pull/415)
* Aug 3, 2021 (v1.13.3)
* zstd: Improve Best compression [#404](https://github.com/klauspost/compress/pull/404)
* zstd: Fix WriteTo error forwarding [#411](https://github.com/klauspost/compress/pull/411)
* gzhttp: Return http.HandlerFunc instead of http.Handler. Unlikely breaking change. [#406](https://github.com/klauspost/compress/pull/406)
* s2sx: Fix max size error [#399](https://github.com/klauspost/compress/pull/399)
* zstd: Add optional stream content size on reset [#401](https://github.com/klauspost/compress/pull/401)
* zstd: use SpeedBestCompression for level >= 10 [#410](https://github.com/klauspost/compress/pull/410)
* Jun 14, 2021 (v1.13.1)
* s2: Add full Snappy output support [#396](https://github.com/klauspost/compress/pull/396)
* zstd: Add configurable [Decoder window](https://pkg.go.dev/github.com/klauspost/compress/zstd#WithDecoderMaxWindow) size [#394](https://github.com/klauspost/compress/pull/394)
* gzhttp: Add header to skip compression [#389](https://github.com/klauspost/compress/pull/389)
* s2: Improve speed with bigger output margin [#395](https://github.com/klauspost/compress/pull/395)
* Jun 3, 2021 (v1.13.0)
* Added [gzhttp](https://github.com/klauspost/compress/tree/master/gzhttp#gzip-handler) which allows wrapping HTTP servers and clients with GZIP compressors.
* zstd: Detect short invalid signatures [#382](https://github.com/klauspost/compress/pull/382)
* zstd: Spawn decoder goroutine only if needed. [#380](https://github.com/klauspost/compress/pull/380)
</details>
<details>
<summary>See changes to v1.12.x</summary>
* May 25, 2021 (v1.12.3)
* deflate: Better/faster Huffman encoding [#374](https://github.com/klauspost/compress/pull/374)
* deflate: Allocate less for history. [#375](https://github.com/klauspost/compress/pull/375)
* zstd: Forward read errors [#373](https://github.com/klauspost/compress/pull/373)
* Apr 27, 2021 (v1.12.2)
* zstd: Improve better/best compression [#360](https://github.com/klauspost/compress/pull/360) [#364](https://github.com/klauspost/compress/pull/364) [#365](https://github.com/klauspost/compress/pull/365)
* zstd: Add helpers to compress/decompress zstd inside zip files [#363](https://github.com/klauspost/compress/pull/363)
* deflate: Improve level 5+6 compression [#367](https://github.com/klauspost/compress/pull/367)
* s2: Improve better/best compression [#358](https://github.com/klauspost/compress/pull/358) [#359](https://github.com/klauspost/compress/pull/358)
* s2: Load after checking src limit on amd64. [#362](https://github.com/klauspost/compress/pull/362)
* s2sx: Limit max executable size [#368](https://github.com/klauspost/compress/pull/368)
* Apr 14, 2021 (v1.12.1)
* snappy package removed. Upstream added as dependency.
* s2: Better compression in "best" mode [#353](https://github.com/klauspost/compress/pull/353)
* s2sx: Add stdin input and detect pre-compressed from signature [#352](https://github.com/klauspost/compress/pull/352)
* s2c/s2d: Add http as possible input [#348](https://github.com/klauspost/compress/pull/348)
* s2c/s2d/s2sx: Always truncate when writing files [#352](https://github.com/klauspost/compress/pull/352)
* zstd: Reduce memory usage further when using [WithLowerEncoderMem](https://pkg.go.dev/github.com/klauspost/compress/zstd#WithLowerEncoderMem) [#346](https://github.com/klauspost/compress/pull/346)
* s2: Fix potential problem with amd64 assembly and profilers [#349](https://github.com/klauspost/compress/pull/349)
</details>
<details>
<summary>See changes to v1.11.x</summary>
* Mar 26, 2021 (v1.11.13)
* zstd: Big speedup on small dictionary encodes [#344](https://github.com/klauspost/compress/pull/344) [#345](https://github.com/klauspost/compress/pull/345)
* zstd: Add [WithLowerEncoderMem](https://pkg.go.dev/github.com/klauspost/compress/zstd#WithLowerEncoderMem) encoder option [#336](https://github.com/klauspost/compress/pull/336)
* deflate: Improve entropy compression [#338](https://github.com/klauspost/compress/pull/338)
* s2: Clean up and minor performance improvement in best [#341](https://github.com/klauspost/compress/pull/341)
* Mar 5, 2021 (v1.11.12)
* s2: Add `s2sx` binary that creates [self extracting archives](https://github.com/klauspost/compress/tree/master/s2#s2sx-self-extracting-archives).
* s2: Speed up decompression on non-assembly platforms [#328](https://github.com/klauspost/compress/pull/328)
* Mar 1, 2021 (v1.11.9)
* s2: Add ARM64 decompression assembly. Around 2x output speed. [#324](https://github.com/klauspost/compress/pull/324)
* s2: Improve "better" speed and efficiency. [#325](https://github.com/klauspost/compress/pull/325)
* s2: Fix binaries.
* Feb 25, 2021 (v1.11.8)
* s2: Fixed occational out-of-bounds write on amd64. Upgrade recommended.
* s2: Add AMD64 assembly for better mode. 25-50% faster. [#315](https://github.com/klauspost/compress/pull/315)
* s2: Less upfront decoder allocation. [#322](https://github.com/klauspost/compress/pull/322)
* zstd: Faster "compression" of incompressible data. [#314](https://github.com/klauspost/compress/pull/314)
* zip: Fix zip64 headers. [#313](https://github.com/klauspost/compress/pull/313)
* Jan 14, 2021 (v1.11.7)
* Use Bytes() interface to get bytes across packages. [#309](https://github.com/klauspost/compress/pull/309)
* s2: Add 'best' compression option. [#310](https://github.com/klauspost/compress/pull/310)
* s2: Add ReaderMaxBlockSize, changes `s2.NewReader` signature to include varargs. [#311](https://github.com/klauspost/compress/pull/311)
* s2: Fix crash on small better buffers. [#308](https://github.com/klauspost/compress/pull/308)
* s2: Clean up decoder. [#312](https://github.com/klauspost/compress/pull/312)
* Jan 7, 2021 (v1.11.6)
* zstd: Make decoder allocations smaller [#306](https://github.com/klauspost/compress/pull/306)
* zstd: Free Decoder resources when Reset is called with a nil io.Reader [#305](https://github.com/klauspost/compress/pull/305)
* Dec 20, 2020 (v1.11.4)
* zstd: Add Best compression mode [#304](https://github.com/klauspost/compress/pull/304)
* Add header decoder [#299](https://github.com/klauspost/compress/pull/299)
* s2: Add uncompressed stream option [#297](https://github.com/klauspost/compress/pull/297)
* Simplify/speed up small blocks with known max size. [#300](https://github.com/klauspost/compress/pull/300)
* zstd: Always reset literal dict encoder [#303](https://github.com/klauspost/compress/pull/303)
* Nov 15, 2020 (v1.11.3)
* inflate: 10-15% faster decompression [#293](https://github.com/klauspost/compress/pull/293)
* zstd: Tweak DecodeAll default allocation [#295](https://github.com/klauspost/compress/pull/295)
* Oct 11, 2020 (v1.11.2)
* s2: Fix out of bounds read in "better" block compression [#291](https://github.com/klauspost/compress/pull/291)
* Oct 1, 2020 (v1.11.1)
* zstd: Set allLitEntropy true in default configuration [#286](https://github.com/klauspost/compress/pull/286)
* Sept 8, 2020 (v1.11.0)
* zstd: Add experimental compression [dictionaries](https://github.com/klauspost/compress/tree/master/zstd#dictionaries) [#281](https://github.com/klauspost/compress/pull/281)
* zstd: Fix mixed Write and ReadFrom calls [#282](https://github.com/klauspost/compress/pull/282)
* inflate/gz: Limit variable shifts, ~5% faster decompression [#274](https://github.com/klauspost/compress/pull/274)
</details>
<details>
<summary>See changes to v1.10.x</summary>
* July 8, 2020 (v1.10.11)
* zstd: Fix extra block when compressing with ReadFrom. [#278](https://github.com/klauspost/compress/pull/278)
* huff0: Also populate compression table when reading decoding table. [#275](https://github.com/klauspost/compress/pull/275)
* June 23, 2020 (v1.10.10)
* zstd: Skip entropy compression in fastest mode when no matches. [#270](https://github.com/klauspost/compress/pull/270)
* June 16, 2020 (v1.10.9):
* zstd: API change for specifying dictionaries. See [#268](https://github.com/klauspost/compress/pull/268)
* zip: update CreateHeaderRaw to handle zip64 fields. [#266](https://github.com/klauspost/compress/pull/266)
* Fuzzit tests removed. The service has been purchased and is no longer available.
* June 5, 2020 (v1.10.8):
* 1.15x faster zstd block decompression. [#265](https://github.com/klauspost/compress/pull/265)
* June 1, 2020 (v1.10.7):
* Added zstd decompression [dictionary support](https://github.com/klauspost/compress/tree/master/zstd#dictionaries)
* Increase zstd decompression speed up to 1.19x. [#259](https://github.com/klauspost/compress/pull/259)
* Remove internal reset call in zstd compression and reduce allocations. [#263](https://github.com/klauspost/compress/pull/263)
* May 21, 2020: (v1.10.6)
* zstd: Reduce allocations while decoding. [#258](https://github.com/klauspost/compress/pull/258), [#252](https://github.com/klauspost/compress/pull/252)
* zstd: Stricter decompression checks.
* April 12, 2020: (v1.10.5)
* s2-commands: Flush output when receiving SIGINT. [#239](https://github.com/klauspost/compress/pull/239)
* Apr 8, 2020: (v1.10.4)
* zstd: Minor/special case optimizations. [#251](https://github.com/klauspost/compress/pull/251), [#250](https://github.com/klauspost/compress/pull/250), [#249](https://github.com/klauspost/compress/pull/249), [#247](https://github.com/klauspost/compress/pull/247)
* Mar 11, 2020: (v1.10.3)
* s2: Use S2 encoder in pure Go mode for Snappy output as well. [#245](https://github.com/klauspost/compress/pull/245)
* s2: Fix pure Go block encoder. [#244](https://github.com/klauspost/compress/pull/244)
* zstd: Added "better compression" mode. [#240](https://github.com/klauspost/compress/pull/240)
* zstd: Improve speed of fastest compression mode by 5-10% [#241](https://github.com/klauspost/compress/pull/241)
* zstd: Skip creating encoders when not needed. [#238](https://github.com/klauspost/compress/pull/238)
* Feb 27, 2020: (v1.10.2)
* Close to 50% speedup in inflate (gzip/zip decompression). [#236](https://github.com/klauspost/compress/pull/236) [#234](https://github.com/klauspost/compress/pull/234) [#232](https://github.com/klauspost/compress/pull/232)
* Reduce deflate level 1-6 memory usage up to 59%. [#227](https://github.com/klauspost/compress/pull/227)
* Feb 18, 2020: (v1.10.1)
* Fix zstd crash when resetting multiple times without sending data. [#226](https://github.com/klauspost/compress/pull/226)
* deflate: Fix dictionary use on level 1-6. [#224](https://github.com/klauspost/compress/pull/224)
* Remove deflate writer reference when closing. [#224](https://github.com/klauspost/compress/pull/224)
* Feb 4, 2020: (v1.10.0)
* Add optional dictionary to [stateless deflate](https://pkg.go.dev/github.com/klauspost/compress/flate?tab=doc#StatelessDeflate). Breaking change, send `nil` for previous behaviour. [#216](https://github.com/klauspost/compress/pull/216)
* Fix buffer overflow on repeated small block deflate. [#218](https://github.com/klauspost/compress/pull/218)
* Allow copying content from an existing ZIP file without decompressing+compressing. [#214](https://github.com/klauspost/compress/pull/214)
* Added [S2](https://github.com/klauspost/compress/tree/master/s2#s2-compression) AMD64 assembler and various optimizations. Stream speed >10GB/s. [#186](https://github.com/klauspost/compress/pull/186)
</details>
<details>
<summary>See changes prior to v1.10.0</summary>
* Jan 20,2020 (v1.9.8) Optimize gzip/deflate with better size estimates and faster table generation. [#207](https://github.com/klauspost/compress/pull/207) by [luyu6056](https://github.com/luyu6056), [#206](https://github.com/klauspost/compress/pull/206).
* Jan 11, 2020: S2 Encode/Decode will use provided buffer if capacity is big enough. [#204](https://github.com/klauspost/compress/pull/204)
* Jan 5, 2020: (v1.9.7) Fix another zstd regression in v1.9.5 - v1.9.6 removed.
* Jan 4, 2020: (v1.9.6) Regression in v1.9.5 fixed causing corrupt zstd encodes in rare cases.
* Jan 4, 2020: Faster IO in [s2c + s2d commandline tools](https://github.com/klauspost/compress/tree/master/s2#commandline-tools) compression/decompression. [#192](https://github.com/klauspost/compress/pull/192)
* Dec 29, 2019: Removed v1.9.5 since fuzz tests showed a compatibility problem with the reference zstandard decoder.
* Dec 29, 2019: (v1.9.5) zstd: 10-20% faster block compression. [#199](https://github.com/klauspost/compress/pull/199)
* Dec 29, 2019: [zip](https://godoc.org/github.com/klauspost/compress/zip) package updated with latest Go features
* Dec 29, 2019: zstd: Single segment flag condintions tweaked. [#197](https://github.com/klauspost/compress/pull/197)
* Dec 18, 2019: s2: Faster compression when ReadFrom is used. [#198](https://github.com/klauspost/compress/pull/198)
* Dec 10, 2019: s2: Fix repeat length output when just above at 16MB limit.
* Dec 10, 2019: zstd: Add function to get decoder as io.ReadCloser. [#191](https://github.com/klauspost/compress/pull/191)
* Dec 3, 2019: (v1.9.4) S2: limit max repeat length. [#188](https://github.com/klauspost/compress/pull/188)
* Dec 3, 2019: Add [WithNoEntropyCompression](https://godoc.org/github.com/klauspost/compress/zstd#WithNoEntropyCompression) to zstd [#187](https://github.com/klauspost/compress/pull/187)
* Dec 3, 2019: Reduce memory use for tests. Check for leaked goroutines.
* Nov 28, 2019 (v1.9.3) Less allocations in stateless deflate.
* Nov 28, 2019: 5-20% Faster huff0 decode. Impacts zstd as well. [#184](https://github.com/klauspost/compress/pull/184)
* Nov 12, 2019 (v1.9.2) Added [Stateless Compression](#stateless-compression) for gzip/deflate.
* Nov 12, 2019: Fixed zstd decompression of large single blocks. [#180](https://github.com/klauspost/compress/pull/180)
* Nov 11, 2019: Set default [s2c](https://github.com/klauspost/compress/tree/master/s2#commandline-tools) block size to 4MB.
* Nov 11, 2019: Reduce inflate memory use by 1KB.
* Nov 10, 2019: Less allocations in deflate bit writer.
* Nov 10, 2019: Fix inconsistent error returned by zstd decoder.
* Oct 28, 2019 (v1.9.1) ztsd: Fix crash when compressing blocks. [#174](https://github.com/klauspost/compress/pull/174)
* Oct 24, 2019 (v1.9.0) zstd: Fix rare data corruption [#173](https://github.com/klauspost/compress/pull/173)
* Oct 24, 2019 zstd: Fix huff0 out of buffer write [#171](https://github.com/klauspost/compress/pull/171) and always return errors [#172](https://github.com/klauspost/compress/pull/172)
* Oct 10, 2019: Big deflate rewrite, 30-40% faster with better compression [#105](https://github.com/klauspost/compress/pull/105)
</details>
<details>
<summary>See changes prior to v1.9.0</summary>
* Oct 10, 2019: (v1.8.6) zstd: Allow partial reads to get flushed data. [#169](https://github.com/klauspost/compress/pull/169)
* Oct 3, 2019: Fix inconsistent results on broken zstd streams.
* Sep 25, 2019: Added `-rm` (remove source files) and `-q` (no output except errors) to `s2c` and `s2d` [commands](https://github.com/klauspost/compress/tree/master/s2#commandline-tools)
* Sep 16, 2019: (v1.8.4) Add `s2c` and `s2d` [commandline tools](https://github.com/klauspost/compress/tree/master/s2#commandline-tools).
* Sep 10, 2019: (v1.8.3) Fix s2 decoder [Skip](https://godoc.org/github.com/klauspost/compress/s2#Reader.Skip).
* Sep 7, 2019: zstd: Added [WithWindowSize](https://godoc.org/github.com/klauspost/compress/zstd#WithWindowSize), contributed by [ianwilkes](https://github.com/ianwilkes).
* Sep 5, 2019: (v1.8.2) Add [WithZeroFrames](https://godoc.org/github.com/klauspost/compress/zstd#WithZeroFrames) which adds full zero payload block encoding option.
* Sep 5, 2019: Lazy initialization of zstandard predefined en/decoder tables.
* Aug 26, 2019: (v1.8.1) S2: 1-2% compression increase in "better" compression mode.
* Aug 26, 2019: zstd: Check maximum size of Huffman 1X compressed literals while decoding.
* Aug 24, 2019: (v1.8.0) Added [S2 compression](https://github.com/klauspost/compress/tree/master/s2#s2-compression), a high performance replacement for Snappy.
* Aug 21, 2019: (v1.7.6) Fixed minor issues found by fuzzer. One could lead to zstd not decompressing.
* Aug 18, 2019: Add [fuzzit](https://fuzzit.dev/) continuous fuzzing.
* Aug 14, 2019: zstd: Skip incompressible data 2x faster. [#147](https://github.com/klauspost/compress/pull/147)
* Aug 4, 2019 (v1.7.5): Better literal compression. [#146](https://github.com/klauspost/compress/pull/146)
* Aug 4, 2019: Faster zstd compression. [#143](https://github.com/klauspost/compress/pull/143) [#144](https://github.com/klauspost/compress/pull/144)
* Aug 4, 2019: Faster zstd decompression. [#145](https://github.com/klauspost/compress/pull/145) [#143](https://github.com/klauspost/compress/pull/143) [#142](https://github.com/klauspost/compress/pull/142)
* July 15, 2019 (v1.7.4): Fix double EOF block in rare cases on zstd encoder.
* July 15, 2019 (v1.7.3): Minor speedup/compression increase in default zstd encoder.
* July 14, 2019: zstd decoder: Fix decompression error on multiple uses with mixed content.
* July 7, 2019 (v1.7.2): Snappy update, zstd decoder potential race fix.
* June 17, 2019: zstd decompression bugfix.
* June 17, 2019: fix 32 bit builds.
* June 17, 2019: Easier use in modules (less dependencies).
* June 9, 2019: New stronger "default" [zstd](https://github.com/klauspost/compress/tree/master/zstd#zstd) compression mode. Matches zstd default compression ratio.
* June 5, 2019: 20-40% throughput in [zstandard](https://github.com/klauspost/compress/tree/master/zstd#zstd) compression and better compression.
* June 5, 2019: deflate/gzip compression: Reduce memory usage of lower compression levels.
* June 2, 2019: Added [zstandard](https://github.com/klauspost/compress/tree/master/zstd#zstd) compression!
* May 25, 2019: deflate/gzip: 10% faster bit writer, mostly visible in lower levels.
* Apr 22, 2019: [zstd](https://github.com/klauspost/compress/tree/master/zstd#zstd) decompression added.
* Aug 1, 2018: Added [huff0 README](https://github.com/klauspost/compress/tree/master/huff0#huff0-entropy-compression).
* Jul 8, 2018: Added [Performance Update 2018](#performance-update-2018) below.
* Jun 23, 2018: Merged [Go 1.11 inflate optimizations](https://go-review.googlesource.com/c/go/+/102235). Go 1.9 is now required. Backwards compatible version tagged with [v1.3.0](https://github.com/klauspost/compress/releases/tag/v1.3.0).
* Apr 2, 2018: Added [huff0](https://godoc.org/github.com/klauspost/compress/huff0) en/decoder. Experimental for now, API may change.
* Mar 4, 2018: Added [FSE Entropy](https://godoc.org/github.com/klauspost/compress/fse) en/decoder. Experimental for now, API may change.
* Nov 3, 2017: Add compression [Estimate](https://godoc.org/github.com/klauspost/compress#Estimate) function.
* May 28, 2017: Reduce allocations when resetting decoder.
* Apr 02, 2017: Change back to official crc32, since changes were merged in Go 1.7.
* Jan 14, 2017: Reduce stack pressure due to array copies. See [Issue #18625](https://github.com/golang/go/issues/18625).
* Oct 25, 2016: Level 2-4 have been rewritten and now offers significantly better performance than before.
* Oct 20, 2016: Port zlib changes from Go 1.7 to fix zlib writer issue. Please update.
* Oct 16, 2016: Go 1.7 changes merged. Apples to apples this package is a few percent faster, but has a significantly better balance between speed and compression per level.
* Mar 24, 2016: Always attempt Huffman encoding on level 4-7. This improves base 64 encoded data compression.
* Mar 24, 2016: Small speedup for level 1-3.
* Feb 19, 2016: Faster bit writer, level -2 is 15% faster, level 1 is 4% faster.
* Feb 19, 2016: Handle small payloads faster in level 1-3.
* Feb 19, 2016: Added faster level 2 + 3 compression modes.
* Feb 19, 2016: [Rebalanced compression levels](https://blog.klauspost.com/rebalancing-deflate-compression-levels/), so there is a more even progresssion in terms of compression. New default level is 5.
* Feb 14, 2016: Snappy: Merge upstream changes.
* Feb 14, 2016: Snappy: Fix aggressive skipping.
* Feb 14, 2016: Snappy: Update benchmark.
* Feb 13, 2016: Deflate: Fixed assembler problem that could lead to sub-optimal compression.
* Feb 12, 2016: Snappy: Added AMD64 SSE 4.2 optimizations to matching, which makes easy to compress material run faster. Typical speedup is around 25%.
* Feb 9, 2016: Added Snappy package fork. This version is 5-7% faster, much more on hard to compress content.
* Jan 30, 2016: Optimize level 1 to 3 by not considering static dictionary or storing uncompressed. ~4-5% speedup.
* Jan 16, 2016: Optimization on deflate level 1,2,3 compression.
* Jan 8 2016: Merge [CL 18317](https://go-review.googlesource.com/#/c/18317): fix reading, writing of zip64 archives.
* Dec 8 2015: Make level 1 and -2 deterministic even if write size differs.
* Dec 8 2015: Split encoding functions, so hashing and matching can potentially be inlined. 1-3% faster on AMD64. 5% faster on other platforms.
* Dec 8 2015: Fixed rare [one byte out-of bounds read](https://github.com/klauspost/compress/issues/20). Please update!
* Nov 23 2015: Optimization on token writer. ~2-4% faster. Contributed by [@dsnet](https://github.com/dsnet).
* Nov 20 2015: Small optimization to bit writer on 64 bit systems.
* Nov 17 2015: Fixed out-of-bound errors if the underlying Writer returned an error. See [#15](https://github.com/klauspost/compress/issues/15).
* Nov 12 2015: Added [io.WriterTo](https://golang.org/pkg/io/#WriterTo) support to gzip/inflate.
* Nov 11 2015: Merged [CL 16669](https://go-review.googlesource.com/#/c/16669/4): archive/zip: enable overriding (de)compressors per file
* Oct 15 2015: Added skipping on uncompressible data. Random data speed up >5x.
</details>
# deflate usage
The packages are drop-in replacements for standard libraries. Simply replace the import path to use them:
| old import | new import | Documentation
|--------------------|-----------------------------------------|--------------------|
| `compress/gzip` | `github.com/klauspost/compress/gzip` | [gzip](https://pkg.go.dev/github.com/klauspost/compress/gzip?tab=doc)
| `compress/zlib` | `github.com/klauspost/compress/zlib` | [zlib](https://pkg.go.dev/github.com/klauspost/compress/zlib?tab=doc)
| `archive/zip` | `github.com/klauspost/compress/zip` | [zip](https://pkg.go.dev/github.com/klauspost/compress/zip?tab=doc)
| `compress/flate` | `github.com/klauspost/compress/flate` | [flate](https://pkg.go.dev/github.com/klauspost/compress/flate?tab=doc)
* Optimized [deflate](https://godoc.org/github.com/klauspost/compress/flate) packages which can be used as a dropin replacement for [gzip](https://godoc.org/github.com/klauspost/compress/gzip), [zip](https://godoc.org/github.com/klauspost/compress/zip) and [zlib](https://godoc.org/github.com/klauspost/compress/zlib).
You may also be interested in [pgzip](https://github.com/klauspost/pgzip), which is a drop in replacement for gzip, which support multithreaded compression on big files and the optimized [crc32](https://github.com/klauspost/crc32) package used by these packages.
The packages contains the same as the standard library, so you can use the godoc for that: [gzip](http://golang.org/pkg/compress/gzip/), [zip](http://golang.org/pkg/archive/zip/), [zlib](http://golang.org/pkg/compress/zlib/), [flate](http://golang.org/pkg/compress/flate/).
Currently there is only minor speedup on decompression (mostly CRC32 calculation).
Memory usage is typically 1MB for a Writer. stdlib is in the same range.
If you expect to have a lot of concurrently allocated Writers consider using
the stateless compress described below.
For compression performance, see: [this spreadsheet](https://docs.google.com/spreadsheets/d/1nuNE2nPfuINCZJRMt6wFWhKpToF95I47XjSsc-1rbPQ/edit?usp=sharing).
To disable all assembly add `-tags=noasm`. This works across all packages.
# Stateless compression
This package offers stateless compression as a special option for gzip/deflate.
It will do compression but without maintaining any state between Write calls.
This means there will be no memory kept between Write calls, but compression and speed will be suboptimal.
This is only relevant in cases where you expect to run many thousands of compressors concurrently,
but with very little activity. This is *not* intended for regular web servers serving individual requests.
Because of this, the size of actual Write calls will affect output size.
In gzip, specify level `-3` / `gzip.StatelessCompression` to enable.
For direct deflate use, NewStatelessWriter and StatelessDeflate are available. See [documentation](https://godoc.org/github.com/klauspost/compress/flate#NewStatelessWriter)
A `bufio.Writer` can of course be used to control write sizes. For example, to use a 4KB buffer:
```go
// replace 'ioutil.Discard' with your output.
gzw, err := gzip.NewWriterLevel(ioutil.Discard, gzip.StatelessCompression)
if err != nil {
return err
}
defer gzw.Close()
w := bufio.NewWriterSize(gzw, 4096)
defer w.Flush()
// Write to 'w'
```
This will only use up to 4KB in memory when the writer is idle.
Compression is almost always worse than the fastest compression level
and each write will allocate (a little) memory.
# Performance Update 2018
It has been a while since we have been looking at the speed of this package compared to the standard library, so I thought I would re-do my tests and give some overall recommendations based on the current state. All benchmarks have been performed with Go 1.10 on my Desktop Intel(R) Core(TM) i7-2600 CPU @3.40GHz. Since I last ran the tests, I have gotten more RAM, which means tests with big files are no longer limited by my SSD.
The raw results are in my [updated spreadsheet](https://docs.google.com/spreadsheets/d/1nuNE2nPfuINCZJRMt6wFWhKpToF95I47XjSsc-1rbPQ/edit?usp=sharing). Due to cgo changes and upstream updates i could not get the cgo version of gzip to compile. Instead I included the [zstd](https://github.com/datadog/zstd) cgo implementation. If I get cgo gzip to work again, I might replace the results in the sheet.
The columns to take note of are: *MB/s* - the throughput. *Reduction* - the data size reduction in percent of the original. *Rel Speed* relative speed compared to the standard library at the same level. *Smaller* - how many percent smaller is the compressed output compared to stdlib. Negative means the output was bigger. *Loss* means the loss (or gain) in compression as a percentage difference of the input.
The `gzstd` (standard library gzip) and `gzkp` (this package gzip) only uses one CPU core. [`pgzip`](https://github.com/klauspost/pgzip), [`bgzf`](https://github.com/biogo/hts/tree/master/bgzf) uses all 4 cores. [`zstd`](https://github.com/DataDog/zstd) uses one core, and is a beast (but not Go, yet).
## Overall differences.
There appears to be a roughly 5-10% speed advantage over the standard library when comparing at similar compression levels.
The biggest difference you will see is the result of [re-balancing](https://blog.klauspost.com/rebalancing-deflate-compression-levels/) the compression levels. I wanted by library to give a smoother transition between the compression levels than the standard library.
This package attempts to provide a more smooth transition, where "1" is taking a lot of shortcuts, "5" is the reasonable trade-off and "9" is the "give me the best compression", and the values in between gives something reasonable in between. The standard library has big differences in levels 1-4, but levels 5-9 having no significant gains - often spending a lot more time than can be justified by the achieved compression.
There are links to all the test data in the [spreadsheet](https://docs.google.com/spreadsheets/d/1nuNE2nPfuINCZJRMt6wFWhKpToF95I47XjSsc-1rbPQ/edit?usp=sharing) in the top left field on each tab.
## Web Content
This test set aims to emulate typical use in a web server. The test-set is 4GB data in 53k files, and is a mixture of (mostly) HTML, JS, CSS.
Since level 1 and 9 are close to being the same code, they are quite close. But looking at the levels in-between the differences are quite big.
Looking at level 6, this package is 88% faster, but will output about 6% more data. For a web server, this means you can serve 88% more data, but have to pay for 6% more bandwidth. You can draw your own conclusions on what would be the most expensive for your case.
## Object files
This test is for typical data files stored on a server. In this case it is a collection of Go precompiled objects. They are very compressible.
The picture is similar to the web content, but with small differences since this is very compressible. Levels 2-3 offer good speed, but is sacrificing quite a bit of compression.
The standard library seems suboptimal on level 3 and 4 - offering both worse compression and speed than level 6 & 7 of this package respectively.
## Highly Compressible File
This is a JSON file with very high redundancy. The reduction starts at 95% on level 1, so in real life terms we are dealing with something like a highly redundant stream of data, etc.
It is definitely visible that we are dealing with specialized content here, so the results are very scattered. This package does not do very well at levels 1-4, but picks up significantly at level 5 and levels 7 and 8 offering great speed for the achieved compression.
So if you know you content is extremely compressible you might want to go slightly higher than the defaults. The standard library has a huge gap between levels 3 and 4 in terms of speed (2.75x slowdown), so it offers little "middle ground".
## Medium-High Compressible
This is a pretty common test corpus: [enwik9](http://mattmahoney.net/dc/textdata.html). It contains the first 10^9 bytes of the English Wikipedia dump on Mar. 3, 2006. This is a very good test of typical text based compression and more data heavy streams.
We see a similar picture here as in "Web Content". On equal levels some compression is sacrificed for more speed. Level 5 seems to be the best trade-off between speed and size, beating stdlib level 3 in both.
## Medium Compressible
I will combine two test sets, one [10GB file set](http://mattmahoney.net/dc/10gb.html) and a VM disk image (~8GB). Both contain different data types and represent a typical backup scenario.
The most notable thing is how quickly the standard library drops to very low compression speeds around level 5-6 without any big gains in compression. Since this type of data is fairly common, this does not seem like good behavior.
## Un-compressible Content
This is mainly a test of how good the algorithms are at detecting un-compressible input. The standard library only offers this feature with very conservative settings at level 1. Obviously there is no reason for the algorithms to try to compress input that cannot be compressed. The only downside is that it might skip some compressible data on false detections.
## Huffman only compression
This compression library adds a special compression level, named `HuffmanOnly`, which allows near linear time compression. This is done by completely disabling matching of previous data, and only reduce the number of bits to represent each character.
This means that often used characters, like 'e' and ' ' (space) in text use the fewest bits to represent, and rare characters like '¤' takes more bits to represent. For more information see [wikipedia](https://en.wikipedia.org/wiki/Huffman_coding) or this nice [video](https://youtu.be/ZdooBTdW5bM).
Since this type of compression has much less variance, the compression speed is mostly unaffected by the input data, and is usually more than *180MB/s* for a single core.
The downside is that the compression ratio is usually considerably worse than even the fastest conventional compression. The compression ratio can never be better than 8:1 (12.5%).
The linear time compression can be used as a "better than nothing" mode, where you cannot risk the encoder to slow down on some content. For comparison, the size of the "Twain" text is *233460 bytes* (+29% vs. level 1) and encode speed is 144MB/s (4.5x level 1). So in this case you trade a 30% size increase for a 4 times speedup.
For more information see my blog post on [Fast Linear Time Compression](http://blog.klauspost.com/constant-time-gzipzip-compression/).
This is implemented on Go 1.7 as "Huffman Only" mode, though not exposed for gzip.
# Other packages
Here are other packages of good quality and pure Go (no cgo wrappers or autoconverted code):
* [github.com/pierrec/lz4](https://github.com/pierrec/lz4) - strong multithreaded LZ4 compression.
* [github.com/cosnicolaou/pbzip2](https://github.com/cosnicolaou/pbzip2) - multithreaded bzip2 decompression.
* [github.com/dsnet/compress](https://github.com/dsnet/compress) - brotli decompression, bzip2 writer.
* [github.com/ronanh/intcomp](https://github.com/ronanh/intcomp) - Integer compression.
* [github.com/spenczar/fpc](https://github.com/spenczar/fpc) - Float compression.
* [github.com/minio/zipindex](https://github.com/minio/zipindex) - External ZIP directory index.
* [github.com/ybirader/pzip](https://github.com/ybirader/pzip) - Fast concurrent zip archiver and extractor.
# license
This code is licensed under the same conditions as the original Go code. See LICENSE file.

25
vendor/github.com/klauspost/compress/SECURITY.md generated vendored Normal file
View File

@@ -0,0 +1,25 @@
# Security Policy
## Supported Versions
Security updates are applied only to the latest release.
## Vulnerability Definition
A security vulnerability is a bug that with certain input triggers a crash or an infinite loop. Most calls will have varying execution time and only in rare cases will slow operation be considered a security vulnerability.
Corrupted output generally is not considered a security vulnerability, unless independent operations are able to affect each other. Note that not all functionality is re-entrant and safe to use concurrently.
Out-of-memory crashes only applies if the en/decoder uses an abnormal amount of memory, with appropriate options applied, to limit maximum window size, concurrency, etc. However, if you are in doubt you are welcome to file a security issue.
It is assumed that all callers are trusted, meaning internal data exposed through reflection or inspection of returned data structures is not considered a vulnerability.
Vulnerabilities resulting from compiler/assembler errors should be reported upstream. Depending on the severity this package may or may not implement a workaround.
## Reporting a Vulnerability
If you have discovered a security vulnerability in this project, please report it privately. **Do not disclose it as a public issue.** This gives us time to work with you to fix the issue before public exposure, reducing the chance that the exploit will be used before a patch is released.
Please disclose it at [security advisory](https://github.com/klauspost/compress/security/advisories/new). If possible please provide a minimal reproducer. If the issue only applies to a single platform, it would be helpful to provide access to that.
This project is maintained by a team of volunteers on a reasonable-effort basis. As such, vulnerabilities will be disclosed in a best effort base.

85
vendor/github.com/klauspost/compress/compressible.go generated vendored Normal file
View File

@@ -0,0 +1,85 @@
package compress
import "math"
// Estimate returns a normalized compressibility estimate of block b.
// Values close to zero are likely uncompressible.
// Values above 0.1 are likely to be compressible.
// Values above 0.5 are very compressible.
// Very small lengths will return 0.
func Estimate(b []byte) float64 {
if len(b) < 16 {
return 0
}
// Correctly predicted order 1
hits := 0
lastMatch := false
var o1 [256]byte
var hist [256]int
c1 := byte(0)
for _, c := range b {
if c == o1[c1] {
// We only count a hit if there was two correct predictions in a row.
if lastMatch {
hits++
}
lastMatch = true
} else {
lastMatch = false
}
o1[c1] = c
c1 = c
hist[c]++
}
// Use x^0.6 to give better spread
prediction := math.Pow(float64(hits)/float64(len(b)), 0.6)
// Calculate histogram distribution
variance := float64(0)
avg := float64(len(b)) / 256
for _, v := range hist {
Δ := float64(v) - avg
variance += Δ * Δ
}
stddev := math.Sqrt(float64(variance)) / float64(len(b))
exp := math.Sqrt(1 / float64(len(b)))
// Subtract expected stddev
stddev -= exp
if stddev < 0 {
stddev = 0
}
stddev *= 1 + exp
// Use x^0.4 to give better spread
entropy := math.Pow(stddev, 0.4)
// 50/50 weight between prediction and histogram distribution
return math.Pow((prediction+entropy)/2, 0.9)
}
// ShannonEntropyBits returns the number of bits minimum required to represent
// an entropy encoding of the input bytes.
// https://en.wiktionary.org/wiki/Shannon_entropy
func ShannonEntropyBits(b []byte) int {
if len(b) == 0 {
return 0
}
var hist [256]int
for _, c := range b {
hist[c]++
}
shannon := float64(0)
invTotal := 1.0 / float64(len(b))
for _, v := range hist[:] {
if v > 0 {
n := float64(v)
shannon += math.Ceil(-math.Log2(n*invTotal) * n)
}
}
return int(math.Ceil(shannon))
}

79
vendor/github.com/klauspost/compress/fse/README.md generated vendored Normal file
View File

@@ -0,0 +1,79 @@
# Finite State Entropy
This package provides Finite State Entropy encoding and decoding.
Finite State Entropy (also referenced as [tANS](https://en.wikipedia.org/wiki/Asymmetric_numeral_systems#tANS))
encoding provides a fast near-optimal symbol encoding/decoding
for byte blocks as implemented in [zstandard](https://github.com/facebook/zstd).
This can be used for compressing input with a lot of similar input values to the smallest number of bytes.
This does not perform any multi-byte [dictionary coding](https://en.wikipedia.org/wiki/Dictionary_coder) as LZ coders,
but it can be used as a secondary step to compressors (like Snappy) that does not do entropy encoding.
* [Godoc documentation](https://godoc.org/github.com/klauspost/compress/fse)
## News
* Feb 2018: First implementation released. Consider this beta software for now.
# Usage
This package provides a low level interface that allows to compress single independent blocks.
Each block is separate, and there is no built in integrity checks.
This means that the caller should keep track of block sizes and also do checksums if needed.
Compressing a block is done via the [`Compress`](https://godoc.org/github.com/klauspost/compress/fse#Compress) function.
You must provide input and will receive the output and maybe an error.
These error values can be returned:
| Error | Description |
|---------------------|-----------------------------------------------------------------------------|
| `<nil>` | Everything ok, output is returned |
| `ErrIncompressible` | Returned when input is judged to be too hard to compress |
| `ErrUseRLE` | Returned from the compressor when the input is a single byte value repeated |
| `(error)` | An internal error occurred. |
As can be seen above there are errors that will be returned even under normal operation so it is important to handle these.
To reduce allocations you can provide a [`Scratch`](https://godoc.org/github.com/klauspost/compress/fse#Scratch) object
that can be re-used for successive calls. Both compression and decompression accepts a `Scratch` object, and the same
object can be used for both.
Be aware, that when re-using a `Scratch` object that the *output* buffer is also re-used, so if you are still using this
you must set the `Out` field in the scratch to nil. The same buffer is used for compression and decompression output.
Decompressing is done by calling the [`Decompress`](https://godoc.org/github.com/klauspost/compress/fse#Decompress) function.
You must provide the output from the compression stage, at exactly the size you got back. If you receive an error back
your input was likely corrupted.
It is important to note that a successful decoding does *not* mean your output matches your original input.
There are no integrity checks, so relying on errors from the decompressor does not assure your data is valid.
For more detailed usage, see examples in the [godoc documentation](https://godoc.org/github.com/klauspost/compress/fse#pkg-examples).
# Performance
A lot of factors are affecting speed. Block sizes and compressibility of the material are primary factors.
All compression functions are currently only running on the calling goroutine so only one core will be used per block.
The compressor is significantly faster if symbols are kept as small as possible. The highest byte value of the input
is used to reduce some of the processing, so if all your input is above byte value 64 for instance, it may be
beneficial to transpose all your input values down by 64.
With moderate block sizes around 64k speed are typically 200MB/s per core for compression and
around 300MB/s decompression speed.
The same hardware typically does Huffman (deflate) encoding at 125MB/s and decompression at 100MB/s.
# Plans
At one point, more internals will be exposed to facilitate more "expert" usage of the components.
A streaming interface is also likely to be implemented. Likely compatible with [FSE stream format](https://github.com/Cyan4973/FiniteStateEntropy/blob/dev/programs/fileio.c#L261).
# Contributing
Contributions are always welcome. Be aware that adding public functions will require good justification and breaking
changes will likely not be accepted. If in doubt open an issue before writing the PR.

122
vendor/github.com/klauspost/compress/fse/bitreader.go generated vendored Normal file
View File

@@ -0,0 +1,122 @@
// Copyright 2018 Klaus Post. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Based on work Copyright (c) 2013, Yann Collet, released under BSD License.
package fse
import (
"encoding/binary"
"errors"
"io"
)
// bitReader reads a bitstream in reverse.
// The last set bit indicates the start of the stream and is used
// for aligning the input.
type bitReader struct {
in []byte
off uint // next byte to read is at in[off - 1]
value uint64
bitsRead uint8
}
// init initializes and resets the bit reader.
func (b *bitReader) init(in []byte) error {
if len(in) < 1 {
return errors.New("corrupt stream: too short")
}
b.in = in
b.off = uint(len(in))
// The highest bit of the last byte indicates where to start
v := in[len(in)-1]
if v == 0 {
return errors.New("corrupt stream, did not find end of stream")
}
b.bitsRead = 64
b.value = 0
if len(in) >= 8 {
b.fillFastStart()
} else {
b.fill()
b.fill()
}
b.bitsRead += 8 - uint8(highBits(uint32(v)))
return nil
}
// getBits will return n bits. n can be 0.
func (b *bitReader) getBits(n uint8) uint16 {
if n == 0 || b.bitsRead >= 64 {
return 0
}
return b.getBitsFast(n)
}
// getBitsFast requires that at least one bit is requested every time.
// There are no checks if the buffer is filled.
func (b *bitReader) getBitsFast(n uint8) uint16 {
const regMask = 64 - 1
v := uint16((b.value << (b.bitsRead & regMask)) >> ((regMask + 1 - n) & regMask))
b.bitsRead += n
return v
}
// fillFast() will make sure at least 32 bits are available.
// There must be at least 4 bytes available.
func (b *bitReader) fillFast() {
if b.bitsRead < 32 {
return
}
// 2 bounds checks.
v := b.in[b.off-4:]
v = v[:4]
low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
b.value = (b.value << 32) | uint64(low)
b.bitsRead -= 32
b.off -= 4
}
// fill() will make sure at least 32 bits are available.
func (b *bitReader) fill() {
if b.bitsRead < 32 {
return
}
if b.off > 4 {
v := b.in[b.off-4:]
v = v[:4]
low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
b.value = (b.value << 32) | uint64(low)
b.bitsRead -= 32
b.off -= 4
return
}
for b.off > 0 {
b.value = (b.value << 8) | uint64(b.in[b.off-1])
b.bitsRead -= 8
b.off--
}
}
// fillFastStart() assumes the bitreader is empty and there is at least 8 bytes to read.
func (b *bitReader) fillFastStart() {
// Do single re-slice to avoid bounds checks.
b.value = binary.LittleEndian.Uint64(b.in[b.off-8:])
b.bitsRead = 0
b.off -= 8
}
// finished returns true if all bits have been read from the bit stream.
func (b *bitReader) finished() bool {
return b.bitsRead >= 64 && b.off == 0
}
// close the bitstream and returns an error if out-of-buffer reads occurred.
func (b *bitReader) close() error {
// Release reference.
b.in = nil
if b.bitsRead > 64 {
return io.ErrUnexpectedEOF
}
return nil
}

167
vendor/github.com/klauspost/compress/fse/bitwriter.go generated vendored Normal file
View File

@@ -0,0 +1,167 @@
// Copyright 2018 Klaus Post. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Based on work Copyright (c) 2013, Yann Collet, released under BSD License.
package fse
import "fmt"
// bitWriter will write bits.
// First bit will be LSB of the first byte of output.
type bitWriter struct {
bitContainer uint64
nBits uint8
out []byte
}
// bitMask16 is bitmasks. Has extra to avoid bounds check.
var bitMask16 = [32]uint16{
0, 1, 3, 7, 0xF, 0x1F,
0x3F, 0x7F, 0xFF, 0x1FF, 0x3FF, 0x7FF,
0xFFF, 0x1FFF, 0x3FFF, 0x7FFF, 0xFFFF, 0xFFFF,
0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF,
0xFFFF, 0xFFFF} /* up to 16 bits */
// addBits16NC will add up to 16 bits.
// It will not check if there is space for them,
// so the caller must ensure that it has flushed recently.
func (b *bitWriter) addBits16NC(value uint16, bits uint8) {
b.bitContainer |= uint64(value&bitMask16[bits&31]) << (b.nBits & 63)
b.nBits += bits
}
// addBits16Clean will add up to 16 bits. value may not contain more set bits than indicated.
// It will not check if there is space for them, so the caller must ensure that it has flushed recently.
func (b *bitWriter) addBits16Clean(value uint16, bits uint8) {
b.bitContainer |= uint64(value) << (b.nBits & 63)
b.nBits += bits
}
// addBits16ZeroNC will add up to 16 bits.
// It will not check if there is space for them,
// so the caller must ensure that it has flushed recently.
// This is fastest if bits can be zero.
func (b *bitWriter) addBits16ZeroNC(value uint16, bits uint8) {
if bits == 0 {
return
}
value <<= (16 - bits) & 15
value >>= (16 - bits) & 15
b.bitContainer |= uint64(value) << (b.nBits & 63)
b.nBits += bits
}
// flush will flush all pending full bytes.
// There will be at least 56 bits available for writing when this has been called.
// Using flush32 is faster, but leaves less space for writing.
func (b *bitWriter) flush() {
v := b.nBits >> 3
switch v {
case 0:
case 1:
b.out = append(b.out,
byte(b.bitContainer),
)
case 2:
b.out = append(b.out,
byte(b.bitContainer),
byte(b.bitContainer>>8),
)
case 3:
b.out = append(b.out,
byte(b.bitContainer),
byte(b.bitContainer>>8),
byte(b.bitContainer>>16),
)
case 4:
b.out = append(b.out,
byte(b.bitContainer),
byte(b.bitContainer>>8),
byte(b.bitContainer>>16),
byte(b.bitContainer>>24),
)
case 5:
b.out = append(b.out,
byte(b.bitContainer),
byte(b.bitContainer>>8),
byte(b.bitContainer>>16),
byte(b.bitContainer>>24),
byte(b.bitContainer>>32),
)
case 6:
b.out = append(b.out,
byte(b.bitContainer),
byte(b.bitContainer>>8),
byte(b.bitContainer>>16),
byte(b.bitContainer>>24),
byte(b.bitContainer>>32),
byte(b.bitContainer>>40),
)
case 7:
b.out = append(b.out,
byte(b.bitContainer),
byte(b.bitContainer>>8),
byte(b.bitContainer>>16),
byte(b.bitContainer>>24),
byte(b.bitContainer>>32),
byte(b.bitContainer>>40),
byte(b.bitContainer>>48),
)
case 8:
b.out = append(b.out,
byte(b.bitContainer),
byte(b.bitContainer>>8),
byte(b.bitContainer>>16),
byte(b.bitContainer>>24),
byte(b.bitContainer>>32),
byte(b.bitContainer>>40),
byte(b.bitContainer>>48),
byte(b.bitContainer>>56),
)
default:
panic(fmt.Errorf("bits (%d) > 64", b.nBits))
}
b.bitContainer >>= v << 3
b.nBits &= 7
}
// flush32 will flush out, so there are at least 32 bits available for writing.
func (b *bitWriter) flush32() {
if b.nBits < 32 {
return
}
b.out = append(b.out,
byte(b.bitContainer),
byte(b.bitContainer>>8),
byte(b.bitContainer>>16),
byte(b.bitContainer>>24))
b.nBits -= 32
b.bitContainer >>= 32
}
// flushAlign will flush remaining full bytes and align to next byte boundary.
func (b *bitWriter) flushAlign() {
nbBytes := (b.nBits + 7) >> 3
for i := uint8(0); i < nbBytes; i++ {
b.out = append(b.out, byte(b.bitContainer>>(i*8)))
}
b.nBits = 0
b.bitContainer = 0
}
// close will write the alignment bit and write the final byte(s)
// to the output.
func (b *bitWriter) close() {
// End mark
b.addBits16Clean(1, 1)
// flush until next byte.
b.flushAlign()
}
// reset and continue writing by appending to out.
func (b *bitWriter) reset(out []byte) {
b.bitContainer = 0
b.nBits = 0
b.out = out
}

47
vendor/github.com/klauspost/compress/fse/bytereader.go generated vendored Normal file
View File

@@ -0,0 +1,47 @@
// Copyright 2018 Klaus Post. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Based on work Copyright (c) 2013, Yann Collet, released under BSD License.
package fse
// byteReader provides a byte reader that reads
// little endian values from a byte stream.
// The input stream is manually advanced.
// The reader performs no bounds checks.
type byteReader struct {
b []byte
off int
}
// init will initialize the reader and set the input.
func (b *byteReader) init(in []byte) {
b.b = in
b.off = 0
}
// advance the stream b n bytes.
func (b *byteReader) advance(n uint) {
b.off += int(n)
}
// Uint32 returns a little endian uint32 starting at current offset.
func (b byteReader) Uint32() uint32 {
b2 := b.b[b.off:]
b2 = b2[:4]
v3 := uint32(b2[3])
v2 := uint32(b2[2])
v1 := uint32(b2[1])
v0 := uint32(b2[0])
return v0 | (v1 << 8) | (v2 << 16) | (v3 << 24)
}
// unread returns the unread portion of the input.
func (b byteReader) unread() []byte {
return b.b[b.off:]
}
// remain will return the number of bytes remaining.
func (b byteReader) remain() int {
return len(b.b) - b.off
}

683
vendor/github.com/klauspost/compress/fse/compress.go generated vendored Normal file
View File

@@ -0,0 +1,683 @@
// Copyright 2018 Klaus Post. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Based on work Copyright (c) 2013, Yann Collet, released under BSD License.
package fse
import (
"errors"
"fmt"
)
// Compress the input bytes. Input must be < 2GB.
// Provide a Scratch buffer to avoid memory allocations.
// Note that the output is also kept in the scratch buffer.
// If input is too hard to compress, ErrIncompressible is returned.
// If input is a single byte value repeated ErrUseRLE is returned.
func Compress(in []byte, s *Scratch) ([]byte, error) {
if len(in) <= 1 {
return nil, ErrIncompressible
}
if len(in) > (2<<30)-1 {
return nil, errors.New("input too big, must be < 2GB")
}
s, err := s.prepare(in)
if err != nil {
return nil, err
}
// Create histogram, if none was provided.
maxCount := s.maxCount
if maxCount == 0 {
maxCount = s.countSimple(in)
}
// Reset for next run.
s.clearCount = true
s.maxCount = 0
if maxCount == len(in) {
// One symbol, use RLE
return nil, ErrUseRLE
}
if maxCount == 1 || maxCount < (len(in)>>7) {
// Each symbol present maximum once or too well distributed.
return nil, ErrIncompressible
}
s.optimalTableLog()
err = s.normalizeCount()
if err != nil {
return nil, err
}
err = s.writeCount()
if err != nil {
return nil, err
}
if false {
err = s.validateNorm()
if err != nil {
return nil, err
}
}
err = s.buildCTable()
if err != nil {
return nil, err
}
err = s.compress(in)
if err != nil {
return nil, err
}
s.Out = s.bw.out
// Check if we compressed.
if len(s.Out) >= len(in) {
return nil, ErrIncompressible
}
return s.Out, nil
}
// cState contains the compression state of a stream.
type cState struct {
bw *bitWriter
stateTable []uint16
state uint16
}
// init will initialize the compression state to the first symbol of the stream.
func (c *cState) init(bw *bitWriter, ct *cTable, tableLog uint8, first symbolTransform) {
c.bw = bw
c.stateTable = ct.stateTable
nbBitsOut := (first.deltaNbBits + (1 << 15)) >> 16
im := int32((nbBitsOut << 16) - first.deltaNbBits)
lu := (im >> nbBitsOut) + first.deltaFindState
c.state = c.stateTable[lu]
}
// encode the output symbol provided and write it to the bitstream.
func (c *cState) encode(symbolTT symbolTransform) {
nbBitsOut := (uint32(c.state) + symbolTT.deltaNbBits) >> 16
dstState := int32(c.state>>(nbBitsOut&15)) + symbolTT.deltaFindState
c.bw.addBits16NC(c.state, uint8(nbBitsOut))
c.state = c.stateTable[dstState]
}
// encode the output symbol provided and write it to the bitstream.
func (c *cState) encodeZero(symbolTT symbolTransform) {
nbBitsOut := (uint32(c.state) + symbolTT.deltaNbBits) >> 16
dstState := int32(c.state>>(nbBitsOut&15)) + symbolTT.deltaFindState
c.bw.addBits16ZeroNC(c.state, uint8(nbBitsOut))
c.state = c.stateTable[dstState]
}
// flush will write the tablelog to the output and flush the remaining full bytes.
func (c *cState) flush(tableLog uint8) {
c.bw.flush32()
c.bw.addBits16NC(c.state, tableLog)
c.bw.flush()
}
// compress is the main compression loop that will encode the input from the last byte to the first.
func (s *Scratch) compress(src []byte) error {
if len(src) <= 2 {
return errors.New("compress: src too small")
}
tt := s.ct.symbolTT[:256]
s.bw.reset(s.Out)
// Our two states each encodes every second byte.
// Last byte encoded (first byte decoded) will always be encoded by c1.
var c1, c2 cState
// Encode so remaining size is divisible by 4.
ip := len(src)
if ip&1 == 1 {
c1.init(&s.bw, &s.ct, s.actualTableLog, tt[src[ip-1]])
c2.init(&s.bw, &s.ct, s.actualTableLog, tt[src[ip-2]])
c1.encodeZero(tt[src[ip-3]])
ip -= 3
} else {
c2.init(&s.bw, &s.ct, s.actualTableLog, tt[src[ip-1]])
c1.init(&s.bw, &s.ct, s.actualTableLog, tt[src[ip-2]])
ip -= 2
}
if ip&2 != 0 {
c2.encodeZero(tt[src[ip-1]])
c1.encodeZero(tt[src[ip-2]])
ip -= 2
}
src = src[:ip]
// Main compression loop.
switch {
case !s.zeroBits && s.actualTableLog <= 8:
// We can encode 4 symbols without requiring a flush.
// We do not need to check if any output is 0 bits.
for ; len(src) >= 4; src = src[:len(src)-4] {
s.bw.flush32()
v3, v2, v1, v0 := src[len(src)-4], src[len(src)-3], src[len(src)-2], src[len(src)-1]
c2.encode(tt[v0])
c1.encode(tt[v1])
c2.encode(tt[v2])
c1.encode(tt[v3])
}
case !s.zeroBits:
// We do not need to check if any output is 0 bits.
for ; len(src) >= 4; src = src[:len(src)-4] {
s.bw.flush32()
v3, v2, v1, v0 := src[len(src)-4], src[len(src)-3], src[len(src)-2], src[len(src)-1]
c2.encode(tt[v0])
c1.encode(tt[v1])
s.bw.flush32()
c2.encode(tt[v2])
c1.encode(tt[v3])
}
case s.actualTableLog <= 8:
// We can encode 4 symbols without requiring a flush
for ; len(src) >= 4; src = src[:len(src)-4] {
s.bw.flush32()
v3, v2, v1, v0 := src[len(src)-4], src[len(src)-3], src[len(src)-2], src[len(src)-1]
c2.encodeZero(tt[v0])
c1.encodeZero(tt[v1])
c2.encodeZero(tt[v2])
c1.encodeZero(tt[v3])
}
default:
for ; len(src) >= 4; src = src[:len(src)-4] {
s.bw.flush32()
v3, v2, v1, v0 := src[len(src)-4], src[len(src)-3], src[len(src)-2], src[len(src)-1]
c2.encodeZero(tt[v0])
c1.encodeZero(tt[v1])
s.bw.flush32()
c2.encodeZero(tt[v2])
c1.encodeZero(tt[v3])
}
}
// Flush final state.
// Used to initialize state when decoding.
c2.flush(s.actualTableLog)
c1.flush(s.actualTableLog)
s.bw.close()
return nil
}
// writeCount will write the normalized histogram count to header.
// This is read back by readNCount.
func (s *Scratch) writeCount() error {
var (
tableLog = s.actualTableLog
tableSize = 1 << tableLog
previous0 bool
charnum uint16
maxHeaderSize = ((int(s.symbolLen)*int(tableLog) + 4 + 2) >> 3) + 3
// Write Table Size
bitStream = uint32(tableLog - minTablelog)
bitCount = uint(4)
remaining = int16(tableSize + 1) /* +1 for extra accuracy */
threshold = int16(tableSize)
nbBits = uint(tableLog + 1)
)
if cap(s.Out) < maxHeaderSize {
s.Out = make([]byte, 0, s.br.remain()+maxHeaderSize)
}
outP := uint(0)
out := s.Out[:maxHeaderSize]
// stops at 1
for remaining > 1 {
if previous0 {
start := charnum
for s.norm[charnum] == 0 {
charnum++
}
for charnum >= start+24 {
start += 24
bitStream += uint32(0xFFFF) << bitCount
out[outP] = byte(bitStream)
out[outP+1] = byte(bitStream >> 8)
outP += 2
bitStream >>= 16
}
for charnum >= start+3 {
start += 3
bitStream += 3 << bitCount
bitCount += 2
}
bitStream += uint32(charnum-start) << bitCount
bitCount += 2
if bitCount > 16 {
out[outP] = byte(bitStream)
out[outP+1] = byte(bitStream >> 8)
outP += 2
bitStream >>= 16
bitCount -= 16
}
}
count := s.norm[charnum]
charnum++
max := (2*threshold - 1) - remaining
if count < 0 {
remaining += count
} else {
remaining -= count
}
count++ // +1 for extra accuracy
if count >= threshold {
count += max // [0..max[ [max..threshold[ (...) [threshold+max 2*threshold[
}
bitStream += uint32(count) << bitCount
bitCount += nbBits
if count < max {
bitCount--
}
previous0 = count == 1
if remaining < 1 {
return errors.New("internal error: remaining<1")
}
for remaining < threshold {
nbBits--
threshold >>= 1
}
if bitCount > 16 {
out[outP] = byte(bitStream)
out[outP+1] = byte(bitStream >> 8)
outP += 2
bitStream >>= 16
bitCount -= 16
}
}
out[outP] = byte(bitStream)
out[outP+1] = byte(bitStream >> 8)
outP += (bitCount + 7) / 8
if charnum > s.symbolLen {
return errors.New("internal error: charnum > s.symbolLen")
}
s.Out = out[:outP]
return nil
}
// symbolTransform contains the state transform for a symbol.
type symbolTransform struct {
deltaFindState int32
deltaNbBits uint32
}
// String prints values as a human readable string.
func (s symbolTransform) String() string {
return fmt.Sprintf("dnbits: %08x, fs:%d", s.deltaNbBits, s.deltaFindState)
}
// cTable contains tables used for compression.
type cTable struct {
tableSymbol []byte
stateTable []uint16
symbolTT []symbolTransform
}
// allocCtable will allocate tables needed for compression.
// If existing tables a re big enough, they are simply re-used.
func (s *Scratch) allocCtable() {
tableSize := 1 << s.actualTableLog
// get tableSymbol that is big enough.
if cap(s.ct.tableSymbol) < tableSize {
s.ct.tableSymbol = make([]byte, tableSize)
}
s.ct.tableSymbol = s.ct.tableSymbol[:tableSize]
ctSize := tableSize
if cap(s.ct.stateTable) < ctSize {
s.ct.stateTable = make([]uint16, ctSize)
}
s.ct.stateTable = s.ct.stateTable[:ctSize]
if cap(s.ct.symbolTT) < 256 {
s.ct.symbolTT = make([]symbolTransform, 256)
}
s.ct.symbolTT = s.ct.symbolTT[:256]
}
// buildCTable will populate the compression table so it is ready to be used.
func (s *Scratch) buildCTable() error {
tableSize := uint32(1 << s.actualTableLog)
highThreshold := tableSize - 1
var cumul [maxSymbolValue + 2]int16
s.allocCtable()
tableSymbol := s.ct.tableSymbol[:tableSize]
// symbol start positions
{
cumul[0] = 0
for ui, v := range s.norm[:s.symbolLen-1] {
u := byte(ui) // one less than reference
if v == -1 {
// Low proba symbol
cumul[u+1] = cumul[u] + 1
tableSymbol[highThreshold] = u
highThreshold--
} else {
cumul[u+1] = cumul[u] + v
}
}
// Encode last symbol separately to avoid overflowing u
u := int(s.symbolLen - 1)
v := s.norm[s.symbolLen-1]
if v == -1 {
// Low proba symbol
cumul[u+1] = cumul[u] + 1
tableSymbol[highThreshold] = byte(u)
highThreshold--
} else {
cumul[u+1] = cumul[u] + v
}
if uint32(cumul[s.symbolLen]) != tableSize {
return fmt.Errorf("internal error: expected cumul[s.symbolLen] (%d) == tableSize (%d)", cumul[s.symbolLen], tableSize)
}
cumul[s.symbolLen] = int16(tableSize) + 1
}
// Spread symbols
s.zeroBits = false
{
step := tableStep(tableSize)
tableMask := tableSize - 1
var position uint32
// if any symbol > largeLimit, we may have 0 bits output.
largeLimit := int16(1 << (s.actualTableLog - 1))
for ui, v := range s.norm[:s.symbolLen] {
symbol := byte(ui)
if v > largeLimit {
s.zeroBits = true
}
for nbOccurrences := int16(0); nbOccurrences < v; nbOccurrences++ {
tableSymbol[position] = symbol
position = (position + step) & tableMask
for position > highThreshold {
position = (position + step) & tableMask
} /* Low proba area */
}
}
// Check if we have gone through all positions
if position != 0 {
return errors.New("position!=0")
}
}
// Build table
table := s.ct.stateTable
{
tsi := int(tableSize)
for u, v := range tableSymbol {
// TableU16 : sorted by symbol order; gives next state value
table[cumul[v]] = uint16(tsi + u)
cumul[v]++
}
}
// Build Symbol Transformation Table
{
total := int16(0)
symbolTT := s.ct.symbolTT[:s.symbolLen]
tableLog := s.actualTableLog
tl := (uint32(tableLog) << 16) - (1 << tableLog)
for i, v := range s.norm[:s.symbolLen] {
switch v {
case 0:
case -1, 1:
symbolTT[i].deltaNbBits = tl
symbolTT[i].deltaFindState = int32(total - 1)
total++
default:
maxBitsOut := uint32(tableLog) - highBits(uint32(v-1))
minStatePlus := uint32(v) << maxBitsOut
symbolTT[i].deltaNbBits = (maxBitsOut << 16) - minStatePlus
symbolTT[i].deltaFindState = int32(total - v)
total += v
}
}
if total != int16(tableSize) {
return fmt.Errorf("total mismatch %d (got) != %d (want)", total, tableSize)
}
}
return nil
}
// countSimple will create a simple histogram in s.count.
// Returns the biggest count.
// Does not update s.clearCount.
func (s *Scratch) countSimple(in []byte) (max int) {
for _, v := range in {
s.count[v]++
}
m, symlen := uint32(0), s.symbolLen
for i, v := range s.count[:] {
if v == 0 {
continue
}
if v > m {
m = v
}
symlen = uint16(i) + 1
}
s.symbolLen = symlen
return int(m)
}
// minTableLog provides the minimum logSize to safely represent a distribution.
func (s *Scratch) minTableLog() uint8 {
minBitsSrc := highBits(uint32(s.br.remain()-1)) + 1
minBitsSymbols := highBits(uint32(s.symbolLen-1)) + 2
if minBitsSrc < minBitsSymbols {
return uint8(minBitsSrc)
}
return uint8(minBitsSymbols)
}
// optimalTableLog calculates and sets the optimal tableLog in s.actualTableLog
func (s *Scratch) optimalTableLog() {
tableLog := s.TableLog
minBits := s.minTableLog()
maxBitsSrc := uint8(highBits(uint32(s.br.remain()-1))) - 2
if maxBitsSrc < tableLog {
// Accuracy can be reduced
tableLog = maxBitsSrc
}
if minBits > tableLog {
tableLog = minBits
}
// Need a minimum to safely represent all symbol values
if tableLog < minTablelog {
tableLog = minTablelog
}
if tableLog > maxTableLog {
tableLog = maxTableLog
}
s.actualTableLog = tableLog
}
var rtbTable = [...]uint32{0, 473195, 504333, 520860, 550000, 700000, 750000, 830000}
// normalizeCount will normalize the count of the symbols so
// the total is equal to the table size.
func (s *Scratch) normalizeCount() error {
var (
tableLog = s.actualTableLog
scale = 62 - uint64(tableLog)
step = (1 << 62) / uint64(s.br.remain())
vStep = uint64(1) << (scale - 20)
stillToDistribute = int16(1 << tableLog)
largest int
largestP int16
lowThreshold = (uint32)(s.br.remain() >> tableLog)
)
for i, cnt := range s.count[:s.symbolLen] {
// already handled
// if (count[s] == s.length) return 0; /* rle special case */
if cnt == 0 {
s.norm[i] = 0
continue
}
if cnt <= lowThreshold {
s.norm[i] = -1
stillToDistribute--
} else {
proba := (int16)((uint64(cnt) * step) >> scale)
if proba < 8 {
restToBeat := vStep * uint64(rtbTable[proba])
v := uint64(cnt)*step - (uint64(proba) << scale)
if v > restToBeat {
proba++
}
}
if proba > largestP {
largestP = proba
largest = i
}
s.norm[i] = proba
stillToDistribute -= proba
}
}
if -stillToDistribute >= (s.norm[largest] >> 1) {
// corner case, need another normalization method
return s.normalizeCount2()
}
s.norm[largest] += stillToDistribute
return nil
}
// Secondary normalization method.
// To be used when primary method fails.
func (s *Scratch) normalizeCount2() error {
const notYetAssigned = -2
var (
distributed uint32
total = uint32(s.br.remain())
tableLog = s.actualTableLog
lowThreshold = total >> tableLog
lowOne = (total * 3) >> (tableLog + 1)
)
for i, cnt := range s.count[:s.symbolLen] {
if cnt == 0 {
s.norm[i] = 0
continue
}
if cnt <= lowThreshold {
s.norm[i] = -1
distributed++
total -= cnt
continue
}
if cnt <= lowOne {
s.norm[i] = 1
distributed++
total -= cnt
continue
}
s.norm[i] = notYetAssigned
}
toDistribute := (1 << tableLog) - distributed
if (total / toDistribute) > lowOne {
// risk of rounding to zero
lowOne = (total * 3) / (toDistribute * 2)
for i, cnt := range s.count[:s.symbolLen] {
if (s.norm[i] == notYetAssigned) && (cnt <= lowOne) {
s.norm[i] = 1
distributed++
total -= cnt
continue
}
}
toDistribute = (1 << tableLog) - distributed
}
if distributed == uint32(s.symbolLen)+1 {
// all values are pretty poor;
// probably incompressible data (should have already been detected);
// find max, then give all remaining points to max
var maxV int
var maxC uint32
for i, cnt := range s.count[:s.symbolLen] {
if cnt > maxC {
maxV = i
maxC = cnt
}
}
s.norm[maxV] += int16(toDistribute)
return nil
}
if total == 0 {
// all of the symbols were low enough for the lowOne or lowThreshold
for i := uint32(0); toDistribute > 0; i = (i + 1) % (uint32(s.symbolLen)) {
if s.norm[i] > 0 {
toDistribute--
s.norm[i]++
}
}
return nil
}
var (
vStepLog = 62 - uint64(tableLog)
mid = uint64((1 << (vStepLog - 1)) - 1)
rStep = (((1 << vStepLog) * uint64(toDistribute)) + mid) / uint64(total) // scale on remaining
tmpTotal = mid
)
for i, cnt := range s.count[:s.symbolLen] {
if s.norm[i] == notYetAssigned {
var (
end = tmpTotal + uint64(cnt)*rStep
sStart = uint32(tmpTotal >> vStepLog)
sEnd = uint32(end >> vStepLog)
weight = sEnd - sStart
)
if weight < 1 {
return errors.New("weight < 1")
}
s.norm[i] = int16(weight)
tmpTotal = end
}
}
return nil
}
// validateNorm validates the normalized histogram table.
func (s *Scratch) validateNorm() (err error) {
var total int
for _, v := range s.norm[:s.symbolLen] {
if v >= 0 {
total += int(v)
} else {
total -= int(v)
}
}
defer func() {
if err == nil {
return
}
fmt.Printf("selected TableLog: %d, Symbol length: %d\n", s.actualTableLog, s.symbolLen)
for i, v := range s.norm[:s.symbolLen] {
fmt.Printf("%3d: %5d -> %4d \n", i, s.count[i], v)
}
}()
if total != (1 << s.actualTableLog) {
return fmt.Errorf("warning: Total == %d != %d", total, 1<<s.actualTableLog)
}
for i, v := range s.count[s.symbolLen:] {
if v != 0 {
return fmt.Errorf("warning: Found symbol out of range, %d after cut", i)
}
}
return nil
}

376
vendor/github.com/klauspost/compress/fse/decompress.go generated vendored Normal file
View File

@@ -0,0 +1,376 @@
package fse
import (
"errors"
"fmt"
)
const (
tablelogAbsoluteMax = 15
)
// Decompress a block of data.
// You can provide a scratch buffer to avoid allocations.
// If nil is provided a temporary one will be allocated.
// It is possible, but by no way guaranteed that corrupt data will
// return an error.
// It is up to the caller to verify integrity of the returned data.
// Use a predefined Scrach to set maximum acceptable output size.
func Decompress(b []byte, s *Scratch) ([]byte, error) {
s, err := s.prepare(b)
if err != nil {
return nil, err
}
s.Out = s.Out[:0]
err = s.readNCount()
if err != nil {
return nil, err
}
err = s.buildDtable()
if err != nil {
return nil, err
}
err = s.decompress()
if err != nil {
return nil, err
}
return s.Out, nil
}
// readNCount will read the symbol distribution so decoding tables can be constructed.
func (s *Scratch) readNCount() error {
var (
charnum uint16
previous0 bool
b = &s.br
)
iend := b.remain()
if iend < 4 {
return errors.New("input too small")
}
bitStream := b.Uint32()
nbBits := uint((bitStream & 0xF) + minTablelog) // extract tableLog
if nbBits > tablelogAbsoluteMax {
return errors.New("tableLog too large")
}
bitStream >>= 4
bitCount := uint(4)
s.actualTableLog = uint8(nbBits)
remaining := int32((1 << nbBits) + 1)
threshold := int32(1 << nbBits)
gotTotal := int32(0)
nbBits++
for remaining > 1 {
if previous0 {
n0 := charnum
for (bitStream & 0xFFFF) == 0xFFFF {
n0 += 24
if b.off < iend-5 {
b.advance(2)
bitStream = b.Uint32() >> bitCount
} else {
bitStream >>= 16
bitCount += 16
}
}
for (bitStream & 3) == 3 {
n0 += 3
bitStream >>= 2
bitCount += 2
}
n0 += uint16(bitStream & 3)
bitCount += 2
if n0 > maxSymbolValue {
return errors.New("maxSymbolValue too small")
}
for charnum < n0 {
s.norm[charnum&0xff] = 0
charnum++
}
if b.off <= iend-7 || b.off+int(bitCount>>3) <= iend-4 {
b.advance(bitCount >> 3)
bitCount &= 7
bitStream = b.Uint32() >> bitCount
} else {
bitStream >>= 2
}
}
max := (2*(threshold) - 1) - (remaining)
var count int32
if (int32(bitStream) & (threshold - 1)) < max {
count = int32(bitStream) & (threshold - 1)
bitCount += nbBits - 1
} else {
count = int32(bitStream) & (2*threshold - 1)
if count >= threshold {
count -= max
}
bitCount += nbBits
}
count-- // extra accuracy
if count < 0 {
// -1 means +1
remaining += count
gotTotal -= count
} else {
remaining -= count
gotTotal += count
}
s.norm[charnum&0xff] = int16(count)
charnum++
previous0 = count == 0
for remaining < threshold {
nbBits--
threshold >>= 1
}
if b.off <= iend-7 || b.off+int(bitCount>>3) <= iend-4 {
b.advance(bitCount >> 3)
bitCount &= 7
} else {
bitCount -= (uint)(8 * (len(b.b) - 4 - b.off))
b.off = len(b.b) - 4
}
bitStream = b.Uint32() >> (bitCount & 31)
}
s.symbolLen = charnum
if s.symbolLen <= 1 {
return fmt.Errorf("symbolLen (%d) too small", s.symbolLen)
}
if s.symbolLen > maxSymbolValue+1 {
return fmt.Errorf("symbolLen (%d) too big", s.symbolLen)
}
if remaining != 1 {
return fmt.Errorf("corruption detected (remaining %d != 1)", remaining)
}
if bitCount > 32 {
return fmt.Errorf("corruption detected (bitCount %d > 32)", bitCount)
}
if gotTotal != 1<<s.actualTableLog {
return fmt.Errorf("corruption detected (total %d != %d)", gotTotal, 1<<s.actualTableLog)
}
b.advance((bitCount + 7) >> 3)
return nil
}
// decSymbol contains information about a state entry,
// Including the state offset base, the output symbol and
// the number of bits to read for the low part of the destination state.
type decSymbol struct {
newState uint16
symbol uint8
nbBits uint8
}
// allocDtable will allocate decoding tables if they are not big enough.
func (s *Scratch) allocDtable() {
tableSize := 1 << s.actualTableLog
if cap(s.decTable) < tableSize {
s.decTable = make([]decSymbol, tableSize)
}
s.decTable = s.decTable[:tableSize]
if cap(s.ct.tableSymbol) < 256 {
s.ct.tableSymbol = make([]byte, 256)
}
s.ct.tableSymbol = s.ct.tableSymbol[:256]
if cap(s.ct.stateTable) < 256 {
s.ct.stateTable = make([]uint16, 256)
}
s.ct.stateTable = s.ct.stateTable[:256]
}
// buildDtable will build the decoding table.
func (s *Scratch) buildDtable() error {
tableSize := uint32(1 << s.actualTableLog)
highThreshold := tableSize - 1
s.allocDtable()
symbolNext := s.ct.stateTable[:256]
// Init, lay down lowprob symbols
s.zeroBits = false
{
largeLimit := int16(1 << (s.actualTableLog - 1))
for i, v := range s.norm[:s.symbolLen] {
if v == -1 {
s.decTable[highThreshold].symbol = uint8(i)
highThreshold--
symbolNext[i] = 1
} else {
if v >= largeLimit {
s.zeroBits = true
}
symbolNext[i] = uint16(v)
}
}
}
// Spread symbols
{
tableMask := tableSize - 1
step := tableStep(tableSize)
position := uint32(0)
for ss, v := range s.norm[:s.symbolLen] {
for i := 0; i < int(v); i++ {
s.decTable[position].symbol = uint8(ss)
position = (position + step) & tableMask
for position > highThreshold {
// lowprob area
position = (position + step) & tableMask
}
}
}
if position != 0 {
// position must reach all cells once, otherwise normalizedCounter is incorrect
return errors.New("corrupted input (position != 0)")
}
}
// Build Decoding table
{
tableSize := uint16(1 << s.actualTableLog)
for u, v := range s.decTable {
symbol := v.symbol
nextState := symbolNext[symbol]
symbolNext[symbol] = nextState + 1
nBits := s.actualTableLog - byte(highBits(uint32(nextState)))
s.decTable[u].nbBits = nBits
newState := (nextState << nBits) - tableSize
if newState >= tableSize {
return fmt.Errorf("newState (%d) outside table size (%d)", newState, tableSize)
}
if newState == uint16(u) && nBits == 0 {
// Seems weird that this is possible with nbits > 0.
return fmt.Errorf("newState (%d) == oldState (%d) and no bits", newState, u)
}
s.decTable[u].newState = newState
}
}
return nil
}
// decompress will decompress the bitstream.
// If the buffer is over-read an error is returned.
func (s *Scratch) decompress() error {
br := &s.bits
if err := br.init(s.br.unread()); err != nil {
return err
}
var s1, s2 decoder
// Initialize and decode first state and symbol.
s1.init(br, s.decTable, s.actualTableLog)
s2.init(br, s.decTable, s.actualTableLog)
// Use temp table to avoid bound checks/append penalty.
var tmp = s.ct.tableSymbol[:256]
var off uint8
// Main part
if !s.zeroBits {
for br.off >= 8 {
br.fillFast()
tmp[off+0] = s1.nextFast()
tmp[off+1] = s2.nextFast()
br.fillFast()
tmp[off+2] = s1.nextFast()
tmp[off+3] = s2.nextFast()
off += 4
// When off is 0, we have overflowed and should write.
if off == 0 {
s.Out = append(s.Out, tmp...)
if len(s.Out) >= s.DecompressLimit {
return fmt.Errorf("output size (%d) > DecompressLimit (%d)", len(s.Out), s.DecompressLimit)
}
}
}
} else {
for br.off >= 8 {
br.fillFast()
tmp[off+0] = s1.next()
tmp[off+1] = s2.next()
br.fillFast()
tmp[off+2] = s1.next()
tmp[off+3] = s2.next()
off += 4
if off == 0 {
s.Out = append(s.Out, tmp...)
// When off is 0, we have overflowed and should write.
if len(s.Out) >= s.DecompressLimit {
return fmt.Errorf("output size (%d) > DecompressLimit (%d)", len(s.Out), s.DecompressLimit)
}
}
}
}
s.Out = append(s.Out, tmp[:off]...)
// Final bits, a bit more expensive check
for {
if s1.finished() {
s.Out = append(s.Out, s1.final(), s2.final())
break
}
br.fill()
s.Out = append(s.Out, s1.next())
if s2.finished() {
s.Out = append(s.Out, s2.final(), s1.final())
break
}
s.Out = append(s.Out, s2.next())
if len(s.Out) >= s.DecompressLimit {
return fmt.Errorf("output size (%d) > DecompressLimit (%d)", len(s.Out), s.DecompressLimit)
}
}
return br.close()
}
// decoder keeps track of the current state and updates it from the bitstream.
type decoder struct {
state uint16
br *bitReader
dt []decSymbol
}
// init will initialize the decoder and read the first state from the stream.
func (d *decoder) init(in *bitReader, dt []decSymbol, tableLog uint8) {
d.dt = dt
d.br = in
d.state = in.getBits(tableLog)
}
// next returns the next symbol and sets the next state.
// At least tablelog bits must be available in the bit reader.
func (d *decoder) next() uint8 {
n := &d.dt[d.state]
lowBits := d.br.getBits(n.nbBits)
d.state = n.newState + lowBits
return n.symbol
}
// finished returns true if all bits have been read from the bitstream
// and the next state would require reading bits from the input.
func (d *decoder) finished() bool {
return d.br.finished() && d.dt[d.state].nbBits > 0
}
// final returns the current state symbol without decoding the next.
func (d *decoder) final() uint8 {
return d.dt[d.state].symbol
}
// nextFast returns the next symbol and sets the next state.
// This can only be used if no symbols are 0 bits.
// At least tablelog bits must be available in the bit reader.
func (d *decoder) nextFast() uint8 {
n := d.dt[d.state]
lowBits := d.br.getBitsFast(n.nbBits)
d.state = n.newState + lowBits
return n.symbol
}

144
vendor/github.com/klauspost/compress/fse/fse.go generated vendored Normal file
View File

@@ -0,0 +1,144 @@
// Copyright 2018 Klaus Post. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Based on work Copyright (c) 2013, Yann Collet, released under BSD License.
// Package fse provides Finite State Entropy encoding and decoding.
//
// Finite State Entropy encoding provides a fast near-optimal symbol encoding/decoding
// for byte blocks as implemented in zstd.
//
// See https://github.com/klauspost/compress/tree/master/fse for more information.
package fse
import (
"errors"
"fmt"
"math/bits"
)
const (
/*!MEMORY_USAGE :
* Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; etc.)
* Increasing memory usage improves compression ratio
* Reduced memory usage can improve speed, due to cache effect
* Recommended max value is 14, for 16KB, which nicely fits into Intel x86 L1 cache */
maxMemoryUsage = 14
defaultMemoryUsage = 13
maxTableLog = maxMemoryUsage - 2
maxTablesize = 1 << maxTableLog
defaultTablelog = defaultMemoryUsage - 2
minTablelog = 5
maxSymbolValue = 255
)
var (
// ErrIncompressible is returned when input is judged to be too hard to compress.
ErrIncompressible = errors.New("input is not compressible")
// ErrUseRLE is returned from the compressor when the input is a single byte value repeated.
ErrUseRLE = errors.New("input is single value repeated")
)
// Scratch provides temporary storage for compression and decompression.
type Scratch struct {
// Private
count [maxSymbolValue + 1]uint32
norm [maxSymbolValue + 1]int16
br byteReader
bits bitReader
bw bitWriter
ct cTable // Compression tables.
decTable []decSymbol // Decompression table.
maxCount int // count of the most probable symbol
// Per block parameters.
// These can be used to override compression parameters of the block.
// Do not touch, unless you know what you are doing.
// Out is output buffer.
// If the scratch is re-used before the caller is done processing the output,
// set this field to nil.
// Otherwise the output buffer will be re-used for next Compression/Decompression step
// and allocation will be avoided.
Out []byte
// DecompressLimit limits the maximum decoded size acceptable.
// If > 0 decompression will stop when approximately this many bytes
// has been decoded.
// If 0, maximum size will be 2GB.
DecompressLimit int
symbolLen uint16 // Length of active part of the symbol table.
actualTableLog uint8 // Selected tablelog.
zeroBits bool // no bits has prob > 50%.
clearCount bool // clear count
// MaxSymbolValue will override the maximum symbol value of the next block.
MaxSymbolValue uint8
// TableLog will attempt to override the tablelog for the next block.
TableLog uint8
}
// Histogram allows to populate the histogram and skip that step in the compression,
// It otherwise allows to inspect the histogram when compression is done.
// To indicate that you have populated the histogram call HistogramFinished
// with the value of the highest populated symbol, as well as the number of entries
// in the most populated entry. These are accepted at face value.
// The returned slice will always be length 256.
func (s *Scratch) Histogram() []uint32 {
return s.count[:]
}
// HistogramFinished can be called to indicate that the histogram has been populated.
// maxSymbol is the index of the highest set symbol of the next data segment.
// maxCount is the number of entries in the most populated entry.
// These are accepted at face value.
func (s *Scratch) HistogramFinished(maxSymbol uint8, maxCount int) {
s.maxCount = maxCount
s.symbolLen = uint16(maxSymbol) + 1
s.clearCount = maxCount != 0
}
// prepare will prepare and allocate scratch tables used for both compression and decompression.
func (s *Scratch) prepare(in []byte) (*Scratch, error) {
if s == nil {
s = &Scratch{}
}
if s.MaxSymbolValue == 0 {
s.MaxSymbolValue = 255
}
if s.TableLog == 0 {
s.TableLog = defaultTablelog
}
if s.TableLog > maxTableLog {
return nil, fmt.Errorf("tableLog (%d) > maxTableLog (%d)", s.TableLog, maxTableLog)
}
if cap(s.Out) == 0 {
s.Out = make([]byte, 0, len(in))
}
if s.clearCount && s.maxCount == 0 {
for i := range s.count {
s.count[i] = 0
}
s.clearCount = false
}
s.br.init(in)
if s.DecompressLimit == 0 {
// Max size 2GB.
s.DecompressLimit = (2 << 30) - 1
}
return s, nil
}
// tableStep returns the next table index.
func tableStep(tableSize uint32) uint32 {
return (tableSize >> 1) + (tableSize >> 3) + 3
}
func highBits(val uint32) (n uint32) {
return uint32(bits.Len32(val) - 1)
}

4
vendor/github.com/klauspost/compress/gen.sh generated vendored Normal file
View File

@@ -0,0 +1,4 @@
#!/bin/sh
cd s2/cmd/_s2sx/ || exit 1
go generate .

380
vendor/github.com/klauspost/compress/gzip/gunzip.go generated vendored Normal file
View File

@@ -0,0 +1,380 @@
// Copyright 2009 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Package gzip implements reading and writing of gzip format compressed files,
// as specified in RFC 1952.
package gzip
import (
"bufio"
"compress/gzip"
"encoding/binary"
"hash/crc32"
"io"
"time"
"github.com/klauspost/compress/flate"
)
const (
gzipID1 = 0x1f
gzipID2 = 0x8b
gzipDeflate = 8
flagText = 1 << 0
flagHdrCrc = 1 << 1
flagExtra = 1 << 2
flagName = 1 << 3
flagComment = 1 << 4
)
var (
// ErrChecksum is returned when reading GZIP data that has an invalid checksum.
ErrChecksum = gzip.ErrChecksum
// ErrHeader is returned when reading GZIP data that has an invalid header.
ErrHeader = gzip.ErrHeader
)
var le = binary.LittleEndian
// noEOF converts io.EOF to io.ErrUnexpectedEOF.
func noEOF(err error) error {
if err == io.EOF {
return io.ErrUnexpectedEOF
}
return err
}
// The gzip file stores a header giving metadata about the compressed file.
// That header is exposed as the fields of the Writer and Reader structs.
//
// Strings must be UTF-8 encoded and may only contain Unicode code points
// U+0001 through U+00FF, due to limitations of the GZIP file format.
type Header struct {
Comment string // comment
Extra []byte // "extra data"
ModTime time.Time // modification time
Name string // file name
OS byte // operating system type
}
// A Reader is an io.Reader that can be read to retrieve
// uncompressed data from a gzip-format compressed file.
//
// In general, a gzip file can be a concatenation of gzip files,
// each with its own header. Reads from the Reader
// return the concatenation of the uncompressed data of each.
// Only the first header is recorded in the Reader fields.
//
// Gzip files store a length and checksum of the uncompressed data.
// The Reader will return a ErrChecksum when Read
// reaches the end of the uncompressed data if it does not
// have the expected length or checksum. Clients should treat data
// returned by Read as tentative until they receive the io.EOF
// marking the end of the data.
type Reader struct {
Header // valid after NewReader or Reader.Reset
r flate.Reader
br *bufio.Reader
decompressor io.ReadCloser
digest uint32 // CRC-32, IEEE polynomial (section 8)
size uint32 // Uncompressed size (section 2.3.1)
buf [512]byte
err error
multistream bool
}
// NewReader creates a new Reader reading the given reader.
// If r does not also implement io.ByteReader,
// the decompressor may read more data than necessary from r.
//
// It is the caller's responsibility to call Close on the Reader when done.
//
// The Reader.Header fields will be valid in the Reader returned.
func NewReader(r io.Reader) (*Reader, error) {
z := new(Reader)
if err := z.Reset(r); err != nil {
return nil, err
}
return z, nil
}
// Reset discards the Reader z's state and makes it equivalent to the
// result of its original state from NewReader, but reading from r instead.
// This permits reusing a Reader rather than allocating a new one.
func (z *Reader) Reset(r io.Reader) error {
*z = Reader{
decompressor: z.decompressor,
multistream: true,
br: z.br,
}
if rr, ok := r.(flate.Reader); ok {
z.r = rr
} else {
// Reuse if we can.
if z.br != nil {
z.br.Reset(r)
} else {
z.br = bufio.NewReader(r)
}
z.r = z.br
}
z.Header, z.err = z.readHeader()
return z.err
}
// Multistream controls whether the reader supports multistream files.
//
// If enabled (the default), the Reader expects the input to be a sequence
// of individually gzipped data streams, each with its own header and
// trailer, ending at EOF. The effect is that the concatenation of a sequence
// of gzipped files is treated as equivalent to the gzip of the concatenation
// of the sequence. This is standard behavior for gzip readers.
//
// Calling Multistream(false) disables this behavior; disabling the behavior
// can be useful when reading file formats that distinguish individual gzip
// data streams or mix gzip data streams with other data streams.
// In this mode, when the Reader reaches the end of the data stream,
// Read returns io.EOF. If the underlying reader implements io.ByteReader,
// it will be left positioned just after the gzip stream.
// To start the next stream, call z.Reset(r) followed by z.Multistream(false).
// If there is no next stream, z.Reset(r) will return io.EOF.
func (z *Reader) Multistream(ok bool) {
z.multistream = ok
}
// readString reads a NUL-terminated string from z.r.
// It treats the bytes read as being encoded as ISO 8859-1 (Latin-1) and
// will output a string encoded using UTF-8.
// This method always updates z.digest with the data read.
func (z *Reader) readString() (string, error) {
var err error
needConv := false
for i := 0; ; i++ {
if i >= len(z.buf) {
return "", ErrHeader
}
z.buf[i], err = z.r.ReadByte()
if err != nil {
return "", err
}
if z.buf[i] > 0x7f {
needConv = true
}
if z.buf[i] == 0 {
// Digest covers the NUL terminator.
z.digest = crc32.Update(z.digest, crc32.IEEETable, z.buf[:i+1])
// Strings are ISO 8859-1, Latin-1 (RFC 1952, section 2.3.1).
if needConv {
s := make([]rune, 0, i)
for _, v := range z.buf[:i] {
s = append(s, rune(v))
}
return string(s), nil
}
return string(z.buf[:i]), nil
}
}
}
// readHeader reads the GZIP header according to section 2.3.1.
// This method does not set z.err.
func (z *Reader) readHeader() (hdr Header, err error) {
if _, err = io.ReadFull(z.r, z.buf[:10]); err != nil {
// RFC 1952, section 2.2, says the following:
// A gzip file consists of a series of "members" (compressed data sets).
//
// Other than this, the specification does not clarify whether a
// "series" is defined as "one or more" or "zero or more". To err on the
// side of caution, Go interprets this to mean "zero or more".
// Thus, it is okay to return io.EOF here.
return hdr, err
}
if z.buf[0] != gzipID1 || z.buf[1] != gzipID2 || z.buf[2] != gzipDeflate {
return hdr, ErrHeader
}
flg := z.buf[3]
hdr.ModTime = time.Unix(int64(le.Uint32(z.buf[4:8])), 0)
// z.buf[8] is XFL and is currently ignored.
hdr.OS = z.buf[9]
z.digest = crc32.ChecksumIEEE(z.buf[:10])
if flg&flagExtra != 0 {
if _, err = io.ReadFull(z.r, z.buf[:2]); err != nil {
return hdr, noEOF(err)
}
z.digest = crc32.Update(z.digest, crc32.IEEETable, z.buf[:2])
data := make([]byte, le.Uint16(z.buf[:2]))
if _, err = io.ReadFull(z.r, data); err != nil {
return hdr, noEOF(err)
}
z.digest = crc32.Update(z.digest, crc32.IEEETable, data)
hdr.Extra = data
}
var s string
if flg&flagName != 0 {
if s, err = z.readString(); err != nil {
return hdr, err
}
hdr.Name = s
}
if flg&flagComment != 0 {
if s, err = z.readString(); err != nil {
return hdr, err
}
hdr.Comment = s
}
if flg&flagHdrCrc != 0 {
if _, err = io.ReadFull(z.r, z.buf[:2]); err != nil {
return hdr, noEOF(err)
}
digest := le.Uint16(z.buf[:2])
if digest != uint16(z.digest) {
return hdr, ErrHeader
}
}
// Reserved FLG bits must be zero.
if flg>>5 != 0 {
return hdr, ErrHeader
}
z.digest = 0
if z.decompressor == nil {
z.decompressor = flate.NewReader(z.r)
} else {
z.decompressor.(flate.Resetter).Reset(z.r, nil)
}
return hdr, nil
}
// Read implements io.Reader, reading uncompressed bytes from its underlying Reader.
func (z *Reader) Read(p []byte) (n int, err error) {
if z.err != nil {
return 0, z.err
}
for n == 0 {
n, z.err = z.decompressor.Read(p)
z.digest = crc32.Update(z.digest, crc32.IEEETable, p[:n])
z.size += uint32(n)
if z.err != io.EOF {
// In the normal case we return here.
return n, z.err
}
// Finished file; check checksum and size.
if _, err := io.ReadFull(z.r, z.buf[:8]); err != nil {
z.err = noEOF(err)
return n, z.err
}
digest := le.Uint32(z.buf[:4])
size := le.Uint32(z.buf[4:8])
if digest != z.digest || size != z.size {
z.err = ErrChecksum
return n, z.err
}
z.digest, z.size = 0, 0
// File is ok; check if there is another.
if !z.multistream {
return n, io.EOF
}
z.err = nil // Remove io.EOF
if _, z.err = z.readHeader(); z.err != nil {
return n, z.err
}
}
return n, nil
}
type crcer interface {
io.Writer
Sum32() uint32
Reset()
}
type crcUpdater struct {
z *Reader
}
func (c *crcUpdater) Write(p []byte) (int, error) {
c.z.digest = crc32.Update(c.z.digest, crc32.IEEETable, p)
return len(p), nil
}
func (c *crcUpdater) Sum32() uint32 {
return c.z.digest
}
func (c *crcUpdater) Reset() {
c.z.digest = 0
}
// WriteTo support the io.WriteTo interface for io.Copy and friends.
func (z *Reader) WriteTo(w io.Writer) (int64, error) {
total := int64(0)
crcWriter := crcer(crc32.NewIEEE())
if z.digest != 0 {
crcWriter = &crcUpdater{z: z}
}
for {
if z.err != nil {
if z.err == io.EOF {
return total, nil
}
return total, z.err
}
// We write both to output and digest.
mw := io.MultiWriter(w, crcWriter)
n, err := z.decompressor.(io.WriterTo).WriteTo(mw)
total += n
z.size += uint32(n)
if err != nil {
z.err = err
return total, z.err
}
// Finished file; check checksum + size.
if _, err := io.ReadFull(z.r, z.buf[0:8]); err != nil {
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
z.err = err
return total, err
}
z.digest = crcWriter.Sum32()
digest := le.Uint32(z.buf[:4])
size := le.Uint32(z.buf[4:8])
if digest != z.digest || size != z.size {
z.err = ErrChecksum
return total, z.err
}
z.digest, z.size = 0, 0
// File is ok; check if there is another.
if !z.multistream {
return total, nil
}
crcWriter.Reset()
z.err = nil // Remove io.EOF
if _, z.err = z.readHeader(); z.err != nil {
if z.err == io.EOF {
return total, nil
}
return total, z.err
}
}
}
// Close closes the Reader. It does not close the underlying io.Reader.
// In order for the GZIP checksum to be verified, the reader must be
// fully consumed until the io.EOF.
func (z *Reader) Close() error { return z.decompressor.Close() }

290
vendor/github.com/klauspost/compress/gzip/gzip.go generated vendored Normal file
View File

@@ -0,0 +1,290 @@
// Copyright 2010 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package gzip
import (
"errors"
"fmt"
"hash/crc32"
"io"
"github.com/klauspost/compress/flate"
)
// These constants are copied from the flate package, so that code that imports
// "compress/gzip" does not also have to import "compress/flate".
const (
NoCompression = flate.NoCompression
BestSpeed = flate.BestSpeed
BestCompression = flate.BestCompression
DefaultCompression = flate.DefaultCompression
ConstantCompression = flate.ConstantCompression
HuffmanOnly = flate.HuffmanOnly
// StatelessCompression will do compression but without maintaining any state
// between Write calls.
// There will be no memory kept between Write calls,
// but compression and speed will be suboptimal.
// Because of this, the size of actual Write calls will affect output size.
StatelessCompression = -3
)
// A Writer is an io.WriteCloser.
// Writes to a Writer are compressed and written to w.
type Writer struct {
Header // written at first call to Write, Flush, or Close
w io.Writer
level int
err error
compressor *flate.Writer
digest uint32 // CRC-32, IEEE polynomial (section 8)
size uint32 // Uncompressed size (section 2.3.1)
wroteHeader bool
closed bool
buf [10]byte
}
// NewWriter returns a new Writer.
// Writes to the returned writer are compressed and written to w.
//
// It is the caller's responsibility to call Close on the WriteCloser when done.
// Writes may be buffered and not flushed until Close.
//
// Callers that wish to set the fields in Writer.Header must do so before
// the first call to Write, Flush, or Close.
func NewWriter(w io.Writer) *Writer {
z, _ := NewWriterLevel(w, DefaultCompression)
return z
}
// NewWriterLevel is like NewWriter but specifies the compression level instead
// of assuming DefaultCompression.
//
// The compression level can be DefaultCompression, NoCompression, or any
// integer value between BestSpeed and BestCompression inclusive. The error
// returned will be nil if the level is valid.
func NewWriterLevel(w io.Writer, level int) (*Writer, error) {
if level < StatelessCompression || level > BestCompression {
return nil, fmt.Errorf("gzip: invalid compression level: %d", level)
}
z := new(Writer)
z.init(w, level)
return z, nil
}
// MinCustomWindowSize is the minimum window size that can be sent to NewWriterWindow.
const MinCustomWindowSize = flate.MinCustomWindowSize
// MaxCustomWindowSize is the maximum custom window that can be sent to NewWriterWindow.
const MaxCustomWindowSize = flate.MaxCustomWindowSize
// NewWriterWindow returns a new Writer compressing data with a custom window size.
// windowSize must be from MinCustomWindowSize to MaxCustomWindowSize.
func NewWriterWindow(w io.Writer, windowSize int) (*Writer, error) {
if windowSize < MinCustomWindowSize {
return nil, errors.New("gzip: requested window size less than MinWindowSize")
}
if windowSize > MaxCustomWindowSize {
return nil, errors.New("gzip: requested window size bigger than MaxCustomWindowSize")
}
z := new(Writer)
z.init(w, -windowSize)
return z, nil
}
func (z *Writer) init(w io.Writer, level int) {
compressor := z.compressor
if level != StatelessCompression {
if compressor != nil {
compressor.Reset(w)
}
}
*z = Writer{
Header: Header{
OS: 255, // unknown
},
w: w,
level: level,
compressor: compressor,
}
}
// Reset discards the Writer z's state and makes it equivalent to the
// result of its original state from NewWriter or NewWriterLevel, but
// writing to w instead. This permits reusing a Writer rather than
// allocating a new one.
func (z *Writer) Reset(w io.Writer) {
z.init(w, z.level)
}
// writeBytes writes a length-prefixed byte slice to z.w.
func (z *Writer) writeBytes(b []byte) error {
if len(b) > 0xffff {
return errors.New("gzip.Write: Extra data is too large")
}
le.PutUint16(z.buf[:2], uint16(len(b)))
_, err := z.w.Write(z.buf[:2])
if err != nil {
return err
}
_, err = z.w.Write(b)
return err
}
// writeString writes a UTF-8 string s in GZIP's format to z.w.
// GZIP (RFC 1952) specifies that strings are NUL-terminated ISO 8859-1 (Latin-1).
func (z *Writer) writeString(s string) (err error) {
// GZIP stores Latin-1 strings; error if non-Latin-1; convert if non-ASCII.
needconv := false
for _, v := range s {
if v == 0 || v > 0xff {
return errors.New("gzip.Write: non-Latin-1 header string")
}
if v > 0x7f {
needconv = true
}
}
if needconv {
b := make([]byte, 0, len(s))
for _, v := range s {
b = append(b, byte(v))
}
_, err = z.w.Write(b)
} else {
_, err = io.WriteString(z.w, s)
}
if err != nil {
return err
}
// GZIP strings are NUL-terminated.
z.buf[0] = 0
_, err = z.w.Write(z.buf[:1])
return err
}
// Write writes a compressed form of p to the underlying io.Writer. The
// compressed bytes are not necessarily flushed until the Writer is closed.
func (z *Writer) Write(p []byte) (int, error) {
if z.err != nil {
return 0, z.err
}
var n int
// Write the GZIP header lazily.
if !z.wroteHeader {
z.wroteHeader = true
z.buf[0] = gzipID1
z.buf[1] = gzipID2
z.buf[2] = gzipDeflate
z.buf[3] = 0
if z.Extra != nil {
z.buf[3] |= 0x04
}
if z.Name != "" {
z.buf[3] |= 0x08
}
if z.Comment != "" {
z.buf[3] |= 0x10
}
le.PutUint32(z.buf[4:8], uint32(z.ModTime.Unix()))
if z.level == BestCompression {
z.buf[8] = 2
} else if z.level == BestSpeed {
z.buf[8] = 4
} else {
z.buf[8] = 0
}
z.buf[9] = z.OS
n, z.err = z.w.Write(z.buf[:10])
if z.err != nil {
return n, z.err
}
if z.Extra != nil {
z.err = z.writeBytes(z.Extra)
if z.err != nil {
return n, z.err
}
}
if z.Name != "" {
z.err = z.writeString(z.Name)
if z.err != nil {
return n, z.err
}
}
if z.Comment != "" {
z.err = z.writeString(z.Comment)
if z.err != nil {
return n, z.err
}
}
if z.compressor == nil && z.level != StatelessCompression {
z.compressor, _ = flate.NewWriter(z.w, z.level)
}
}
z.size += uint32(len(p))
z.digest = crc32.Update(z.digest, crc32.IEEETable, p)
if z.level == StatelessCompression {
return len(p), flate.StatelessDeflate(z.w, p, false, nil)
}
n, z.err = z.compressor.Write(p)
return n, z.err
}
// Flush flushes any pending compressed data to the underlying writer.
//
// It is useful mainly in compressed network protocols, to ensure that
// a remote reader has enough data to reconstruct a packet. Flush does
// not return until the data has been written. If the underlying
// writer returns an error, Flush returns that error.
//
// In the terminology of the zlib library, Flush is equivalent to Z_SYNC_FLUSH.
func (z *Writer) Flush() error {
if z.err != nil {
return z.err
}
if z.closed || z.level == StatelessCompression {
return nil
}
if !z.wroteHeader {
z.Write(nil)
if z.err != nil {
return z.err
}
}
z.err = z.compressor.Flush()
return z.err
}
// Close closes the Writer, flushing any unwritten data to the underlying
// io.Writer, but does not close the underlying io.Writer.
func (z *Writer) Close() error {
if z.err != nil {
return z.err
}
if z.closed {
return nil
}
z.closed = true
if !z.wroteHeader {
z.Write(nil)
if z.err != nil {
return z.err
}
}
if z.level == StatelessCompression {
z.err = flate.StatelessDeflate(z.w, nil, true, nil)
} else {
z.err = z.compressor.Close()
}
if z.err != nil {
return z.err
}
le.PutUint32(z.buf[:4], z.digest)
le.PutUint32(z.buf[4:8], z.size)
_, z.err = z.w.Write(z.buf[:8])
return z.err
}

1
vendor/github.com/klauspost/compress/huff0/.gitignore generated vendored Normal file
View File

@@ -0,0 +1 @@
/huff0-fuzz.zip

89
vendor/github.com/klauspost/compress/huff0/README.md generated vendored Normal file
View File

@@ -0,0 +1,89 @@
# Huff0 entropy compression
This package provides Huff0 encoding and decoding as used in zstd.
[Huff0](https://github.com/Cyan4973/FiniteStateEntropy#new-generation-entropy-coders),
a Huffman codec designed for modern CPU, featuring OoO (Out of Order) operations on multiple ALU
(Arithmetic Logic Unit), achieving extremely fast compression and decompression speeds.
This can be used for compressing input with a lot of similar input values to the smallest number of bytes.
This does not perform any multi-byte [dictionary coding](https://en.wikipedia.org/wiki/Dictionary_coder) as LZ coders,
but it can be used as a secondary step to compressors (like Snappy) that does not do entropy encoding.
* [Godoc documentation](https://godoc.org/github.com/klauspost/compress/huff0)
## News
This is used as part of the [zstandard](https://github.com/klauspost/compress/tree/master/zstd#zstd) compression and decompression package.
This ensures that most functionality is well tested.
# Usage
This package provides a low level interface that allows to compress single independent blocks.
Each block is separate, and there is no built in integrity checks.
This means that the caller should keep track of block sizes and also do checksums if needed.
Compressing a block is done via the [`Compress1X`](https://godoc.org/github.com/klauspost/compress/huff0#Compress1X) and
[`Compress4X`](https://godoc.org/github.com/klauspost/compress/huff0#Compress4X) functions.
You must provide input and will receive the output and maybe an error.
These error values can be returned:
| Error | Description |
|---------------------|-----------------------------------------------------------------------------|
| `<nil>` | Everything ok, output is returned |
| `ErrIncompressible` | Returned when input is judged to be too hard to compress |
| `ErrUseRLE` | Returned from the compressor when the input is a single byte value repeated |
| `ErrTooBig` | Returned if the input block exceeds the maximum allowed size (128 Kib) |
| `(error)` | An internal error occurred. |
As can be seen above some of there are errors that will be returned even under normal operation so it is important to handle these.
To reduce allocations you can provide a [`Scratch`](https://godoc.org/github.com/klauspost/compress/huff0#Scratch) object
that can be re-used for successive calls. Both compression and decompression accepts a `Scratch` object, and the same
object can be used for both.
Be aware, that when re-using a `Scratch` object that the *output* buffer is also re-used, so if you are still using this
you must set the `Out` field in the scratch to nil. The same buffer is used for compression and decompression output.
The `Scratch` object will retain state that allows to re-use previous tables for encoding and decoding.
## Tables and re-use
Huff0 allows for reusing tables from the previous block to save space if that is expected to give better/faster results.
The Scratch object allows you to set a [`ReusePolicy`](https://godoc.org/github.com/klauspost/compress/huff0#ReusePolicy)
that controls this behaviour. See the documentation for details. This can be altered between each block.
Do however note that this information is *not* stored in the output block and it is up to the users of the package to
record whether [`ReadTable`](https://godoc.org/github.com/klauspost/compress/huff0#ReadTable) should be called,
based on the boolean reported back from the CompressXX call.
If you want to store the table separate from the data, you can access them as `OutData` and `OutTable` on the
[`Scratch`](https://godoc.org/github.com/klauspost/compress/huff0#Scratch) object.
## Decompressing
The first part of decoding is to initialize the decoding table through [`ReadTable`](https://godoc.org/github.com/klauspost/compress/huff0#ReadTable).
This will initialize the decoding tables.
You can supply the complete block to `ReadTable` and it will return the data part of the block
which can be given to the decompressor.
Decompressing is done by calling the [`Decompress1X`](https://godoc.org/github.com/klauspost/compress/huff0#Scratch.Decompress1X)
or [`Decompress4X`](https://godoc.org/github.com/klauspost/compress/huff0#Scratch.Decompress4X) function.
For concurrently decompressing content with a fixed table a stateless [`Decoder`](https://godoc.org/github.com/klauspost/compress/huff0#Decoder) can be requested which will remain correct as long as the scratch is unchanged. The capacity of the provided slice indicates the expected output size.
You must provide the output from the compression stage, at exactly the size you got back. If you receive an error back
your input was likely corrupted.
It is important to note that a successful decoding does *not* mean your output matches your original input.
There are no integrity checks, so relying on errors from the decompressor does not assure your data is valid.
# Contributing
Contributions are always welcome. Be aware that adding public functions will require good justification and breaking
changes will likely not be accepted. If in doubt open an issue before writing the PR.

229
vendor/github.com/klauspost/compress/huff0/bitreader.go generated vendored Normal file
View File

@@ -0,0 +1,229 @@
// Copyright 2018 Klaus Post. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Based on work Copyright (c) 2013, Yann Collet, released under BSD License.
package huff0
import (
"encoding/binary"
"errors"
"fmt"
"io"
)
// bitReader reads a bitstream in reverse.
// The last set bit indicates the start of the stream and is used
// for aligning the input.
type bitReaderBytes struct {
in []byte
off uint // next byte to read is at in[off - 1]
value uint64
bitsRead uint8
}
// init initializes and resets the bit reader.
func (b *bitReaderBytes) init(in []byte) error {
if len(in) < 1 {
return errors.New("corrupt stream: too short")
}
b.in = in
b.off = uint(len(in))
// The highest bit of the last byte indicates where to start
v := in[len(in)-1]
if v == 0 {
return errors.New("corrupt stream, did not find end of stream")
}
b.bitsRead = 64
b.value = 0
if len(in) >= 8 {
b.fillFastStart()
} else {
b.fill()
b.fill()
}
b.advance(8 - uint8(highBit32(uint32(v))))
return nil
}
// peekBitsFast requires that at least one bit is requested every time.
// There are no checks if the buffer is filled.
func (b *bitReaderBytes) peekByteFast() uint8 {
got := uint8(b.value >> 56)
return got
}
func (b *bitReaderBytes) advance(n uint8) {
b.bitsRead += n
b.value <<= n & 63
}
// fillFast() will make sure at least 32 bits are available.
// There must be at least 4 bytes available.
func (b *bitReaderBytes) fillFast() {
if b.bitsRead < 32 {
return
}
// 2 bounds checks.
v := b.in[b.off-4 : b.off]
low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
b.value |= uint64(low) << (b.bitsRead - 32)
b.bitsRead -= 32
b.off -= 4
}
// fillFastStart() assumes the bitReaderBytes is empty and there is at least 8 bytes to read.
func (b *bitReaderBytes) fillFastStart() {
// Do single re-slice to avoid bounds checks.
b.value = binary.LittleEndian.Uint64(b.in[b.off-8:])
b.bitsRead = 0
b.off -= 8
}
// fill() will make sure at least 32 bits are available.
func (b *bitReaderBytes) fill() {
if b.bitsRead < 32 {
return
}
if b.off > 4 {
v := b.in[b.off-4 : b.off]
low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
b.value |= uint64(low) << (b.bitsRead - 32)
b.bitsRead -= 32
b.off -= 4
return
}
for b.off > 0 {
b.value |= uint64(b.in[b.off-1]) << (b.bitsRead - 8)
b.bitsRead -= 8
b.off--
}
}
// finished returns true if all bits have been read from the bit stream.
func (b *bitReaderBytes) finished() bool {
return b.off == 0 && b.bitsRead >= 64
}
func (b *bitReaderBytes) remaining() uint {
return b.off*8 + uint(64-b.bitsRead)
}
// close the bitstream and returns an error if out-of-buffer reads occurred.
func (b *bitReaderBytes) close() error {
// Release reference.
b.in = nil
if b.remaining() > 0 {
return fmt.Errorf("corrupt input: %d bits remain on stream", b.remaining())
}
if b.bitsRead > 64 {
return io.ErrUnexpectedEOF
}
return nil
}
// bitReaderShifted reads a bitstream in reverse.
// The last set bit indicates the start of the stream and is used
// for aligning the input.
type bitReaderShifted struct {
in []byte
off uint // next byte to read is at in[off - 1]
value uint64
bitsRead uint8
}
// init initializes and resets the bit reader.
func (b *bitReaderShifted) init(in []byte) error {
if len(in) < 1 {
return errors.New("corrupt stream: too short")
}
b.in = in
b.off = uint(len(in))
// The highest bit of the last byte indicates where to start
v := in[len(in)-1]
if v == 0 {
return errors.New("corrupt stream, did not find end of stream")
}
b.bitsRead = 64
b.value = 0
if len(in) >= 8 {
b.fillFastStart()
} else {
b.fill()
b.fill()
}
b.advance(8 - uint8(highBit32(uint32(v))))
return nil
}
// peekBitsFast requires that at least one bit is requested every time.
// There are no checks if the buffer is filled.
func (b *bitReaderShifted) peekBitsFast(n uint8) uint16 {
return uint16(b.value >> ((64 - n) & 63))
}
func (b *bitReaderShifted) advance(n uint8) {
b.bitsRead += n
b.value <<= n & 63
}
// fillFast() will make sure at least 32 bits are available.
// There must be at least 4 bytes available.
func (b *bitReaderShifted) fillFast() {
if b.bitsRead < 32 {
return
}
// 2 bounds checks.
v := b.in[b.off-4 : b.off]
low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
b.value |= uint64(low) << ((b.bitsRead - 32) & 63)
b.bitsRead -= 32
b.off -= 4
}
// fillFastStart() assumes the bitReaderShifted is empty and there is at least 8 bytes to read.
func (b *bitReaderShifted) fillFastStart() {
// Do single re-slice to avoid bounds checks.
b.value = binary.LittleEndian.Uint64(b.in[b.off-8:])
b.bitsRead = 0
b.off -= 8
}
// fill() will make sure at least 32 bits are available.
func (b *bitReaderShifted) fill() {
if b.bitsRead < 32 {
return
}
if b.off > 4 {
v := b.in[b.off-4 : b.off]
low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
b.value |= uint64(low) << ((b.bitsRead - 32) & 63)
b.bitsRead -= 32
b.off -= 4
return
}
for b.off > 0 {
b.value |= uint64(b.in[b.off-1]) << ((b.bitsRead - 8) & 63)
b.bitsRead -= 8
b.off--
}
}
func (b *bitReaderShifted) remaining() uint {
return b.off*8 + uint(64-b.bitsRead)
}
// close the bitstream and returns an error if out-of-buffer reads occurred.
func (b *bitReaderShifted) close() error {
// Release reference.
b.in = nil
if b.remaining() > 0 {
return fmt.Errorf("corrupt input: %d bits remain on stream", b.remaining())
}
if b.bitsRead > 64 {
return io.ErrUnexpectedEOF
}
return nil
}

102
vendor/github.com/klauspost/compress/huff0/bitwriter.go generated vendored Normal file
View File

@@ -0,0 +1,102 @@
// Copyright 2018 Klaus Post. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Based on work Copyright (c) 2013, Yann Collet, released under BSD License.
package huff0
// bitWriter will write bits.
// First bit will be LSB of the first byte of output.
type bitWriter struct {
bitContainer uint64
nBits uint8
out []byte
}
// addBits16Clean will add up to 16 bits. value may not contain more set bits than indicated.
// It will not check if there is space for them, so the caller must ensure that it has flushed recently.
func (b *bitWriter) addBits16Clean(value uint16, bits uint8) {
b.bitContainer |= uint64(value) << (b.nBits & 63)
b.nBits += bits
}
// encSymbol will add up to 16 bits. value may not contain more set bits than indicated.
// It will not check if there is space for them, so the caller must ensure that it has flushed recently.
func (b *bitWriter) encSymbol(ct cTable, symbol byte) {
enc := ct[symbol]
b.bitContainer |= uint64(enc.val) << (b.nBits & 63)
if false {
if enc.nBits == 0 {
panic("nbits 0")
}
}
b.nBits += enc.nBits
}
// encTwoSymbols will add up to 32 bits. value may not contain more set bits than indicated.
// It will not check if there is space for them, so the caller must ensure that it has flushed recently.
func (b *bitWriter) encTwoSymbols(ct cTable, av, bv byte) {
encA := ct[av]
encB := ct[bv]
sh := b.nBits & 63
combined := uint64(encA.val) | (uint64(encB.val) << (encA.nBits & 63))
b.bitContainer |= combined << sh
if false {
if encA.nBits == 0 {
panic("nbitsA 0")
}
if encB.nBits == 0 {
panic("nbitsB 0")
}
}
b.nBits += encA.nBits + encB.nBits
}
// encFourSymbols adds up to 32 bits from four symbols.
// It will not check if there is space for them,
// so the caller must ensure that b has been flushed recently.
func (b *bitWriter) encFourSymbols(encA, encB, encC, encD cTableEntry) {
bitsA := encA.nBits
bitsB := bitsA + encB.nBits
bitsC := bitsB + encC.nBits
bitsD := bitsC + encD.nBits
combined := uint64(encA.val) |
(uint64(encB.val) << (bitsA & 63)) |
(uint64(encC.val) << (bitsB & 63)) |
(uint64(encD.val) << (bitsC & 63))
b.bitContainer |= combined << (b.nBits & 63)
b.nBits += bitsD
}
// flush32 will flush out, so there are at least 32 bits available for writing.
func (b *bitWriter) flush32() {
if b.nBits < 32 {
return
}
b.out = append(b.out,
byte(b.bitContainer),
byte(b.bitContainer>>8),
byte(b.bitContainer>>16),
byte(b.bitContainer>>24))
b.nBits -= 32
b.bitContainer >>= 32
}
// flushAlign will flush remaining full bytes and align to next byte boundary.
func (b *bitWriter) flushAlign() {
nbBytes := (b.nBits + 7) >> 3
for i := uint8(0); i < nbBytes; i++ {
b.out = append(b.out, byte(b.bitContainer>>(i*8)))
}
b.nBits = 0
b.bitContainer = 0
}
// close will write the alignment bit and write the final byte(s)
// to the output.
func (b *bitWriter) close() {
// End mark
b.addBits16Clean(1, 1)
// flush until next byte.
b.flushAlign()
}

742
vendor/github.com/klauspost/compress/huff0/compress.go generated vendored Normal file
View File

@@ -0,0 +1,742 @@
package huff0
import (
"fmt"
"math"
"runtime"
"sync"
)
// Compress1X will compress the input.
// The output can be decoded using Decompress1X.
// Supply a Scratch object. The scratch object contains state about re-use,
// So when sharing across independent encodes, be sure to set the re-use policy.
func Compress1X(in []byte, s *Scratch) (out []byte, reUsed bool, err error) {
s, err = s.prepare(in)
if err != nil {
return nil, false, err
}
return compress(in, s, s.compress1X)
}
// Compress4X will compress the input. The input is split into 4 independent blocks
// and compressed similar to Compress1X.
// The output can be decoded using Decompress4X.
// Supply a Scratch object. The scratch object contains state about re-use,
// So when sharing across independent encodes, be sure to set the re-use policy.
func Compress4X(in []byte, s *Scratch) (out []byte, reUsed bool, err error) {
s, err = s.prepare(in)
if err != nil {
return nil, false, err
}
if false {
// TODO: compress4Xp only slightly faster.
const parallelThreshold = 8 << 10
if len(in) < parallelThreshold || runtime.GOMAXPROCS(0) == 1 {
return compress(in, s, s.compress4X)
}
return compress(in, s, s.compress4Xp)
}
return compress(in, s, s.compress4X)
}
func compress(in []byte, s *Scratch, compressor func(src []byte) ([]byte, error)) (out []byte, reUsed bool, err error) {
// Nuke previous table if we cannot reuse anyway.
if s.Reuse == ReusePolicyNone {
s.prevTable = s.prevTable[:0]
}
// Create histogram, if none was provided.
maxCount := s.maxCount
var canReuse = false
if maxCount == 0 {
maxCount, canReuse = s.countSimple(in)
} else {
canReuse = s.canUseTable(s.prevTable)
}
// We want the output size to be less than this:
wantSize := len(in)
if s.WantLogLess > 0 {
wantSize -= wantSize >> s.WantLogLess
}
// Reset for next run.
s.clearCount = true
s.maxCount = 0
if maxCount >= len(in) {
if maxCount > len(in) {
return nil, false, fmt.Errorf("maxCount (%d) > length (%d)", maxCount, len(in))
}
if len(in) == 1 {
return nil, false, ErrIncompressible
}
// One symbol, use RLE
return nil, false, ErrUseRLE
}
if maxCount == 1 || maxCount < (len(in)>>7) {
// Each symbol present maximum once or too well distributed.
return nil, false, ErrIncompressible
}
if s.Reuse == ReusePolicyMust && !canReuse {
// We must reuse, but we can't.
return nil, false, ErrIncompressible
}
if (s.Reuse == ReusePolicyPrefer || s.Reuse == ReusePolicyMust) && canReuse {
keepTable := s.cTable
keepTL := s.actualTableLog
s.cTable = s.prevTable
s.actualTableLog = s.prevTableLog
s.Out, err = compressor(in)
s.cTable = keepTable
s.actualTableLog = keepTL
if err == nil && len(s.Out) < wantSize {
s.OutData = s.Out
return s.Out, true, nil
}
if s.Reuse == ReusePolicyMust {
return nil, false, ErrIncompressible
}
// Do not attempt to re-use later.
s.prevTable = s.prevTable[:0]
}
// Calculate new table.
err = s.buildCTable()
if err != nil {
return nil, false, err
}
if false && !s.canUseTable(s.cTable) {
panic("invalid table generated")
}
if s.Reuse == ReusePolicyAllow && canReuse {
hSize := len(s.Out)
oldSize := s.prevTable.estimateSize(s.count[:s.symbolLen])
newSize := s.cTable.estimateSize(s.count[:s.symbolLen])
if oldSize <= hSize+newSize || hSize+12 >= wantSize {
// Retain cTable even if we re-use.
keepTable := s.cTable
keepTL := s.actualTableLog
s.cTable = s.prevTable
s.actualTableLog = s.prevTableLog
s.Out, err = compressor(in)
// Restore ctable.
s.cTable = keepTable
s.actualTableLog = keepTL
if err != nil {
return nil, false, err
}
if len(s.Out) >= wantSize {
return nil, false, ErrIncompressible
}
s.OutData = s.Out
return s.Out, true, nil
}
}
// Use new table
err = s.cTable.write(s)
if err != nil {
s.OutTable = nil
return nil, false, err
}
s.OutTable = s.Out
// Compress using new table
s.Out, err = compressor(in)
if err != nil {
s.OutTable = nil
return nil, false, err
}
if len(s.Out) >= wantSize {
s.OutTable = nil
return nil, false, ErrIncompressible
}
// Move current table into previous.
s.prevTable, s.prevTableLog, s.cTable = s.cTable, s.actualTableLog, s.prevTable[:0]
s.OutData = s.Out[len(s.OutTable):]
return s.Out, false, nil
}
// EstimateSizes will estimate the data sizes
func EstimateSizes(in []byte, s *Scratch) (tableSz, dataSz, reuseSz int, err error) {
s, err = s.prepare(in)
if err != nil {
return 0, 0, 0, err
}
// Create histogram, if none was provided.
tableSz, dataSz, reuseSz = -1, -1, -1
maxCount := s.maxCount
var canReuse = false
if maxCount == 0 {
maxCount, canReuse = s.countSimple(in)
} else {
canReuse = s.canUseTable(s.prevTable)
}
// We want the output size to be less than this:
wantSize := len(in)
if s.WantLogLess > 0 {
wantSize -= wantSize >> s.WantLogLess
}
// Reset for next run.
s.clearCount = true
s.maxCount = 0
if maxCount >= len(in) {
if maxCount > len(in) {
return 0, 0, 0, fmt.Errorf("maxCount (%d) > length (%d)", maxCount, len(in))
}
if len(in) == 1 {
return 0, 0, 0, ErrIncompressible
}
// One symbol, use RLE
return 0, 0, 0, ErrUseRLE
}
if maxCount == 1 || maxCount < (len(in)>>7) {
// Each symbol present maximum once or too well distributed.
return 0, 0, 0, ErrIncompressible
}
// Calculate new table.
err = s.buildCTable()
if err != nil {
return 0, 0, 0, err
}
if false && !s.canUseTable(s.cTable) {
panic("invalid table generated")
}
tableSz, err = s.cTable.estTableSize(s)
if err != nil {
return 0, 0, 0, err
}
if canReuse {
reuseSz = s.prevTable.estimateSize(s.count[:s.symbolLen])
}
dataSz = s.cTable.estimateSize(s.count[:s.symbolLen])
// Restore
return tableSz, dataSz, reuseSz, nil
}
func (s *Scratch) compress1X(src []byte) ([]byte, error) {
return s.compress1xDo(s.Out, src), nil
}
func (s *Scratch) compress1xDo(dst, src []byte) []byte {
var bw = bitWriter{out: dst}
// N is length divisible by 4.
n := len(src)
n -= n & 3
cTable := s.cTable[:256]
// Encode last bytes.
for i := len(src) & 3; i > 0; i-- {
bw.encSymbol(cTable, src[n+i-1])
}
n -= 4
if s.actualTableLog <= 8 {
for ; n >= 0; n -= 4 {
tmp := src[n : n+4]
// tmp should be len 4
bw.flush32()
bw.encFourSymbols(cTable[tmp[3]], cTable[tmp[2]], cTable[tmp[1]], cTable[tmp[0]])
}
} else {
for ; n >= 0; n -= 4 {
tmp := src[n : n+4]
// tmp should be len 4
bw.flush32()
bw.encTwoSymbols(cTable, tmp[3], tmp[2])
bw.flush32()
bw.encTwoSymbols(cTable, tmp[1], tmp[0])
}
}
bw.close()
return bw.out
}
var sixZeros [6]byte
func (s *Scratch) compress4X(src []byte) ([]byte, error) {
if len(src) < 12 {
return nil, ErrIncompressible
}
segmentSize := (len(src) + 3) / 4
// Add placeholder for output length
offsetIdx := len(s.Out)
s.Out = append(s.Out, sixZeros[:]...)
for i := 0; i < 4; i++ {
toDo := src
if len(toDo) > segmentSize {
toDo = toDo[:segmentSize]
}
src = src[len(toDo):]
idx := len(s.Out)
s.Out = s.compress1xDo(s.Out, toDo)
if len(s.Out)-idx > math.MaxUint16 {
// We cannot store the size in the jump table
return nil, ErrIncompressible
}
// Write compressed length as little endian before block.
if i < 3 {
// Last length is not written.
length := len(s.Out) - idx
s.Out[i*2+offsetIdx] = byte(length)
s.Out[i*2+offsetIdx+1] = byte(length >> 8)
}
}
return s.Out, nil
}
// compress4Xp will compress 4 streams using separate goroutines.
func (s *Scratch) compress4Xp(src []byte) ([]byte, error) {
if len(src) < 12 {
return nil, ErrIncompressible
}
// Add placeholder for output length
s.Out = s.Out[:6]
segmentSize := (len(src) + 3) / 4
var wg sync.WaitGroup
wg.Add(4)
for i := 0; i < 4; i++ {
toDo := src
if len(toDo) > segmentSize {
toDo = toDo[:segmentSize]
}
src = src[len(toDo):]
// Separate goroutine for each block.
go func(i int) {
s.tmpOut[i] = s.compress1xDo(s.tmpOut[i][:0], toDo)
wg.Done()
}(i)
}
wg.Wait()
for i := 0; i < 4; i++ {
o := s.tmpOut[i]
if len(o) > math.MaxUint16 {
// We cannot store the size in the jump table
return nil, ErrIncompressible
}
// Write compressed length as little endian before block.
if i < 3 {
// Last length is not written.
s.Out[i*2] = byte(len(o))
s.Out[i*2+1] = byte(len(o) >> 8)
}
// Write output.
s.Out = append(s.Out, o...)
}
return s.Out, nil
}
// countSimple will create a simple histogram in s.count.
// Returns the biggest count.
// Does not update s.clearCount.
func (s *Scratch) countSimple(in []byte) (max int, reuse bool) {
reuse = true
_ = s.count // Assert that s != nil to speed up the following loop.
for _, v := range in {
s.count[v]++
}
m := uint32(0)
if len(s.prevTable) > 0 {
for i, v := range s.count[:] {
if v == 0 {
continue
}
if v > m {
m = v
}
s.symbolLen = uint16(i) + 1
if i >= len(s.prevTable) {
reuse = false
} else if s.prevTable[i].nBits == 0 {
reuse = false
}
}
return int(m), reuse
}
for i, v := range s.count[:] {
if v == 0 {
continue
}
if v > m {
m = v
}
s.symbolLen = uint16(i) + 1
}
return int(m), false
}
func (s *Scratch) canUseTable(c cTable) bool {
if len(c) < int(s.symbolLen) {
return false
}
for i, v := range s.count[:s.symbolLen] {
if v != 0 && c[i].nBits == 0 {
return false
}
}
return true
}
//lint:ignore U1000 used for debugging
func (s *Scratch) validateTable(c cTable) bool {
if len(c) < int(s.symbolLen) {
return false
}
for i, v := range s.count[:s.symbolLen] {
if v != 0 {
if c[i].nBits == 0 {
return false
}
if c[i].nBits > s.actualTableLog {
return false
}
}
}
return true
}
// minTableLog provides the minimum logSize to safely represent a distribution.
func (s *Scratch) minTableLog() uint8 {
minBitsSrc := highBit32(uint32(s.srcLen)) + 1
minBitsSymbols := highBit32(uint32(s.symbolLen-1)) + 2
if minBitsSrc < minBitsSymbols {
return uint8(minBitsSrc)
}
return uint8(minBitsSymbols)
}
// optimalTableLog calculates and sets the optimal tableLog in s.actualTableLog
func (s *Scratch) optimalTableLog() {
tableLog := s.TableLog
minBits := s.minTableLog()
maxBitsSrc := uint8(highBit32(uint32(s.srcLen-1))) - 1
if maxBitsSrc < tableLog {
// Accuracy can be reduced
tableLog = maxBitsSrc
}
if minBits > tableLog {
tableLog = minBits
}
// Need a minimum to safely represent all symbol values
if tableLog < minTablelog {
tableLog = minTablelog
}
if tableLog > tableLogMax {
tableLog = tableLogMax
}
s.actualTableLog = tableLog
}
type cTableEntry struct {
val uint16
nBits uint8
// We have 8 bits extra
}
const huffNodesMask = huffNodesLen - 1
func (s *Scratch) buildCTable() error {
s.optimalTableLog()
s.huffSort()
if cap(s.cTable) < maxSymbolValue+1 {
s.cTable = make([]cTableEntry, s.symbolLen, maxSymbolValue+1)
} else {
s.cTable = s.cTable[:s.symbolLen]
for i := range s.cTable {
s.cTable[i] = cTableEntry{}
}
}
var startNode = int16(s.symbolLen)
nonNullRank := s.symbolLen - 1
nodeNb := startNode
huffNode := s.nodes[1 : huffNodesLen+1]
// This overlays the slice above, but allows "-1" index lookups.
// Different from reference implementation.
huffNode0 := s.nodes[0 : huffNodesLen+1]
for huffNode[nonNullRank].count() == 0 {
nonNullRank--
}
lowS := int16(nonNullRank)
nodeRoot := nodeNb + lowS - 1
lowN := nodeNb
huffNode[nodeNb].setCount(huffNode[lowS].count() + huffNode[lowS-1].count())
huffNode[lowS].setParent(nodeNb)
huffNode[lowS-1].setParent(nodeNb)
nodeNb++
lowS -= 2
for n := nodeNb; n <= nodeRoot; n++ {
huffNode[n].setCount(1 << 30)
}
// fake entry, strong barrier
huffNode0[0].setCount(1 << 31)
// create parents
for nodeNb <= nodeRoot {
var n1, n2 int16
if huffNode0[lowS+1].count() < huffNode0[lowN+1].count() {
n1 = lowS
lowS--
} else {
n1 = lowN
lowN++
}
if huffNode0[lowS+1].count() < huffNode0[lowN+1].count() {
n2 = lowS
lowS--
} else {
n2 = lowN
lowN++
}
huffNode[nodeNb].setCount(huffNode0[n1+1].count() + huffNode0[n2+1].count())
huffNode0[n1+1].setParent(nodeNb)
huffNode0[n2+1].setParent(nodeNb)
nodeNb++
}
// distribute weights (unlimited tree height)
huffNode[nodeRoot].setNbBits(0)
for n := nodeRoot - 1; n >= startNode; n-- {
huffNode[n].setNbBits(huffNode[huffNode[n].parent()].nbBits() + 1)
}
for n := uint16(0); n <= nonNullRank; n++ {
huffNode[n].setNbBits(huffNode[huffNode[n].parent()].nbBits() + 1)
}
s.actualTableLog = s.setMaxHeight(int(nonNullRank))
maxNbBits := s.actualTableLog
// fill result into tree (val, nbBits)
if maxNbBits > tableLogMax {
return fmt.Errorf("internal error: maxNbBits (%d) > tableLogMax (%d)", maxNbBits, tableLogMax)
}
var nbPerRank [tableLogMax + 1]uint16
var valPerRank [16]uint16
for _, v := range huffNode[:nonNullRank+1] {
nbPerRank[v.nbBits()]++
}
// determine stating value per rank
{
min := uint16(0)
for n := maxNbBits; n > 0; n-- {
// get starting value within each rank
valPerRank[n] = min
min += nbPerRank[n]
min >>= 1
}
}
// push nbBits per symbol, symbol order
for _, v := range huffNode[:nonNullRank+1] {
s.cTable[v.symbol()].nBits = v.nbBits()
}
// assign value within rank, symbol order
t := s.cTable[:s.symbolLen]
for n, val := range t {
nbits := val.nBits & 15
v := valPerRank[nbits]
t[n].val = v
valPerRank[nbits] = v + 1
}
return nil
}
// huffSort will sort symbols, decreasing order.
func (s *Scratch) huffSort() {
type rankPos struct {
base uint32
current uint32
}
// Clear nodes
nodes := s.nodes[:huffNodesLen+1]
s.nodes = nodes
nodes = nodes[1 : huffNodesLen+1]
// Sort into buckets based on length of symbol count.
var rank [32]rankPos
for _, v := range s.count[:s.symbolLen] {
r := highBit32(v+1) & 31
rank[r].base++
}
// maxBitLength is log2(BlockSizeMax) + 1
const maxBitLength = 18 + 1
for n := maxBitLength; n > 0; n-- {
rank[n-1].base += rank[n].base
}
for n := range rank[:maxBitLength] {
rank[n].current = rank[n].base
}
for n, c := range s.count[:s.symbolLen] {
r := (highBit32(c+1) + 1) & 31
pos := rank[r].current
rank[r].current++
prev := nodes[(pos-1)&huffNodesMask]
for pos > rank[r].base && c > prev.count() {
nodes[pos&huffNodesMask] = prev
pos--
prev = nodes[(pos-1)&huffNodesMask]
}
nodes[pos&huffNodesMask] = makeNodeElt(c, byte(n))
}
}
func (s *Scratch) setMaxHeight(lastNonNull int) uint8 {
maxNbBits := s.actualTableLog
huffNode := s.nodes[1 : huffNodesLen+1]
//huffNode = huffNode[: huffNodesLen]
largestBits := huffNode[lastNonNull].nbBits()
// early exit : no elt > maxNbBits
if largestBits <= maxNbBits {
return largestBits
}
totalCost := int(0)
baseCost := int(1) << (largestBits - maxNbBits)
n := uint32(lastNonNull)
for huffNode[n].nbBits() > maxNbBits {
totalCost += baseCost - (1 << (largestBits - huffNode[n].nbBits()))
huffNode[n].setNbBits(maxNbBits)
n--
}
// n stops at huffNode[n].nbBits <= maxNbBits
for huffNode[n].nbBits() == maxNbBits {
n--
}
// n end at index of smallest symbol using < maxNbBits
// renorm totalCost
totalCost >>= largestBits - maxNbBits /* note : totalCost is necessarily a multiple of baseCost */
// repay normalized cost
{
const noSymbol = 0xF0F0F0F0
var rankLast [tableLogMax + 2]uint32
for i := range rankLast[:] {
rankLast[i] = noSymbol
}
// Get pos of last (smallest) symbol per rank
{
currentNbBits := maxNbBits
for pos := int(n); pos >= 0; pos-- {
if huffNode[pos].nbBits() >= currentNbBits {
continue
}
currentNbBits = huffNode[pos].nbBits() // < maxNbBits
rankLast[maxNbBits-currentNbBits] = uint32(pos)
}
}
for totalCost > 0 {
nBitsToDecrease := uint8(highBit32(uint32(totalCost))) + 1
for ; nBitsToDecrease > 1; nBitsToDecrease-- {
highPos := rankLast[nBitsToDecrease]
lowPos := rankLast[nBitsToDecrease-1]
if highPos == noSymbol {
continue
}
if lowPos == noSymbol {
break
}
highTotal := huffNode[highPos].count()
lowTotal := 2 * huffNode[lowPos].count()
if highTotal <= lowTotal {
break
}
}
// only triggered when no more rank 1 symbol left => find closest one (note : there is necessarily at least one !)
// HUF_MAX_TABLELOG test just to please gcc 5+; but it should not be necessary
// FIXME: try to remove
for (nBitsToDecrease <= tableLogMax) && (rankLast[nBitsToDecrease] == noSymbol) {
nBitsToDecrease++
}
totalCost -= 1 << (nBitsToDecrease - 1)
if rankLast[nBitsToDecrease-1] == noSymbol {
// this rank is no longer empty
rankLast[nBitsToDecrease-1] = rankLast[nBitsToDecrease]
}
huffNode[rankLast[nBitsToDecrease]].setNbBits(1 +
huffNode[rankLast[nBitsToDecrease]].nbBits())
if rankLast[nBitsToDecrease] == 0 {
/* special case, reached largest symbol */
rankLast[nBitsToDecrease] = noSymbol
} else {
rankLast[nBitsToDecrease]--
if huffNode[rankLast[nBitsToDecrease]].nbBits() != maxNbBits-nBitsToDecrease {
rankLast[nBitsToDecrease] = noSymbol /* this rank is now empty */
}
}
}
for totalCost < 0 { /* Sometimes, cost correction overshoot */
if rankLast[1] == noSymbol { /* special case : no rank 1 symbol (using maxNbBits-1); let's create one from largest rank 0 (using maxNbBits) */
for huffNode[n].nbBits() == maxNbBits {
n--
}
huffNode[n+1].setNbBits(huffNode[n+1].nbBits() - 1)
rankLast[1] = n + 1
totalCost++
continue
}
huffNode[rankLast[1]+1].setNbBits(huffNode[rankLast[1]+1].nbBits() - 1)
rankLast[1]++
totalCost++
}
}
return maxNbBits
}
// A nodeElt is the fields
//
// count uint32
// parent uint16
// symbol byte
// nbBits uint8
//
// in some order, all squashed into an integer so that the compiler
// always loads and stores entire nodeElts instead of separate fields.
type nodeElt uint64
func makeNodeElt(count uint32, symbol byte) nodeElt {
return nodeElt(count) | nodeElt(symbol)<<48
}
func (e *nodeElt) count() uint32 { return uint32(*e) }
func (e *nodeElt) parent() uint16 { return uint16(*e >> 32) }
func (e *nodeElt) symbol() byte { return byte(*e >> 48) }
func (e *nodeElt) nbBits() uint8 { return uint8(*e >> 56) }
func (e *nodeElt) setCount(c uint32) { *e = (*e)&0xffffffff00000000 | nodeElt(c) }
func (e *nodeElt) setParent(p int16) { *e = (*e)&0xffff0000ffffffff | nodeElt(uint16(p))<<32 }
func (e *nodeElt) setNbBits(n uint8) { *e = (*e)&0x00ffffffffffffff | nodeElt(n)<<56 }

1167
vendor/github.com/klauspost/compress/huff0/decompress.go generated vendored Normal file
View File

File diff suppressed because it is too large Load Diff

226
vendor/github.com/klauspost/compress/huff0/decompress_amd64.go generated vendored Normal file
View File

@@ -0,0 +1,226 @@
//go:build amd64 && !appengine && !noasm && gc
// +build amd64,!appengine,!noasm,gc
// This file contains the specialisation of Decoder.Decompress4X
// and Decoder.Decompress1X that use an asm implementation of thir main loops.
package huff0
import (
"errors"
"fmt"
"github.com/klauspost/compress/internal/cpuinfo"
)
// decompress4x_main_loop_x86 is an x86 assembler implementation
// of Decompress4X when tablelog > 8.
//
//go:noescape
func decompress4x_main_loop_amd64(ctx *decompress4xContext)
// decompress4x_8b_loop_x86 is an x86 assembler implementation
// of Decompress4X when tablelog <= 8 which decodes 4 entries
// per loop.
//
//go:noescape
func decompress4x_8b_main_loop_amd64(ctx *decompress4xContext)
// fallback8BitSize is the size where using Go version is faster.
const fallback8BitSize = 800
type decompress4xContext struct {
pbr *[4]bitReaderShifted
peekBits uint8
out *byte
dstEvery int
tbl *dEntrySingle
decoded int
limit *byte
}
// Decompress4X will decompress a 4X encoded stream.
// The length of the supplied input must match the end of a block exactly.
// The *capacity* of the dst slice must match the destination size of
// the uncompressed data exactly.
func (d *Decoder) Decompress4X(dst, src []byte) ([]byte, error) {
if len(d.dt.single) == 0 {
return nil, errors.New("no table loaded")
}
if len(src) < 6+(4*1) {
return nil, errors.New("input too small")
}
use8BitTables := d.actualTableLog <= 8
if cap(dst) < fallback8BitSize && use8BitTables {
return d.decompress4X8bit(dst, src)
}
var br [4]bitReaderShifted
// Decode "jump table"
start := 6
for i := 0; i < 3; i++ {
length := int(src[i*2]) | (int(src[i*2+1]) << 8)
if start+length >= len(src) {
return nil, errors.New("truncated input (or invalid offset)")
}
err := br[i].init(src[start : start+length])
if err != nil {
return nil, err
}
start += length
}
err := br[3].init(src[start:])
if err != nil {
return nil, err
}
// destination, offset to match first output
dstSize := cap(dst)
dst = dst[:dstSize]
out := dst
dstEvery := (dstSize + 3) / 4
const tlSize = 1 << tableLogMax
const tlMask = tlSize - 1
single := d.dt.single[:tlSize]
var decoded int
if len(out) > 4*4 && !(br[0].off < 4 || br[1].off < 4 || br[2].off < 4 || br[3].off < 4) {
ctx := decompress4xContext{
pbr: &br,
peekBits: uint8((64 - d.actualTableLog) & 63), // see: bitReaderShifted.peekBitsFast()
out: &out[0],
dstEvery: dstEvery,
tbl: &single[0],
limit: &out[dstEvery-4], // Always stop decoding when first buffer gets here to avoid writing OOB on last.
}
if use8BitTables {
decompress4x_8b_main_loop_amd64(&ctx)
} else {
decompress4x_main_loop_amd64(&ctx)
}
decoded = ctx.decoded
out = out[decoded/4:]
}
// Decode remaining.
remainBytes := dstEvery - (decoded / 4)
for i := range br {
offset := dstEvery * i
endsAt := offset + remainBytes
if endsAt > len(out) {
endsAt = len(out)
}
br := &br[i]
bitsLeft := br.remaining()
for bitsLeft > 0 {
br.fill()
if offset >= endsAt {
return nil, errors.New("corruption detected: stream overrun 4")
}
// Read value and increment offset.
val := br.peekBitsFast(d.actualTableLog)
v := single[val&tlMask].entry
nBits := uint8(v)
br.advance(nBits)
bitsLeft -= uint(nBits)
out[offset] = uint8(v >> 8)
offset++
}
if offset != endsAt {
return nil, fmt.Errorf("corruption detected: short output block %d, end %d != %d", i, offset, endsAt)
}
decoded += offset - dstEvery*i
err = br.close()
if err != nil {
return nil, err
}
}
if dstSize != decoded {
return nil, errors.New("corruption detected: short output block")
}
return dst, nil
}
// decompress4x_main_loop_x86 is an x86 assembler implementation
// of Decompress1X when tablelog > 8.
//
//go:noescape
func decompress1x_main_loop_amd64(ctx *decompress1xContext)
// decompress4x_main_loop_x86 is an x86 with BMI2 assembler implementation
// of Decompress1X when tablelog > 8.
//
//go:noescape
func decompress1x_main_loop_bmi2(ctx *decompress1xContext)
type decompress1xContext struct {
pbr *bitReaderShifted
peekBits uint8
out *byte
outCap int
tbl *dEntrySingle
decoded int
}
// Error reported by asm implementations
const error_max_decoded_size_exeeded = -1
// Decompress1X will decompress a 1X encoded stream.
// The cap of the output buffer will be the maximum decompressed size.
// The length of the supplied input must match the end of a block exactly.
func (d *Decoder) Decompress1X(dst, src []byte) ([]byte, error) {
if len(d.dt.single) == 0 {
return nil, errors.New("no table loaded")
}
var br bitReaderShifted
err := br.init(src)
if err != nil {
return dst, err
}
maxDecodedSize := cap(dst)
dst = dst[:maxDecodedSize]
const tlSize = 1 << tableLogMax
const tlMask = tlSize - 1
if maxDecodedSize >= 4 {
ctx := decompress1xContext{
pbr: &br,
out: &dst[0],
outCap: maxDecodedSize,
peekBits: uint8((64 - d.actualTableLog) & 63), // see: bitReaderShifted.peekBitsFast()
tbl: &d.dt.single[0],
}
if cpuinfo.HasBMI2() {
decompress1x_main_loop_bmi2(&ctx)
} else {
decompress1x_main_loop_amd64(&ctx)
}
if ctx.decoded == error_max_decoded_size_exeeded {
return nil, ErrMaxDecodedSizeExceeded
}
dst = dst[:ctx.decoded]
}
// br < 8, so uint8 is fine
bitsLeft := uint8(br.off)*8 + 64 - br.bitsRead
for bitsLeft > 0 {
br.fill()
if len(dst) >= maxDecodedSize {
br.close()
return nil, ErrMaxDecodedSizeExceeded
}
v := d.dt.single[br.peekBitsFast(d.actualTableLog)&tlMask]
nBits := uint8(v.entry)
br.advance(nBits)
bitsLeft -= nBits
dst = append(dst, uint8(v.entry>>8))
}
return dst, br.close()
}

830
vendor/github.com/klauspost/compress/huff0/decompress_amd64.s generated vendored Normal file
View File

@@ -0,0 +1,830 @@
// Code generated by command: go run gen.go -out ../decompress_amd64.s -pkg=huff0. DO NOT EDIT.
//go:build amd64 && !appengine && !noasm && gc
// func decompress4x_main_loop_amd64(ctx *decompress4xContext)
TEXT ·decompress4x_main_loop_amd64(SB), $0-8
// Preload values
MOVQ ctx+0(FP), AX
MOVBQZX 8(AX), DI
MOVQ 16(AX), BX
MOVQ 48(AX), SI
MOVQ 24(AX), R8
MOVQ 32(AX), R9
MOVQ (AX), R10
// Main loop
main_loop:
XORL DX, DX
CMPQ BX, SI
SETGE DL
// br0.fillFast32()
MOVQ 32(R10), R11
MOVBQZX 40(R10), R12
CMPQ R12, $0x20
JBE skip_fill0
MOVQ 24(R10), AX
SUBQ $0x20, R12
SUBQ $0x04, AX
MOVQ (R10), R13
// b.value |= uint64(low) << (b.bitsRead & 63)
MOVL (AX)(R13*1), R13
MOVQ R12, CX
SHLQ CL, R13
MOVQ AX, 24(R10)
ORQ R13, R11
// exhausted += (br0.off < 4)
CMPQ AX, $0x04
ADCB $+0, DL
skip_fill0:
// val0 := br0.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v0 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br0.advance(uint8(v0.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
// val1 := br0.peekTopBits(peekBits)
MOVQ DI, CX
MOVQ R11, R13
SHRQ CL, R13
// v1 := table[val1&mask]
MOVW (R9)(R13*2), CX
// br0.advance(uint8(v1.entry))
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
// these two writes get coalesced
// out[id * dstEvery + 0] = uint8(v0.entry >> 8)
// out[id * dstEvery + 1] = uint8(v1.entry >> 8)
MOVW AX, (BX)
// update the bitreader structure
MOVQ R11, 32(R10)
MOVB R12, 40(R10)
// br1.fillFast32()
MOVQ 80(R10), R11
MOVBQZX 88(R10), R12
CMPQ R12, $0x20
JBE skip_fill1
MOVQ 72(R10), AX
SUBQ $0x20, R12
SUBQ $0x04, AX
MOVQ 48(R10), R13
// b.value |= uint64(low) << (b.bitsRead & 63)
MOVL (AX)(R13*1), R13
MOVQ R12, CX
SHLQ CL, R13
MOVQ AX, 72(R10)
ORQ R13, R11
// exhausted += (br1.off < 4)
CMPQ AX, $0x04
ADCB $+0, DL
skip_fill1:
// val0 := br1.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v0 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br1.advance(uint8(v0.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
// val1 := br1.peekTopBits(peekBits)
MOVQ DI, CX
MOVQ R11, R13
SHRQ CL, R13
// v1 := table[val1&mask]
MOVW (R9)(R13*2), CX
// br1.advance(uint8(v1.entry))
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
// these two writes get coalesced
// out[id * dstEvery + 0] = uint8(v0.entry >> 8)
// out[id * dstEvery + 1] = uint8(v1.entry >> 8)
MOVW AX, (BX)(R8*1)
// update the bitreader structure
MOVQ R11, 80(R10)
MOVB R12, 88(R10)
// br2.fillFast32()
MOVQ 128(R10), R11
MOVBQZX 136(R10), R12
CMPQ R12, $0x20
JBE skip_fill2
MOVQ 120(R10), AX
SUBQ $0x20, R12
SUBQ $0x04, AX
MOVQ 96(R10), R13
// b.value |= uint64(low) << (b.bitsRead & 63)
MOVL (AX)(R13*1), R13
MOVQ R12, CX
SHLQ CL, R13
MOVQ AX, 120(R10)
ORQ R13, R11
// exhausted += (br2.off < 4)
CMPQ AX, $0x04
ADCB $+0, DL
skip_fill2:
// val0 := br2.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v0 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br2.advance(uint8(v0.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
// val1 := br2.peekTopBits(peekBits)
MOVQ DI, CX
MOVQ R11, R13
SHRQ CL, R13
// v1 := table[val1&mask]
MOVW (R9)(R13*2), CX
// br2.advance(uint8(v1.entry))
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
// these two writes get coalesced
// out[id * dstEvery + 0] = uint8(v0.entry >> 8)
// out[id * dstEvery + 1] = uint8(v1.entry >> 8)
MOVW AX, (BX)(R8*2)
// update the bitreader structure
MOVQ R11, 128(R10)
MOVB R12, 136(R10)
// br3.fillFast32()
MOVQ 176(R10), R11
MOVBQZX 184(R10), R12
CMPQ R12, $0x20
JBE skip_fill3
MOVQ 168(R10), AX
SUBQ $0x20, R12
SUBQ $0x04, AX
MOVQ 144(R10), R13
// b.value |= uint64(low) << (b.bitsRead & 63)
MOVL (AX)(R13*1), R13
MOVQ R12, CX
SHLQ CL, R13
MOVQ AX, 168(R10)
ORQ R13, R11
// exhausted += (br3.off < 4)
CMPQ AX, $0x04
ADCB $+0, DL
skip_fill3:
// val0 := br3.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v0 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br3.advance(uint8(v0.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
// val1 := br3.peekTopBits(peekBits)
MOVQ DI, CX
MOVQ R11, R13
SHRQ CL, R13
// v1 := table[val1&mask]
MOVW (R9)(R13*2), CX
// br3.advance(uint8(v1.entry))
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
// these two writes get coalesced
// out[id * dstEvery + 0] = uint8(v0.entry >> 8)
// out[id * dstEvery + 1] = uint8(v1.entry >> 8)
LEAQ (R8)(R8*2), CX
MOVW AX, (BX)(CX*1)
// update the bitreader structure
MOVQ R11, 176(R10)
MOVB R12, 184(R10)
ADDQ $0x02, BX
TESTB DL, DL
JZ main_loop
MOVQ ctx+0(FP), AX
SUBQ 16(AX), BX
SHLQ $0x02, BX
MOVQ BX, 40(AX)
RET
// func decompress4x_8b_main_loop_amd64(ctx *decompress4xContext)
TEXT ·decompress4x_8b_main_loop_amd64(SB), $0-8
// Preload values
MOVQ ctx+0(FP), CX
MOVBQZX 8(CX), DI
MOVQ 16(CX), BX
MOVQ 48(CX), SI
MOVQ 24(CX), R8
MOVQ 32(CX), R9
MOVQ (CX), R10
// Main loop
main_loop:
XORL DX, DX
CMPQ BX, SI
SETGE DL
// br0.fillFast32()
MOVQ 32(R10), R11
MOVBQZX 40(R10), R12
CMPQ R12, $0x20
JBE skip_fill0
MOVQ 24(R10), R13
SUBQ $0x20, R12
SUBQ $0x04, R13
MOVQ (R10), R14
// b.value |= uint64(low) << (b.bitsRead & 63)
MOVL (R13)(R14*1), R14
MOVQ R12, CX
SHLQ CL, R14
MOVQ R13, 24(R10)
ORQ R14, R11
// exhausted += (br0.off < 4)
CMPQ R13, $0x04
ADCB $+0, DL
skip_fill0:
// val0 := br0.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v0 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br0.advance(uint8(v0.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
// val1 := br0.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v1 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br0.advance(uint8(v1.entry)
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
BSWAPL AX
// val2 := br0.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v2 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br0.advance(uint8(v2.entry)
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
// val3 := br0.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v3 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br0.advance(uint8(v3.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
BSWAPL AX
// these four writes get coalesced
// out[id * dstEvery + 0] = uint8(v0.entry >> 8)
// out[id * dstEvery + 1] = uint8(v1.entry >> 8)
// out[id * dstEvery + 3] = uint8(v2.entry >> 8)
// out[id * dstEvery + 4] = uint8(v3.entry >> 8)
MOVL AX, (BX)
// update the bitreader structure
MOVQ R11, 32(R10)
MOVB R12, 40(R10)
// br1.fillFast32()
MOVQ 80(R10), R11
MOVBQZX 88(R10), R12
CMPQ R12, $0x20
JBE skip_fill1
MOVQ 72(R10), R13
SUBQ $0x20, R12
SUBQ $0x04, R13
MOVQ 48(R10), R14
// b.value |= uint64(low) << (b.bitsRead & 63)
MOVL (R13)(R14*1), R14
MOVQ R12, CX
SHLQ CL, R14
MOVQ R13, 72(R10)
ORQ R14, R11
// exhausted += (br1.off < 4)
CMPQ R13, $0x04
ADCB $+0, DL
skip_fill1:
// val0 := br1.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v0 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br1.advance(uint8(v0.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
// val1 := br1.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v1 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br1.advance(uint8(v1.entry)
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
BSWAPL AX
// val2 := br1.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v2 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br1.advance(uint8(v2.entry)
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
// val3 := br1.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v3 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br1.advance(uint8(v3.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
BSWAPL AX
// these four writes get coalesced
// out[id * dstEvery + 0] = uint8(v0.entry >> 8)
// out[id * dstEvery + 1] = uint8(v1.entry >> 8)
// out[id * dstEvery + 3] = uint8(v2.entry >> 8)
// out[id * dstEvery + 4] = uint8(v3.entry >> 8)
MOVL AX, (BX)(R8*1)
// update the bitreader structure
MOVQ R11, 80(R10)
MOVB R12, 88(R10)
// br2.fillFast32()
MOVQ 128(R10), R11
MOVBQZX 136(R10), R12
CMPQ R12, $0x20
JBE skip_fill2
MOVQ 120(R10), R13
SUBQ $0x20, R12
SUBQ $0x04, R13
MOVQ 96(R10), R14
// b.value |= uint64(low) << (b.bitsRead & 63)
MOVL (R13)(R14*1), R14
MOVQ R12, CX
SHLQ CL, R14
MOVQ R13, 120(R10)
ORQ R14, R11
// exhausted += (br2.off < 4)
CMPQ R13, $0x04
ADCB $+0, DL
skip_fill2:
// val0 := br2.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v0 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br2.advance(uint8(v0.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
// val1 := br2.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v1 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br2.advance(uint8(v1.entry)
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
BSWAPL AX
// val2 := br2.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v2 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br2.advance(uint8(v2.entry)
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
// val3 := br2.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v3 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br2.advance(uint8(v3.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
BSWAPL AX
// these four writes get coalesced
// out[id * dstEvery + 0] = uint8(v0.entry >> 8)
// out[id * dstEvery + 1] = uint8(v1.entry >> 8)
// out[id * dstEvery + 3] = uint8(v2.entry >> 8)
// out[id * dstEvery + 4] = uint8(v3.entry >> 8)
MOVL AX, (BX)(R8*2)
// update the bitreader structure
MOVQ R11, 128(R10)
MOVB R12, 136(R10)
// br3.fillFast32()
MOVQ 176(R10), R11
MOVBQZX 184(R10), R12
CMPQ R12, $0x20
JBE skip_fill3
MOVQ 168(R10), R13
SUBQ $0x20, R12
SUBQ $0x04, R13
MOVQ 144(R10), R14
// b.value |= uint64(low) << (b.bitsRead & 63)
MOVL (R13)(R14*1), R14
MOVQ R12, CX
SHLQ CL, R14
MOVQ R13, 168(R10)
ORQ R14, R11
// exhausted += (br3.off < 4)
CMPQ R13, $0x04
ADCB $+0, DL
skip_fill3:
// val0 := br3.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v0 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br3.advance(uint8(v0.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
// val1 := br3.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v1 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br3.advance(uint8(v1.entry)
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
BSWAPL AX
// val2 := br3.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v2 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br3.advance(uint8(v2.entry)
MOVB CH, AH
SHLQ CL, R11
ADDB CL, R12
// val3 := br3.peekTopBits(peekBits)
MOVQ R11, R13
MOVQ DI, CX
SHRQ CL, R13
// v3 := table[val0&mask]
MOVW (R9)(R13*2), CX
// br3.advance(uint8(v3.entry)
MOVB CH, AL
SHLQ CL, R11
ADDB CL, R12
BSWAPL AX
// these four writes get coalesced
// out[id * dstEvery + 0] = uint8(v0.entry >> 8)
// out[id * dstEvery + 1] = uint8(v1.entry >> 8)
// out[id * dstEvery + 3] = uint8(v2.entry >> 8)
// out[id * dstEvery + 4] = uint8(v3.entry >> 8)
LEAQ (R8)(R8*2), CX
MOVL AX, (BX)(CX*1)
// update the bitreader structure
MOVQ R11, 176(R10)
MOVB R12, 184(R10)
ADDQ $0x04, BX
TESTB DL, DL
JZ main_loop
MOVQ ctx+0(FP), AX
SUBQ 16(AX), BX
SHLQ $0x02, BX
MOVQ BX, 40(AX)
RET
// func decompress1x_main_loop_amd64(ctx *decompress1xContext)
TEXT ·decompress1x_main_loop_amd64(SB), $0-8
MOVQ ctx+0(FP), CX
MOVQ 16(CX), DX
MOVQ 24(CX), BX
CMPQ BX, $0x04
JB error_max_decoded_size_exceeded
LEAQ (DX)(BX*1), BX
MOVQ (CX), SI
MOVQ (SI), R8
MOVQ 24(SI), R9
MOVQ 32(SI), R10
MOVBQZX 40(SI), R11
MOVQ 32(CX), SI
MOVBQZX 8(CX), DI
JMP loop_condition
main_loop:
// Check if we have room for 4 bytes in the output buffer
LEAQ 4(DX), CX
CMPQ CX, BX
JGE error_max_decoded_size_exceeded
// Decode 4 values
CMPQ R11, $0x20
JL bitReader_fillFast_1_end
SUBQ $0x20, R11
SUBQ $0x04, R9
MOVL (R8)(R9*1), R12
MOVQ R11, CX
SHLQ CL, R12
ORQ R12, R10
bitReader_fillFast_1_end:
MOVQ DI, CX
MOVQ R10, R12
SHRQ CL, R12
MOVW (SI)(R12*2), CX
MOVB CH, AL
MOVBQZX CL, CX
ADDQ CX, R11
SHLQ CL, R10
MOVQ DI, CX
MOVQ R10, R12
SHRQ CL, R12
MOVW (SI)(R12*2), CX
MOVB CH, AH
MOVBQZX CL, CX
ADDQ CX, R11
SHLQ CL, R10
BSWAPL AX
CMPQ R11, $0x20
JL bitReader_fillFast_2_end
SUBQ $0x20, R11
SUBQ $0x04, R9
MOVL (R8)(R9*1), R12
MOVQ R11, CX
SHLQ CL, R12
ORQ R12, R10
bitReader_fillFast_2_end:
MOVQ DI, CX
MOVQ R10, R12
SHRQ CL, R12
MOVW (SI)(R12*2), CX
MOVB CH, AH
MOVBQZX CL, CX
ADDQ CX, R11
SHLQ CL, R10
MOVQ DI, CX
MOVQ R10, R12
SHRQ CL, R12
MOVW (SI)(R12*2), CX
MOVB CH, AL
MOVBQZX CL, CX
ADDQ CX, R11
SHLQ CL, R10
BSWAPL AX
// Store the decoded values
MOVL AX, (DX)
ADDQ $0x04, DX
loop_condition:
CMPQ R9, $0x08
JGE main_loop
// Update ctx structure
MOVQ ctx+0(FP), AX
SUBQ 16(AX), DX
MOVQ DX, 40(AX)
MOVQ (AX), AX
MOVQ R9, 24(AX)
MOVQ R10, 32(AX)
MOVB R11, 40(AX)
RET
// Report error
error_max_decoded_size_exceeded:
MOVQ ctx+0(FP), AX
MOVQ $-1, CX
MOVQ CX, 40(AX)
RET
// func decompress1x_main_loop_bmi2(ctx *decompress1xContext)
// Requires: BMI2
TEXT ·decompress1x_main_loop_bmi2(SB), $0-8
MOVQ ctx+0(FP), CX
MOVQ 16(CX), DX
MOVQ 24(CX), BX
CMPQ BX, $0x04
JB error_max_decoded_size_exceeded
LEAQ (DX)(BX*1), BX
MOVQ (CX), SI
MOVQ (SI), R8
MOVQ 24(SI), R9
MOVQ 32(SI), R10
MOVBQZX 40(SI), R11
MOVQ 32(CX), SI
MOVBQZX 8(CX), DI
JMP loop_condition
main_loop:
// Check if we have room for 4 bytes in the output buffer
LEAQ 4(DX), CX
CMPQ CX, BX
JGE error_max_decoded_size_exceeded
// Decode 4 values
CMPQ R11, $0x20
JL bitReader_fillFast_1_end
SUBQ $0x20, R11
SUBQ $0x04, R9
MOVL (R8)(R9*1), CX
SHLXQ R11, CX, CX
ORQ CX, R10
bitReader_fillFast_1_end:
SHRXQ DI, R10, CX
MOVW (SI)(CX*2), CX
MOVB CH, AL
MOVBQZX CL, CX
ADDQ CX, R11
SHLXQ CX, R10, R10
SHRXQ DI, R10, CX
MOVW (SI)(CX*2), CX
MOVB CH, AH
MOVBQZX CL, CX
ADDQ CX, R11
SHLXQ CX, R10, R10
BSWAPL AX
CMPQ R11, $0x20
JL bitReader_fillFast_2_end
SUBQ $0x20, R11
SUBQ $0x04, R9
MOVL (R8)(R9*1), CX
SHLXQ R11, CX, CX
ORQ CX, R10
bitReader_fillFast_2_end:
SHRXQ DI, R10, CX
MOVW (SI)(CX*2), CX
MOVB CH, AH
MOVBQZX CL, CX
ADDQ CX, R11
SHLXQ CX, R10, R10
SHRXQ DI, R10, CX
MOVW (SI)(CX*2), CX
MOVB CH, AL
MOVBQZX CL, CX
ADDQ CX, R11
SHLXQ CX, R10, R10
BSWAPL AX
// Store the decoded values
MOVL AX, (DX)
ADDQ $0x04, DX
loop_condition:
CMPQ R9, $0x08
JGE main_loop
// Update ctx structure
MOVQ ctx+0(FP), AX
SUBQ 16(AX), DX
MOVQ DX, 40(AX)
MOVQ (AX), AX
MOVQ R9, 24(AX)
MOVQ R10, 32(AX)
MOVB R11, 40(AX)
RET
// Report error
error_max_decoded_size_exceeded:
MOVQ ctx+0(FP), AX
MOVQ $-1, CX
MOVQ CX, 40(AX)
RET

299
vendor/github.com/klauspost/compress/huff0/decompress_generic.go generated vendored Normal file
View File

@@ -0,0 +1,299 @@
//go:build !amd64 || appengine || !gc || noasm
// +build !amd64 appengine !gc noasm
// This file contains a generic implementation of Decoder.Decompress4X.
package huff0
import (
"errors"
"fmt"
)
// Decompress4X will decompress a 4X encoded stream.
// The length of the supplied input must match the end of a block exactly.
// The *capacity* of the dst slice must match the destination size of
// the uncompressed data exactly.
func (d *Decoder) Decompress4X(dst, src []byte) ([]byte, error) {
if len(d.dt.single) == 0 {
return nil, errors.New("no table loaded")
}
if len(src) < 6+(4*1) {
return nil, errors.New("input too small")
}
if use8BitTables && d.actualTableLog <= 8 {
return d.decompress4X8bit(dst, src)
}
var br [4]bitReaderShifted
// Decode "jump table"
start := 6
for i := 0; i < 3; i++ {
length := int(src[i*2]) | (int(src[i*2+1]) << 8)
if start+length >= len(src) {
return nil, errors.New("truncated input (or invalid offset)")
}
err := br[i].init(src[start : start+length])
if err != nil {
return nil, err
}
start += length
}
err := br[3].init(src[start:])
if err != nil {
return nil, err
}
// destination, offset to match first output
dstSize := cap(dst)
dst = dst[:dstSize]
out := dst
dstEvery := (dstSize + 3) / 4
const tlSize = 1 << tableLogMax
const tlMask = tlSize - 1
single := d.dt.single[:tlSize]
// Use temp table to avoid bound checks/append penalty.
buf := d.buffer()
var off uint8
var decoded int
// Decode 2 values from each decoder/loop.
const bufoff = 256
for {
if br[0].off < 4 || br[1].off < 4 || br[2].off < 4 || br[3].off < 4 {
break
}
{
const stream = 0
const stream2 = 1
br[stream].fillFast()
br[stream2].fillFast()
val := br[stream].peekBitsFast(d.actualTableLog)
val2 := br[stream2].peekBitsFast(d.actualTableLog)
v := single[val&tlMask]
v2 := single[val2&tlMask]
br[stream].advance(uint8(v.entry))
br[stream2].advance(uint8(v2.entry))
buf[stream][off] = uint8(v.entry >> 8)
buf[stream2][off] = uint8(v2.entry >> 8)
val = br[stream].peekBitsFast(d.actualTableLog)
val2 = br[stream2].peekBitsFast(d.actualTableLog)
v = single[val&tlMask]
v2 = single[val2&tlMask]
br[stream].advance(uint8(v.entry))
br[stream2].advance(uint8(v2.entry))
buf[stream][off+1] = uint8(v.entry >> 8)
buf[stream2][off+1] = uint8(v2.entry >> 8)
}
{
const stream = 2
const stream2 = 3
br[stream].fillFast()
br[stream2].fillFast()
val := br[stream].peekBitsFast(d.actualTableLog)
val2 := br[stream2].peekBitsFast(d.actualTableLog)
v := single[val&tlMask]
v2 := single[val2&tlMask]
br[stream].advance(uint8(v.entry))
br[stream2].advance(uint8(v2.entry))
buf[stream][off] = uint8(v.entry >> 8)
buf[stream2][off] = uint8(v2.entry >> 8)
val = br[stream].peekBitsFast(d.actualTableLog)
val2 = br[stream2].peekBitsFast(d.actualTableLog)
v = single[val&tlMask]
v2 = single[val2&tlMask]
br[stream].advance(uint8(v.entry))
br[stream2].advance(uint8(v2.entry))
buf[stream][off+1] = uint8(v.entry >> 8)
buf[stream2][off+1] = uint8(v2.entry >> 8)
}
off += 2
if off == 0 {
if bufoff > dstEvery {
d.bufs.Put(buf)
return nil, errors.New("corruption detected: stream overrun 1")
}
// There must at least be 3 buffers left.
if len(out)-bufoff < dstEvery*3 {
d.bufs.Put(buf)
return nil, errors.New("corruption detected: stream overrun 2")
}
//copy(out, buf[0][:])
//copy(out[dstEvery:], buf[1][:])
//copy(out[dstEvery*2:], buf[2][:])
//copy(out[dstEvery*3:], buf[3][:])
*(*[bufoff]byte)(out) = buf[0]
*(*[bufoff]byte)(out[dstEvery:]) = buf[1]
*(*[bufoff]byte)(out[dstEvery*2:]) = buf[2]
*(*[bufoff]byte)(out[dstEvery*3:]) = buf[3]
out = out[bufoff:]
decoded += bufoff * 4
}
}
if off > 0 {
ioff := int(off)
if len(out) < dstEvery*3+ioff {
d.bufs.Put(buf)
return nil, errors.New("corruption detected: stream overrun 3")
}
copy(out, buf[0][:off])
copy(out[dstEvery:], buf[1][:off])
copy(out[dstEvery*2:], buf[2][:off])
copy(out[dstEvery*3:], buf[3][:off])
decoded += int(off) * 4
out = out[off:]
}
// Decode remaining.
remainBytes := dstEvery - (decoded / 4)
for i := range br {
offset := dstEvery * i
endsAt := offset + remainBytes
if endsAt > len(out) {
endsAt = len(out)
}
br := &br[i]
bitsLeft := br.remaining()
for bitsLeft > 0 {
br.fill()
if offset >= endsAt {
d.bufs.Put(buf)
return nil, errors.New("corruption detected: stream overrun 4")
}
// Read value and increment offset.
val := br.peekBitsFast(d.actualTableLog)
v := single[val&tlMask].entry
nBits := uint8(v)
br.advance(nBits)
bitsLeft -= uint(nBits)
out[offset] = uint8(v >> 8)
offset++
}
if offset != endsAt {
d.bufs.Put(buf)
return nil, fmt.Errorf("corruption detected: short output block %d, end %d != %d", i, offset, endsAt)
}
decoded += offset - dstEvery*i
err = br.close()
if err != nil {
return nil, err
}
}
d.bufs.Put(buf)
if dstSize != decoded {
return nil, errors.New("corruption detected: short output block")
}
return dst, nil
}
// Decompress1X will decompress a 1X encoded stream.
// The cap of the output buffer will be the maximum decompressed size.
// The length of the supplied input must match the end of a block exactly.
func (d *Decoder) Decompress1X(dst, src []byte) ([]byte, error) {
if len(d.dt.single) == 0 {
return nil, errors.New("no table loaded")
}
if use8BitTables && d.actualTableLog <= 8 {
return d.decompress1X8Bit(dst, src)
}
var br bitReaderShifted
err := br.init(src)
if err != nil {
return dst, err
}
maxDecodedSize := cap(dst)
dst = dst[:0]
// Avoid bounds check by always having full sized table.
const tlSize = 1 << tableLogMax
const tlMask = tlSize - 1
dt := d.dt.single[:tlSize]
// Use temp table to avoid bound checks/append penalty.
bufs := d.buffer()
buf := &bufs[0]
var off uint8
for br.off >= 8 {
br.fillFast()
v := dt[br.peekBitsFast(d.actualTableLog)&tlMask]
br.advance(uint8(v.entry))
buf[off+0] = uint8(v.entry >> 8)
v = dt[br.peekBitsFast(d.actualTableLog)&tlMask]
br.advance(uint8(v.entry))
buf[off+1] = uint8(v.entry >> 8)
// Refill
br.fillFast()
v = dt[br.peekBitsFast(d.actualTableLog)&tlMask]
br.advance(uint8(v.entry))
buf[off+2] = uint8(v.entry >> 8)
v = dt[br.peekBitsFast(d.actualTableLog)&tlMask]
br.advance(uint8(v.entry))
buf[off+3] = uint8(v.entry >> 8)
off += 4
if off == 0 {
if len(dst)+256 > maxDecodedSize {
br.close()
d.bufs.Put(bufs)
return nil, ErrMaxDecodedSizeExceeded
}
dst = append(dst, buf[:]...)
}
}
if len(dst)+int(off) > maxDecodedSize {
d.bufs.Put(bufs)
br.close()
return nil, ErrMaxDecodedSizeExceeded
}
dst = append(dst, buf[:off]...)
// br < 8, so uint8 is fine
bitsLeft := uint8(br.off)*8 + 64 - br.bitsRead
for bitsLeft > 0 {
br.fill()
if false && br.bitsRead >= 32 {
if br.off >= 4 {
v := br.in[br.off-4:]
v = v[:4]
low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
br.value = (br.value << 32) | uint64(low)
br.bitsRead -= 32
br.off -= 4
} else {
for br.off > 0 {
br.value = (br.value << 8) | uint64(br.in[br.off-1])
br.bitsRead -= 8
br.off--
}
}
}
if len(dst) >= maxDecodedSize {
d.bufs.Put(bufs)
br.close()
return nil, ErrMaxDecodedSizeExceeded
}
v := d.dt.single[br.peekBitsFast(d.actualTableLog)&tlMask]
nBits := uint8(v.entry)
br.advance(nBits)
bitsLeft -= nBits
dst = append(dst, uint8(v.entry>>8))
}
d.bufs.Put(bufs)
return dst, br.close()
}

337
vendor/github.com/klauspost/compress/huff0/huff0.go generated vendored Normal file
View File

@@ -0,0 +1,337 @@
// Package huff0 provides fast huffman encoding as used in zstd.
//
// See README.md at https://github.com/klauspost/compress/tree/master/huff0 for details.
package huff0
import (
"errors"
"fmt"
"math"
"math/bits"
"sync"
"github.com/klauspost/compress/fse"
)
const (
maxSymbolValue = 255
// zstandard limits tablelog to 11, see:
// https://github.com/facebook/zstd/blob/dev/doc/zstd_compression_format.md#huffman-tree-description
tableLogMax = 11
tableLogDefault = 11
minTablelog = 5
huffNodesLen = 512
// BlockSizeMax is maximum input size for a single block uncompressed.
BlockSizeMax = 1<<18 - 1
)
var (
// ErrIncompressible is returned when input is judged to be too hard to compress.
ErrIncompressible = errors.New("input is not compressible")
// ErrUseRLE is returned from the compressor when the input is a single byte value repeated.
ErrUseRLE = errors.New("input is single value repeated")
// ErrTooBig is return if input is too large for a single block.
ErrTooBig = errors.New("input too big")
// ErrMaxDecodedSizeExceeded is return if input is too large for a single block.
ErrMaxDecodedSizeExceeded = errors.New("maximum output size exceeded")
)
type ReusePolicy uint8
const (
// ReusePolicyAllow will allow reuse if it produces smaller output.
ReusePolicyAllow ReusePolicy = iota
// ReusePolicyPrefer will re-use aggressively if possible.
// This will not check if a new table will produce smaller output,
// except if the current table is impossible to use or
// compressed output is bigger than input.
ReusePolicyPrefer
// ReusePolicyNone will disable re-use of tables.
// This is slightly faster than ReusePolicyAllow but may produce larger output.
ReusePolicyNone
// ReusePolicyMust must allow reuse and produce smaller output.
ReusePolicyMust
)
type Scratch struct {
count [maxSymbolValue + 1]uint32
// Per block parameters.
// These can be used to override compression parameters of the block.
// Do not touch, unless you know what you are doing.
// Out is output buffer.
// If the scratch is re-used before the caller is done processing the output,
// set this field to nil.
// Otherwise the output buffer will be re-used for next Compression/Decompression step
// and allocation will be avoided.
Out []byte
// OutTable will contain the table data only, if a new table has been generated.
// Slice of the returned data.
OutTable []byte
// OutData will contain the compressed data.
// Slice of the returned data.
OutData []byte
// MaxDecodedSize will set the maximum allowed output size.
// This value will automatically be set to BlockSizeMax if not set.
// Decoders will return ErrMaxDecodedSizeExceeded is this limit is exceeded.
MaxDecodedSize int
srcLen int
// MaxSymbolValue will override the maximum symbol value of the next block.
MaxSymbolValue uint8
// TableLog will attempt to override the tablelog for the next block.
// Must be <= 11 and >= 5.
TableLog uint8
// Reuse will specify the reuse policy
Reuse ReusePolicy
// WantLogLess allows to specify a log 2 reduction that should at least be achieved,
// otherwise the block will be returned as incompressible.
// The reduction should then at least be (input size >> WantLogLess)
// If WantLogLess == 0 any improvement will do.
WantLogLess uint8
symbolLen uint16 // Length of active part of the symbol table.
maxCount int // count of the most probable symbol
clearCount bool // clear count
actualTableLog uint8 // Selected tablelog.
prevTableLog uint8 // Tablelog for previous table
prevTable cTable // Table used for previous compression.
cTable cTable // compression table
dt dTable // decompression table
nodes []nodeElt
tmpOut [4][]byte
fse *fse.Scratch
decPool sync.Pool // *[4][256]byte buffers.
huffWeight [maxSymbolValue + 1]byte
}
// TransferCTable will transfer the previously used compression table.
func (s *Scratch) TransferCTable(src *Scratch) {
if cap(s.prevTable) < len(src.prevTable) {
s.prevTable = make(cTable, 0, maxSymbolValue+1)
}
s.prevTable = s.prevTable[:len(src.prevTable)]
copy(s.prevTable, src.prevTable)
s.prevTableLog = src.prevTableLog
}
func (s *Scratch) prepare(in []byte) (*Scratch, error) {
if len(in) > BlockSizeMax {
return nil, ErrTooBig
}
if s == nil {
s = &Scratch{}
}
if s.MaxSymbolValue == 0 {
s.MaxSymbolValue = maxSymbolValue
}
if s.TableLog == 0 {
s.TableLog = tableLogDefault
}
if s.TableLog > tableLogMax || s.TableLog < minTablelog {
return nil, fmt.Errorf(" invalid tableLog %d (%d -> %d)", s.TableLog, minTablelog, tableLogMax)
}
if s.MaxDecodedSize <= 0 || s.MaxDecodedSize > BlockSizeMax {
s.MaxDecodedSize = BlockSizeMax
}
if s.clearCount && s.maxCount == 0 {
for i := range s.count {
s.count[i] = 0
}
s.clearCount = false
}
if cap(s.Out) == 0 {
s.Out = make([]byte, 0, len(in))
}
s.Out = s.Out[:0]
s.OutTable = nil
s.OutData = nil
if cap(s.nodes) < huffNodesLen+1 {
s.nodes = make([]nodeElt, 0, huffNodesLen+1)
}
s.nodes = s.nodes[:0]
if s.fse == nil {
s.fse = &fse.Scratch{}
}
s.srcLen = len(in)
return s, nil
}
type cTable []cTableEntry
func (c cTable) write(s *Scratch) error {
var (
// precomputed conversion table
bitsToWeight [tableLogMax + 1]byte
huffLog = s.actualTableLog
// last weight is not saved.
maxSymbolValue = uint8(s.symbolLen - 1)
huffWeight = s.huffWeight[:256]
)
const (
maxFSETableLog = 6
)
// convert to weight
bitsToWeight[0] = 0
for n := uint8(1); n < huffLog+1; n++ {
bitsToWeight[n] = huffLog + 1 - n
}
// Acquire histogram for FSE.
hist := s.fse.Histogram()
hist = hist[:256]
for i := range hist[:16] {
hist[i] = 0
}
for n := uint8(0); n < maxSymbolValue; n++ {
v := bitsToWeight[c[n].nBits] & 15
huffWeight[n] = v
hist[v]++
}
// FSE compress if feasible.
if maxSymbolValue >= 2 {
huffMaxCnt := uint32(0)
huffMax := uint8(0)
for i, v := range hist[:16] {
if v == 0 {
continue
}
huffMax = byte(i)
if v > huffMaxCnt {
huffMaxCnt = v
}
}
s.fse.HistogramFinished(huffMax, int(huffMaxCnt))
s.fse.TableLog = maxFSETableLog
b, err := fse.Compress(huffWeight[:maxSymbolValue], s.fse)
if err == nil && len(b) < int(s.symbolLen>>1) {
s.Out = append(s.Out, uint8(len(b)))
s.Out = append(s.Out, b...)
return nil
}
// Unable to compress (RLE/uncompressible)
}
// write raw values as 4-bits (max : 15)
if maxSymbolValue > (256 - 128) {
// should not happen : likely means source cannot be compressed
return ErrIncompressible
}
op := s.Out
// special case, pack weights 4 bits/weight.
op = append(op, 128|(maxSymbolValue-1))
// be sure it doesn't cause msan issue in final combination
huffWeight[maxSymbolValue] = 0
for n := uint16(0); n < uint16(maxSymbolValue); n += 2 {
op = append(op, (huffWeight[n]<<4)|huffWeight[n+1])
}
s.Out = op
return nil
}
func (c cTable) estTableSize(s *Scratch) (sz int, err error) {
var (
// precomputed conversion table
bitsToWeight [tableLogMax + 1]byte
huffLog = s.actualTableLog
// last weight is not saved.
maxSymbolValue = uint8(s.symbolLen - 1)
huffWeight = s.huffWeight[:256]
)
const (
maxFSETableLog = 6
)
// convert to weight
bitsToWeight[0] = 0
for n := uint8(1); n < huffLog+1; n++ {
bitsToWeight[n] = huffLog + 1 - n
}
// Acquire histogram for FSE.
hist := s.fse.Histogram()
hist = hist[:256]
for i := range hist[:16] {
hist[i] = 0
}
for n := uint8(0); n < maxSymbolValue; n++ {
v := bitsToWeight[c[n].nBits] & 15
huffWeight[n] = v
hist[v]++
}
// FSE compress if feasible.
if maxSymbolValue >= 2 {
huffMaxCnt := uint32(0)
huffMax := uint8(0)
for i, v := range hist[:16] {
if v == 0 {
continue
}
huffMax = byte(i)
if v > huffMaxCnt {
huffMaxCnt = v
}
}
s.fse.HistogramFinished(huffMax, int(huffMaxCnt))
s.fse.TableLog = maxFSETableLog
b, err := fse.Compress(huffWeight[:maxSymbolValue], s.fse)
if err == nil && len(b) < int(s.symbolLen>>1) {
sz += 1 + len(b)
return sz, nil
}
// Unable to compress (RLE/uncompressible)
}
// write raw values as 4-bits (max : 15)
if maxSymbolValue > (256 - 128) {
// should not happen : likely means source cannot be compressed
return 0, ErrIncompressible
}
// special case, pack weights 4 bits/weight.
sz += 1 + int(maxSymbolValue/2)
return sz, nil
}
// estimateSize returns the estimated size in bytes of the input represented in the
// histogram supplied.
func (c cTable) estimateSize(hist []uint32) int {
nbBits := uint32(7)
for i, v := range c[:len(hist)] {
nbBits += uint32(v.nBits) * hist[i]
}
return int(nbBits >> 3)
}
// minSize returns the minimum possible size considering the shannon limit.
func (s *Scratch) minSize(total int) int {
nbBits := float64(7)
fTotal := float64(total)
for _, v := range s.count[:s.symbolLen] {
n := float64(v)
if n > 0 {
nbBits += math.Log2(fTotal/n) * n
}
}
return int(nbBits) >> 3
}
func highBit32(val uint32) (n uint32) {
return uint32(bits.Len32(val) - 1)
}

34
vendor/github.com/klauspost/compress/internal/cpuinfo/cpuinfo.go generated vendored Normal file
View File

@@ -0,0 +1,34 @@
// Package cpuinfo gives runtime info about the current CPU.
//
// This is a very limited module meant for use internally
// in this project. For more versatile solution check
// https://github.com/klauspost/cpuid.
package cpuinfo
// HasBMI1 checks whether an x86 CPU supports the BMI1 extension.
func HasBMI1() bool {
return hasBMI1
}
// HasBMI2 checks whether an x86 CPU supports the BMI2 extension.
func HasBMI2() bool {
return hasBMI2
}
// DisableBMI2 will disable BMI2, for testing purposes.
// Call returned function to restore previous state.
func DisableBMI2() func() {
old := hasBMI2
hasBMI2 = false
return func() {
hasBMI2 = old
}
}
// HasBMI checks whether an x86 CPU supports both BMI1 and BMI2 extensions.
func HasBMI() bool {
return HasBMI1() && HasBMI2()
}
var hasBMI1 bool
var hasBMI2 bool

11
vendor/github.com/klauspost/compress/internal/cpuinfo/cpuinfo_amd64.go generated vendored Normal file
View File

@@ -0,0 +1,11 @@
//go:build amd64 && !appengine && !noasm && gc
// +build amd64,!appengine,!noasm,gc
package cpuinfo
// go:noescape
func x86extensions() (bmi1, bmi2 bool)
func init() {
hasBMI1, hasBMI2 = x86extensions()
}

36
vendor/github.com/klauspost/compress/internal/cpuinfo/cpuinfo_amd64.s generated vendored Normal file
View File

@@ -0,0 +1,36 @@
// +build !appengine
// +build gc
// +build !noasm
#include "textflag.h"
#include "funcdata.h"
#include "go_asm.h"
TEXT ·x86extensions(SB), NOSPLIT, $0
// 1. determine max EAX value
XORQ AX, AX
CPUID
CMPQ AX, $7
JB unsupported
// 2. EAX = 7, ECX = 0 --- see Table 3-8 "Information Returned by CPUID Instruction"
MOVQ $7, AX
MOVQ $0, CX
CPUID
BTQ $3, BX // bit 3 = BMI1
SETCS AL
BTQ $8, BX // bit 8 = BMI2
SETCS AH
MOVB AL, bmi1+0(FP)
MOVB AH, bmi2+1(FP)
RET
unsupported:
XORQ AX, AX
MOVB AL, bmi1+0(FP)
MOVB AL, bmi2+1(FP)
RET

27
vendor/github.com/klauspost/compress/internal/snapref/LICENSE generated vendored Normal file
View File

@@ -0,0 +1,27 @@
Copyright (c) 2011 The Snappy-Go Authors. All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above
copyright notice, this list of conditions and the following disclaimer
in the documentation and/or other materials provided with the
distribution.
* Neither the name of Google Inc. nor the names of its
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

264
vendor/github.com/klauspost/compress/internal/snapref/decode.go generated vendored Normal file
View File

@@ -0,0 +1,264 @@
// Copyright 2011 The Snappy-Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package snapref
import (
"encoding/binary"
"errors"
"io"
)
var (
// ErrCorrupt reports that the input is invalid.
ErrCorrupt = errors.New("snappy: corrupt input")
// ErrTooLarge reports that the uncompressed length is too large.
ErrTooLarge = errors.New("snappy: decoded block is too large")
// ErrUnsupported reports that the input isn't supported.
ErrUnsupported = errors.New("snappy: unsupported input")
errUnsupportedLiteralLength = errors.New("snappy: unsupported literal length")
)
// DecodedLen returns the length of the decoded block.
func DecodedLen(src []byte) (int, error) {
v, _, err := decodedLen(src)
return v, err
}
// decodedLen returns the length of the decoded block and the number of bytes
// that the length header occupied.
func decodedLen(src []byte) (blockLen, headerLen int, err error) {
v, n := binary.Uvarint(src)
if n <= 0 || v > 0xffffffff {
return 0, 0, ErrCorrupt
}
const wordSize = 32 << (^uint(0) >> 32 & 1)
if wordSize == 32 && v > 0x7fffffff {
return 0, 0, ErrTooLarge
}
return int(v), n, nil
}
const (
decodeErrCodeCorrupt = 1
decodeErrCodeUnsupportedLiteralLength = 2
)
// Decode returns the decoded form of src. The returned slice may be a sub-
// slice of dst if dst was large enough to hold the entire decoded block.
// Otherwise, a newly allocated slice will be returned.
//
// The dst and src must not overlap. It is valid to pass a nil dst.
//
// Decode handles the Snappy block format, not the Snappy stream format.
func Decode(dst, src []byte) ([]byte, error) {
dLen, s, err := decodedLen(src)
if err != nil {
return nil, err
}
if dLen <= len(dst) {
dst = dst[:dLen]
} else {
dst = make([]byte, dLen)
}
switch decode(dst, src[s:]) {
case 0:
return dst, nil
case decodeErrCodeUnsupportedLiteralLength:
return nil, errUnsupportedLiteralLength
}
return nil, ErrCorrupt
}
// NewReader returns a new Reader that decompresses from r, using the framing
// format described at
// https://github.com/google/snappy/blob/master/framing_format.txt
func NewReader(r io.Reader) *Reader {
return &Reader{
r: r,
decoded: make([]byte, maxBlockSize),
buf: make([]byte, maxEncodedLenOfMaxBlockSize+checksumSize),
}
}
// Reader is an io.Reader that can read Snappy-compressed bytes.
//
// Reader handles the Snappy stream format, not the Snappy block format.
type Reader struct {
r io.Reader
err error
decoded []byte
buf []byte
// decoded[i:j] contains decoded bytes that have not yet been passed on.
i, j int
readHeader bool
}
// Reset discards any buffered data, resets all state, and switches the Snappy
// reader to read from r. This permits reusing a Reader rather than allocating
// a new one.
func (r *Reader) Reset(reader io.Reader) {
r.r = reader
r.err = nil
r.i = 0
r.j = 0
r.readHeader = false
}
func (r *Reader) readFull(p []byte, allowEOF bool) (ok bool) {
if _, r.err = io.ReadFull(r.r, p); r.err != nil {
if r.err == io.ErrUnexpectedEOF || (r.err == io.EOF && !allowEOF) {
r.err = ErrCorrupt
}
return false
}
return true
}
func (r *Reader) fill() error {
for r.i >= r.j {
if !r.readFull(r.buf[:4], true) {
return r.err
}
chunkType := r.buf[0]
if !r.readHeader {
if chunkType != chunkTypeStreamIdentifier {
r.err = ErrCorrupt
return r.err
}
r.readHeader = true
}
chunkLen := int(r.buf[1]) | int(r.buf[2])<<8 | int(r.buf[3])<<16
if chunkLen > len(r.buf) {
r.err = ErrUnsupported
return r.err
}
// The chunk types are specified at
// https://github.com/google/snappy/blob/master/framing_format.txt
switch chunkType {
case chunkTypeCompressedData:
// Section 4.2. Compressed data (chunk type 0x00).
if chunkLen < checksumSize {
r.err = ErrCorrupt
return r.err
}
buf := r.buf[:chunkLen]
if !r.readFull(buf, false) {
return r.err
}
checksum := uint32(buf[0]) | uint32(buf[1])<<8 | uint32(buf[2])<<16 | uint32(buf[3])<<24
buf = buf[checksumSize:]
n, err := DecodedLen(buf)
if err != nil {
r.err = err
return r.err
}
if n > len(r.decoded) {
r.err = ErrCorrupt
return r.err
}
if _, err := Decode(r.decoded, buf); err != nil {
r.err = err
return r.err
}
if crc(r.decoded[:n]) != checksum {
r.err = ErrCorrupt
return r.err
}
r.i, r.j = 0, n
continue
case chunkTypeUncompressedData:
// Section 4.3. Uncompressed data (chunk type 0x01).
if chunkLen < checksumSize {
r.err = ErrCorrupt
return r.err
}
buf := r.buf[:checksumSize]
if !r.readFull(buf, false) {
return r.err
}
checksum := uint32(buf[0]) | uint32(buf[1])<<8 | uint32(buf[2])<<16 | uint32(buf[3])<<24
// Read directly into r.decoded instead of via r.buf.
n := chunkLen - checksumSize
if n > len(r.decoded) {
r.err = ErrCorrupt
return r.err
}
if !r.readFull(r.decoded[:n], false) {
return r.err
}
if crc(r.decoded[:n]) != checksum {
r.err = ErrCorrupt
return r.err
}
r.i, r.j = 0, n
continue
case chunkTypeStreamIdentifier:
// Section 4.1. Stream identifier (chunk type 0xff).
if chunkLen != len(magicBody) {
r.err = ErrCorrupt
return r.err
}
if !r.readFull(r.buf[:len(magicBody)], false) {
return r.err
}
for i := 0; i < len(magicBody); i++ {
if r.buf[i] != magicBody[i] {
r.err = ErrCorrupt
return r.err
}
}
continue
}
if chunkType <= 0x7f {
// Section 4.5. Reserved unskippable chunks (chunk types 0x02-0x7f).
r.err = ErrUnsupported
return r.err
}
// Section 4.4 Padding (chunk type 0xfe).
// Section 4.6. Reserved skippable chunks (chunk types 0x80-0xfd).
if !r.readFull(r.buf[:chunkLen], false) {
return r.err
}
}
return nil
}
// Read satisfies the io.Reader interface.
func (r *Reader) Read(p []byte) (int, error) {
if r.err != nil {
return 0, r.err
}
if err := r.fill(); err != nil {
return 0, err
}
n := copy(p, r.decoded[r.i:r.j])
r.i += n
return n, nil
}
// ReadByte satisfies the io.ByteReader interface.
func (r *Reader) ReadByte() (byte, error) {
if r.err != nil {
return 0, r.err
}
if err := r.fill(); err != nil {
return 0, err
}
c := r.decoded[r.i]
r.i++
return c, nil
}

113
vendor/github.com/klauspost/compress/internal/snapref/decode_other.go generated vendored Normal file
View File

@@ -0,0 +1,113 @@
// Copyright 2016 The Snappy-Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package snapref
// decode writes the decoding of src to dst. It assumes that the varint-encoded
// length of the decompressed bytes has already been read, and that len(dst)
// equals that length.
//
// It returns 0 on success or a decodeErrCodeXxx error code on failure.
func decode(dst, src []byte) int {
var d, s, offset, length int
for s < len(src) {
switch src[s] & 0x03 {
case tagLiteral:
x := uint32(src[s] >> 2)
switch {
case x < 60:
s++
case x == 60:
s += 2
if uint(s) > uint(len(src)) { // The uint conversions catch overflow from the previous line.
return decodeErrCodeCorrupt
}
x = uint32(src[s-1])
case x == 61:
s += 3
if uint(s) > uint(len(src)) { // The uint conversions catch overflow from the previous line.
return decodeErrCodeCorrupt
}
x = uint32(src[s-2]) | uint32(src[s-1])<<8
case x == 62:
s += 4
if uint(s) > uint(len(src)) { // The uint conversions catch overflow from the previous line.
return decodeErrCodeCorrupt
}
x = uint32(src[s-3]) | uint32(src[s-2])<<8 | uint32(src[s-1])<<16
case x == 63:
s += 5
if uint(s) > uint(len(src)) { // The uint conversions catch overflow from the previous line.
return decodeErrCodeCorrupt
}
x = uint32(src[s-4]) | uint32(src[s-3])<<8 | uint32(src[s-2])<<16 | uint32(src[s-1])<<24
}
length = int(x) + 1
if length <= 0 {
return decodeErrCodeUnsupportedLiteralLength
}
if length > len(dst)-d || length > len(src)-s {
return decodeErrCodeCorrupt
}
copy(dst[d:], src[s:s+length])
d += length
s += length
continue
case tagCopy1:
s += 2
if uint(s) > uint(len(src)) { // The uint conversions catch overflow from the previous line.
return decodeErrCodeCorrupt
}
length = 4 + int(src[s-2])>>2&0x7
offset = int(uint32(src[s-2])&0xe0<<3 | uint32(src[s-1]))
case tagCopy2:
s += 3
if uint(s) > uint(len(src)) { // The uint conversions catch overflow from the previous line.
return decodeErrCodeCorrupt
}
length = 1 + int(src[s-3])>>2
offset = int(uint32(src[s-2]) | uint32(src[s-1])<<8)
case tagCopy4:
s += 5
if uint(s) > uint(len(src)) { // The uint conversions catch overflow from the previous line.
return decodeErrCodeCorrupt
}
length = 1 + int(src[s-5])>>2
offset = int(uint32(src[s-4]) | uint32(src[s-3])<<8 | uint32(src[s-2])<<16 | uint32(src[s-1])<<24)
}
if offset <= 0 || d < offset || length > len(dst)-d {
return decodeErrCodeCorrupt
}
// Copy from an earlier sub-slice of dst to a later sub-slice.
// If no overlap, use the built-in copy:
if offset >= length {
copy(dst[d:d+length], dst[d-offset:])
d += length
continue
}
// Unlike the built-in copy function, this byte-by-byte copy always runs
// forwards, even if the slices overlap. Conceptually, this is:
//
// d += forwardCopy(dst[d:d+length], dst[d-offset:])
//
// We align the slices into a and b and show the compiler they are the same size.
// This allows the loop to run without bounds checks.
a := dst[d : d+length]
b := dst[d-offset:]
b = b[:len(a)]
for i := range a {
a[i] = b[i]
}
d += length
}
if d != len(dst) {
return decodeErrCodeCorrupt
}
return 0
}

289
vendor/github.com/klauspost/compress/internal/snapref/encode.go generated vendored Normal file
View File

@@ -0,0 +1,289 @@
// Copyright 2011 The Snappy-Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package snapref
import (
"encoding/binary"
"errors"
"io"
)
// Encode returns the encoded form of src. The returned slice may be a sub-
// slice of dst if dst was large enough to hold the entire encoded block.
// Otherwise, a newly allocated slice will be returned.
//
// The dst and src must not overlap. It is valid to pass a nil dst.
//
// Encode handles the Snappy block format, not the Snappy stream format.
func Encode(dst, src []byte) []byte {
if n := MaxEncodedLen(len(src)); n < 0 {
panic(ErrTooLarge)
} else if len(dst) < n {
dst = make([]byte, n)
}
// The block starts with the varint-encoded length of the decompressed bytes.
d := binary.PutUvarint(dst, uint64(len(src)))
for len(src) > 0 {
p := src
src = nil
if len(p) > maxBlockSize {
p, src = p[:maxBlockSize], p[maxBlockSize:]
}
if len(p) < minNonLiteralBlockSize {
d += emitLiteral(dst[d:], p)
} else {
d += encodeBlock(dst[d:], p)
}
}
return dst[:d]
}
// inputMargin is the minimum number of extra input bytes to keep, inside
// encodeBlock's inner loop. On some architectures, this margin lets us
// implement a fast path for emitLiteral, where the copy of short (<= 16 byte)
// literals can be implemented as a single load to and store from a 16-byte
// register. That literal's actual length can be as short as 1 byte, so this
// can copy up to 15 bytes too much, but that's OK as subsequent iterations of
// the encoding loop will fix up the copy overrun, and this inputMargin ensures
// that we don't overrun the dst and src buffers.
const inputMargin = 16 - 1
// minNonLiteralBlockSize is the minimum size of the input to encodeBlock that
// could be encoded with a copy tag. This is the minimum with respect to the
// algorithm used by encodeBlock, not a minimum enforced by the file format.
//
// The encoded output must start with at least a 1 byte literal, as there are
// no previous bytes to copy. A minimal (1 byte) copy after that, generated
// from an emitCopy call in encodeBlock's main loop, would require at least
// another inputMargin bytes, for the reason above: we want any emitLiteral
// calls inside encodeBlock's main loop to use the fast path if possible, which
// requires being able to overrun by inputMargin bytes. Thus,
// minNonLiteralBlockSize equals 1 + 1 + inputMargin.
//
// The C++ code doesn't use this exact threshold, but it could, as discussed at
// https://groups.google.com/d/topic/snappy-compression/oGbhsdIJSJ8/discussion
// The difference between Go (2+inputMargin) and C++ (inputMargin) is purely an
// optimization. It should not affect the encoded form. This is tested by
// TestSameEncodingAsCppShortCopies.
const minNonLiteralBlockSize = 1 + 1 + inputMargin
// MaxEncodedLen returns the maximum length of a snappy block, given its
// uncompressed length.
//
// It will return a negative value if srcLen is too large to encode.
func MaxEncodedLen(srcLen int) int {
n := uint64(srcLen)
if n > 0xffffffff {
return -1
}
// Compressed data can be defined as:
// compressed := item* literal*
// item := literal* copy
//
// The trailing literal sequence has a space blowup of at most 62/60
// since a literal of length 60 needs one tag byte + one extra byte
// for length information.
//
// Item blowup is trickier to measure. Suppose the "copy" op copies
// 4 bytes of data. Because of a special check in the encoding code,
// we produce a 4-byte copy only if the offset is < 65536. Therefore
// the copy op takes 3 bytes to encode, and this type of item leads
// to at most the 62/60 blowup for representing literals.
//
// Suppose the "copy" op copies 5 bytes of data. If the offset is big
// enough, it will take 5 bytes to encode the copy op. Therefore the
// worst case here is a one-byte literal followed by a five-byte copy.
// That is, 6 bytes of input turn into 7 bytes of "compressed" data.
//
// This last factor dominates the blowup, so the final estimate is:
n = 32 + n + n/6
if n > 0xffffffff {
return -1
}
return int(n)
}
var errClosed = errors.New("snappy: Writer is closed")
// NewWriter returns a new Writer that compresses to w.
//
// The Writer returned does not buffer writes. There is no need to Flush or
// Close such a Writer.
//
// Deprecated: the Writer returned is not suitable for many small writes, only
// for few large writes. Use NewBufferedWriter instead, which is efficient
// regardless of the frequency and shape of the writes, and remember to Close
// that Writer when done.
func NewWriter(w io.Writer) *Writer {
return &Writer{
w: w,
obuf: make([]byte, obufLen),
}
}
// NewBufferedWriter returns a new Writer that compresses to w, using the
// framing format described at
// https://github.com/google/snappy/blob/master/framing_format.txt
//
// The Writer returned buffers writes. Users must call Close to guarantee all
// data has been forwarded to the underlying io.Writer. They may also call
// Flush zero or more times before calling Close.
func NewBufferedWriter(w io.Writer) *Writer {
return &Writer{
w: w,
ibuf: make([]byte, 0, maxBlockSize),
obuf: make([]byte, obufLen),
}
}
// Writer is an io.Writer that can write Snappy-compressed bytes.
//
// Writer handles the Snappy stream format, not the Snappy block format.
type Writer struct {
w io.Writer
err error
// ibuf is a buffer for the incoming (uncompressed) bytes.
//
// Its use is optional. For backwards compatibility, Writers created by the
// NewWriter function have ibuf == nil, do not buffer incoming bytes, and
// therefore do not need to be Flush'ed or Close'd.
ibuf []byte
// obuf is a buffer for the outgoing (compressed) bytes.
obuf []byte
// wroteStreamHeader is whether we have written the stream header.
wroteStreamHeader bool
}
// Reset discards the writer's state and switches the Snappy writer to write to
// w. This permits reusing a Writer rather than allocating a new one.
func (w *Writer) Reset(writer io.Writer) {
w.w = writer
w.err = nil
if w.ibuf != nil {
w.ibuf = w.ibuf[:0]
}
w.wroteStreamHeader = false
}
// Write satisfies the io.Writer interface.
func (w *Writer) Write(p []byte) (nRet int, errRet error) {
if w.ibuf == nil {
// Do not buffer incoming bytes. This does not perform or compress well
// if the caller of Writer.Write writes many small slices. This
// behavior is therefore deprecated, but still supported for backwards
// compatibility with code that doesn't explicitly Flush or Close.
return w.write(p)
}
// The remainder of this method is based on bufio.Writer.Write from the
// standard library.
for len(p) > (cap(w.ibuf)-len(w.ibuf)) && w.err == nil {
var n int
if len(w.ibuf) == 0 {
// Large write, empty buffer.
// Write directly from p to avoid copy.
n, _ = w.write(p)
} else {
n = copy(w.ibuf[len(w.ibuf):cap(w.ibuf)], p)
w.ibuf = w.ibuf[:len(w.ibuf)+n]
w.Flush()
}
nRet += n
p = p[n:]
}
if w.err != nil {
return nRet, w.err
}
n := copy(w.ibuf[len(w.ibuf):cap(w.ibuf)], p)
w.ibuf = w.ibuf[:len(w.ibuf)+n]
nRet += n
return nRet, nil
}
func (w *Writer) write(p []byte) (nRet int, errRet error) {
if w.err != nil {
return 0, w.err
}
for len(p) > 0 {
obufStart := len(magicChunk)
if !w.wroteStreamHeader {
w.wroteStreamHeader = true
copy(w.obuf, magicChunk)
obufStart = 0
}
var uncompressed []byte
if len(p) > maxBlockSize {
uncompressed, p = p[:maxBlockSize], p[maxBlockSize:]
} else {
uncompressed, p = p, nil
}
checksum := crc(uncompressed)
// Compress the buffer, discarding the result if the improvement
// isn't at least 12.5%.
compressed := Encode(w.obuf[obufHeaderLen:], uncompressed)
chunkType := uint8(chunkTypeCompressedData)
chunkLen := 4 + len(compressed)
obufEnd := obufHeaderLen + len(compressed)
if len(compressed) >= len(uncompressed)-len(uncompressed)/8 {
chunkType = chunkTypeUncompressedData
chunkLen = 4 + len(uncompressed)
obufEnd = obufHeaderLen
}
// Fill in the per-chunk header that comes before the body.
w.obuf[len(magicChunk)+0] = chunkType
w.obuf[len(magicChunk)+1] = uint8(chunkLen >> 0)
w.obuf[len(magicChunk)+2] = uint8(chunkLen >> 8)
w.obuf[len(magicChunk)+3] = uint8(chunkLen >> 16)
w.obuf[len(magicChunk)+4] = uint8(checksum >> 0)
w.obuf[len(magicChunk)+5] = uint8(checksum >> 8)
w.obuf[len(magicChunk)+6] = uint8(checksum >> 16)
w.obuf[len(magicChunk)+7] = uint8(checksum >> 24)
if _, err := w.w.Write(w.obuf[obufStart:obufEnd]); err != nil {
w.err = err
return nRet, err
}
if chunkType == chunkTypeUncompressedData {
if _, err := w.w.Write(uncompressed); err != nil {
w.err = err
return nRet, err
}
}
nRet += len(uncompressed)
}
return nRet, nil
}
// Flush flushes the Writer to its underlying io.Writer.
func (w *Writer) Flush() error {
if w.err != nil {
return w.err
}
if len(w.ibuf) == 0 {
return nil
}
w.write(w.ibuf)
w.ibuf = w.ibuf[:0]
return w.err
}
// Close calls Flush and then closes the Writer.
func (w *Writer) Close() error {
w.Flush()
ret := w.err
if w.err == nil {
w.err = errClosed
}
return ret
}

250
vendor/github.com/klauspost/compress/internal/snapref/encode_other.go generated vendored Normal file
View File

@@ -0,0 +1,250 @@
// Copyright 2016 The Snappy-Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package snapref
func load32(b []byte, i int) uint32 {
b = b[i : i+4 : len(b)] // Help the compiler eliminate bounds checks on the next line.
return uint32(b[0]) | uint32(b[1])<<8 | uint32(b[2])<<16 | uint32(b[3])<<24
}
func load64(b []byte, i int) uint64 {
b = b[i : i+8 : len(b)] // Help the compiler eliminate bounds checks on the next line.
return uint64(b[0]) | uint64(b[1])<<8 | uint64(b[2])<<16 | uint64(b[3])<<24 |
uint64(b[4])<<32 | uint64(b[5])<<40 | uint64(b[6])<<48 | uint64(b[7])<<56
}
// emitLiteral writes a literal chunk and returns the number of bytes written.
//
// It assumes that:
//
// dst is long enough to hold the encoded bytes
// 1 <= len(lit) && len(lit) <= 65536
func emitLiteral(dst, lit []byte) int {
i, n := 0, uint(len(lit)-1)
switch {
case n < 60:
dst[0] = uint8(n)<<2 | tagLiteral
i = 1
case n < 1<<8:
dst[0] = 60<<2 | tagLiteral
dst[1] = uint8(n)
i = 2
default:
dst[0] = 61<<2 | tagLiteral
dst[1] = uint8(n)
dst[2] = uint8(n >> 8)
i = 3
}
return i + copy(dst[i:], lit)
}
// emitCopy writes a copy chunk and returns the number of bytes written.
//
// It assumes that:
//
// dst is long enough to hold the encoded bytes
// 1 <= offset && offset <= 65535
// 4 <= length && length <= 65535
func emitCopy(dst []byte, offset, length int) int {
i := 0
// The maximum length for a single tagCopy1 or tagCopy2 op is 64 bytes. The
// threshold for this loop is a little higher (at 68 = 64 + 4), and the
// length emitted down below is a little lower (at 60 = 64 - 4), because
// it's shorter to encode a length 67 copy as a length 60 tagCopy2 followed
// by a length 7 tagCopy1 (which encodes as 3+2 bytes) than to encode it as
// a length 64 tagCopy2 followed by a length 3 tagCopy2 (which encodes as
// 3+3 bytes). The magic 4 in the 64±4 is because the minimum length for a
// tagCopy1 op is 4 bytes, which is why a length 3 copy has to be an
// encodes-as-3-bytes tagCopy2 instead of an encodes-as-2-bytes tagCopy1.
for length >= 68 {
// Emit a length 64 copy, encoded as 3 bytes.
dst[i+0] = 63<<2 | tagCopy2
dst[i+1] = uint8(offset)
dst[i+2] = uint8(offset >> 8)
i += 3
length -= 64
}
if length > 64 {
// Emit a length 60 copy, encoded as 3 bytes.
dst[i+0] = 59<<2 | tagCopy2
dst[i+1] = uint8(offset)
dst[i+2] = uint8(offset >> 8)
i += 3
length -= 60
}
if length >= 12 || offset >= 2048 {
// Emit the remaining copy, encoded as 3 bytes.
dst[i+0] = uint8(length-1)<<2 | tagCopy2
dst[i+1] = uint8(offset)
dst[i+2] = uint8(offset >> 8)
return i + 3
}
// Emit the remaining copy, encoded as 2 bytes.
dst[i+0] = uint8(offset>>8)<<5 | uint8(length-4)<<2 | tagCopy1
dst[i+1] = uint8(offset)
return i + 2
}
func hash(u, shift uint32) uint32 {
return (u * 0x1e35a7bd) >> shift
}
// EncodeBlockInto exposes encodeBlock but checks dst size.
func EncodeBlockInto(dst, src []byte) (d int) {
if MaxEncodedLen(len(src)) > len(dst) {
return 0
}
// encodeBlock breaks on too big blocks, so split.
for len(src) > 0 {
p := src
src = nil
if len(p) > maxBlockSize {
p, src = p[:maxBlockSize], p[maxBlockSize:]
}
if len(p) < minNonLiteralBlockSize {
d += emitLiteral(dst[d:], p)
} else {
d += encodeBlock(dst[d:], p)
}
}
return d
}
// encodeBlock encodes a non-empty src to a guaranteed-large-enough dst. It
// assumes that the varint-encoded length of the decompressed bytes has already
// been written.
//
// It also assumes that:
//
// len(dst) >= MaxEncodedLen(len(src)) &&
// minNonLiteralBlockSize <= len(src) && len(src) <= maxBlockSize
func encodeBlock(dst, src []byte) (d int) {
// Initialize the hash table. Its size ranges from 1<<8 to 1<<14 inclusive.
// The table element type is uint16, as s < sLimit and sLimit < len(src)
// and len(src) <= maxBlockSize and maxBlockSize == 65536.
const (
maxTableSize = 1 << 14
// tableMask is redundant, but helps the compiler eliminate bounds
// checks.
tableMask = maxTableSize - 1
)
shift := uint32(32 - 8)
for tableSize := 1 << 8; tableSize < maxTableSize && tableSize < len(src); tableSize *= 2 {
shift--
}
// In Go, all array elements are zero-initialized, so there is no advantage
// to a smaller tableSize per se. However, it matches the C++ algorithm,
// and in the asm versions of this code, we can get away with zeroing only
// the first tableSize elements.
var table [maxTableSize]uint16
// sLimit is when to stop looking for offset/length copies. The inputMargin
// lets us use a fast path for emitLiteral in the main loop, while we are
// looking for copies.
sLimit := len(src) - inputMargin
// nextEmit is where in src the next emitLiteral should start from.
nextEmit := 0
// The encoded form must start with a literal, as there are no previous
// bytes to copy, so we start looking for hash matches at s == 1.
s := 1
nextHash := hash(load32(src, s), shift)
for {
// Copied from the C++ snappy implementation:
//
// Heuristic match skipping: If 32 bytes are scanned with no matches
// found, start looking only at every other byte. If 32 more bytes are
// scanned (or skipped), look at every third byte, etc.. When a match
// is found, immediately go back to looking at every byte. This is a
// small loss (~5% performance, ~0.1% density) for compressible data
// due to more bookkeeping, but for non-compressible data (such as
// JPEG) it's a huge win since the compressor quickly "realizes" the
// data is incompressible and doesn't bother looking for matches
// everywhere.
//
// The "skip" variable keeps track of how many bytes there are since
// the last match; dividing it by 32 (ie. right-shifting by five) gives
// the number of bytes to move ahead for each iteration.
skip := 32
nextS := s
candidate := 0
for {
s = nextS
bytesBetweenHashLookups := skip >> 5
nextS = s + bytesBetweenHashLookups
skip += bytesBetweenHashLookups
if nextS > sLimit {
goto emitRemainder
}
candidate = int(table[nextHash&tableMask])
table[nextHash&tableMask] = uint16(s)
nextHash = hash(load32(src, nextS), shift)
if load32(src, s) == load32(src, candidate) {
break
}
}
// A 4-byte match has been found. We'll later see if more than 4 bytes
// match. But, prior to the match, src[nextEmit:s] are unmatched. Emit
// them as literal bytes.
d += emitLiteral(dst[d:], src[nextEmit:s])
// Call emitCopy, and then see if another emitCopy could be our next
// move. Repeat until we find no match for the input immediately after
// what was consumed by the last emitCopy call.
//
// If we exit this loop normally then we need to call emitLiteral next,
// though we don't yet know how big the literal will be. We handle that
// by proceeding to the next iteration of the main loop. We also can
// exit this loop via goto if we get close to exhausting the input.
for {
// Invariant: we have a 4-byte match at s, and no need to emit any
// literal bytes prior to s.
base := s
// Extend the 4-byte match as long as possible.
//
// This is an inlined version of:
// s = extendMatch(src, candidate+4, s+4)
s += 4
for i := candidate + 4; s < len(src) && src[i] == src[s]; i, s = i+1, s+1 {
}
d += emitCopy(dst[d:], base-candidate, s-base)
nextEmit = s
if s >= sLimit {
goto emitRemainder
}
// We could immediately start working at s now, but to improve
// compression we first update the hash table at s-1 and at s. If
// another emitCopy is not our next move, also calculate nextHash
// at s+1. At least on GOARCH=amd64, these three hash calculations
// are faster as one load64 call (with some shifts) instead of
// three load32 calls.
x := load64(src, s-1)
prevHash := hash(uint32(x>>0), shift)
table[prevHash&tableMask] = uint16(s - 1)
currHash := hash(uint32(x>>8), shift)
candidate = int(table[currHash&tableMask])
table[currHash&tableMask] = uint16(s)
if uint32(x>>8) != load32(src, candidate) {
nextHash = hash(uint32(x>>16), shift)
s++
break
}
}
}
emitRemainder:
if nextEmit < len(src) {
d += emitLiteral(dst[d:], src[nextEmit:])
}
return d
}

98
vendor/github.com/klauspost/compress/internal/snapref/snappy.go generated vendored Normal file
View File

@@ -0,0 +1,98 @@
// Copyright 2011 The Snappy-Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Package snapref implements the Snappy compression format. It aims for very
// high speeds and reasonable compression.
//
// There are actually two Snappy formats: block and stream. They are related,
// but different: trying to decompress block-compressed data as a Snappy stream
// will fail, and vice versa. The block format is the Decode and Encode
// functions and the stream format is the Reader and Writer types.
//
// The block format, the more common case, is used when the complete size (the
// number of bytes) of the original data is known upfront, at the time
// compression starts. The stream format, also known as the framing format, is
// for when that isn't always true.
//
// The canonical, C++ implementation is at https://github.com/google/snappy and
// it only implements the block format.
package snapref
import (
"hash/crc32"
)
/*
Each encoded block begins with the varint-encoded length of the decoded data,
followed by a sequence of chunks. Chunks begin and end on byte boundaries. The
first byte of each chunk is broken into its 2 least and 6 most significant bits
called l and m: l ranges in [0, 4) and m ranges in [0, 64). l is the chunk tag.
Zero means a literal tag. All other values mean a copy tag.
For literal tags:
- If m < 60, the next 1 + m bytes are literal bytes.
- Otherwise, let n be the little-endian unsigned integer denoted by the next
m - 59 bytes. The next 1 + n bytes after that are literal bytes.
For copy tags, length bytes are copied from offset bytes ago, in the style of
Lempel-Ziv compression algorithms. In particular:
- For l == 1, the offset ranges in [0, 1<<11) and the length in [4, 12).
The length is 4 + the low 3 bits of m. The high 3 bits of m form bits 8-10
of the offset. The next byte is bits 0-7 of the offset.
- For l == 2, the offset ranges in [0, 1<<16) and the length in [1, 65).
The length is 1 + m. The offset is the little-endian unsigned integer
denoted by the next 2 bytes.
- For l == 3, this tag is a legacy format that is no longer issued by most
encoders. Nonetheless, the offset ranges in [0, 1<<32) and the length in
[1, 65). The length is 1 + m. The offset is the little-endian unsigned
integer denoted by the next 4 bytes.
*/
const (
tagLiteral = 0x00
tagCopy1 = 0x01
tagCopy2 = 0x02
tagCopy4 = 0x03
)
const (
checksumSize = 4
chunkHeaderSize = 4
magicChunk = "\xff\x06\x00\x00" + magicBody
magicBody = "sNaPpY"
// maxBlockSize is the maximum size of the input to encodeBlock. It is not
// part of the wire format per se, but some parts of the encoder assume
// that an offset fits into a uint16.
//
// Also, for the framing format (Writer type instead of Encode function),
// https://github.com/google/snappy/blob/master/framing_format.txt says
// that "the uncompressed data in a chunk must be no longer than 65536
// bytes".
maxBlockSize = 65536
// maxEncodedLenOfMaxBlockSize equals MaxEncodedLen(maxBlockSize), but is
// hard coded to be a const instead of a variable, so that obufLen can also
// be a const. Their equivalence is confirmed by
// TestMaxEncodedLenOfMaxBlockSize.
maxEncodedLenOfMaxBlockSize = 76490
obufHeaderLen = len(magicChunk) + checksumSize + chunkHeaderSize
obufLen = obufHeaderLen + maxEncodedLenOfMaxBlockSize
)
const (
chunkTypeCompressedData = 0x00
chunkTypeUncompressedData = 0x01
chunkTypePadding = 0xfe
chunkTypeStreamIdentifier = 0xff
)
var crcTable = crc32.MakeTable(crc32.Castagnoli)
// crc implements the checksum specified in section 3 of
// https://github.com/google/snappy/blob/master/framing_format.txt
func crc(b []byte) uint32 {
c := crc32.Update(0, crcTable, b)
return uint32(c>>15|c<<17) + 0xa282ead8
}

4
vendor/github.com/klauspost/compress/s2sx.mod generated vendored Normal file
View File

@@ -0,0 +1,4 @@
module github.com/klauspost/compress
go 1.19

0
vendor/github.com/klauspost/compress/s2sx.sum generated vendored Normal file
View File

16
vendor/github.com/klauspost/compress/snappy/.gitignore generated vendored Normal file
View File

@@ -0,0 +1,16 @@
cmd/snappytool/snappytool
testdata/bench
# These explicitly listed benchmark data files are for an obsolete version of
# snappy_test.go.
testdata/alice29.txt
testdata/asyoulik.txt
testdata/fireworks.jpeg
testdata/geo.protodata
testdata/html
testdata/html_x_4
testdata/kppkn.gtb
testdata/lcet10.txt
testdata/paper-100k.pdf
testdata/plrabn12.txt
testdata/urls.10K

18
vendor/github.com/klauspost/compress/snappy/AUTHORS generated vendored Normal file
View File

@@ -0,0 +1,18 @@
# This is the official list of Snappy-Go authors for copyright purposes.
# This file is distinct from the CONTRIBUTORS files.
# See the latter for an explanation.
# Names should be added to this file as
# Name or Organization <email address>
# The email address is not required for organizations.
# Please keep the list sorted.
Amazon.com, Inc
Damian Gryski <dgryski@gmail.com>
Eric Buth <eric@topos.com>
Google Inc.
Jan Mercl <0xjnml@gmail.com>
Klaus Post <klauspost@gmail.com>
Rodolfo Carvalho <rhcarvalho@gmail.com>
Sebastien Binet <seb.binet@gmail.com>

41
vendor/github.com/klauspost/compress/snappy/CONTRIBUTORS generated vendored Normal file
View File

@@ -0,0 +1,41 @@
# This is the official list of people who can contribute
# (and typically have contributed) code to the Snappy-Go repository.
# The AUTHORS file lists the copyright holders; this file
# lists people. For example, Google employees are listed here
# but not in AUTHORS, because Google holds the copyright.
#
# The submission process automatically checks to make sure
# that people submitting code are listed in this file (by email address).
#
# Names should be added to this file only after verifying that
# the individual or the individual's organization has agreed to
# the appropriate Contributor License Agreement, found here:
#
# http://code.google.com/legal/individual-cla-v1.0.html
# http://code.google.com/legal/corporate-cla-v1.0.html
#
# The agreement for individuals can be filled out on the web.
#
# When adding J Random Contributor's name to this file,
# either J's name or J's organization's name should be
# added to the AUTHORS file, depending on whether the
# individual or corporate CLA was used.
# Names should be added to this file like so:
# Name <email address>
# Please keep the list sorted.
Alex Legg <alexlegg@google.com>
Damian Gryski <dgryski@gmail.com>
Eric Buth <eric@topos.com>
Jan Mercl <0xjnml@gmail.com>
Jonathan Swinney <jswinney@amazon.com>
Kai Backman <kaib@golang.org>
Klaus Post <klauspost@gmail.com>
Marc-Antoine Ruel <maruel@chromium.org>
Nigel Tao <nigeltao@golang.org>
Rob Pike <r@golang.org>
Rodolfo Carvalho <rhcarvalho@gmail.com>
Russ Cox <rsc@golang.org>
Sebastien Binet <seb.binet@gmail.com>

27
vendor/github.com/klauspost/compress/snappy/LICENSE generated vendored Normal file
View File

@@ -0,0 +1,27 @@
Copyright (c) 2011 The Snappy-Go Authors. All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above
copyright notice, this list of conditions and the following disclaimer
in the documentation and/or other materials provided with the
distribution.
* Neither the name of Google Inc. nor the names of its
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

17
vendor/github.com/klauspost/compress/snappy/README.md generated vendored Normal file
View File

@@ -0,0 +1,17 @@
# snappy
The Snappy compression format in the Go programming language.
This is a drop-in replacement for `github.com/golang/snappy`.
It provides a full, compatible replacement of the Snappy package by simply changing imports.
See [Snappy Compatibility](https://github.com/klauspost/compress/tree/master/s2#snappy-compatibility) in the S2 documentation.
"Better" compression mode is used. For buffered streams concurrent compression is used.
For more options use the [s2 package](https://pkg.go.dev/github.com/klauspost/compress/s2).
# usage
Replace imports `github.com/golang/snappy` with `github.com/klauspost/compress/snappy`.

60
vendor/github.com/klauspost/compress/snappy/decode.go generated vendored Normal file
View File

@@ -0,0 +1,60 @@
// Copyright 2011 The Snappy-Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package snappy
import (
"io"
"github.com/klauspost/compress/s2"
)
var (
// ErrCorrupt reports that the input is invalid.
ErrCorrupt = s2.ErrCorrupt
// ErrTooLarge reports that the uncompressed length is too large.
ErrTooLarge = s2.ErrTooLarge
// ErrUnsupported reports that the input isn't supported.
ErrUnsupported = s2.ErrUnsupported
)
const (
// maxBlockSize is the maximum size of the input to encodeBlock. It is not
// part of the wire format per se, but some parts of the encoder assume
// that an offset fits into a uint16.
//
// Also, for the framing format (Writer type instead of Encode function),
// https://github.com/google/snappy/blob/master/framing_format.txt says
// that "the uncompressed data in a chunk must be no longer than 65536
// bytes".
maxBlockSize = 65536
)
// DecodedLen returns the length of the decoded block.
func DecodedLen(src []byte) (int, error) {
return s2.DecodedLen(src)
}
// Decode returns the decoded form of src. The returned slice may be a sub-
// slice of dst if dst was large enough to hold the entire decoded block.
// Otherwise, a newly allocated slice will be returned.
//
// The dst and src must not overlap. It is valid to pass a nil dst.
//
// Decode handles the Snappy block format, not the Snappy stream format.
func Decode(dst, src []byte) ([]byte, error) {
return s2.Decode(dst, src)
}
// NewReader returns a new Reader that decompresses from r, using the framing
// format described at
// https://github.com/google/snappy/blob/master/framing_format.txt
func NewReader(r io.Reader) *Reader {
return s2.NewReader(r, s2.ReaderMaxBlockSize(maxBlockSize))
}
// Reader is an io.Reader that can read Snappy-compressed bytes.
//
// Reader handles the Snappy stream format, not the Snappy block format.
type Reader = s2.Reader

59
vendor/github.com/klauspost/compress/snappy/encode.go generated vendored Normal file
View File

@@ -0,0 +1,59 @@
// Copyright 2011 The Snappy-Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package snappy
import (
"io"
"github.com/klauspost/compress/s2"
)
// Encode returns the encoded form of src. The returned slice may be a sub-
// slice of dst if dst was large enough to hold the entire encoded block.
// Otherwise, a newly allocated slice will be returned.
//
// The dst and src must not overlap. It is valid to pass a nil dst.
//
// Encode handles the Snappy block format, not the Snappy stream format.
func Encode(dst, src []byte) []byte {
return s2.EncodeSnappyBetter(dst, src)
}
// MaxEncodedLen returns the maximum length of a snappy block, given its
// uncompressed length.
//
// It will return a negative value if srcLen is too large to encode.
func MaxEncodedLen(srcLen int) int {
return s2.MaxEncodedLen(srcLen)
}
// NewWriter returns a new Writer that compresses to w.
//
// The Writer returned does not buffer writes. There is no need to Flush or
// Close such a Writer.
//
// Deprecated: the Writer returned is not suitable for many small writes, only
// for few large writes. Use NewBufferedWriter instead, which is efficient
// regardless of the frequency and shape of the writes, and remember to Close
// that Writer when done.
func NewWriter(w io.Writer) *Writer {
return s2.NewWriter(w, s2.WriterSnappyCompat(), s2.WriterBetterCompression(), s2.WriterFlushOnWrite(), s2.WriterConcurrency(1))
}
// NewBufferedWriter returns a new Writer that compresses to w, using the
// framing format described at
// https://github.com/google/snappy/blob/master/framing_format.txt
//
// The Writer returned buffers writes. Users must call Close to guarantee all
// data has been forwarded to the underlying io.Writer. They may also call
// Flush zero or more times before calling Close.
func NewBufferedWriter(w io.Writer) *Writer {
return s2.NewWriter(w, s2.WriterSnappyCompat(), s2.WriterBetterCompression())
}
// Writer is an io.Writer that can write Snappy-compressed bytes.
//
// Writer handles the Snappy stream format, not the Snappy block format.
type Writer = s2.Writer

46
vendor/github.com/klauspost/compress/snappy/snappy.go generated vendored Normal file
View File

@@ -0,0 +1,46 @@
// Copyright 2011 The Snappy-Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Package snappy implements the Snappy compression format. It aims for very
// high speeds and reasonable compression.
//
// There are actually two Snappy formats: block and stream. They are related,
// but different: trying to decompress block-compressed data as a Snappy stream
// will fail, and vice versa. The block format is the Decode and Encode
// functions and the stream format is the Reader and Writer types.
//
// The block format, the more common case, is used when the complete size (the
// number of bytes) of the original data is known upfront, at the time
// compression starts. The stream format, also known as the framing format, is
// for when that isn't always true.
//
// The canonical, C++ implementation is at https://github.com/google/snappy and
// it only implements the block format.
package snappy
/*
Each encoded block begins with the varint-encoded length of the decoded data,
followed by a sequence of chunks. Chunks begin and end on byte boundaries. The
first byte of each chunk is broken into its 2 least and 6 most significant bits
called l and m: l ranges in [0, 4) and m ranges in [0, 64). l is the chunk tag.
Zero means a literal tag. All other values mean a copy tag.
For literal tags:
- If m < 60, the next 1 + m bytes are literal bytes.
- Otherwise, let n be the little-endian unsigned integer denoted by the next
m - 59 bytes. The next 1 + n bytes after that are literal bytes.
For copy tags, length bytes are copied from offset bytes ago, in the style of
Lempel-Ziv compression algorithms. In particular:
- For l == 1, the offset ranges in [0, 1<<11) and the length in [4, 12).
The length is 4 + the low 3 bits of m. The high 3 bits of m form bits 8-10
of the offset. The next byte is bits 0-7 of the offset.
- For l == 2, the offset ranges in [0, 1<<16) and the length in [1, 65).
The length is 1 + m. The offset is the little-endian unsigned integer
denoted by the next 2 bytes.
- For l == 3, this tag is a legacy format that is no longer issued by most
encoders. Nonetheless, the offset ranges in [0, 1<<32) and the length in
[1, 65). The length is 1 + m. The offset is the little-endian unsigned
integer denoted by the next 4 bytes.
*/

441
vendor/github.com/klauspost/compress/zstd/README.md generated vendored Normal file
View File

@@ -0,0 +1,441 @@
# zstd
[Zstandard](https://facebook.github.io/zstd/) is a real-time compression algorithm, providing high compression ratios.
It offers a very wide range of compression / speed trade-off, while being backed by a very fast decoder.
A high performance compression algorithm is implemented. For now focused on speed.
This package provides [compression](#Compressor) to and [decompression](#Decompressor) of Zstandard content.
This package is pure Go and without use of "unsafe".
The `zstd` package is provided as open source software using a Go standard license.
Currently the package is heavily optimized for 64 bit processors and will be significantly slower on 32 bit processors.
For seekable zstd streams, see [this excellent package](https://github.com/SaveTheRbtz/zstd-seekable-format-go).
## Installation
Install using `go get -u github.com/klauspost/compress`. The package is located in `github.com/klauspost/compress/zstd`.
[![Go Reference](https://pkg.go.dev/badge/github.com/klauspost/compress/zstd.svg)](https://pkg.go.dev/github.com/klauspost/compress/zstd)
## Compressor
### Status:
STABLE - there may always be subtle bugs, a wide variety of content has been tested and the library is actively
used by several projects. This library is being [fuzz-tested](https://github.com/klauspost/compress-fuzz) for all updates.
There may still be specific combinations of data types/size/settings that could lead to edge cases,
so as always, testing is recommended.
For now, a high speed (fastest) and medium-fast (default) compressor has been implemented.
* The "Fastest" compression ratio is roughly equivalent to zstd level 1.
* The "Default" compression ratio is roughly equivalent to zstd level 3 (default).
* The "Better" compression ratio is roughly equivalent to zstd level 7.
* The "Best" compression ratio is roughly equivalent to zstd level 11.
In terms of speed, it is typically 2x as fast as the stdlib deflate/gzip in its fastest mode.
The compression ratio compared to stdlib is around level 3, but usually 3x as fast.
### Usage
An Encoder can be used for either compressing a stream via the
`io.WriteCloser` interface supported by the Encoder or as multiple independent
tasks via the `EncodeAll` function.
Smaller encodes are encouraged to use the EncodeAll function.
Use `NewWriter` to create a new instance that can be used for both.
To create a writer with default options, do like this:
```Go
// Compress input to output.
func Compress(in io.Reader, out io.Writer) error {
enc, err := zstd.NewWriter(out)
if err != nil {
return err
}
_, err = io.Copy(enc, in)
if err != nil {
enc.Close()
return err
}
return enc.Close()
}
```
Now you can encode by writing data to `enc`. The output will be finished writing when `Close()` is called.
Even if your encode fails, you should still call `Close()` to release any resources that may be held up.
The above is fine for big encodes. However, whenever possible try to *reuse* the writer.
To reuse the encoder, you can use the `Reset(io.Writer)` function to change to another output.
This will allow the encoder to reuse all resources and avoid wasteful allocations.
Currently stream encoding has 'light' concurrency, meaning up to 2 goroutines can be working on part
of a stream. This is independent of the `WithEncoderConcurrency(n)`, but that is likely to change
in the future. So if you want to limit concurrency for future updates, specify the concurrency
you would like.
If you would like stream encoding to be done without spawning async goroutines, use `WithEncoderConcurrency(1)`
which will compress input as each block is completed, blocking on writes until each has completed.
You can specify your desired compression level using `WithEncoderLevel()` option. Currently only pre-defined
compression settings can be specified.
#### Future Compatibility Guarantees
This will be an evolving project. When using this package it is important to note that both the compression efficiency and speed may change.
The goal will be to keep the default efficiency at the default zstd (level 3).
However the encoding should never be assumed to remain the same,
and you should not use hashes of compressed output for similarity checks.
The Encoder can be assumed to produce the same output from the exact same code version.
However, the may be modes in the future that break this,
although they will not be enabled without an explicit option.
This encoder is not designed to (and will probably never) output the exact same bitstream as the reference encoder.
Also note, that the cgo decompressor currently does not [report all errors on invalid input](https://github.com/DataDog/zstd/issues/59),
[omits error checks](https://github.com/DataDog/zstd/issues/61), [ignores checksums](https://github.com/DataDog/zstd/issues/43)
and seems to ignore concatenated streams, even though [it is part of the spec](https://github.com/facebook/zstd/blob/dev/doc/zstd_compression_format.md#frames).
#### Blocks
For compressing small blocks, the returned encoder has a function called `EncodeAll(src, dst []byte) []byte`.
`EncodeAll` will encode all input in src and append it to dst.
This function can be called concurrently.
Each call will only run on a same goroutine as the caller.
Encoded blocks can be concatenated and the result will be the combined input stream.
Data compressed with EncodeAll can be decoded with the Decoder, using either a stream or `DecodeAll`.
Especially when encoding blocks you should take special care to reuse the encoder.
This will effectively make it run without allocations after a warmup period.
To make it run completely without allocations, supply a destination buffer with space for all content.
```Go
import "github.com/klauspost/compress/zstd"
// Create a writer that caches compressors.
// For this operation type we supply a nil Reader.
var encoder, _ = zstd.NewWriter(nil)
// Compress a buffer.
// If you have a destination buffer, the allocation in the call can also be eliminated.
func Compress(src []byte) []byte {
return encoder.EncodeAll(src, make([]byte, 0, len(src)))
}
```
You can control the maximum number of concurrent encodes using the `WithEncoderConcurrency(n)`
option when creating the writer.
Using the Encoder for both a stream and individual blocks concurrently is safe.
### Performance
I have collected some speed examples to compare speed and compression against other compressors.
* `file` is the input file.
* `out` is the compressor used. `zskp` is this package. `zstd` is the Datadog cgo library. `gzstd/gzkp` is gzip standard and this library.
* `level` is the compression level used. For `zskp` level 1 is "fastest", level 2 is "default"; 3 is "better", 4 is "best".
* `insize`/`outsize` is the input/output size.
* `millis` is the number of milliseconds used for compression.
* `mb/s` is megabytes (2^20 bytes) per second.
```
Silesia Corpus:
http://sun.aei.polsl.pl/~sdeor/corpus/silesia.zip
This package:
file out level insize outsize millis mb/s
silesia.tar zskp 1 211947520 73821326 634 318.47
silesia.tar zskp 2 211947520 67655404 1508 133.96
silesia.tar zskp 3 211947520 64746933 3000 67.37
silesia.tar zskp 4 211947520 60073508 16926 11.94
cgo zstd:
silesia.tar zstd 1 211947520 73605392 543 371.56
silesia.tar zstd 3 211947520 66793289 864 233.68
silesia.tar zstd 6 211947520 62916450 1913 105.66
silesia.tar zstd 9 211947520 60212393 5063 39.92
gzip, stdlib/this package:
silesia.tar gzstd 1 211947520 80007735 1498 134.87
silesia.tar gzkp 1 211947520 80088272 1009 200.31
GOB stream of binary data. Highly compressible.
https://files.klauspost.com/compress/gob-stream.7z
file out level insize outsize millis mb/s
gob-stream zskp 1 1911399616 233948096 3230 564.34
gob-stream zskp 2 1911399616 203997694 4997 364.73
gob-stream zskp 3 1911399616 173526523 13435 135.68
gob-stream zskp 4 1911399616 162195235 47559 38.33
gob-stream zstd 1 1911399616 249810424 2637 691.26
gob-stream zstd 3 1911399616 208192146 3490 522.31
gob-stream zstd 6 1911399616 193632038 6687 272.56
gob-stream zstd 9 1911399616 177620386 16175 112.70
gob-stream gzstd 1 1911399616 357382013 9046 201.49
gob-stream gzkp 1 1911399616 359136669 4885 373.08
The test data for the Large Text Compression Benchmark is the first
10^9 bytes of the English Wikipedia dump on Mar. 3, 2006.
http://mattmahoney.net/dc/textdata.html
file out level insize outsize millis mb/s
enwik9 zskp 1 1000000000 343833605 3687 258.64
enwik9 zskp 2 1000000000 317001237 7672 124.29
enwik9 zskp 3 1000000000 291915823 15923 59.89
enwik9 zskp 4 1000000000 261710291 77697 12.27
enwik9 zstd 1 1000000000 358072021 3110 306.65
enwik9 zstd 3 1000000000 313734672 4784 199.35
enwik9 zstd 6 1000000000 295138875 10290 92.68
enwik9 zstd 9 1000000000 278348700 28549 33.40
enwik9 gzstd 1 1000000000 382578136 8608 110.78
enwik9 gzkp 1 1000000000 382781160 5628 169.45
Highly compressible JSON file.
https://files.klauspost.com/compress/github-june-2days-2019.json.zst
file out level insize outsize millis mb/s
github-june-2days-2019.json zskp 1 6273951764 697439532 9789 611.17
github-june-2days-2019.json zskp 2 6273951764 610876538 18553 322.49
github-june-2days-2019.json zskp 3 6273951764 517662858 44186 135.41
github-june-2days-2019.json zskp 4 6273951764 464617114 165373 36.18
github-june-2days-2019.json zstd 1 6273951764 766284037 8450 708.00
github-june-2days-2019.json zstd 3 6273951764 661889476 10927 547.57
github-june-2days-2019.json zstd 6 6273951764 642756859 22996 260.18
github-june-2days-2019.json zstd 9 6273951764 601974523 52413 114.16
github-june-2days-2019.json gzstd 1 6273951764 1164397768 26793 223.32
github-june-2days-2019.json gzkp 1 6273951764 1120631856 17693 338.16
VM Image, Linux mint with a few installed applications:
https://files.klauspost.com/compress/rawstudio-mint14.7z
file out level insize outsize millis mb/s
rawstudio-mint14.tar zskp 1 8558382592 3718400221 18206 448.29
rawstudio-mint14.tar zskp 2 8558382592 3326118337 37074 220.15
rawstudio-mint14.tar zskp 3 8558382592 3163842361 87306 93.49
rawstudio-mint14.tar zskp 4 8558382592 2970480650 783862 10.41
rawstudio-mint14.tar zstd 1 8558382592 3609250104 17136 476.27
rawstudio-mint14.tar zstd 3 8558382592 3341679997 29262 278.92
rawstudio-mint14.tar zstd 6 8558382592 3235846406 77904 104.77
rawstudio-mint14.tar zstd 9 8558382592 3160778861 140946 57.91
rawstudio-mint14.tar gzstd 1 8558382592 3926234992 51345 158.96
rawstudio-mint14.tar gzkp 1 8558382592 3960117298 36722 222.26
CSV data:
https://files.klauspost.com/compress/nyc-taxi-data-10M.csv.zst
file out level insize outsize millis mb/s
nyc-taxi-data-10M.csv zskp 1 3325605752 641319332 9462 335.17
nyc-taxi-data-10M.csv zskp 2 3325605752 588976126 17570 180.50
nyc-taxi-data-10M.csv zskp 3 3325605752 529329260 32432 97.79
nyc-taxi-data-10M.csv zskp 4 3325605752 474949772 138025 22.98
nyc-taxi-data-10M.csv zstd 1 3325605752 687399637 8233 385.18
nyc-taxi-data-10M.csv zstd 3 3325605752 598514411 10065 315.07
nyc-taxi-data-10M.csv zstd 6 3325605752 570522953 20038 158.27
nyc-taxi-data-10M.csv zstd 9 3325605752 517554797 64565 49.12
nyc-taxi-data-10M.csv gzstd 1 3325605752 928654908 21270 149.11
nyc-taxi-data-10M.csv gzkp 1 3325605752 922273214 13929 227.68
```
## Decompressor
Status: STABLE - there may still be subtle bugs, but a wide variety of content has been tested.
This library is being continuously [fuzz-tested](https://github.com/klauspost/compress-fuzz),
kindly supplied by [fuzzit.dev](https://fuzzit.dev/).
The main purpose of the fuzz testing is to ensure that it is not possible to crash the decoder,
or run it past its limits with ANY input provided.
### Usage
The package has been designed for two main usages, big streams of data and smaller in-memory buffers.
There are two main usages of the package for these. Both of them are accessed by creating a `Decoder`.
For streaming use a simple setup could look like this:
```Go
import "github.com/klauspost/compress/zstd"
func Decompress(in io.Reader, out io.Writer) error {
d, err := zstd.NewReader(in)
if err != nil {
return err
}
defer d.Close()
// Copy content...
_, err = io.Copy(out, d)
return err
}
```
It is important to use the "Close" function when you no longer need the Reader to stop running goroutines,
when running with default settings.
Goroutines will exit once an error has been returned, including `io.EOF` at the end of a stream.
Streams are decoded concurrently in 4 asynchronous stages to give the best possible throughput.
However, if you prefer synchronous decompression, use `WithDecoderConcurrency(1)` which will decompress data
as it is being requested only.
For decoding buffers, it could look something like this:
```Go
import "github.com/klauspost/compress/zstd"
// Create a reader that caches decompressors.
// For this operation type we supply a nil Reader.
var decoder, _ = zstd.NewReader(nil, zstd.WithDecoderConcurrency(0))
// Decompress a buffer. We don't supply a destination buffer,
// so it will be allocated by the decoder.
func Decompress(src []byte) ([]byte, error) {
return decoder.DecodeAll(src, nil)
}
```
Both of these cases should provide the functionality needed.
The decoder can be used for *concurrent* decompression of multiple buffers.
By default 4 decompressors will be created.
It will only allow a certain number of concurrent operations to run.
To tweak that yourself use the `WithDecoderConcurrency(n)` option when creating the decoder.
It is possible to use `WithDecoderConcurrency(0)` to create GOMAXPROCS decoders.
### Dictionaries
Data compressed with [dictionaries](https://github.com/facebook/zstd#the-case-for-small-data-compression) can be decompressed.
Dictionaries are added individually to Decoders.
Dictionaries are generated by the `zstd --train` command and contains an initial state for the decoder.
To add a dictionary use the `WithDecoderDicts(dicts ...[]byte)` option with the dictionary data.
Several dictionaries can be added at once.
The dictionary will be used automatically for the data that specifies them.
A re-used Decoder will still contain the dictionaries registered.
When registering multiple dictionaries with the same ID, the last one will be used.
It is possible to use dictionaries when compressing data.
To enable a dictionary use `WithEncoderDict(dict []byte)`. Here only one dictionary will be used
and it will likely be used even if it doesn't improve compression.
The used dictionary must be used to decompress the content.
For any real gains, the dictionary should be built with similar data.
If an unsuitable dictionary is used the output may be slightly larger than using no dictionary.
Use the [zstd commandline tool](https://github.com/facebook/zstd/releases) to build a dictionary from sample data.
For information see [zstd dictionary information](https://github.com/facebook/zstd#the-case-for-small-data-compression).
For now there is a fixed startup performance penalty for compressing content with dictionaries.
This will likely be improved over time. Just be aware to test performance when implementing.
### Allocation-less operation
The decoder has been designed to operate without allocations after a warmup.
This means that you should *store* the decoder for best performance.
To re-use a stream decoder, use the `Reset(r io.Reader) error` to switch to another stream.
A decoder can safely be re-used even if the previous stream failed.
To release the resources, you must call the `Close()` function on a decoder.
After this it can *no longer be reused*, but all running goroutines will be stopped.
So you *must* use this if you will no longer need the Reader.
For decompressing smaller buffers a single decoder can be used.
When decoding buffers, you can supply a destination slice with length 0 and your expected capacity.
In this case no unneeded allocations should be made.
### Concurrency
The buffer decoder does everything on the same goroutine and does nothing concurrently.
It can however decode several buffers concurrently. Use `WithDecoderConcurrency(n)` to limit that.
The stream decoder will create goroutines that:
1) Reads input and splits the input into blocks.
2) Decompression of literals.
3) Decompression of sequences.
4) Reconstruction of output stream.
So effectively this also means the decoder will "read ahead" and prepare data to always be available for output.
The concurrency level will, for streams, determine how many blocks ahead the compression will start.
Since "blocks" are quite dependent on the output of the previous block stream decoding will only have limited concurrency.
In practice this means that concurrency is often limited to utilizing about 3 cores effectively.
### Benchmarks
The first two are streaming decodes and the last are smaller inputs.
Running on AMD Ryzen 9 3950X 16-Core Processor. AMD64 assembly used.
```
BenchmarkDecoderSilesia-32 5 206878840 ns/op 1024.50 MB/s 49808 B/op 43 allocs/op
BenchmarkDecoderEnwik9-32 1 1271809000 ns/op 786.28 MB/s 72048 B/op 52 allocs/op
Concurrent blocks, performance:
BenchmarkDecoder_DecodeAllParallel/kppkn.gtb.zst-32 67356 17857 ns/op 10321.96 MB/s 22.48 pct 102 B/op 0 allocs/op
BenchmarkDecoder_DecodeAllParallel/geo.protodata.zst-32 266656 4421 ns/op 26823.21 MB/s 11.89 pct 19 B/op 0 allocs/op
BenchmarkDecoder_DecodeAllParallel/plrabn12.txt.zst-32 20992 56842 ns/op 8477.17 MB/s 39.90 pct 754 B/op 0 allocs/op
BenchmarkDecoder_DecodeAllParallel/lcet10.txt.zst-32 27456 43932 ns/op 9714.01 MB/s 33.27 pct 524 B/op 0 allocs/op
BenchmarkDecoder_DecodeAllParallel/asyoulik.txt.zst-32 78432 15047 ns/op 8319.15 MB/s 40.34 pct 66 B/op 0 allocs/op
BenchmarkDecoder_DecodeAllParallel/alice29.txt.zst-32 65800 18436 ns/op 8249.63 MB/s 37.75 pct 88 B/op 0 allocs/op
BenchmarkDecoder_DecodeAllParallel/html_x_4.zst-32 102993 11523 ns/op 35546.09 MB/s 3.637 pct 143 B/op 0 allocs/op
BenchmarkDecoder_DecodeAllParallel/paper-100k.pdf.zst-32 1000000 1070 ns/op 95720.98 MB/s 80.53 pct 3 B/op 0 allocs/op
BenchmarkDecoder_DecodeAllParallel/fireworks.jpeg.zst-32 749802 1752 ns/op 70272.35 MB/s 100.0 pct 5 B/op 0 allocs/op
BenchmarkDecoder_DecodeAllParallel/urls.10K.zst-32 22640 52934 ns/op 13263.37 MB/s 26.25 pct 1014 B/op 0 allocs/op
BenchmarkDecoder_DecodeAllParallel/html.zst-32 226412 5232 ns/op 19572.27 MB/s 14.49 pct 20 B/op 0 allocs/op
BenchmarkDecoder_DecodeAllParallel/comp-data.bin.zst-32 923041 1276 ns/op 3194.71 MB/s 31.26 pct 0 B/op 0 allocs/op
```
This reflects the performance around May 2022, but this may be out of date.
## Zstd inside ZIP files
It is possible to use zstandard to compress individual files inside zip archives.
While this isn't widely supported it can be useful for internal files.
To support the compression and decompression of these files you must register a compressor and decompressor.
It is highly recommended registering the (de)compressors on individual zip Reader/Writer and NOT
use the global registration functions. The main reason for this is that 2 registrations from
different packages will result in a panic.
It is a good idea to only have a single compressor and decompressor, since they can be used for multiple zip
files concurrently, and using a single instance will allow reusing some resources.
See [this example](https://pkg.go.dev/github.com/klauspost/compress/zstd#example-ZipCompressor) for
how to compress and decompress files inside zip archives.
# Contributions
Contributions are always welcome.
For new features/fixes, remember to add tests and for performance enhancements include benchmarks.
For general feedback and experience reports, feel free to open an issue or write me on [Twitter](https://twitter.com/sh0dan).
This package includes the excellent [`github.com/cespare/xxhash`](https://github.com/cespare/xxhash) package Copyright (c) 2016 Caleb Spare.

136
vendor/github.com/klauspost/compress/zstd/bitreader.go generated vendored Normal file
View File

@@ -0,0 +1,136 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"encoding/binary"
"errors"
"fmt"
"io"
"math/bits"
)
// bitReader reads a bitstream in reverse.
// The last set bit indicates the start of the stream and is used
// for aligning the input.
type bitReader struct {
in []byte
value uint64 // Maybe use [16]byte, but shifting is awkward.
bitsRead uint8
}
// init initializes and resets the bit reader.
func (b *bitReader) init(in []byte) error {
if len(in) < 1 {
return errors.New("corrupt stream: too short")
}
b.in = in
// The highest bit of the last byte indicates where to start
v := in[len(in)-1]
if v == 0 {
return errors.New("corrupt stream, did not find end of stream")
}
b.bitsRead = 64
b.value = 0
if len(in) >= 8 {
b.fillFastStart()
} else {
b.fill()
b.fill()
}
b.bitsRead += 8 - uint8(highBits(uint32(v)))
return nil
}
// getBits will return n bits. n can be 0.
func (b *bitReader) getBits(n uint8) int {
if n == 0 /*|| b.bitsRead >= 64 */ {
return 0
}
return int(b.get32BitsFast(n))
}
// get32BitsFast requires that at least one bit is requested every time.
// There are no checks if the buffer is filled.
func (b *bitReader) get32BitsFast(n uint8) uint32 {
const regMask = 64 - 1
v := uint32((b.value << (b.bitsRead & regMask)) >> ((regMask + 1 - n) & regMask))
b.bitsRead += n
return v
}
// fillFast() will make sure at least 32 bits are available.
// There must be at least 4 bytes available.
func (b *bitReader) fillFast() {
if b.bitsRead < 32 {
return
}
v := b.in[len(b.in)-4:]
b.in = b.in[:len(b.in)-4]
low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
b.value = (b.value << 32) | uint64(low)
b.bitsRead -= 32
}
// fillFastStart() assumes the bitreader is empty and there is at least 8 bytes to read.
func (b *bitReader) fillFastStart() {
v := b.in[len(b.in)-8:]
b.in = b.in[:len(b.in)-8]
b.value = binary.LittleEndian.Uint64(v)
b.bitsRead = 0
}
// fill() will make sure at least 32 bits are available.
func (b *bitReader) fill() {
if b.bitsRead < 32 {
return
}
if len(b.in) >= 4 {
v := b.in[len(b.in)-4:]
b.in = b.in[:len(b.in)-4]
low := (uint32(v[0])) | (uint32(v[1]) << 8) | (uint32(v[2]) << 16) | (uint32(v[3]) << 24)
b.value = (b.value << 32) | uint64(low)
b.bitsRead -= 32
return
}
b.bitsRead -= uint8(8 * len(b.in))
for len(b.in) > 0 {
b.value = (b.value << 8) | uint64(b.in[len(b.in)-1])
b.in = b.in[:len(b.in)-1]
}
}
// finished returns true if all bits have been read from the bit stream.
func (b *bitReader) finished() bool {
return len(b.in) == 0 && b.bitsRead >= 64
}
// overread returns true if more bits have been requested than is on the stream.
func (b *bitReader) overread() bool {
return b.bitsRead > 64
}
// remain returns the number of bits remaining.
func (b *bitReader) remain() uint {
return 8*uint(len(b.in)) + 64 - uint(b.bitsRead)
}
// close the bitstream and returns an error if out-of-buffer reads occurred.
func (b *bitReader) close() error {
// Release reference.
b.in = nil
if !b.finished() {
return fmt.Errorf("%d extra bits on block, should be 0", b.remain())
}
if b.bitsRead > 64 {
return io.ErrUnexpectedEOF
}
return nil
}
func highBits(val uint32) (n uint32) {
return uint32(bits.Len32(val) - 1)
}

112
vendor/github.com/klauspost/compress/zstd/bitwriter.go generated vendored Normal file
View File

@@ -0,0 +1,112 @@
// Copyright 2018 Klaus Post. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Based on work Copyright (c) 2013, Yann Collet, released under BSD License.
package zstd
// bitWriter will write bits.
// First bit will be LSB of the first byte of output.
type bitWriter struct {
bitContainer uint64
nBits uint8
out []byte
}
// bitMask16 is bitmasks. Has extra to avoid bounds check.
var bitMask16 = [32]uint16{
0, 1, 3, 7, 0xF, 0x1F,
0x3F, 0x7F, 0xFF, 0x1FF, 0x3FF, 0x7FF,
0xFFF, 0x1FFF, 0x3FFF, 0x7FFF, 0xFFFF, 0xFFFF,
0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF,
0xFFFF, 0xFFFF} /* up to 16 bits */
var bitMask32 = [32]uint32{
0, 1, 3, 7, 0xF, 0x1F, 0x3F, 0x7F, 0xFF,
0x1FF, 0x3FF, 0x7FF, 0xFFF, 0x1FFF, 0x3FFF, 0x7FFF, 0xFFFF,
0x1ffff, 0x3ffff, 0x7FFFF, 0xfFFFF, 0x1fFFFF, 0x3fFFFF, 0x7fFFFF, 0xffFFFF,
0x1ffFFFF, 0x3ffFFFF, 0x7ffFFFF, 0xfffFFFF, 0x1fffFFFF, 0x3fffFFFF, 0x7fffFFFF,
} // up to 32 bits
// addBits16NC will add up to 16 bits.
// It will not check if there is space for them,
// so the caller must ensure that it has flushed recently.
func (b *bitWriter) addBits16NC(value uint16, bits uint8) {
b.bitContainer |= uint64(value&bitMask16[bits&31]) << (b.nBits & 63)
b.nBits += bits
}
// addBits32NC will add up to 31 bits.
// It will not check if there is space for them,
// so the caller must ensure that it has flushed recently.
func (b *bitWriter) addBits32NC(value uint32, bits uint8) {
b.bitContainer |= uint64(value&bitMask32[bits&31]) << (b.nBits & 63)
b.nBits += bits
}
// addBits64NC will add up to 64 bits.
// There must be space for 32 bits.
func (b *bitWriter) addBits64NC(value uint64, bits uint8) {
if bits <= 31 {
b.addBits32Clean(uint32(value), bits)
return
}
b.addBits32Clean(uint32(value), 32)
b.flush32()
b.addBits32Clean(uint32(value>>32), bits-32)
}
// addBits32Clean will add up to 32 bits.
// It will not check if there is space for them.
// The input must not contain more bits than specified.
func (b *bitWriter) addBits32Clean(value uint32, bits uint8) {
b.bitContainer |= uint64(value) << (b.nBits & 63)
b.nBits += bits
}
// addBits16Clean will add up to 16 bits. value may not contain more set bits than indicated.
// It will not check if there is space for them, so the caller must ensure that it has flushed recently.
func (b *bitWriter) addBits16Clean(value uint16, bits uint8) {
b.bitContainer |= uint64(value) << (b.nBits & 63)
b.nBits += bits
}
// flush32 will flush out, so there are at least 32 bits available for writing.
func (b *bitWriter) flush32() {
if b.nBits < 32 {
return
}
b.out = append(b.out,
byte(b.bitContainer),
byte(b.bitContainer>>8),
byte(b.bitContainer>>16),
byte(b.bitContainer>>24))
b.nBits -= 32
b.bitContainer >>= 32
}
// flushAlign will flush remaining full bytes and align to next byte boundary.
func (b *bitWriter) flushAlign() {
nbBytes := (b.nBits + 7) >> 3
for i := uint8(0); i < nbBytes; i++ {
b.out = append(b.out, byte(b.bitContainer>>(i*8)))
}
b.nBits = 0
b.bitContainer = 0
}
// close will write the alignment bit and write the final byte(s)
// to the output.
func (b *bitWriter) close() {
// End mark
b.addBits16Clean(1, 1)
// flush until next byte.
b.flushAlign()
}
// reset and continue writing by appending to out.
func (b *bitWriter) reset(out []byte) {
b.bitContainer = 0
b.nBits = 0
b.out = out
}

729
vendor/github.com/klauspost/compress/zstd/blockdec.go generated vendored Normal file
View File

@@ -0,0 +1,729 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"bytes"
"encoding/binary"
"errors"
"fmt"
"hash/crc32"
"io"
"os"
"path/filepath"
"sync"
"github.com/klauspost/compress/huff0"
"github.com/klauspost/compress/zstd/internal/xxhash"
)
type blockType uint8
//go:generate stringer -type=blockType,literalsBlockType,seqCompMode,tableIndex
const (
blockTypeRaw blockType = iota
blockTypeRLE
blockTypeCompressed
blockTypeReserved
)
type literalsBlockType uint8
const (
literalsBlockRaw literalsBlockType = iota
literalsBlockRLE
literalsBlockCompressed
literalsBlockTreeless
)
const (
// maxCompressedBlockSize is the biggest allowed compressed block size (128KB)
maxCompressedBlockSize = 128 << 10
compressedBlockOverAlloc = 16
maxCompressedBlockSizeAlloc = 128<<10 + compressedBlockOverAlloc
// Maximum possible block size (all Raw+Uncompressed).
maxBlockSize = (1 << 21) - 1
maxMatchLen = 131074
maxSequences = 0x7f00 + 0xffff
// We support slightly less than the reference decoder to be able to
// use ints on 32 bit archs.
maxOffsetBits = 30
)
var (
huffDecoderPool = sync.Pool{New: func() interface{} {
return &huff0.Scratch{}
}}
fseDecoderPool = sync.Pool{New: func() interface{} {
return &fseDecoder{}
}}
)
type blockDec struct {
// Raw source data of the block.
data []byte
dataStorage []byte
// Destination of the decoded data.
dst []byte
// Buffer for literals data.
literalBuf []byte
// Window size of the block.
WindowSize uint64
err error
// Check against this crc, if hasCRC is true.
checkCRC uint32
hasCRC bool
// Frame to use for singlethreaded decoding.
// Should not be used by the decoder itself since parent may be another frame.
localFrame *frameDec
sequence []seqVals
async struct {
newHist *history
literals []byte
seqData []byte
seqSize int // Size of uncompressed sequences
fcs uint64
}
// Block is RLE, this is the size.
RLESize uint32
Type blockType
// Is this the last block of a frame?
Last bool
// Use less memory
lowMem bool
}
func (b *blockDec) String() string {
if b == nil {
return "<nil>"
}
return fmt.Sprintf("Steam Size: %d, Type: %v, Last: %t, Window: %d", len(b.data), b.Type, b.Last, b.WindowSize)
}
func newBlockDec(lowMem bool) *blockDec {
b := blockDec{
lowMem: lowMem,
}
return &b
}
// reset will reset the block.
// Input must be a start of a block and will be at the end of the block when returned.
func (b *blockDec) reset(br byteBuffer, windowSize uint64) error {
b.WindowSize = windowSize
tmp, err := br.readSmall(3)
if err != nil {
println("Reading block header:", err)
return err
}
bh := uint32(tmp[0]) | (uint32(tmp[1]) << 8) | (uint32(tmp[2]) << 16)
b.Last = bh&1 != 0
b.Type = blockType((bh >> 1) & 3)
// find size.
cSize := int(bh >> 3)
maxSize := maxCompressedBlockSizeAlloc
switch b.Type {
case blockTypeReserved:
return ErrReservedBlockType
case blockTypeRLE:
if cSize > maxCompressedBlockSize || cSize > int(b.WindowSize) {
if debugDecoder {
printf("rle block too big: csize:%d block: %+v\n", uint64(cSize), b)
}
return ErrWindowSizeExceeded
}
b.RLESize = uint32(cSize)
if b.lowMem {
maxSize = cSize
}
cSize = 1
case blockTypeCompressed:
if debugDecoder {
println("Data size on stream:", cSize)
}
b.RLESize = 0
maxSize = maxCompressedBlockSizeAlloc
if windowSize < maxCompressedBlockSize && b.lowMem {
maxSize = int(windowSize) + compressedBlockOverAlloc
}
if cSize > maxCompressedBlockSize || uint64(cSize) > b.WindowSize {
if debugDecoder {
printf("compressed block too big: csize:%d block: %+v\n", uint64(cSize), b)
}
return ErrCompressedSizeTooBig
}
// Empty compressed blocks must at least be 2 bytes
// for Literals_Block_Type and one for Sequences_Section_Header.
if cSize < 2 {
return ErrBlockTooSmall
}
case blockTypeRaw:
if cSize > maxCompressedBlockSize || cSize > int(b.WindowSize) {
if debugDecoder {
printf("rle block too big: csize:%d block: %+v\n", uint64(cSize), b)
}
return ErrWindowSizeExceeded
}
b.RLESize = 0
// We do not need a destination for raw blocks.
maxSize = -1
default:
panic("Invalid block type")
}
// Read block data.
if _, ok := br.(*byteBuf); !ok && cap(b.dataStorage) < cSize {
// byteBuf doesn't need a destination buffer.
if b.lowMem || cSize > maxCompressedBlockSize {
b.dataStorage = make([]byte, 0, cSize+compressedBlockOverAlloc)
} else {
b.dataStorage = make([]byte, 0, maxCompressedBlockSizeAlloc)
}
}
b.data, err = br.readBig(cSize, b.dataStorage)
if err != nil {
if debugDecoder {
println("Reading block:", err, "(", cSize, ")", len(b.data))
printf("%T", br)
}
return err
}
if cap(b.dst) <= maxSize {
b.dst = make([]byte, 0, maxSize+1)
}
return nil
}
// sendEOF will make the decoder send EOF on this frame.
func (b *blockDec) sendErr(err error) {
b.Last = true
b.Type = blockTypeReserved
b.err = err
}
// Close will release resources.
// Closed blockDec cannot be reset.
func (b *blockDec) Close() {
}
// decodeBuf
func (b *blockDec) decodeBuf(hist *history) error {
switch b.Type {
case blockTypeRLE:
if cap(b.dst) < int(b.RLESize) {
if b.lowMem {
b.dst = make([]byte, b.RLESize)
} else {
b.dst = make([]byte, maxCompressedBlockSize)
}
}
b.dst = b.dst[:b.RLESize]
v := b.data[0]
for i := range b.dst {
b.dst[i] = v
}
hist.appendKeep(b.dst)
return nil
case blockTypeRaw:
hist.appendKeep(b.data)
return nil
case blockTypeCompressed:
saved := b.dst
// Append directly to history
if hist.ignoreBuffer == 0 {
b.dst = hist.b
hist.b = nil
} else {
b.dst = b.dst[:0]
}
err := b.decodeCompressed(hist)
if debugDecoder {
println("Decompressed to total", len(b.dst), "bytes, hash:", xxhash.Sum64(b.dst), "error:", err)
}
if hist.ignoreBuffer == 0 {
hist.b = b.dst
b.dst = saved
} else {
hist.appendKeep(b.dst)
}
return err
case blockTypeReserved:
// Used for returning errors.
return b.err
default:
panic("Invalid block type")
}
}
func (b *blockDec) decodeLiterals(in []byte, hist *history) (remain []byte, err error) {
// There must be at least one byte for Literals_Block_Type and one for Sequences_Section_Header
if len(in) < 2 {
return in, ErrBlockTooSmall
}
litType := literalsBlockType(in[0] & 3)
var litRegenSize int
var litCompSize int
sizeFormat := (in[0] >> 2) & 3
var fourStreams bool
var literals []byte
switch litType {
case literalsBlockRaw, literalsBlockRLE:
switch sizeFormat {
case 0, 2:
// Regenerated_Size uses 5 bits (0-31). Literals_Section_Header uses 1 byte.
litRegenSize = int(in[0] >> 3)
in = in[1:]
case 1:
// Regenerated_Size uses 12 bits (0-4095). Literals_Section_Header uses 2 bytes.
litRegenSize = int(in[0]>>4) + (int(in[1]) << 4)
in = in[2:]
case 3:
// Regenerated_Size uses 20 bits (0-1048575). Literals_Section_Header uses 3 bytes.
if len(in) < 3 {
println("too small: litType:", litType, " sizeFormat", sizeFormat, len(in))
return in, ErrBlockTooSmall
}
litRegenSize = int(in[0]>>4) + (int(in[1]) << 4) + (int(in[2]) << 12)
in = in[3:]
}
case literalsBlockCompressed, literalsBlockTreeless:
switch sizeFormat {
case 0, 1:
// Both Regenerated_Size and Compressed_Size use 10 bits (0-1023).
if len(in) < 3 {
println("too small: litType:", litType, " sizeFormat", sizeFormat, len(in))
return in, ErrBlockTooSmall
}
n := uint64(in[0]>>4) + (uint64(in[1]) << 4) + (uint64(in[2]) << 12)
litRegenSize = int(n & 1023)
litCompSize = int(n >> 10)
fourStreams = sizeFormat == 1
in = in[3:]
case 2:
fourStreams = true
if len(in) < 4 {
println("too small: litType:", litType, " sizeFormat", sizeFormat, len(in))
return in, ErrBlockTooSmall
}
n := uint64(in[0]>>4) + (uint64(in[1]) << 4) + (uint64(in[2]) << 12) + (uint64(in[3]) << 20)
litRegenSize = int(n & 16383)
litCompSize = int(n >> 14)
in = in[4:]
case 3:
fourStreams = true
if len(in) < 5 {
println("too small: litType:", litType, " sizeFormat", sizeFormat, len(in))
return in, ErrBlockTooSmall
}
n := uint64(in[0]>>4) + (uint64(in[1]) << 4) + (uint64(in[2]) << 12) + (uint64(in[3]) << 20) + (uint64(in[4]) << 28)
litRegenSize = int(n & 262143)
litCompSize = int(n >> 18)
in = in[5:]
}
}
if debugDecoder {
println("literals type:", litType, "litRegenSize:", litRegenSize, "litCompSize:", litCompSize, "sizeFormat:", sizeFormat, "4X:", fourStreams)
}
if litRegenSize > int(b.WindowSize) || litRegenSize > maxCompressedBlockSize {
return in, ErrWindowSizeExceeded
}
switch litType {
case literalsBlockRaw:
if len(in) < litRegenSize {
println("too small: litType:", litType, " sizeFormat", sizeFormat, "remain:", len(in), "want:", litRegenSize)
return in, ErrBlockTooSmall
}
literals = in[:litRegenSize]
in = in[litRegenSize:]
//printf("Found %d uncompressed literals\n", litRegenSize)
case literalsBlockRLE:
if len(in) < 1 {
println("too small: litType:", litType, " sizeFormat", sizeFormat, "remain:", len(in), "want:", 1)
return in, ErrBlockTooSmall
}
if cap(b.literalBuf) < litRegenSize {
if b.lowMem {
b.literalBuf = make([]byte, litRegenSize, litRegenSize+compressedBlockOverAlloc)
} else {
b.literalBuf = make([]byte, litRegenSize, maxCompressedBlockSize+compressedBlockOverAlloc)
}
}
literals = b.literalBuf[:litRegenSize]
v := in[0]
for i := range literals {
literals[i] = v
}
in = in[1:]
if debugDecoder {
printf("Found %d RLE compressed literals\n", litRegenSize)
}
case literalsBlockTreeless:
if len(in) < litCompSize {
println("too small: litType:", litType, " sizeFormat", sizeFormat, "remain:", len(in), "want:", litCompSize)
return in, ErrBlockTooSmall
}
// Store compressed literals, so we defer decoding until we get history.
literals = in[:litCompSize]
in = in[litCompSize:]
if debugDecoder {
printf("Found %d compressed literals\n", litCompSize)
}
huff := hist.huffTree
if huff == nil {
return in, errors.New("literal block was treeless, but no history was defined")
}
// Ensure we have space to store it.
if cap(b.literalBuf) < litRegenSize {
if b.lowMem {
b.literalBuf = make([]byte, 0, litRegenSize+compressedBlockOverAlloc)
} else {
b.literalBuf = make([]byte, 0, maxCompressedBlockSize+compressedBlockOverAlloc)
}
}
var err error
// Use our out buffer.
huff.MaxDecodedSize = litRegenSize
if fourStreams {
literals, err = huff.Decoder().Decompress4X(b.literalBuf[:0:litRegenSize], literals)
} else {
literals, err = huff.Decoder().Decompress1X(b.literalBuf[:0:litRegenSize], literals)
}
// Make sure we don't leak our literals buffer
if err != nil {
println("decompressing literals:", err)
return in, err
}
if len(literals) != litRegenSize {
return in, fmt.Errorf("literal output size mismatch want %d, got %d", litRegenSize, len(literals))
}
case literalsBlockCompressed:
if len(in) < litCompSize {
println("too small: litType:", litType, " sizeFormat", sizeFormat, "remain:", len(in), "want:", litCompSize)
return in, ErrBlockTooSmall
}
literals = in[:litCompSize]
in = in[litCompSize:]
// Ensure we have space to store it.
if cap(b.literalBuf) < litRegenSize {
if b.lowMem {
b.literalBuf = make([]byte, 0, litRegenSize+compressedBlockOverAlloc)
} else {
b.literalBuf = make([]byte, 0, maxCompressedBlockSize+compressedBlockOverAlloc)
}
}
huff := hist.huffTree
if huff == nil || (hist.dict != nil && huff == hist.dict.litEnc) {
huff = huffDecoderPool.Get().(*huff0.Scratch)
if huff == nil {
huff = &huff0.Scratch{}
}
}
var err error
if debugDecoder {
println("huff table input:", len(literals), "CRC:", crc32.ChecksumIEEE(literals))
}
huff, literals, err = huff0.ReadTable(literals, huff)
if err != nil {
println("reading huffman table:", err)
return in, err
}
hist.huffTree = huff
huff.MaxDecodedSize = litRegenSize
// Use our out buffer.
if fourStreams {
literals, err = huff.Decoder().Decompress4X(b.literalBuf[:0:litRegenSize], literals)
} else {
literals, err = huff.Decoder().Decompress1X(b.literalBuf[:0:litRegenSize], literals)
}
if err != nil {
println("decoding compressed literals:", err)
return in, err
}
// Make sure we don't leak our literals buffer
if len(literals) != litRegenSize {
return in, fmt.Errorf("literal output size mismatch want %d, got %d", litRegenSize, len(literals))
}
// Re-cap to get extra size.
literals = b.literalBuf[:len(literals)]
if debugDecoder {
printf("Decompressed %d literals into %d bytes\n", litCompSize, litRegenSize)
}
}
hist.decoders.literals = literals
return in, nil
}
// decodeCompressed will start decompressing a block.
func (b *blockDec) decodeCompressed(hist *history) error {
in := b.data
in, err := b.decodeLiterals(in, hist)
if err != nil {
return err
}
err = b.prepareSequences(in, hist)
if err != nil {
return err
}
if hist.decoders.nSeqs == 0 {
b.dst = append(b.dst, hist.decoders.literals...)
return nil
}
before := len(hist.decoders.out)
err = hist.decoders.decodeSync(hist.b[hist.ignoreBuffer:])
if err != nil {
return err
}
if hist.decoders.maxSyncLen > 0 {
hist.decoders.maxSyncLen += uint64(before)
hist.decoders.maxSyncLen -= uint64(len(hist.decoders.out))
}
b.dst = hist.decoders.out
hist.recentOffsets = hist.decoders.prevOffset
return nil
}
func (b *blockDec) prepareSequences(in []byte, hist *history) (err error) {
if debugDecoder {
printf("prepareSequences: %d byte(s) input\n", len(in))
}
// Decode Sequences
// https://github.com/facebook/zstd/blob/dev/doc/zstd_compression_format.md#sequences-section
if len(in) < 1 {
return ErrBlockTooSmall
}
var nSeqs int
seqHeader := in[0]
switch {
case seqHeader < 128:
nSeqs = int(seqHeader)
in = in[1:]
case seqHeader < 255:
if len(in) < 2 {
return ErrBlockTooSmall
}
nSeqs = int(seqHeader-128)<<8 | int(in[1])
in = in[2:]
case seqHeader == 255:
if len(in) < 3 {
return ErrBlockTooSmall
}
nSeqs = 0x7f00 + int(in[1]) + (int(in[2]) << 8)
in = in[3:]
}
if nSeqs == 0 && len(in) != 0 {
// When no sequences, there should not be any more data...
if debugDecoder {
printf("prepareSequences: 0 sequences, but %d byte(s) left on stream\n", len(in))
}
return ErrUnexpectedBlockSize
}
var seqs = &hist.decoders
seqs.nSeqs = nSeqs
if nSeqs > 0 {
if len(in) < 1 {
return ErrBlockTooSmall
}
br := byteReader{b: in, off: 0}
compMode := br.Uint8()
br.advance(1)
if debugDecoder {
printf("Compression modes: 0b%b", compMode)
}
if compMode&3 != 0 {
return errors.New("corrupt block: reserved bits not zero")
}
for i := uint(0); i < 3; i++ {
mode := seqCompMode((compMode >> (6 - i*2)) & 3)
if debugDecoder {
println("Table", tableIndex(i), "is", mode)
}
var seq *sequenceDec
switch tableIndex(i) {
case tableLiteralLengths:
seq = &seqs.litLengths
case tableOffsets:
seq = &seqs.offsets
case tableMatchLengths:
seq = &seqs.matchLengths
default:
panic("unknown table")
}
switch mode {
case compModePredefined:
if seq.fse != nil && !seq.fse.preDefined {
fseDecoderPool.Put(seq.fse)
}
seq.fse = &fsePredef[i]
case compModeRLE:
if br.remain() < 1 {
return ErrBlockTooSmall
}
v := br.Uint8()
br.advance(1)
if seq.fse == nil || seq.fse.preDefined {
seq.fse = fseDecoderPool.Get().(*fseDecoder)
}
symb, err := decSymbolValue(v, symbolTableX[i])
if err != nil {
printf("RLE Transform table (%v) error: %v", tableIndex(i), err)
return err
}
seq.fse.setRLE(symb)
if debugDecoder {
printf("RLE set to 0x%x, code: %v", symb, v)
}
case compModeFSE:
println("Reading table for", tableIndex(i))
if seq.fse == nil || seq.fse.preDefined {
seq.fse = fseDecoderPool.Get().(*fseDecoder)
}
err := seq.fse.readNCount(&br, uint16(maxTableSymbol[i]))
if err != nil {
println("Read table error:", err)
return err
}
err = seq.fse.transform(symbolTableX[i])
if err != nil {
println("Transform table error:", err)
return err
}
if debugDecoder {
println("Read table ok", "symbolLen:", seq.fse.symbolLen)
}
case compModeRepeat:
seq.repeat = true
}
if br.overread() {
return io.ErrUnexpectedEOF
}
}
in = br.unread()
}
if debugDecoder {
println("Literals:", len(seqs.literals), "hash:", xxhash.Sum64(seqs.literals), "and", seqs.nSeqs, "sequences.")
}
if nSeqs == 0 {
if len(b.sequence) > 0 {
b.sequence = b.sequence[:0]
}
return nil
}
br := seqs.br
if br == nil {
br = &bitReader{}
}
if err := br.init(in); err != nil {
return err
}
if err := seqs.initialize(br, hist, b.dst); err != nil {
println("initializing sequences:", err)
return err
}
// Extract blocks...
if false && hist.dict == nil {
fatalErr := func(err error) {
if err != nil {
panic(err)
}
}
fn := fmt.Sprintf("n-%d-lits-%d-prev-%d-%d-%d-win-%d.blk", hist.decoders.nSeqs, len(hist.decoders.literals), hist.recentOffsets[0], hist.recentOffsets[1], hist.recentOffsets[2], hist.windowSize)
var buf bytes.Buffer
fatalErr(binary.Write(&buf, binary.LittleEndian, hist.decoders.litLengths.fse))
fatalErr(binary.Write(&buf, binary.LittleEndian, hist.decoders.matchLengths.fse))
fatalErr(binary.Write(&buf, binary.LittleEndian, hist.decoders.offsets.fse))
buf.Write(in)
os.WriteFile(filepath.Join("testdata", "seqs", fn), buf.Bytes(), os.ModePerm)
}
return nil
}
func (b *blockDec) decodeSequences(hist *history) error {
if cap(b.sequence) < hist.decoders.nSeqs {
if b.lowMem {
b.sequence = make([]seqVals, 0, hist.decoders.nSeqs)
} else {
b.sequence = make([]seqVals, 0, 0x7F00+0xffff)
}
}
b.sequence = b.sequence[:hist.decoders.nSeqs]
if hist.decoders.nSeqs == 0 {
hist.decoders.seqSize = len(hist.decoders.literals)
return nil
}
hist.decoders.windowSize = hist.windowSize
hist.decoders.prevOffset = hist.recentOffsets
err := hist.decoders.decode(b.sequence)
hist.recentOffsets = hist.decoders.prevOffset
return err
}
func (b *blockDec) executeSequences(hist *history) error {
hbytes := hist.b
if len(hbytes) > hist.windowSize {
hbytes = hbytes[len(hbytes)-hist.windowSize:]
// We do not need history anymore.
if hist.dict != nil {
hist.dict.content = nil
}
}
hist.decoders.windowSize = hist.windowSize
hist.decoders.out = b.dst[:0]
err := hist.decoders.execute(b.sequence, hbytes)
if err != nil {
return err
}
return b.updateHistory(hist)
}
func (b *blockDec) updateHistory(hist *history) error {
if len(b.data) > maxCompressedBlockSize {
return fmt.Errorf("compressed block size too large (%d)", len(b.data))
}
// Set output and release references.
b.dst = hist.decoders.out
hist.recentOffsets = hist.decoders.prevOffset
if b.Last {
// if last block we don't care about history.
println("Last block, no history returned")
hist.b = hist.b[:0]
return nil
} else {
hist.append(b.dst)
if debugDecoder {
println("Finished block with ", len(b.sequence), "sequences. Added", len(b.dst), "to history, now length", len(hist.b))
}
}
hist.decoders.out, hist.decoders.literals = nil, nil
return nil
}

909
vendor/github.com/klauspost/compress/zstd/blockenc.go generated vendored Normal file
View File

@@ -0,0 +1,909 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"errors"
"fmt"
"math"
"math/bits"
"github.com/klauspost/compress/huff0"
)
type blockEnc struct {
size int
literals []byte
sequences []seq
coders seqCoders
litEnc *huff0.Scratch
dictLitEnc *huff0.Scratch
wr bitWriter
extraLits int
output []byte
recentOffsets [3]uint32
prevRecentOffsets [3]uint32
last bool
lowMem bool
}
// init should be used once the block has been created.
// If called more than once, the effect is the same as calling reset.
func (b *blockEnc) init() {
if b.lowMem {
// 1K literals
if cap(b.literals) < 1<<10 {
b.literals = make([]byte, 0, 1<<10)
}
const defSeqs = 20
if cap(b.sequences) < defSeqs {
b.sequences = make([]seq, 0, defSeqs)
}
// 1K
if cap(b.output) < 1<<10 {
b.output = make([]byte, 0, 1<<10)
}
} else {
if cap(b.literals) < maxCompressedBlockSize {
b.literals = make([]byte, 0, maxCompressedBlockSize)
}
const defSeqs = 2000
if cap(b.sequences) < defSeqs {
b.sequences = make([]seq, 0, defSeqs)
}
if cap(b.output) < maxCompressedBlockSize {
b.output = make([]byte, 0, maxCompressedBlockSize)
}
}
if b.coders.mlEnc == nil {
b.coders.mlEnc = &fseEncoder{}
b.coders.mlPrev = &fseEncoder{}
b.coders.ofEnc = &fseEncoder{}
b.coders.ofPrev = &fseEncoder{}
b.coders.llEnc = &fseEncoder{}
b.coders.llPrev = &fseEncoder{}
}
b.litEnc = &huff0.Scratch{WantLogLess: 4}
b.reset(nil)
}
// initNewEncode can be used to reset offsets and encoders to the initial state.
func (b *blockEnc) initNewEncode() {
b.recentOffsets = [3]uint32{1, 4, 8}
b.litEnc.Reuse = huff0.ReusePolicyNone
b.coders.setPrev(nil, nil, nil)
}
// reset will reset the block for a new encode, but in the same stream,
// meaning that state will be carried over, but the block content is reset.
// If a previous block is provided, the recent offsets are carried over.
func (b *blockEnc) reset(prev *blockEnc) {
b.extraLits = 0
b.literals = b.literals[:0]
b.size = 0
b.sequences = b.sequences[:0]
b.output = b.output[:0]
b.last = false
if prev != nil {
b.recentOffsets = prev.prevRecentOffsets
}
b.dictLitEnc = nil
}
// reset will reset the block for a new encode, but in the same stream,
// meaning that state will be carried over, but the block content is reset.
// If a previous block is provided, the recent offsets are carried over.
func (b *blockEnc) swapEncoders(prev *blockEnc) {
b.coders.swap(&prev.coders)
b.litEnc, prev.litEnc = prev.litEnc, b.litEnc
}
// blockHeader contains the information for a block header.
type blockHeader uint32
// setLast sets the 'last' indicator on a block.
func (h *blockHeader) setLast(b bool) {
if b {
*h = *h | 1
} else {
const mask = (1 << 24) - 2
*h = *h & mask
}
}
// setSize will store the compressed size of a block.
func (h *blockHeader) setSize(v uint32) {
const mask = 7
*h = (*h)&mask | blockHeader(v<<3)
}
// setType sets the block type.
func (h *blockHeader) setType(t blockType) {
const mask = 1 | (((1 << 24) - 1) ^ 7)
*h = (*h & mask) | blockHeader(t<<1)
}
// appendTo will append the block header to a slice.
func (h blockHeader) appendTo(b []byte) []byte {
return append(b, uint8(h), uint8(h>>8), uint8(h>>16))
}
// String returns a string representation of the block.
func (h blockHeader) String() string {
return fmt.Sprintf("Type: %d, Size: %d, Last:%t", (h>>1)&3, h>>3, h&1 == 1)
}
// literalsHeader contains literals header information.
type literalsHeader uint64
// setType can be used to set the type of literal block.
func (h *literalsHeader) setType(t literalsBlockType) {
const mask = math.MaxUint64 - 3
*h = (*h & mask) | literalsHeader(t)
}
// setSize can be used to set a single size, for uncompressed and RLE content.
func (h *literalsHeader) setSize(regenLen int) {
inBits := bits.Len32(uint32(regenLen))
// Only retain 2 bits
const mask = 3
lh := uint64(*h & mask)
switch {
case inBits < 5:
lh |= (uint64(regenLen) << 3) | (1 << 60)
if debugEncoder {
got := int(lh>>3) & 0xff
if got != regenLen {
panic(fmt.Sprint("litRegenSize = ", regenLen, "(want) != ", got, "(got)"))
}
}
case inBits < 12:
lh |= (1 << 2) | (uint64(regenLen) << 4) | (2 << 60)
case inBits < 20:
lh |= (3 << 2) | (uint64(regenLen) << 4) | (3 << 60)
default:
panic(fmt.Errorf("internal error: block too big (%d)", regenLen))
}
*h = literalsHeader(lh)
}
// setSizes will set the size of a compressed literals section and the input length.
func (h *literalsHeader) setSizes(compLen, inLen int, single bool) {
compBits, inBits := bits.Len32(uint32(compLen)), bits.Len32(uint32(inLen))
// Only retain 2 bits
const mask = 3
lh := uint64(*h & mask)
switch {
case compBits <= 10 && inBits <= 10:
if !single {
lh |= 1 << 2
}
lh |= (uint64(inLen) << 4) | (uint64(compLen) << (10 + 4)) | (3 << 60)
if debugEncoder {
const mmask = (1 << 24) - 1
n := (lh >> 4) & mmask
if int(n&1023) != inLen {
panic(fmt.Sprint("regensize:", int(n&1023), "!=", inLen, inBits))
}
if int(n>>10) != compLen {
panic(fmt.Sprint("compsize:", int(n>>10), "!=", compLen, compBits))
}
}
case compBits <= 14 && inBits <= 14:
lh |= (2 << 2) | (uint64(inLen) << 4) | (uint64(compLen) << (14 + 4)) | (4 << 60)
if single {
panic("single stream used with more than 10 bits length.")
}
case compBits <= 18 && inBits <= 18:
lh |= (3 << 2) | (uint64(inLen) << 4) | (uint64(compLen) << (18 + 4)) | (5 << 60)
if single {
panic("single stream used with more than 10 bits length.")
}
default:
panic("internal error: block too big")
}
*h = literalsHeader(lh)
}
// appendTo will append the literals header to a byte slice.
func (h literalsHeader) appendTo(b []byte) []byte {
size := uint8(h >> 60)
switch size {
case 1:
b = append(b, uint8(h))
case 2:
b = append(b, uint8(h), uint8(h>>8))
case 3:
b = append(b, uint8(h), uint8(h>>8), uint8(h>>16))
case 4:
b = append(b, uint8(h), uint8(h>>8), uint8(h>>16), uint8(h>>24))
case 5:
b = append(b, uint8(h), uint8(h>>8), uint8(h>>16), uint8(h>>24), uint8(h>>32))
default:
panic(fmt.Errorf("internal error: literalsHeader has invalid size (%d)", size))
}
return b
}
// size returns the output size with currently set values.
func (h literalsHeader) size() int {
return int(h >> 60)
}
func (h literalsHeader) String() string {
return fmt.Sprintf("Type: %d, SizeFormat: %d, Size: 0x%d, Bytes:%d", literalsBlockType(h&3), (h>>2)&3, h&((1<<60)-1)>>4, h>>60)
}
// pushOffsets will push the recent offsets to the backup store.
func (b *blockEnc) pushOffsets() {
b.prevRecentOffsets = b.recentOffsets
}
// pushOffsets will push the recent offsets to the backup store.
func (b *blockEnc) popOffsets() {
b.recentOffsets = b.prevRecentOffsets
}
// matchOffset will adjust recent offsets and return the adjusted one,
// if it matches a previous offset.
func (b *blockEnc) matchOffset(offset, lits uint32) uint32 {
// Check if offset is one of the recent offsets.
// Adjusts the output offset accordingly.
// Gives a tiny bit of compression, typically around 1%.
if true {
if lits > 0 {
switch offset {
case b.recentOffsets[0]:
offset = 1
case b.recentOffsets[1]:
b.recentOffsets[1] = b.recentOffsets[0]
b.recentOffsets[0] = offset
offset = 2
case b.recentOffsets[2]:
b.recentOffsets[2] = b.recentOffsets[1]
b.recentOffsets[1] = b.recentOffsets[0]
b.recentOffsets[0] = offset
offset = 3
default:
b.recentOffsets[2] = b.recentOffsets[1]
b.recentOffsets[1] = b.recentOffsets[0]
b.recentOffsets[0] = offset
offset += 3
}
} else {
switch offset {
case b.recentOffsets[1]:
b.recentOffsets[1] = b.recentOffsets[0]
b.recentOffsets[0] = offset
offset = 1
case b.recentOffsets[2]:
b.recentOffsets[2] = b.recentOffsets[1]
b.recentOffsets[1] = b.recentOffsets[0]
b.recentOffsets[0] = offset
offset = 2
case b.recentOffsets[0] - 1:
b.recentOffsets[2] = b.recentOffsets[1]
b.recentOffsets[1] = b.recentOffsets[0]
b.recentOffsets[0] = offset
offset = 3
default:
b.recentOffsets[2] = b.recentOffsets[1]
b.recentOffsets[1] = b.recentOffsets[0]
b.recentOffsets[0] = offset
offset += 3
}
}
} else {
offset += 3
}
return offset
}
// encodeRaw can be used to set the output to a raw representation of supplied bytes.
func (b *blockEnc) encodeRaw(a []byte) {
var bh blockHeader
bh.setLast(b.last)
bh.setSize(uint32(len(a)))
bh.setType(blockTypeRaw)
b.output = bh.appendTo(b.output[:0])
b.output = append(b.output, a...)
if debugEncoder {
println("Adding RAW block, length", len(a), "last:", b.last)
}
}
// encodeRaw can be used to set the output to a raw representation of supplied bytes.
func (b *blockEnc) encodeRawTo(dst, src []byte) []byte {
var bh blockHeader
bh.setLast(b.last)
bh.setSize(uint32(len(src)))
bh.setType(blockTypeRaw)
dst = bh.appendTo(dst)
dst = append(dst, src...)
if debugEncoder {
println("Adding RAW block, length", len(src), "last:", b.last)
}
return dst
}
// encodeLits can be used if the block is only litLen.
func (b *blockEnc) encodeLits(lits []byte, raw bool) error {
var bh blockHeader
bh.setLast(b.last)
bh.setSize(uint32(len(lits)))
// Don't compress extremely small blocks
if len(lits) < 8 || (len(lits) < 32 && b.dictLitEnc == nil) || raw {
if debugEncoder {
println("Adding RAW block, length", len(lits), "last:", b.last)
}
bh.setType(blockTypeRaw)
b.output = bh.appendTo(b.output)
b.output = append(b.output, lits...)
return nil
}
var (
out []byte
reUsed, single bool
err error
)
if b.dictLitEnc != nil {
b.litEnc.TransferCTable(b.dictLitEnc)
b.litEnc.Reuse = huff0.ReusePolicyAllow
b.dictLitEnc = nil
}
if len(lits) >= 1024 {
// Use 4 Streams.
out, reUsed, err = huff0.Compress4X(lits, b.litEnc)
} else if len(lits) > 16 {
// Use 1 stream
single = true
out, reUsed, err = huff0.Compress1X(lits, b.litEnc)
} else {
err = huff0.ErrIncompressible
}
if err == nil && len(out)+5 > len(lits) {
// If we are close, we may still be worse or equal to raw.
var lh literalsHeader
lh.setSizes(len(out), len(lits), single)
if len(out)+lh.size() >= len(lits) {
err = huff0.ErrIncompressible
}
}
switch err {
case huff0.ErrIncompressible:
if debugEncoder {
println("Adding RAW block, length", len(lits), "last:", b.last)
}
bh.setType(blockTypeRaw)
b.output = bh.appendTo(b.output)
b.output = append(b.output, lits...)
return nil
case huff0.ErrUseRLE:
if debugEncoder {
println("Adding RLE block, length", len(lits))
}
bh.setType(blockTypeRLE)
b.output = bh.appendTo(b.output)
b.output = append(b.output, lits[0])
return nil
case nil:
default:
return err
}
// Compressed...
// Now, allow reuse
b.litEnc.Reuse = huff0.ReusePolicyAllow
bh.setType(blockTypeCompressed)
var lh literalsHeader
if reUsed {
if debugEncoder {
println("Reused tree, compressed to", len(out))
}
lh.setType(literalsBlockTreeless)
} else {
if debugEncoder {
println("New tree, compressed to", len(out), "tree size:", len(b.litEnc.OutTable))
}
lh.setType(literalsBlockCompressed)
}
// Set sizes
lh.setSizes(len(out), len(lits), single)
bh.setSize(uint32(len(out) + lh.size() + 1))
// Write block headers.
b.output = bh.appendTo(b.output)
b.output = lh.appendTo(b.output)
// Add compressed data.
b.output = append(b.output, out...)
// No sequences.
b.output = append(b.output, 0)
return nil
}
// encodeRLE will encode an RLE block.
func (b *blockEnc) encodeRLE(val byte, length uint32) {
var bh blockHeader
bh.setLast(b.last)
bh.setSize(length)
bh.setType(blockTypeRLE)
b.output = bh.appendTo(b.output)
b.output = append(b.output, val)
}
// fuzzFseEncoder can be used to fuzz the FSE encoder.
func fuzzFseEncoder(data []byte) int {
if len(data) > maxSequences || len(data) < 2 {
return 0
}
enc := fseEncoder{}
hist := enc.Histogram()
maxSym := uint8(0)
for i, v := range data {
v = v & 63
data[i] = v
hist[v]++
if v > maxSym {
maxSym = v
}
}
if maxSym == 0 {
// All 0
return 0
}
maxCount := func(a []uint32) int {
var max uint32
for _, v := range a {
if v > max {
max = v
}
}
return int(max)
}
cnt := maxCount(hist[:maxSym])
if cnt == len(data) {
// RLE
return 0
}
enc.HistogramFinished(maxSym, cnt)
err := enc.normalizeCount(len(data))
if err != nil {
return 0
}
_, err = enc.writeCount(nil)
if err != nil {
panic(err)
}
return 1
}
// encode will encode the block and append the output in b.output.
// Previous offset codes must be pushed if more blocks are expected.
func (b *blockEnc) encode(org []byte, raw, rawAllLits bool) error {
if len(b.sequences) == 0 {
return b.encodeLits(b.literals, rawAllLits)
}
if len(b.sequences) == 1 && len(org) > 0 && len(b.literals) <= 1 {
// Check common RLE cases.
seq := b.sequences[0]
if seq.litLen == uint32(len(b.literals)) && seq.offset-3 == 1 {
// Offset == 1 and 0 or 1 literals.
b.encodeRLE(org[0], b.sequences[0].matchLen+zstdMinMatch+seq.litLen)
return nil
}
}
// We want some difference to at least account for the headers.
saved := b.size - len(b.literals) - (b.size >> 6)
if saved < 16 {
if org == nil {
return errIncompressible
}
b.popOffsets()
return b.encodeLits(org, rawAllLits)
}
var bh blockHeader
var lh literalsHeader
bh.setLast(b.last)
bh.setType(blockTypeCompressed)
// Store offset of the block header. Needed when we know the size.
bhOffset := len(b.output)
b.output = bh.appendTo(b.output)
var (
out []byte
reUsed, single bool
err error
)
if b.dictLitEnc != nil {
b.litEnc.TransferCTable(b.dictLitEnc)
b.litEnc.Reuse = huff0.ReusePolicyAllow
b.dictLitEnc = nil
}
if len(b.literals) >= 1024 && !raw {
// Use 4 Streams.
out, reUsed, err = huff0.Compress4X(b.literals, b.litEnc)
} else if len(b.literals) > 16 && !raw {
// Use 1 stream
single = true
out, reUsed, err = huff0.Compress1X(b.literals, b.litEnc)
} else {
err = huff0.ErrIncompressible
}
if err == nil && len(out)+5 > len(b.literals) {
// If we are close, we may still be worse or equal to raw.
var lh literalsHeader
lh.setSize(len(b.literals))
szRaw := lh.size()
lh.setSizes(len(out), len(b.literals), single)
szComp := lh.size()
if len(out)+szComp >= len(b.literals)+szRaw {
err = huff0.ErrIncompressible
}
}
switch err {
case huff0.ErrIncompressible:
lh.setType(literalsBlockRaw)
lh.setSize(len(b.literals))
b.output = lh.appendTo(b.output)
b.output = append(b.output, b.literals...)
if debugEncoder {
println("Adding literals RAW, length", len(b.literals))
}
case huff0.ErrUseRLE:
lh.setType(literalsBlockRLE)
lh.setSize(len(b.literals))
b.output = lh.appendTo(b.output)
b.output = append(b.output, b.literals[0])
if debugEncoder {
println("Adding literals RLE")
}
case nil:
// Compressed litLen...
if reUsed {
if debugEncoder {
println("reused tree")
}
lh.setType(literalsBlockTreeless)
} else {
if debugEncoder {
println("new tree, size:", len(b.litEnc.OutTable))
}
lh.setType(literalsBlockCompressed)
if debugEncoder {
_, _, err := huff0.ReadTable(out, nil)
if err != nil {
panic(err)
}
}
}
lh.setSizes(len(out), len(b.literals), single)
if debugEncoder {
printf("Compressed %d literals to %d bytes", len(b.literals), len(out))
println("Adding literal header:", lh)
}
b.output = lh.appendTo(b.output)
b.output = append(b.output, out...)
b.litEnc.Reuse = huff0.ReusePolicyAllow
if debugEncoder {
println("Adding literals compressed")
}
default:
if debugEncoder {
println("Adding literals ERROR:", err)
}
return err
}
// Sequence compression
// Write the number of sequences
switch {
case len(b.sequences) < 128:
b.output = append(b.output, uint8(len(b.sequences)))
case len(b.sequences) < 0x7f00: // TODO: this could be wrong
n := len(b.sequences)
b.output = append(b.output, 128+uint8(n>>8), uint8(n))
default:
n := len(b.sequences) - 0x7f00
b.output = append(b.output, 255, uint8(n), uint8(n>>8))
}
if debugEncoder {
println("Encoding", len(b.sequences), "sequences")
}
b.genCodes()
llEnc := b.coders.llEnc
ofEnc := b.coders.ofEnc
mlEnc := b.coders.mlEnc
err = llEnc.normalizeCount(len(b.sequences))
if err != nil {
return err
}
err = ofEnc.normalizeCount(len(b.sequences))
if err != nil {
return err
}
err = mlEnc.normalizeCount(len(b.sequences))
if err != nil {
return err
}
// Choose the best compression mode for each type.
// Will evaluate the new vs predefined and previous.
chooseComp := func(cur, prev, preDef *fseEncoder) (*fseEncoder, seqCompMode) {
// See if predefined/previous is better
hist := cur.count[:cur.symbolLen]
nSize := cur.approxSize(hist) + cur.maxHeaderSize()
predefSize := preDef.approxSize(hist)
prevSize := prev.approxSize(hist)
// Add a small penalty for new encoders.
// Don't bother with extremely small (<2 byte gains).
nSize = nSize + (nSize+2*8*16)>>4
switch {
case predefSize <= prevSize && predefSize <= nSize || forcePreDef:
if debugEncoder {
println("Using predefined", predefSize>>3, "<=", nSize>>3)
}
return preDef, compModePredefined
case prevSize <= nSize:
if debugEncoder {
println("Using previous", prevSize>>3, "<=", nSize>>3)
}
return prev, compModeRepeat
default:
if debugEncoder {
println("Using new, predef", predefSize>>3, ". previous:", prevSize>>3, ">", nSize>>3, "header max:", cur.maxHeaderSize()>>3, "bytes")
println("tl:", cur.actualTableLog, "symbolLen:", cur.symbolLen, "norm:", cur.norm[:cur.symbolLen], "hist", cur.count[:cur.symbolLen])
}
return cur, compModeFSE
}
}
// Write compression mode
var mode uint8
if llEnc.useRLE {
mode |= uint8(compModeRLE) << 6
llEnc.setRLE(b.sequences[0].llCode)
if debugEncoder {
println("llEnc.useRLE")
}
} else {
var m seqCompMode
llEnc, m = chooseComp(llEnc, b.coders.llPrev, &fsePredefEnc[tableLiteralLengths])
mode |= uint8(m) << 6
}
if ofEnc.useRLE {
mode |= uint8(compModeRLE) << 4
ofEnc.setRLE(b.sequences[0].ofCode)
if debugEncoder {
println("ofEnc.useRLE")
}
} else {
var m seqCompMode
ofEnc, m = chooseComp(ofEnc, b.coders.ofPrev, &fsePredefEnc[tableOffsets])
mode |= uint8(m) << 4
}
if mlEnc.useRLE {
mode |= uint8(compModeRLE) << 2
mlEnc.setRLE(b.sequences[0].mlCode)
if debugEncoder {
println("mlEnc.useRLE, code: ", b.sequences[0].mlCode, "value", b.sequences[0].matchLen)
}
} else {
var m seqCompMode
mlEnc, m = chooseComp(mlEnc, b.coders.mlPrev, &fsePredefEnc[tableMatchLengths])
mode |= uint8(m) << 2
}
b.output = append(b.output, mode)
if debugEncoder {
printf("Compression modes: 0b%b", mode)
}
b.output, err = llEnc.writeCount(b.output)
if err != nil {
return err
}
start := len(b.output)
b.output, err = ofEnc.writeCount(b.output)
if err != nil {
return err
}
if false {
println("block:", b.output[start:], "tablelog", ofEnc.actualTableLog, "maxcount:", ofEnc.maxCount)
fmt.Printf("selected TableLog: %d, Symbol length: %d\n", ofEnc.actualTableLog, ofEnc.symbolLen)
for i, v := range ofEnc.norm[:ofEnc.symbolLen] {
fmt.Printf("%3d: %5d -> %4d \n", i, ofEnc.count[i], v)
}
}
b.output, err = mlEnc.writeCount(b.output)
if err != nil {
return err
}
// Maybe in block?
wr := &b.wr
wr.reset(b.output)
var ll, of, ml cState
// Current sequence
seq := len(b.sequences) - 1
s := b.sequences[seq]
llEnc.setBits(llBitsTable[:])
mlEnc.setBits(mlBitsTable[:])
ofEnc.setBits(nil)
llTT, ofTT, mlTT := llEnc.ct.symbolTT[:256], ofEnc.ct.symbolTT[:256], mlEnc.ct.symbolTT[:256]
// We have 3 bounds checks here (and in the loop).
// Since we are iterating backwards it is kinda hard to avoid.
llB, ofB, mlB := llTT[s.llCode], ofTT[s.ofCode], mlTT[s.mlCode]
ll.init(wr, &llEnc.ct, llB)
of.init(wr, &ofEnc.ct, ofB)
wr.flush32()
ml.init(wr, &mlEnc.ct, mlB)
// Each of these lookups also generates a bounds check.
wr.addBits32NC(s.litLen, llB.outBits)
wr.addBits32NC(s.matchLen, mlB.outBits)
wr.flush32()
wr.addBits32NC(s.offset, ofB.outBits)
if debugSequences {
println("Encoded seq", seq, s, "codes:", s.llCode, s.mlCode, s.ofCode, "states:", ll.state, ml.state, of.state, "bits:", llB, mlB, ofB)
}
seq--
// Store sequences in reverse...
for seq >= 0 {
s = b.sequences[seq]
ofB := ofTT[s.ofCode]
wr.flush32() // tablelog max is below 8 for each, so it will fill max 24 bits.
//of.encode(ofB)
nbBitsOut := (uint32(of.state) + ofB.deltaNbBits) >> 16
dstState := int32(of.state>>(nbBitsOut&15)) + int32(ofB.deltaFindState)
wr.addBits16NC(of.state, uint8(nbBitsOut))
of.state = of.stateTable[dstState]
// Accumulate extra bits.
outBits := ofB.outBits & 31
extraBits := uint64(s.offset & bitMask32[outBits])
extraBitsN := outBits
mlB := mlTT[s.mlCode]
//ml.encode(mlB)
nbBitsOut = (uint32(ml.state) + mlB.deltaNbBits) >> 16
dstState = int32(ml.state>>(nbBitsOut&15)) + int32(mlB.deltaFindState)
wr.addBits16NC(ml.state, uint8(nbBitsOut))
ml.state = ml.stateTable[dstState]
outBits = mlB.outBits & 31
extraBits = extraBits<<outBits | uint64(s.matchLen&bitMask32[outBits])
extraBitsN += outBits
llB := llTT[s.llCode]
//ll.encode(llB)
nbBitsOut = (uint32(ll.state) + llB.deltaNbBits) >> 16
dstState = int32(ll.state>>(nbBitsOut&15)) + int32(llB.deltaFindState)
wr.addBits16NC(ll.state, uint8(nbBitsOut))
ll.state = ll.stateTable[dstState]
outBits = llB.outBits & 31
extraBits = extraBits<<outBits | uint64(s.litLen&bitMask32[outBits])
extraBitsN += outBits
wr.flush32()
wr.addBits64NC(extraBits, extraBitsN)
if debugSequences {
println("Encoded seq", seq, s)
}
seq--
}
ml.flush(mlEnc.actualTableLog)
of.flush(ofEnc.actualTableLog)
ll.flush(llEnc.actualTableLog)
wr.close()
b.output = wr.out
// Maybe even add a bigger margin.
if len(b.output)-3-bhOffset >= b.size {
// Discard and encode as raw block.
b.output = b.encodeRawTo(b.output[:bhOffset], org)
b.popOffsets()
b.litEnc.Reuse = huff0.ReusePolicyNone
return nil
}
// Size is output minus block header.
bh.setSize(uint32(len(b.output)-bhOffset) - 3)
if debugEncoder {
println("Rewriting block header", bh)
}
_ = bh.appendTo(b.output[bhOffset:bhOffset])
b.coders.setPrev(llEnc, mlEnc, ofEnc)
return nil
}
var errIncompressible = errors.New("incompressible")
func (b *blockEnc) genCodes() {
if len(b.sequences) == 0 {
// nothing to do
return
}
if len(b.sequences) > math.MaxUint16 {
panic("can only encode up to 64K sequences")
}
// No bounds checks after here:
llH := b.coders.llEnc.Histogram()
ofH := b.coders.ofEnc.Histogram()
mlH := b.coders.mlEnc.Histogram()
for i := range llH {
llH[i] = 0
}
for i := range ofH {
ofH[i] = 0
}
for i := range mlH {
mlH[i] = 0
}
var llMax, ofMax, mlMax uint8
for i := range b.sequences {
seq := &b.sequences[i]
v := llCode(seq.litLen)
seq.llCode = v
llH[v]++
if v > llMax {
llMax = v
}
v = ofCode(seq.offset)
seq.ofCode = v
ofH[v]++
if v > ofMax {
ofMax = v
}
v = mlCode(seq.matchLen)
seq.mlCode = v
mlH[v]++
if v > mlMax {
mlMax = v
if debugAsserts && mlMax > maxMatchLengthSymbol {
panic(fmt.Errorf("mlMax > maxMatchLengthSymbol (%d), matchlen: %d", mlMax, seq.matchLen))
}
}
}
maxCount := func(a []uint32) int {
var max uint32
for _, v := range a {
if v > max {
max = v
}
}
return int(max)
}
if debugAsserts && mlMax > maxMatchLengthSymbol {
panic(fmt.Errorf("mlMax > maxMatchLengthSymbol (%d)", mlMax))
}
if debugAsserts && ofMax > maxOffsetBits {
panic(fmt.Errorf("ofMax > maxOffsetBits (%d)", ofMax))
}
if debugAsserts && llMax > maxLiteralLengthSymbol {
panic(fmt.Errorf("llMax > maxLiteralLengthSymbol (%d)", llMax))
}
b.coders.mlEnc.HistogramFinished(mlMax, maxCount(mlH[:mlMax+1]))
b.coders.ofEnc.HistogramFinished(ofMax, maxCount(ofH[:ofMax+1]))
b.coders.llEnc.HistogramFinished(llMax, maxCount(llH[:llMax+1]))
}

85
vendor/github.com/klauspost/compress/zstd/blocktype_string.go generated vendored Normal file
View File

@@ -0,0 +1,85 @@
// Code generated by "stringer -type=blockType,literalsBlockType,seqCompMode,tableIndex"; DO NOT EDIT.
package zstd
import "strconv"
func _() {
// An "invalid array index" compiler error signifies that the constant values have changed.
// Re-run the stringer command to generate them again.
var x [1]struct{}
_ = x[blockTypeRaw-0]
_ = x[blockTypeRLE-1]
_ = x[blockTypeCompressed-2]
_ = x[blockTypeReserved-3]
}
const _blockType_name = "blockTypeRawblockTypeRLEblockTypeCompressedblockTypeReserved"
var _blockType_index = [...]uint8{0, 12, 24, 43, 60}
func (i blockType) String() string {
if i >= blockType(len(_blockType_index)-1) {
return "blockType(" + strconv.FormatInt(int64(i), 10) + ")"
}
return _blockType_name[_blockType_index[i]:_blockType_index[i+1]]
}
func _() {
// An "invalid array index" compiler error signifies that the constant values have changed.
// Re-run the stringer command to generate them again.
var x [1]struct{}
_ = x[literalsBlockRaw-0]
_ = x[literalsBlockRLE-1]
_ = x[literalsBlockCompressed-2]
_ = x[literalsBlockTreeless-3]
}
const _literalsBlockType_name = "literalsBlockRawliteralsBlockRLEliteralsBlockCompressedliteralsBlockTreeless"
var _literalsBlockType_index = [...]uint8{0, 16, 32, 55, 76}
func (i literalsBlockType) String() string {
if i >= literalsBlockType(len(_literalsBlockType_index)-1) {
return "literalsBlockType(" + strconv.FormatInt(int64(i), 10) + ")"
}
return _literalsBlockType_name[_literalsBlockType_index[i]:_literalsBlockType_index[i+1]]
}
func _() {
// An "invalid array index" compiler error signifies that the constant values have changed.
// Re-run the stringer command to generate them again.
var x [1]struct{}
_ = x[compModePredefined-0]
_ = x[compModeRLE-1]
_ = x[compModeFSE-2]
_ = x[compModeRepeat-3]
}
const _seqCompMode_name = "compModePredefinedcompModeRLEcompModeFSEcompModeRepeat"
var _seqCompMode_index = [...]uint8{0, 18, 29, 40, 54}
func (i seqCompMode) String() string {
if i >= seqCompMode(len(_seqCompMode_index)-1) {
return "seqCompMode(" + strconv.FormatInt(int64(i), 10) + ")"
}
return _seqCompMode_name[_seqCompMode_index[i]:_seqCompMode_index[i+1]]
}
func _() {
// An "invalid array index" compiler error signifies that the constant values have changed.
// Re-run the stringer command to generate them again.
var x [1]struct{}
_ = x[tableLiteralLengths-0]
_ = x[tableOffsets-1]
_ = x[tableMatchLengths-2]
}
const _tableIndex_name = "tableLiteralLengthstableOffsetstableMatchLengths"
var _tableIndex_index = [...]uint8{0, 19, 31, 48}
func (i tableIndex) String() string {
if i >= tableIndex(len(_tableIndex_index)-1) {
return "tableIndex(" + strconv.FormatInt(int64(i), 10) + ")"
}
return _tableIndex_name[_tableIndex_index[i]:_tableIndex_index[i+1]]
}

131
vendor/github.com/klauspost/compress/zstd/bytebuf.go generated vendored Normal file
View File

@@ -0,0 +1,131 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"fmt"
"io"
)
type byteBuffer interface {
// Read up to 8 bytes.
// Returns io.ErrUnexpectedEOF if this cannot be satisfied.
readSmall(n int) ([]byte, error)
// Read >8 bytes.
// MAY use the destination slice.
readBig(n int, dst []byte) ([]byte, error)
// Read a single byte.
readByte() (byte, error)
// Skip n bytes.
skipN(n int64) error
}
// in-memory buffer
type byteBuf []byte
func (b *byteBuf) readSmall(n int) ([]byte, error) {
if debugAsserts && n > 8 {
panic(fmt.Errorf("small read > 8 (%d). use readBig", n))
}
bb := *b
if len(bb) < n {
return nil, io.ErrUnexpectedEOF
}
r := bb[:n]
*b = bb[n:]
return r, nil
}
func (b *byteBuf) readBig(n int, dst []byte) ([]byte, error) {
bb := *b
if len(bb) < n {
return nil, io.ErrUnexpectedEOF
}
r := bb[:n]
*b = bb[n:]
return r, nil
}
func (b *byteBuf) readByte() (byte, error) {
bb := *b
if len(bb) < 1 {
return 0, io.ErrUnexpectedEOF
}
r := bb[0]
*b = bb[1:]
return r, nil
}
func (b *byteBuf) skipN(n int64) error {
bb := *b
if n < 0 {
return fmt.Errorf("negative skip (%d) requested", n)
}
if int64(len(bb)) < n {
return io.ErrUnexpectedEOF
}
*b = bb[n:]
return nil
}
// wrapper around a reader.
type readerWrapper struct {
r io.Reader
tmp [8]byte
}
func (r *readerWrapper) readSmall(n int) ([]byte, error) {
if debugAsserts && n > 8 {
panic(fmt.Errorf("small read > 8 (%d). use readBig", n))
}
n2, err := io.ReadFull(r.r, r.tmp[:n])
// We only really care about the actual bytes read.
if err != nil {
if err == io.EOF {
return nil, io.ErrUnexpectedEOF
}
if debugDecoder {
println("readSmall: got", n2, "want", n, "err", err)
}
return nil, err
}
return r.tmp[:n], nil
}
func (r *readerWrapper) readBig(n int, dst []byte) ([]byte, error) {
if cap(dst) < n {
dst = make([]byte, n)
}
n2, err := io.ReadFull(r.r, dst[:n])
if err == io.EOF && n > 0 {
err = io.ErrUnexpectedEOF
}
return dst[:n2], err
}
func (r *readerWrapper) readByte() (byte, error) {
n2, err := io.ReadFull(r.r, r.tmp[:1])
if err != nil {
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return 0, err
}
if n2 != 1 {
return 0, io.ErrUnexpectedEOF
}
return r.tmp[0], nil
}
func (r *readerWrapper) skipN(n int64) error {
n2, err := io.CopyN(io.Discard, r.r, n)
if n2 != n {
err = io.ErrUnexpectedEOF
}
return err
}

82
vendor/github.com/klauspost/compress/zstd/bytereader.go generated vendored Normal file
View File

@@ -0,0 +1,82 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
// byteReader provides a byte reader that reads
// little endian values from a byte stream.
// The input stream is manually advanced.
// The reader performs no bounds checks.
type byteReader struct {
b []byte
off int
}
// advance the stream b n bytes.
func (b *byteReader) advance(n uint) {
b.off += int(n)
}
// overread returns whether we have advanced too far.
func (b *byteReader) overread() bool {
return b.off > len(b.b)
}
// Int32 returns a little endian int32 starting at current offset.
func (b byteReader) Int32() int32 {
b2 := b.b[b.off:]
b2 = b2[:4]
v3 := int32(b2[3])
v2 := int32(b2[2])
v1 := int32(b2[1])
v0 := int32(b2[0])
return v0 | (v1 << 8) | (v2 << 16) | (v3 << 24)
}
// Uint8 returns the next byte
func (b *byteReader) Uint8() uint8 {
v := b.b[b.off]
return v
}
// Uint32 returns a little endian uint32 starting at current offset.
func (b byteReader) Uint32() uint32 {
if r := b.remain(); r < 4 {
// Very rare
v := uint32(0)
for i := 1; i <= r; i++ {
v = (v << 8) | uint32(b.b[len(b.b)-i])
}
return v
}
b2 := b.b[b.off:]
b2 = b2[:4]
v3 := uint32(b2[3])
v2 := uint32(b2[2])
v1 := uint32(b2[1])
v0 := uint32(b2[0])
return v0 | (v1 << 8) | (v2 << 16) | (v3 << 24)
}
// Uint32NC returns a little endian uint32 starting at current offset.
// The caller must be sure if there are at least 4 bytes left.
func (b byteReader) Uint32NC() uint32 {
b2 := b.b[b.off:]
b2 = b2[:4]
v3 := uint32(b2[3])
v2 := uint32(b2[2])
v1 := uint32(b2[1])
v0 := uint32(b2[0])
return v0 | (v1 << 8) | (v2 << 16) | (v3 << 24)
}
// unread returns the unread portion of the input.
func (b byteReader) unread() []byte {
return b.b[b.off:]
}
// remain will return the number of bytes remaining.
func (b byteReader) remain() int {
return len(b.b) - b.off
}

261
vendor/github.com/klauspost/compress/zstd/decodeheader.go generated vendored Normal file
View File

@@ -0,0 +1,261 @@
// Copyright 2020+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
package zstd
import (
"encoding/binary"
"errors"
"io"
)
// HeaderMaxSize is the maximum size of a Frame and Block Header.
// If less is sent to Header.Decode it *may* still contain enough information.
const HeaderMaxSize = 14 + 3
// Header contains information about the first frame and block within that.
type Header struct {
// SingleSegment specifies whether the data is to be decompressed into a
// single contiguous memory segment.
// It implies that WindowSize is invalid and that FrameContentSize is valid.
SingleSegment bool
// WindowSize is the window of data to keep while decoding.
// Will only be set if SingleSegment is false.
WindowSize uint64
// Dictionary ID.
// If 0, no dictionary.
DictionaryID uint32
// HasFCS specifies whether FrameContentSize has a valid value.
HasFCS bool
// FrameContentSize is the expected uncompressed size of the entire frame.
FrameContentSize uint64
// Skippable will be true if the frame is meant to be skipped.
// This implies that FirstBlock.OK is false.
Skippable bool
// SkippableID is the user-specific ID for the skippable frame.
// Valid values are between 0 to 15, inclusive.
SkippableID int
// SkippableSize is the length of the user data to skip following
// the header.
SkippableSize uint32
// HeaderSize is the raw size of the frame header.
//
// For normal frames, it includes the size of the magic number and
// the size of the header (per section 3.1.1.1).
// It does not include the size for any data blocks (section 3.1.1.2) nor
// the size for the trailing content checksum.
//
// For skippable frames, this counts the size of the magic number
// along with the size of the size field of the payload.
// It does not include the size of the skippable payload itself.
// The total frame size is the HeaderSize plus the SkippableSize.
HeaderSize int
// First block information.
FirstBlock struct {
// OK will be set if first block could be decoded.
OK bool
// Is this the last block of a frame?
Last bool
// Is the data compressed?
// If true CompressedSize will be populated.
// Unfortunately DecompressedSize cannot be determined
// without decoding the blocks.
Compressed bool
// DecompressedSize is the expected decompressed size of the block.
// Will be 0 if it cannot be determined.
DecompressedSize int
// CompressedSize of the data in the block.
// Does not include the block header.
// Will be equal to DecompressedSize if not Compressed.
CompressedSize int
}
// If set there is a checksum present for the block content.
// The checksum field at the end is always 4 bytes long.
HasCheckSum bool
}
// Decode the header from the beginning of the stream.
// This will decode the frame header and the first block header if enough bytes are provided.
// It is recommended to provide at least HeaderMaxSize bytes.
// If the frame header cannot be read an error will be returned.
// If there isn't enough input, io.ErrUnexpectedEOF is returned.
// The FirstBlock.OK will indicate if enough information was available to decode the first block header.
func (h *Header) Decode(in []byte) error {
_, err := h.DecodeAndStrip(in)
return err
}
// DecodeAndStrip will decode the header from the beginning of the stream
// and on success return the remaining bytes.
// This will decode the frame header and the first block header if enough bytes are provided.
// It is recommended to provide at least HeaderMaxSize bytes.
// If the frame header cannot be read an error will be returned.
// If there isn't enough input, io.ErrUnexpectedEOF is returned.
// The FirstBlock.OK will indicate if enough information was available to decode the first block header.
func (h *Header) DecodeAndStrip(in []byte) (remain []byte, err error) {
*h = Header{}
if len(in) < 4 {
return nil, io.ErrUnexpectedEOF
}
h.HeaderSize += 4
b, in := in[:4], in[4:]
if string(b) != frameMagic {
if string(b[1:4]) != skippableFrameMagic || b[0]&0xf0 != 0x50 {
return nil, ErrMagicMismatch
}
if len(in) < 4 {
return nil, io.ErrUnexpectedEOF
}
h.HeaderSize += 4
h.Skippable = true
h.SkippableID = int(b[0] & 0xf)
h.SkippableSize = binary.LittleEndian.Uint32(in)
return in[4:], nil
}
// Read Window_Descriptor
// https://github.com/facebook/zstd/blob/dev/doc/zstd_compression_format.md#window_descriptor
if len(in) < 1 {
return nil, io.ErrUnexpectedEOF
}
fhd, in := in[0], in[1:]
h.HeaderSize++
h.SingleSegment = fhd&(1<<5) != 0
h.HasCheckSum = fhd&(1<<2) != 0
if fhd&(1<<3) != 0 {
return nil, errors.New("reserved bit set on frame header")
}
if !h.SingleSegment {
if len(in) < 1 {
return nil, io.ErrUnexpectedEOF
}
var wd byte
wd, in = in[0], in[1:]
h.HeaderSize++
windowLog := 10 + (wd >> 3)
windowBase := uint64(1) << windowLog
windowAdd := (windowBase / 8) * uint64(wd&0x7)
h.WindowSize = windowBase + windowAdd
}
// Read Dictionary_ID
// https://github.com/facebook/zstd/blob/dev/doc/zstd_compression_format.md#dictionary_id
if size := fhd & 3; size != 0 {
if size == 3 {
size = 4
}
if len(in) < int(size) {
return nil, io.ErrUnexpectedEOF
}
b, in = in[:size], in[size:]
h.HeaderSize += int(size)
switch len(b) {
case 1:
h.DictionaryID = uint32(b[0])
case 2:
h.DictionaryID = uint32(b[0]) | (uint32(b[1]) << 8)
case 4:
h.DictionaryID = uint32(b[0]) | (uint32(b[1]) << 8) | (uint32(b[2]) << 16) | (uint32(b[3]) << 24)
}
}
// Read Frame_Content_Size
// https://github.com/facebook/zstd/blob/dev/doc/zstd_compression_format.md#frame_content_size
var fcsSize int
v := fhd >> 6
switch v {
case 0:
if h.SingleSegment {
fcsSize = 1
}
default:
fcsSize = 1 << v
}
if fcsSize > 0 {
h.HasFCS = true
if len(in) < fcsSize {
return nil, io.ErrUnexpectedEOF
}
b, in = in[:fcsSize], in[fcsSize:]
h.HeaderSize += int(fcsSize)
switch len(b) {
case 1:
h.FrameContentSize = uint64(b[0])
case 2:
// When FCS_Field_Size is 2, the offset of 256 is added.
h.FrameContentSize = uint64(b[0]) | (uint64(b[1]) << 8) + 256
case 4:
h.FrameContentSize = uint64(b[0]) | (uint64(b[1]) << 8) | (uint64(b[2]) << 16) | (uint64(b[3]) << 24)
case 8:
d1 := uint32(b[0]) | (uint32(b[1]) << 8) | (uint32(b[2]) << 16) | (uint32(b[3]) << 24)
d2 := uint32(b[4]) | (uint32(b[5]) << 8) | (uint32(b[6]) << 16) | (uint32(b[7]) << 24)
h.FrameContentSize = uint64(d1) | (uint64(d2) << 32)
}
}
// Frame Header done, we will not fail from now on.
if len(in) < 3 {
return in, nil
}
tmp := in[:3]
bh := uint32(tmp[0]) | (uint32(tmp[1]) << 8) | (uint32(tmp[2]) << 16)
h.FirstBlock.Last = bh&1 != 0
blockType := blockType((bh >> 1) & 3)
// find size.
cSize := int(bh >> 3)
switch blockType {
case blockTypeReserved:
return in, nil
case blockTypeRLE:
h.FirstBlock.Compressed = true
h.FirstBlock.DecompressedSize = cSize
h.FirstBlock.CompressedSize = 1
case blockTypeCompressed:
h.FirstBlock.Compressed = true
h.FirstBlock.CompressedSize = cSize
case blockTypeRaw:
h.FirstBlock.DecompressedSize = cSize
h.FirstBlock.CompressedSize = cSize
default:
panic("Invalid block type")
}
h.FirstBlock.OK = true
return in, nil
}
// AppendTo will append the encoded header to the dst slice.
// There is no error checking performed on the header values.
func (h *Header) AppendTo(dst []byte) ([]byte, error) {
if h.Skippable {
magic := [4]byte{0x50, 0x2a, 0x4d, 0x18}
magic[0] |= byte(h.SkippableID & 0xf)
dst = append(dst, magic[:]...)
f := h.SkippableSize
return append(dst, uint8(f), uint8(f>>8), uint8(f>>16), uint8(f>>24)), nil
}
f := frameHeader{
ContentSize: h.FrameContentSize,
WindowSize: uint32(h.WindowSize),
SingleSegment: h.SingleSegment,
Checksum: h.HasCheckSum,
DictID: h.DictionaryID,
}
return f.appendTo(dst), nil
}

948
vendor/github.com/klauspost/compress/zstd/decoder.go generated vendored Normal file
View File

@@ -0,0 +1,948 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"context"
"encoding/binary"
"io"
"sync"
"github.com/klauspost/compress/zstd/internal/xxhash"
)
// Decoder provides decoding of zstandard streams.
// The decoder has been designed to operate without allocations after a warmup.
// This means that you should store the decoder for best performance.
// To re-use a stream decoder, use the Reset(r io.Reader) error to switch to another stream.
// A decoder can safely be re-used even if the previous stream failed.
// To release the resources, you must call the Close() function on a decoder.
type Decoder struct {
o decoderOptions
// Unreferenced decoders, ready for use.
decoders chan *blockDec
// Current read position used for Reader functionality.
current decoderState
// sync stream decoding
syncStream struct {
decodedFrame uint64
br readerWrapper
enabled bool
inFrame bool
dstBuf []byte
}
frame *frameDec
// Custom dictionaries.
dicts map[uint32]*dict
// streamWg is the waitgroup for all streams
streamWg sync.WaitGroup
}
// decoderState is used for maintaining state when the decoder
// is used for streaming.
type decoderState struct {
// current block being written to stream.
decodeOutput
// output in order to be written to stream.
output chan decodeOutput
// cancel remaining output.
cancel context.CancelFunc
// crc of current frame
crc *xxhash.Digest
flushed bool
}
var (
// Check the interfaces we want to support.
_ = io.WriterTo(&Decoder{})
_ = io.Reader(&Decoder{})
)
// NewReader creates a new decoder.
// A nil Reader can be provided in which case Reset can be used to start a decode.
//
// A Decoder can be used in two modes:
//
// 1) As a stream, or
// 2) For stateless decoding using DecodeAll.
//
// Only a single stream can be decoded concurrently, but the same decoder
// can run multiple concurrent stateless decodes. It is even possible to
// use stateless decodes while a stream is being decoded.
//
// The Reset function can be used to initiate a new stream, which will considerably
// reduce the allocations normally caused by NewReader.
func NewReader(r io.Reader, opts ...DOption) (*Decoder, error) {
initPredefined()
var d Decoder
d.o.setDefault()
for _, o := range opts {
err := o(&d.o)
if err != nil {
return nil, err
}
}
d.current.crc = xxhash.New()
d.current.flushed = true
if r == nil {
d.current.err = ErrDecoderNilInput
}
// Transfer option dicts.
d.dicts = make(map[uint32]*dict, len(d.o.dicts))
for _, dc := range d.o.dicts {
d.dicts[dc.id] = dc
}
d.o.dicts = nil
// Create decoders
d.decoders = make(chan *blockDec, d.o.concurrent)
for i := 0; i < d.o.concurrent; i++ {
dec := newBlockDec(d.o.lowMem)
dec.localFrame = newFrameDec(d.o)
d.decoders <- dec
}
if r == nil {
return &d, nil
}
return &d, d.Reset(r)
}
// Read bytes from the decompressed stream into p.
// Returns the number of bytes written and any error that occurred.
// When the stream is done, io.EOF will be returned.
func (d *Decoder) Read(p []byte) (int, error) {
var n int
for {
if len(d.current.b) > 0 {
filled := copy(p, d.current.b)
p = p[filled:]
d.current.b = d.current.b[filled:]
n += filled
}
if len(p) == 0 {
break
}
if len(d.current.b) == 0 {
// We have an error and no more data
if d.current.err != nil {
break
}
if !d.nextBlock(n == 0) {
return n, d.current.err
}
}
}
if len(d.current.b) > 0 {
if debugDecoder {
println("returning", n, "still bytes left:", len(d.current.b))
}
// Only return error at end of block
return n, nil
}
if d.current.err != nil {
d.drainOutput()
}
if debugDecoder {
println("returning", n, d.current.err, len(d.decoders))
}
return n, d.current.err
}
// Reset will reset the decoder the supplied stream after the current has finished processing.
// Note that this functionality cannot be used after Close has been called.
// Reset can be called with a nil reader to release references to the previous reader.
// After being called with a nil reader, no other operations than Reset or DecodeAll or Close
// should be used.
func (d *Decoder) Reset(r io.Reader) error {
if d.current.err == ErrDecoderClosed {
return d.current.err
}
d.drainOutput()
d.syncStream.br.r = nil
if r == nil {
d.current.err = ErrDecoderNilInput
if len(d.current.b) > 0 {
d.current.b = d.current.b[:0]
}
d.current.flushed = true
return nil
}
// If bytes buffer and < 5MB, do sync decoding anyway.
if bb, ok := r.(byter); ok && bb.Len() < d.o.decodeBufsBelow && !d.o.limitToCap {
bb2 := bb
if debugDecoder {
println("*bytes.Buffer detected, doing sync decode, len:", bb.Len())
}
b := bb2.Bytes()
var dst []byte
if cap(d.syncStream.dstBuf) > 0 {
dst = d.syncStream.dstBuf[:0]
}
dst, err := d.DecodeAll(b, dst)
if err == nil {
err = io.EOF
}
// Save output buffer
d.syncStream.dstBuf = dst
d.current.b = dst
d.current.err = err
d.current.flushed = true
if debugDecoder {
println("sync decode to", len(dst), "bytes, err:", err)
}
return nil
}
// Remove current block.
d.stashDecoder()
d.current.decodeOutput = decodeOutput{}
d.current.err = nil
d.current.flushed = false
d.current.d = nil
d.syncStream.dstBuf = nil
// Ensure no-one else is still running...
d.streamWg.Wait()
if d.frame == nil {
d.frame = newFrameDec(d.o)
}
if d.o.concurrent == 1 {
return d.startSyncDecoder(r)
}
d.current.output = make(chan decodeOutput, d.o.concurrent)
ctx, cancel := context.WithCancel(context.Background())
d.current.cancel = cancel
d.streamWg.Add(1)
go d.startStreamDecoder(ctx, r, d.current.output)
return nil
}
// drainOutput will drain the output until errEndOfStream is sent.
func (d *Decoder) drainOutput() {
if d.current.cancel != nil {
if debugDecoder {
println("cancelling current")
}
d.current.cancel()
d.current.cancel = nil
}
if d.current.d != nil {
if debugDecoder {
printf("re-adding current decoder %p, decoders: %d", d.current.d, len(d.decoders))
}
d.decoders <- d.current.d
d.current.d = nil
d.current.b = nil
}
if d.current.output == nil || d.current.flushed {
println("current already flushed")
return
}
for v := range d.current.output {
if v.d != nil {
if debugDecoder {
printf("re-adding decoder %p", v.d)
}
d.decoders <- v.d
}
}
d.current.output = nil
d.current.flushed = true
}
// WriteTo writes data to w until there's no more data to write or when an error occurs.
// The return value n is the number of bytes written.
// Any error encountered during the write is also returned.
func (d *Decoder) WriteTo(w io.Writer) (int64, error) {
var n int64
for {
if len(d.current.b) > 0 {
n2, err2 := w.Write(d.current.b)
n += int64(n2)
if err2 != nil && (d.current.err == nil || d.current.err == io.EOF) {
d.current.err = err2
} else if n2 != len(d.current.b) {
d.current.err = io.ErrShortWrite
}
}
if d.current.err != nil {
break
}
d.nextBlock(true)
}
err := d.current.err
if err != nil {
d.drainOutput()
}
if err == io.EOF {
err = nil
}
return n, err
}
// DecodeAll allows stateless decoding of a blob of bytes.
// Output will be appended to dst, so if the destination size is known
// you can pre-allocate the destination slice to avoid allocations.
// DecodeAll can be used concurrently.
// The Decoder concurrency limits will be respected.
func (d *Decoder) DecodeAll(input, dst []byte) ([]byte, error) {
if d.decoders == nil {
return dst, ErrDecoderClosed
}
// Grab a block decoder and frame decoder.
block := <-d.decoders
frame := block.localFrame
initialSize := len(dst)
defer func() {
if debugDecoder {
printf("re-adding decoder: %p", block)
}
frame.rawInput = nil
frame.bBuf = nil
if frame.history.decoders.br != nil {
frame.history.decoders.br.in = nil
}
d.decoders <- block
}()
frame.bBuf = input
for {
frame.history.reset()
err := frame.reset(&frame.bBuf)
if err != nil {
if err == io.EOF {
if debugDecoder {
println("frame reset return EOF")
}
return dst, nil
}
return dst, err
}
if err = d.setDict(frame); err != nil {
return nil, err
}
if frame.WindowSize > d.o.maxWindowSize {
if debugDecoder {
println("window size exceeded:", frame.WindowSize, ">", d.o.maxWindowSize)
}
return dst, ErrWindowSizeExceeded
}
if frame.FrameContentSize != fcsUnknown {
if frame.FrameContentSize > d.o.maxDecodedSize-uint64(len(dst)-initialSize) {
if debugDecoder {
println("decoder size exceeded; fcs:", frame.FrameContentSize, "> mcs:", d.o.maxDecodedSize-uint64(len(dst)-initialSize), "len:", len(dst))
}
return dst, ErrDecoderSizeExceeded
}
if d.o.limitToCap && frame.FrameContentSize > uint64(cap(dst)-len(dst)) {
if debugDecoder {
println("decoder size exceeded; fcs:", frame.FrameContentSize, "> (cap-len)", cap(dst)-len(dst))
}
return dst, ErrDecoderSizeExceeded
}
if cap(dst)-len(dst) < int(frame.FrameContentSize) {
dst2 := make([]byte, len(dst), len(dst)+int(frame.FrameContentSize)+compressedBlockOverAlloc)
copy(dst2, dst)
dst = dst2
}
}
if cap(dst) == 0 && !d.o.limitToCap {
// Allocate len(input) * 2 by default if nothing is provided
// and we didn't get frame content size.
size := len(input) * 2
// Cap to 1 MB.
if size > 1<<20 {
size = 1 << 20
}
if uint64(size) > d.o.maxDecodedSize {
size = int(d.o.maxDecodedSize)
}
dst = make([]byte, 0, size)
}
dst, err = frame.runDecoder(dst, block)
if err != nil {
return dst, err
}
if uint64(len(dst)-initialSize) > d.o.maxDecodedSize {
return dst, ErrDecoderSizeExceeded
}
if len(frame.bBuf) == 0 {
if debugDecoder {
println("frame dbuf empty")
}
break
}
}
return dst, nil
}
// nextBlock returns the next block.
// If an error occurs d.err will be set.
// Optionally the function can block for new output.
// If non-blocking mode is used the returned boolean will be false
// if no data was available without blocking.
func (d *Decoder) nextBlock(blocking bool) (ok bool) {
if d.current.err != nil {
// Keep error state.
return false
}
d.current.b = d.current.b[:0]
// SYNC:
if d.syncStream.enabled {
if !blocking {
return false
}
ok = d.nextBlockSync()
if !ok {
d.stashDecoder()
}
return ok
}
//ASYNC:
d.stashDecoder()
if blocking {
d.current.decodeOutput, ok = <-d.current.output
} else {
select {
case d.current.decodeOutput, ok = <-d.current.output:
default:
return false
}
}
if !ok {
// This should not happen, so signal error state...
d.current.err = io.ErrUnexpectedEOF
return false
}
next := d.current.decodeOutput
if next.d != nil && next.d.async.newHist != nil {
d.current.crc.Reset()
}
if debugDecoder {
var tmp [4]byte
binary.LittleEndian.PutUint32(tmp[:], uint32(xxhash.Sum64(next.b)))
println("got", len(d.current.b), "bytes, error:", d.current.err, "data crc:", tmp)
}
if d.o.ignoreChecksum {
return true
}
if len(next.b) > 0 {
d.current.crc.Write(next.b)
}
if next.err == nil && next.d != nil && next.d.hasCRC {
got := uint32(d.current.crc.Sum64())
if got != next.d.checkCRC {
if debugDecoder {
printf("CRC Check Failed: %08x (got) != %08x (on stream)\n", got, next.d.checkCRC)
}
d.current.err = ErrCRCMismatch
} else {
if debugDecoder {
printf("CRC ok %08x\n", got)
}
}
}
return true
}
func (d *Decoder) nextBlockSync() (ok bool) {
if d.current.d == nil {
d.current.d = <-d.decoders
}
for len(d.current.b) == 0 {
if !d.syncStream.inFrame {
d.frame.history.reset()
d.current.err = d.frame.reset(&d.syncStream.br)
if d.current.err == nil {
d.current.err = d.setDict(d.frame)
}
if d.current.err != nil {
return false
}
if d.frame.WindowSize > d.o.maxDecodedSize || d.frame.WindowSize > d.o.maxWindowSize {
d.current.err = ErrDecoderSizeExceeded
return false
}
d.syncStream.decodedFrame = 0
d.syncStream.inFrame = true
}
d.current.err = d.frame.next(d.current.d)
if d.current.err != nil {
return false
}
d.frame.history.ensureBlock()
if debugDecoder {
println("History trimmed:", len(d.frame.history.b), "decoded already:", d.syncStream.decodedFrame)
}
histBefore := len(d.frame.history.b)
d.current.err = d.current.d.decodeBuf(&d.frame.history)
if d.current.err != nil {
println("error after:", d.current.err)
return false
}
d.current.b = d.frame.history.b[histBefore:]
if debugDecoder {
println("history after:", len(d.frame.history.b))
}
// Check frame size (before CRC)
d.syncStream.decodedFrame += uint64(len(d.current.b))
if d.syncStream.decodedFrame > d.frame.FrameContentSize {
if debugDecoder {
printf("DecodedFrame (%d) > FrameContentSize (%d)\n", d.syncStream.decodedFrame, d.frame.FrameContentSize)
}
d.current.err = ErrFrameSizeExceeded
return false
}
// Check FCS
if d.current.d.Last && d.frame.FrameContentSize != fcsUnknown && d.syncStream.decodedFrame != d.frame.FrameContentSize {
if debugDecoder {
printf("DecodedFrame (%d) != FrameContentSize (%d)\n", d.syncStream.decodedFrame, d.frame.FrameContentSize)
}
d.current.err = ErrFrameSizeMismatch
return false
}
// Update/Check CRC
if d.frame.HasCheckSum {
if !d.o.ignoreChecksum {
d.frame.crc.Write(d.current.b)
}
if d.current.d.Last {
if !d.o.ignoreChecksum {
d.current.err = d.frame.checkCRC()
} else {
d.current.err = d.frame.consumeCRC()
}
if d.current.err != nil {
println("CRC error:", d.current.err)
return false
}
}
}
d.syncStream.inFrame = !d.current.d.Last
}
return true
}
func (d *Decoder) stashDecoder() {
if d.current.d != nil {
if debugDecoder {
printf("re-adding current decoder %p", d.current.d)
}
d.decoders <- d.current.d
d.current.d = nil
}
}
// Close will release all resources.
// It is NOT possible to reuse the decoder after this.
func (d *Decoder) Close() {
if d.current.err == ErrDecoderClosed {
return
}
d.drainOutput()
if d.current.cancel != nil {
d.current.cancel()
d.streamWg.Wait()
d.current.cancel = nil
}
if d.decoders != nil {
close(d.decoders)
for dec := range d.decoders {
dec.Close()
}
d.decoders = nil
}
if d.current.d != nil {
d.current.d.Close()
d.current.d = nil
}
d.current.err = ErrDecoderClosed
}
// IOReadCloser returns the decoder as an io.ReadCloser for convenience.
// Any changes to the decoder will be reflected, so the returned ReadCloser
// can be reused along with the decoder.
// io.WriterTo is also supported by the returned ReadCloser.
func (d *Decoder) IOReadCloser() io.ReadCloser {
return closeWrapper{d: d}
}
// closeWrapper wraps a function call as a closer.
type closeWrapper struct {
d *Decoder
}
// WriteTo forwards WriteTo calls to the decoder.
func (c closeWrapper) WriteTo(w io.Writer) (n int64, err error) {
return c.d.WriteTo(w)
}
// Read forwards read calls to the decoder.
func (c closeWrapper) Read(p []byte) (n int, err error) {
return c.d.Read(p)
}
// Close closes the decoder.
func (c closeWrapper) Close() error {
c.d.Close()
return nil
}
type decodeOutput struct {
d *blockDec
b []byte
err error
}
func (d *Decoder) startSyncDecoder(r io.Reader) error {
d.frame.history.reset()
d.syncStream.br = readerWrapper{r: r}
d.syncStream.inFrame = false
d.syncStream.enabled = true
d.syncStream.decodedFrame = 0
return nil
}
// Create Decoder:
// ASYNC:
// Spawn 3 go routines.
// 0: Read frames and decode block literals.
// 1: Decode sequences.
// 2: Execute sequences, send to output.
func (d *Decoder) startStreamDecoder(ctx context.Context, r io.Reader, output chan decodeOutput) {
defer d.streamWg.Done()
br := readerWrapper{r: r}
var seqDecode = make(chan *blockDec, d.o.concurrent)
var seqExecute = make(chan *blockDec, d.o.concurrent)
// Async 1: Decode sequences...
go func() {
var hist history
var hasErr bool
for block := range seqDecode {
if hasErr {
if block != nil {
seqExecute <- block
}
continue
}
if block.async.newHist != nil {
if debugDecoder {
println("Async 1: new history, recent:", block.async.newHist.recentOffsets)
}
hist.reset()
hist.decoders = block.async.newHist.decoders
hist.recentOffsets = block.async.newHist.recentOffsets
hist.windowSize = block.async.newHist.windowSize
if block.async.newHist.dict != nil {
hist.setDict(block.async.newHist.dict)
}
}
if block.err != nil || block.Type != blockTypeCompressed {
hasErr = block.err != nil
seqExecute <- block
continue
}
hist.decoders.literals = block.async.literals
block.err = block.prepareSequences(block.async.seqData, &hist)
if debugDecoder && block.err != nil {
println("prepareSequences returned:", block.err)
}
hasErr = block.err != nil
if block.err == nil {
block.err = block.decodeSequences(&hist)
if debugDecoder && block.err != nil {
println("decodeSequences returned:", block.err)
}
hasErr = block.err != nil
// block.async.sequence = hist.decoders.seq[:hist.decoders.nSeqs]
block.async.seqSize = hist.decoders.seqSize
}
seqExecute <- block
}
close(seqExecute)
hist.reset()
}()
var wg sync.WaitGroup
wg.Add(1)
// Async 3: Execute sequences...
frameHistCache := d.frame.history.b
go func() {
var hist history
var decodedFrame uint64
var fcs uint64
var hasErr bool
for block := range seqExecute {
out := decodeOutput{err: block.err, d: block}
if block.err != nil || hasErr {
hasErr = true
output <- out
continue
}
if block.async.newHist != nil {
if debugDecoder {
println("Async 2: new history")
}
hist.reset()
hist.windowSize = block.async.newHist.windowSize
hist.allocFrameBuffer = block.async.newHist.allocFrameBuffer
if block.async.newHist.dict != nil {
hist.setDict(block.async.newHist.dict)
}
if cap(hist.b) < hist.allocFrameBuffer {
if cap(frameHistCache) >= hist.allocFrameBuffer {
hist.b = frameHistCache
} else {
hist.b = make([]byte, 0, hist.allocFrameBuffer)
println("Alloc history sized", hist.allocFrameBuffer)
}
}
hist.b = hist.b[:0]
fcs = block.async.fcs
decodedFrame = 0
}
do := decodeOutput{err: block.err, d: block}
switch block.Type {
case blockTypeRLE:
if debugDecoder {
println("add rle block length:", block.RLESize)
}
if cap(block.dst) < int(block.RLESize) {
if block.lowMem {
block.dst = make([]byte, block.RLESize)
} else {
block.dst = make([]byte, maxCompressedBlockSize)
}
}
block.dst = block.dst[:block.RLESize]
v := block.data[0]
for i := range block.dst {
block.dst[i] = v
}
hist.append(block.dst)
do.b = block.dst
case blockTypeRaw:
if debugDecoder {
println("add raw block length:", len(block.data))
}
hist.append(block.data)
do.b = block.data
case blockTypeCompressed:
if debugDecoder {
println("execute with history length:", len(hist.b), "window:", hist.windowSize)
}
hist.decoders.seqSize = block.async.seqSize
hist.decoders.literals = block.async.literals
do.err = block.executeSequences(&hist)
hasErr = do.err != nil
if debugDecoder && hasErr {
println("executeSequences returned:", do.err)
}
do.b = block.dst
}
if !hasErr {
decodedFrame += uint64(len(do.b))
if decodedFrame > fcs {
println("fcs exceeded", block.Last, fcs, decodedFrame)
do.err = ErrFrameSizeExceeded
hasErr = true
} else if block.Last && fcs != fcsUnknown && decodedFrame != fcs {
do.err = ErrFrameSizeMismatch
hasErr = true
} else {
if debugDecoder {
println("fcs ok", block.Last, fcs, decodedFrame)
}
}
}
output <- do
}
close(output)
frameHistCache = hist.b
wg.Done()
if debugDecoder {
println("decoder goroutines finished")
}
hist.reset()
}()
var hist history
decodeStream:
for {
var hasErr bool
hist.reset()
decodeBlock := func(block *blockDec) {
if hasErr {
if block != nil {
seqDecode <- block
}
return
}
if block.err != nil || block.Type != blockTypeCompressed {
hasErr = block.err != nil
seqDecode <- block
return
}
remain, err := block.decodeLiterals(block.data, &hist)
block.err = err
hasErr = block.err != nil
if err == nil {
block.async.literals = hist.decoders.literals
block.async.seqData = remain
} else if debugDecoder {
println("decodeLiterals error:", err)
}
seqDecode <- block
}
frame := d.frame
if debugDecoder {
println("New frame...")
}
var historySent bool
frame.history.reset()
err := frame.reset(&br)
if debugDecoder && err != nil {
println("Frame decoder returned", err)
}
if err == nil {
err = d.setDict(frame)
}
if err == nil && d.frame.WindowSize > d.o.maxWindowSize {
if debugDecoder {
println("decoder size exceeded, fws:", d.frame.WindowSize, "> mws:", d.o.maxWindowSize)
}
err = ErrDecoderSizeExceeded
}
if err != nil {
select {
case <-ctx.Done():
case dec := <-d.decoders:
dec.sendErr(err)
decodeBlock(dec)
}
break decodeStream
}
// Go through all blocks of the frame.
for {
var dec *blockDec
select {
case <-ctx.Done():
break decodeStream
case dec = <-d.decoders:
// Once we have a decoder, we MUST return it.
}
err := frame.next(dec)
if !historySent {
h := frame.history
if debugDecoder {
println("Alloc History:", h.allocFrameBuffer)
}
hist.reset()
if h.dict != nil {
hist.setDict(h.dict)
}
dec.async.newHist = &h
dec.async.fcs = frame.FrameContentSize
historySent = true
} else {
dec.async.newHist = nil
}
if debugDecoder && err != nil {
println("next block returned error:", err)
}
dec.err = err
dec.hasCRC = false
if dec.Last && frame.HasCheckSum && err == nil {
crc, err := frame.rawInput.readSmall(4)
if len(crc) < 4 {
if err == nil {
err = io.ErrUnexpectedEOF
}
println("CRC missing?", err)
dec.err = err
} else {
dec.checkCRC = binary.LittleEndian.Uint32(crc)
dec.hasCRC = true
if debugDecoder {
printf("found crc to check: %08x\n", dec.checkCRC)
}
}
}
err = dec.err
last := dec.Last
decodeBlock(dec)
if err != nil {
break decodeStream
}
if last {
break
}
}
}
close(seqDecode)
wg.Wait()
hist.reset()
d.frame.history.b = frameHistCache
}
func (d *Decoder) setDict(frame *frameDec) (err error) {
dict, ok := d.dicts[frame.DictionaryID]
if ok {
if debugDecoder {
println("setting dict", frame.DictionaryID)
}
frame.history.setDict(dict)
} else if frame.DictionaryID != 0 {
// A zero or missing dictionary id is ambiguous:
// either dictionary zero, or no dictionary. In particular,
// zstd --patch-from uses this id for the source file,
// so only return an error if the dictionary id is not zero.
err = ErrUnknownDictionary
}
return err
}

169
vendor/github.com/klauspost/compress/zstd/decoder_options.go generated vendored Normal file
View File

@@ -0,0 +1,169 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"errors"
"fmt"
"math/bits"
"runtime"
)
// DOption is an option for creating a decoder.
type DOption func(*decoderOptions) error
// options retains accumulated state of multiple options.
type decoderOptions struct {
lowMem bool
concurrent int
maxDecodedSize uint64
maxWindowSize uint64
dicts []*dict
ignoreChecksum bool
limitToCap bool
decodeBufsBelow int
}
func (o *decoderOptions) setDefault() {
*o = decoderOptions{
// use less ram: true for now, but may change.
lowMem: true,
concurrent: runtime.GOMAXPROCS(0),
maxWindowSize: MaxWindowSize,
decodeBufsBelow: 128 << 10,
}
if o.concurrent > 4 {
o.concurrent = 4
}
o.maxDecodedSize = 64 << 30
}
// WithDecoderLowmem will set whether to use a lower amount of memory,
// but possibly have to allocate more while running.
func WithDecoderLowmem(b bool) DOption {
return func(o *decoderOptions) error { o.lowMem = b; return nil }
}
// WithDecoderConcurrency sets the number of created decoders.
// When decoding block with DecodeAll, this will limit the number
// of possible concurrently running decodes.
// When decoding streams, this will limit the number of
// inflight blocks.
// When decoding streams and setting maximum to 1,
// no async decoding will be done.
// When a value of 0 is provided GOMAXPROCS will be used.
// By default this will be set to 4 or GOMAXPROCS, whatever is lower.
func WithDecoderConcurrency(n int) DOption {
return func(o *decoderOptions) error {
if n < 0 {
return errors.New("concurrency must be at least 1")
}
if n == 0 {
o.concurrent = runtime.GOMAXPROCS(0)
} else {
o.concurrent = n
}
return nil
}
}
// WithDecoderMaxMemory allows to set a maximum decoded size for in-memory
// non-streaming operations or maximum window size for streaming operations.
// This can be used to control memory usage of potentially hostile content.
// Maximum is 1 << 63 bytes. Default is 64GiB.
func WithDecoderMaxMemory(n uint64) DOption {
return func(o *decoderOptions) error {
if n == 0 {
return errors.New("WithDecoderMaxMemory must be at least 1")
}
if n > 1<<63 {
return errors.New("WithDecoderMaxmemory must be less than 1 << 63")
}
o.maxDecodedSize = n
return nil
}
}
// WithDecoderDicts allows to register one or more dictionaries for the decoder.
//
// Each slice in dict must be in the [dictionary format] produced by
// "zstd --train" from the Zstandard reference implementation.
//
// If several dictionaries with the same ID are provided, the last one will be used.
//
// [dictionary format]: https://github.com/facebook/zstd/blob/dev/doc/zstd_compression_format.md#dictionary-format
func WithDecoderDicts(dicts ...[]byte) DOption {
return func(o *decoderOptions) error {
for _, b := range dicts {
d, err := loadDict(b)
if err != nil {
return err
}
o.dicts = append(o.dicts, d)
}
return nil
}
}
// WithDecoderDictRaw registers a dictionary that may be used by the decoder.
// The slice content can be arbitrary data.
func WithDecoderDictRaw(id uint32, content []byte) DOption {
return func(o *decoderOptions) error {
if bits.UintSize > 32 && uint(len(content)) > dictMaxLength {
return fmt.Errorf("dictionary of size %d > 2GiB too large", len(content))
}
o.dicts = append(o.dicts, &dict{id: id, content: content, offsets: [3]int{1, 4, 8}})
return nil
}
}
// WithDecoderMaxWindow allows to set a maximum window size for decodes.
// This allows rejecting packets that will cause big memory usage.
// The Decoder will likely allocate more memory based on the WithDecoderLowmem setting.
// If WithDecoderMaxMemory is set to a lower value, that will be used.
// Default is 512MB, Maximum is ~3.75 TB as per zstandard spec.
func WithDecoderMaxWindow(size uint64) DOption {
return func(o *decoderOptions) error {
if size < MinWindowSize {
return errors.New("WithMaxWindowSize must be at least 1KB, 1024 bytes")
}
if size > (1<<41)+7*(1<<38) {
return errors.New("WithMaxWindowSize must be less than (1<<41) + 7*(1<<38) ~ 3.75TB")
}
o.maxWindowSize = size
return nil
}
}
// WithDecodeAllCapLimit will limit DecodeAll to decoding cap(dst)-len(dst) bytes,
// or any size set in WithDecoderMaxMemory.
// This can be used to limit decoding to a specific maximum output size.
// Disabled by default.
func WithDecodeAllCapLimit(b bool) DOption {
return func(o *decoderOptions) error {
o.limitToCap = b
return nil
}
}
// WithDecodeBuffersBelow will fully decode readers that have a
// `Bytes() []byte` and `Len() int` interface similar to bytes.Buffer.
// This typically uses less allocations but will have the full decompressed object in memory.
// Note that DecodeAllCapLimit will disable this, as well as giving a size of 0 or less.
// Default is 128KiB.
func WithDecodeBuffersBelow(size int) DOption {
return func(o *decoderOptions) error {
o.decodeBufsBelow = size
return nil
}
}
// IgnoreChecksum allows to forcibly ignore checksum checking.
func IgnoreChecksum(b bool) DOption {
return func(o *decoderOptions) error {
o.ignoreChecksum = b
return nil
}
}

534
vendor/github.com/klauspost/compress/zstd/dict.go generated vendored Normal file
View File

@@ -0,0 +1,534 @@
package zstd
import (
"bytes"
"encoding/binary"
"errors"
"fmt"
"io"
"math"
"sort"
"github.com/klauspost/compress/huff0"
)
type dict struct {
id uint32
litEnc *huff0.Scratch
llDec, ofDec, mlDec sequenceDec
offsets [3]int
content []byte
}
const dictMagic = "\x37\xa4\x30\xec"
// Maximum dictionary size for the reference implementation (1.5.3) is 2 GiB.
const dictMaxLength = 1 << 31
// ID returns the dictionary id or 0 if d is nil.
func (d *dict) ID() uint32 {
if d == nil {
return 0
}
return d.id
}
// ContentSize returns the dictionary content size or 0 if d is nil.
func (d *dict) ContentSize() int {
if d == nil {
return 0
}
return len(d.content)
}
// Content returns the dictionary content.
func (d *dict) Content() []byte {
if d == nil {
return nil
}
return d.content
}
// Offsets returns the initial offsets.
func (d *dict) Offsets() [3]int {
if d == nil {
return [3]int{}
}
return d.offsets
}
// LitEncoder returns the literal encoder.
func (d *dict) LitEncoder() *huff0.Scratch {
if d == nil {
return nil
}
return d.litEnc
}
// Load a dictionary as described in
// https://github.com/facebook/zstd/blob/master/doc/zstd_compression_format.md#dictionary-format
func loadDict(b []byte) (*dict, error) {
// Check static field size.
if len(b) <= 8+(3*4) {
return nil, io.ErrUnexpectedEOF
}
d := dict{
llDec: sequenceDec{fse: &fseDecoder{}},
ofDec: sequenceDec{fse: &fseDecoder{}},
mlDec: sequenceDec{fse: &fseDecoder{}},
}
if string(b[:4]) != dictMagic {
return nil, ErrMagicMismatch
}
d.id = binary.LittleEndian.Uint32(b[4:8])
if d.id == 0 {
return nil, errors.New("dictionaries cannot have ID 0")
}
// Read literal table
var err error
d.litEnc, b, err = huff0.ReadTable(b[8:], nil)
if err != nil {
return nil, fmt.Errorf("loading literal table: %w", err)
}
d.litEnc.Reuse = huff0.ReusePolicyMust
br := byteReader{
b: b,
off: 0,
}
readDec := func(i tableIndex, dec *fseDecoder) error {
if err := dec.readNCount(&br, uint16(maxTableSymbol[i])); err != nil {
return err
}
if br.overread() {
return io.ErrUnexpectedEOF
}
err = dec.transform(symbolTableX[i])
if err != nil {
println("Transform table error:", err)
return err
}
if debugDecoder || debugEncoder {
println("Read table ok", "symbolLen:", dec.symbolLen)
}
// Set decoders as predefined so they aren't reused.
dec.preDefined = true
return nil
}
if err := readDec(tableOffsets, d.ofDec.fse); err != nil {
return nil, err
}
if err := readDec(tableMatchLengths, d.mlDec.fse); err != nil {
return nil, err
}
if err := readDec(tableLiteralLengths, d.llDec.fse); err != nil {
return nil, err
}
if br.remain() < 12 {
return nil, io.ErrUnexpectedEOF
}
d.offsets[0] = int(br.Uint32())
br.advance(4)
d.offsets[1] = int(br.Uint32())
br.advance(4)
d.offsets[2] = int(br.Uint32())
br.advance(4)
if d.offsets[0] <= 0 || d.offsets[1] <= 0 || d.offsets[2] <= 0 {
return nil, errors.New("invalid offset in dictionary")
}
d.content = make([]byte, br.remain())
copy(d.content, br.unread())
if d.offsets[0] > len(d.content) || d.offsets[1] > len(d.content) || d.offsets[2] > len(d.content) {
return nil, fmt.Errorf("initial offset bigger than dictionary content size %d, offsets: %v", len(d.content), d.offsets)
}
return &d, nil
}
// InspectDictionary loads a zstd dictionary and provides functions to inspect the content.
func InspectDictionary(b []byte) (interface {
ID() uint32
ContentSize() int
Content() []byte
Offsets() [3]int
LitEncoder() *huff0.Scratch
}, error) {
initPredefined()
d, err := loadDict(b)
return d, err
}
type BuildDictOptions struct {
// Dictionary ID.
ID uint32
// Content to use to create dictionary tables.
Contents [][]byte
// History to use for all blocks.
History []byte
// Offsets to use.
Offsets [3]int
// CompatV155 will make the dictionary compatible with Zstd v1.5.5 and earlier.
// See https://github.com/facebook/zstd/issues/3724
CompatV155 bool
// Use the specified encoder level.
// The dictionary will be built using the specified encoder level,
// which will reflect speed and make the dictionary tailored for that level.
// If not set SpeedBestCompression will be used.
Level EncoderLevel
// DebugOut will write stats and other details here if set.
DebugOut io.Writer
}
func BuildDict(o BuildDictOptions) ([]byte, error) {
initPredefined()
hist := o.History
contents := o.Contents
debug := o.DebugOut != nil
println := func(args ...interface{}) {
if o.DebugOut != nil {
fmt.Fprintln(o.DebugOut, args...)
}
}
printf := func(s string, args ...interface{}) {
if o.DebugOut != nil {
fmt.Fprintf(o.DebugOut, s, args...)
}
}
print := func(args ...interface{}) {
if o.DebugOut != nil {
fmt.Fprint(o.DebugOut, args...)
}
}
if int64(len(hist)) > dictMaxLength {
return nil, fmt.Errorf("dictionary of size %d > %d", len(hist), int64(dictMaxLength))
}
if len(hist) < 8 {
return nil, fmt.Errorf("dictionary of size %d < %d", len(hist), 8)
}
if len(contents) == 0 {
return nil, errors.New("no content provided")
}
d := dict{
id: o.ID,
litEnc: nil,
llDec: sequenceDec{},
ofDec: sequenceDec{},
mlDec: sequenceDec{},
offsets: o.Offsets,
content: hist,
}
block := blockEnc{lowMem: false}
block.init()
enc := encoder(&bestFastEncoder{fastBase: fastBase{maxMatchOff: int32(maxMatchLen), bufferReset: math.MaxInt32 - int32(maxMatchLen*2), lowMem: false}})
if o.Level != 0 {
eOpts := encoderOptions{
level: o.Level,
blockSize: maxMatchLen,
windowSize: maxMatchLen,
dict: &d,
lowMem: false,
}
enc = eOpts.encoder()
} else {
o.Level = SpeedBestCompression
}
var (
remain [256]int
ll [256]int
ml [256]int
of [256]int
)
addValues := func(dst *[256]int, src []byte) {
for _, v := range src {
dst[v]++
}
}
addHist := func(dst *[256]int, src *[256]uint32) {
for i, v := range src {
dst[i] += int(v)
}
}
seqs := 0
nUsed := 0
litTotal := 0
newOffsets := make(map[uint32]int, 1000)
for _, b := range contents {
block.reset(nil)
if len(b) < 8 {
continue
}
nUsed++
enc.Reset(&d, true)
enc.Encode(&block, b)
addValues(&remain, block.literals)
litTotal += len(block.literals)
seqs += len(block.sequences)
block.genCodes()
addHist(&ll, block.coders.llEnc.Histogram())
addHist(&ml, block.coders.mlEnc.Histogram())
addHist(&of, block.coders.ofEnc.Histogram())
for i, seq := range block.sequences {
if i > 3 {
break
}
offset := seq.offset
if offset == 0 {
continue
}
if offset > 3 {
newOffsets[offset-3]++
} else {
newOffsets[uint32(o.Offsets[offset-1])]++
}
}
}
// Find most used offsets.
var sortedOffsets []uint32
for k := range newOffsets {
sortedOffsets = append(sortedOffsets, k)
}
sort.Slice(sortedOffsets, func(i, j int) bool {
a, b := sortedOffsets[i], sortedOffsets[j]
if a == b {
// Prefer the longer offset
return sortedOffsets[i] > sortedOffsets[j]
}
return newOffsets[sortedOffsets[i]] > newOffsets[sortedOffsets[j]]
})
if len(sortedOffsets) > 3 {
if debug {
print("Offsets:")
for i, v := range sortedOffsets {
if i > 20 {
break
}
printf("[%d: %d],", v, newOffsets[v])
}
println("")
}
sortedOffsets = sortedOffsets[:3]
}
for i, v := range sortedOffsets {
o.Offsets[i] = int(v)
}
if debug {
println("New repeat offsets", o.Offsets)
}
if nUsed == 0 || seqs == 0 {
return nil, fmt.Errorf("%d blocks, %d sequences found", nUsed, seqs)
}
if debug {
println("Sequences:", seqs, "Blocks:", nUsed, "Literals:", litTotal)
}
if seqs/nUsed < 512 {
// Use 512 as minimum.
nUsed = seqs / 512
}
copyHist := func(dst *fseEncoder, src *[256]int) ([]byte, error) {
hist := dst.Histogram()
var maxSym uint8
var maxCount int
var fakeLength int
for i, v := range src {
if v > 0 {
v = v / nUsed
if v == 0 {
v = 1
}
}
if v > maxCount {
maxCount = v
}
if v != 0 {
maxSym = uint8(i)
}
fakeLength += v
hist[i] = uint32(v)
}
dst.HistogramFinished(maxSym, maxCount)
dst.reUsed = false
dst.useRLE = false
err := dst.normalizeCount(fakeLength)
if err != nil {
return nil, err
}
if debug {
println("RAW:", dst.count[:maxSym+1], "NORM:", dst.norm[:maxSym+1], "LEN:", fakeLength)
}
return dst.writeCount(nil)
}
if debug {
print("Literal lengths: ")
}
llTable, err := copyHist(block.coders.llEnc, &ll)
if err != nil {
return nil, err
}
if debug {
print("Match lengths: ")
}
mlTable, err := copyHist(block.coders.mlEnc, &ml)
if err != nil {
return nil, err
}
if debug {
print("Offsets: ")
}
ofTable, err := copyHist(block.coders.ofEnc, &of)
if err != nil {
return nil, err
}
// Literal table
avgSize := litTotal
if avgSize > huff0.BlockSizeMax/2 {
avgSize = huff0.BlockSizeMax / 2
}
huffBuff := make([]byte, 0, avgSize)
// Target size
div := litTotal / avgSize
if div < 1 {
div = 1
}
if debug {
println("Huffman weights:")
}
for i, n := range remain[:] {
if n > 0 {
n = n / div
// Allow all entries to be represented.
if n == 0 {
n = 1
}
huffBuff = append(huffBuff, bytes.Repeat([]byte{byte(i)}, n)...)
if debug {
printf("[%d: %d], ", i, n)
}
}
}
if o.CompatV155 && remain[255]/div == 0 {
huffBuff = append(huffBuff, 255)
}
scratch := &huff0.Scratch{TableLog: 11}
for tries := 0; tries < 255; tries++ {
scratch = &huff0.Scratch{TableLog: 11}
_, _, err = huff0.Compress1X(huffBuff, scratch)
if err == nil {
break
}
if debug {
printf("Try %d: Huffman error: %v\n", tries+1, err)
}
huffBuff = huffBuff[:0]
if tries == 250 {
if debug {
println("Huffman: Bailing out with predefined table")
}
// Bail out.... Just generate something
huffBuff = append(huffBuff, bytes.Repeat([]byte{255}, 10000)...)
for i := 0; i < 128; i++ {
huffBuff = append(huffBuff, byte(i))
}
continue
}
if errors.Is(err, huff0.ErrIncompressible) {
// Try truncating least common.
for i, n := range remain[:] {
if n > 0 {
n = n / (div * (i + 1))
if n > 0 {
huffBuff = append(huffBuff, bytes.Repeat([]byte{byte(i)}, n)...)
}
}
}
if o.CompatV155 && len(huffBuff) > 0 && huffBuff[len(huffBuff)-1] != 255 {
huffBuff = append(huffBuff, 255)
}
if len(huffBuff) == 0 {
huffBuff = append(huffBuff, 0, 255)
}
}
if errors.Is(err, huff0.ErrUseRLE) {
for i, n := range remain[:] {
n = n / (div * (i + 1))
// Allow all entries to be represented.
if n == 0 {
n = 1
}
huffBuff = append(huffBuff, bytes.Repeat([]byte{byte(i)}, n)...)
}
}
}
var out bytes.Buffer
out.Write([]byte(dictMagic))
out.Write(binary.LittleEndian.AppendUint32(nil, o.ID))
out.Write(scratch.OutTable)
if debug {
println("huff table:", len(scratch.OutTable), "bytes")
println("of table:", len(ofTable), "bytes")
println("ml table:", len(mlTable), "bytes")
println("ll table:", len(llTable), "bytes")
}
out.Write(ofTable)
out.Write(mlTable)
out.Write(llTable)
out.Write(binary.LittleEndian.AppendUint32(nil, uint32(o.Offsets[0])))
out.Write(binary.LittleEndian.AppendUint32(nil, uint32(o.Offsets[1])))
out.Write(binary.LittleEndian.AppendUint32(nil, uint32(o.Offsets[2])))
out.Write(hist)
if debug {
_, err := loadDict(out.Bytes())
if err != nil {
panic(err)
}
i, err := InspectDictionary(out.Bytes())
if err != nil {
panic(err)
}
println("ID:", i.ID())
println("Content size:", i.ContentSize())
println("Encoder:", i.LitEncoder() != nil)
println("Offsets:", i.Offsets())
var totalSize int
for _, b := range contents {
totalSize += len(b)
}
encWith := func(opts ...EOption) int {
enc, err := NewWriter(nil, opts...)
if err != nil {
panic(err)
}
defer enc.Close()
var dst []byte
var totalSize int
for _, b := range contents {
dst = enc.EncodeAll(b, dst[:0])
totalSize += len(dst)
}
return totalSize
}
plain := encWith(WithEncoderLevel(o.Level))
withDict := encWith(WithEncoderLevel(o.Level), WithEncoderDict(out.Bytes()))
println("Input size:", totalSize)
println("Plain Compressed:", plain)
println("Dict Compressed:", withDict)
println("Saved:", plain-withDict, (plain-withDict)/len(contents), "bytes per input (rounded down)")
}
return out.Bytes(), nil
}

173
vendor/github.com/klauspost/compress/zstd/enc_base.go generated vendored Normal file
View File

@@ -0,0 +1,173 @@
package zstd
import (
"fmt"
"math/bits"
"github.com/klauspost/compress/zstd/internal/xxhash"
)
const (
dictShardBits = 6
)
type fastBase struct {
// cur is the offset at the start of hist
cur int32
// maximum offset. Should be at least 2x block size.
maxMatchOff int32
bufferReset int32
hist []byte
crc *xxhash.Digest
tmp [8]byte
blk *blockEnc
lastDictID uint32
lowMem bool
}
// CRC returns the underlying CRC writer.
func (e *fastBase) CRC() *xxhash.Digest {
return e.crc
}
// AppendCRC will append the CRC to the destination slice and return it.
func (e *fastBase) AppendCRC(dst []byte) []byte {
crc := e.crc.Sum(e.tmp[:0])
dst = append(dst, crc[7], crc[6], crc[5], crc[4])
return dst
}
// WindowSize returns the window size of the encoder,
// or a window size small enough to contain the input size, if > 0.
func (e *fastBase) WindowSize(size int64) int32 {
if size > 0 && size < int64(e.maxMatchOff) {
b := int32(1) << uint(bits.Len(uint(size)))
// Keep minimum window.
if b < 1024 {
b = 1024
}
return b
}
return e.maxMatchOff
}
// Block returns the current block.
func (e *fastBase) Block() *blockEnc {
return e.blk
}
func (e *fastBase) addBlock(src []byte) int32 {
if debugAsserts && e.cur > e.bufferReset {
panic(fmt.Sprintf("ecur (%d) > buffer reset (%d)", e.cur, e.bufferReset))
}
// check if we have space already
if len(e.hist)+len(src) > cap(e.hist) {
if cap(e.hist) == 0 {
e.ensureHist(len(src))
} else {
if cap(e.hist) < int(e.maxMatchOff+maxCompressedBlockSize) {
panic(fmt.Errorf("unexpected buffer cap %d, want at least %d with window %d", cap(e.hist), e.maxMatchOff+maxCompressedBlockSize, e.maxMatchOff))
}
// Move down
offset := int32(len(e.hist)) - e.maxMatchOff
copy(e.hist[0:e.maxMatchOff], e.hist[offset:])
e.cur += offset
e.hist = e.hist[:e.maxMatchOff]
}
}
s := int32(len(e.hist))
e.hist = append(e.hist, src...)
return s
}
// ensureHist will ensure that history can keep at least this many bytes.
func (e *fastBase) ensureHist(n int) {
if cap(e.hist) >= n {
return
}
l := e.maxMatchOff
if (e.lowMem && e.maxMatchOff > maxCompressedBlockSize) || e.maxMatchOff <= maxCompressedBlockSize {
l += maxCompressedBlockSize
} else {
l += e.maxMatchOff
}
// Make it at least 1MB.
if l < 1<<20 && !e.lowMem {
l = 1 << 20
}
// Make it at least the requested size.
if l < int32(n) {
l = int32(n)
}
e.hist = make([]byte, 0, l)
}
// useBlock will replace the block with the provided one,
// but transfer recent offsets from the previous.
func (e *fastBase) UseBlock(enc *blockEnc) {
enc.reset(e.blk)
e.blk = enc
}
func (e *fastBase) matchlen(s, t int32, src []byte) int32 {
if debugAsserts {
if s < 0 {
err := fmt.Sprintf("s (%d) < 0", s)
panic(err)
}
if t < 0 {
err := fmt.Sprintf("s (%d) < 0", s)
panic(err)
}
if s-t > e.maxMatchOff {
err := fmt.Sprintf("s (%d) - t (%d) > maxMatchOff (%d)", s, t, e.maxMatchOff)
panic(err)
}
if len(src)-int(s) > maxCompressedBlockSize {
panic(fmt.Sprintf("len(src)-s (%d) > maxCompressedBlockSize (%d)", len(src)-int(s), maxCompressedBlockSize))
}
}
return int32(matchLen(src[s:], src[t:]))
}
// Reset the encoding table.
func (e *fastBase) resetBase(d *dict, singleBlock bool) {
if e.blk == nil {
e.blk = &blockEnc{lowMem: e.lowMem}
e.blk.init()
} else {
e.blk.reset(nil)
}
e.blk.initNewEncode()
if e.crc == nil {
e.crc = xxhash.New()
} else {
e.crc.Reset()
}
e.blk.dictLitEnc = nil
if d != nil {
low := e.lowMem
if singleBlock {
e.lowMem = true
}
e.ensureHist(d.ContentSize() + maxCompressedBlockSize)
e.lowMem = low
}
// We offset current position so everything will be out of reach.
// If above reset line, history will be purged.
if e.cur < e.bufferReset {
e.cur += e.maxMatchOff + int32(len(e.hist))
}
e.hist = e.hist[:0]
if d != nil {
// Set offsets (currently not used)
for i, off := range d.offsets {
e.blk.recentOffsets[i] = uint32(off)
e.blk.prevRecentOffsets[i] = e.blk.recentOffsets[i]
}
// Transfer litenc.
e.blk.dictLitEnc = d.litEnc
e.hist = append(e.hist, d.content...)
}
}

560
vendor/github.com/klauspost/compress/zstd/enc_best.go generated vendored Normal file
View File

@@ -0,0 +1,560 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"bytes"
"fmt"
"github.com/klauspost/compress"
)
const (
bestLongTableBits = 22 // Bits used in the long match table
bestLongTableSize = 1 << bestLongTableBits // Size of the table
bestLongLen = 8 // Bytes used for table hash
// Note: Increasing the short table bits or making the hash shorter
// can actually lead to compression degradation since it will 'steal' more from the
// long match table and match offsets are quite big.
// This greatly depends on the type of input.
bestShortTableBits = 18 // Bits used in the short match table
bestShortTableSize = 1 << bestShortTableBits // Size of the table
bestShortLen = 4 // Bytes used for table hash
)
type match struct {
offset int32
s int32
length int32
rep int32
est int32
}
const highScore = maxMatchLen * 8
// estBits will estimate output bits from predefined tables.
func (m *match) estBits(bitsPerByte int32) {
mlc := mlCode(uint32(m.length - zstdMinMatch))
var ofc uint8
if m.rep < 0 {
ofc = ofCode(uint32(m.s-m.offset) + 3)
} else {
ofc = ofCode(uint32(m.rep) & 3)
}
// Cost, excluding
ofTT, mlTT := fsePredefEnc[tableOffsets].ct.symbolTT[ofc], fsePredefEnc[tableMatchLengths].ct.symbolTT[mlc]
// Add cost of match encoding...
m.est = int32(ofTT.outBits + mlTT.outBits)
m.est += int32(ofTT.deltaNbBits>>16 + mlTT.deltaNbBits>>16)
// Subtract savings compared to literal encoding...
m.est -= (m.length * bitsPerByte) >> 10
if m.est > 0 {
// Unlikely gain..
m.length = 0
m.est = highScore
}
}
// bestFastEncoder uses 2 tables, one for short matches (5 bytes) and one for long matches.
// The long match table contains the previous entry with the same hash,
// effectively making it a "chain" of length 2.
// When we find a long match we choose between the two values and select the longest.
// When we find a short match, after checking the long, we check if we can find a long at n+1
// and that it is longer (lazy matching).
type bestFastEncoder struct {
fastBase
table [bestShortTableSize]prevEntry
longTable [bestLongTableSize]prevEntry
dictTable []prevEntry
dictLongTable []prevEntry
}
// Encode improves compression...
func (e *bestFastEncoder) Encode(blk *blockEnc, src []byte) {
const (
// Input margin is the number of bytes we read (8)
// and the maximum we will read ahead (2)
inputMargin = 8 + 4
minNonLiteralBlockSize = 16
)
// Protect against e.cur wraparound.
for e.cur >= e.bufferReset-int32(len(e.hist)) {
if len(e.hist) == 0 {
e.table = [bestShortTableSize]prevEntry{}
e.longTable = [bestLongTableSize]prevEntry{}
e.cur = e.maxMatchOff
break
}
// Shift down everything in the table that isn't already too far away.
minOff := e.cur + int32(len(e.hist)) - e.maxMatchOff
for i := range e.table[:] {
v := e.table[i].offset
v2 := e.table[i].prev
if v < minOff {
v = 0
v2 = 0
} else {
v = v - e.cur + e.maxMatchOff
if v2 < minOff {
v2 = 0
} else {
v2 = v2 - e.cur + e.maxMatchOff
}
}
e.table[i] = prevEntry{
offset: v,
prev: v2,
}
}
for i := range e.longTable[:] {
v := e.longTable[i].offset
v2 := e.longTable[i].prev
if v < minOff {
v = 0
v2 = 0
} else {
v = v - e.cur + e.maxMatchOff
if v2 < minOff {
v2 = 0
} else {
v2 = v2 - e.cur + e.maxMatchOff
}
}
e.longTable[i] = prevEntry{
offset: v,
prev: v2,
}
}
e.cur = e.maxMatchOff
break
}
// Add block to history
s := e.addBlock(src)
blk.size = len(src)
// Check RLE first
if len(src) > zstdMinMatch {
ml := matchLen(src[1:], src)
if ml == len(src)-1 {
blk.literals = append(blk.literals, src[0])
blk.sequences = append(blk.sequences, seq{litLen: 1, matchLen: uint32(len(src)-1) - zstdMinMatch, offset: 1 + 3})
return
}
}
if len(src) < minNonLiteralBlockSize {
blk.extraLits = len(src)
blk.literals = blk.literals[:len(src)]
copy(blk.literals, src)
return
}
// Use this to estimate literal cost.
// Scaled by 10 bits.
bitsPerByte := int32((compress.ShannonEntropyBits(src) * 1024) / len(src))
// Huffman can never go < 1 bit/byte
if bitsPerByte < 1024 {
bitsPerByte = 1024
}
// Override src
src = e.hist
sLimit := int32(len(src)) - inputMargin
const kSearchStrength = 10
// nextEmit is where in src the next emitLiteral should start from.
nextEmit := s
// Relative offsets
offset1 := int32(blk.recentOffsets[0])
offset2 := int32(blk.recentOffsets[1])
offset3 := int32(blk.recentOffsets[2])
addLiterals := func(s *seq, until int32) {
if until == nextEmit {
return
}
blk.literals = append(blk.literals, src[nextEmit:until]...)
s.litLen = uint32(until - nextEmit)
}
if debugEncoder {
println("recent offsets:", blk.recentOffsets)
}
encodeLoop:
for {
// We allow the encoder to optionally turn off repeat offsets across blocks
canRepeat := len(blk.sequences) > 2
if debugAsserts && canRepeat && offset1 == 0 {
panic("offset0 was 0")
}
const goodEnough = 250
cv := load6432(src, s)
nextHashL := hashLen(cv, bestLongTableBits, bestLongLen)
nextHashS := hashLen(cv, bestShortTableBits, bestShortLen)
candidateL := e.longTable[nextHashL]
candidateS := e.table[nextHashS]
// Set m to a match at offset if it looks like that will improve compression.
improve := func(m *match, offset int32, s int32, first uint32, rep int32) {
delta := s - offset
if delta >= e.maxMatchOff || delta <= 0 || load3232(src, offset) != first {
return
}
// Try to quick reject if we already have a long match.
if m.length > 16 {
left := len(src) - int(m.s+m.length)
// If we are too close to the end, keep as is.
if left <= 0 {
return
}
checkLen := m.length - (s - m.s) - 8
if left > 2 && checkLen > 4 {
// Check 4 bytes, 4 bytes from the end of the current match.
a := load3232(src, offset+checkLen)
b := load3232(src, s+checkLen)
if a != b {
return
}
}
}
l := 4 + e.matchlen(s+4, offset+4, src)
if m.rep <= 0 {
// Extend candidate match backwards as far as possible.
// Do not extend repeats as we can assume they are optimal
// and offsets change if s == nextEmit.
tMin := s - e.maxMatchOff
if tMin < 0 {
tMin = 0
}
for offset > tMin && s > nextEmit && src[offset-1] == src[s-1] && l < maxMatchLength {
s--
offset--
l++
}
}
if debugAsserts {
if offset >= s {
panic(fmt.Sprintf("offset: %d - s:%d - rep: %d - cur :%d - max: %d", offset, s, rep, e.cur, e.maxMatchOff))
}
if !bytes.Equal(src[s:s+l], src[offset:offset+l]) {
panic(fmt.Sprintf("second match mismatch: %v != %v, first: %08x", src[s:s+4], src[offset:offset+4], first))
}
}
cand := match{offset: offset, s: s, length: l, rep: rep}
cand.estBits(bitsPerByte)
if m.est >= highScore || cand.est-m.est+(cand.s-m.s)*bitsPerByte>>10 < 0 {
*m = cand
}
}
best := match{s: s, est: highScore}
improve(&best, candidateL.offset-e.cur, s, uint32(cv), -1)
improve(&best, candidateL.prev-e.cur, s, uint32(cv), -1)
improve(&best, candidateS.offset-e.cur, s, uint32(cv), -1)
improve(&best, candidateS.prev-e.cur, s, uint32(cv), -1)
if canRepeat && best.length < goodEnough {
if s == nextEmit {
// Check repeats straight after a match.
improve(&best, s-offset2, s, uint32(cv), 1|4)
improve(&best, s-offset3, s, uint32(cv), 2|4)
if offset1 > 1 {
improve(&best, s-(offset1-1), s, uint32(cv), 3|4)
}
}
// If either no match or a non-repeat match, check at + 1
if best.rep <= 0 {
cv32 := uint32(cv >> 8)
spp := s + 1
improve(&best, spp-offset1, spp, cv32, 1)
improve(&best, spp-offset2, spp, cv32, 2)
improve(&best, spp-offset3, spp, cv32, 3)
if best.rep < 0 {
cv32 = uint32(cv >> 24)
spp += 2
improve(&best, spp-offset1, spp, cv32, 1)
improve(&best, spp-offset2, spp, cv32, 2)
improve(&best, spp-offset3, spp, cv32, 3)
}
}
}
// Load next and check...
e.longTable[nextHashL] = prevEntry{offset: s + e.cur, prev: candidateL.offset}
e.table[nextHashS] = prevEntry{offset: s + e.cur, prev: candidateS.offset}
index0 := s + 1
// Look far ahead, unless we have a really long match already...
if best.length < goodEnough {
// No match found, move forward on input, no need to check forward...
if best.length < 4 {
s += 1 + (s-nextEmit)>>(kSearchStrength-1)
if s >= sLimit {
break encodeLoop
}
continue
}
candidateS = e.table[hashLen(cv>>8, bestShortTableBits, bestShortLen)]
cv = load6432(src, s+1)
cv2 := load6432(src, s+2)
candidateL = e.longTable[hashLen(cv, bestLongTableBits, bestLongLen)]
candidateL2 := e.longTable[hashLen(cv2, bestLongTableBits, bestLongLen)]
// Short at s+1
improve(&best, candidateS.offset-e.cur, s+1, uint32(cv), -1)
// Long at s+1, s+2
improve(&best, candidateL.offset-e.cur, s+1, uint32(cv), -1)
improve(&best, candidateL.prev-e.cur, s+1, uint32(cv), -1)
improve(&best, candidateL2.offset-e.cur, s+2, uint32(cv2), -1)
improve(&best, candidateL2.prev-e.cur, s+2, uint32(cv2), -1)
if false {
// Short at s+3.
// Too often worse...
improve(&best, e.table[hashLen(cv2>>8, bestShortTableBits, bestShortLen)].offset-e.cur, s+3, uint32(cv2>>8), -1)
}
// Start check at a fixed offset to allow for a few mismatches.
// For this compression level 2 yields the best results.
// We cannot do this if we have already indexed this position.
const skipBeginning = 2
if best.s > s-skipBeginning {
// See if we can find a better match by checking where the current best ends.
// Use that offset to see if we can find a better full match.
if sAt := best.s + best.length; sAt < sLimit {
nextHashL := hashLen(load6432(src, sAt), bestLongTableBits, bestLongLen)
candidateEnd := e.longTable[nextHashL]
if off := candidateEnd.offset - e.cur - best.length + skipBeginning; off >= 0 {
improve(&best, off, best.s+skipBeginning, load3232(src, best.s+skipBeginning), -1)
if off := candidateEnd.prev - e.cur - best.length + skipBeginning; off >= 0 {
improve(&best, off, best.s+skipBeginning, load3232(src, best.s+skipBeginning), -1)
}
}
}
}
}
if debugAsserts {
if best.offset >= best.s {
panic(fmt.Sprintf("best.offset > s: %d >= %d", best.offset, best.s))
}
if best.s < nextEmit {
panic(fmt.Sprintf("s %d < nextEmit %d", best.s, nextEmit))
}
if best.offset < s-e.maxMatchOff {
panic(fmt.Sprintf("best.offset < s-e.maxMatchOff: %d < %d", best.offset, s-e.maxMatchOff))
}
if !bytes.Equal(src[best.s:best.s+best.length], src[best.offset:best.offset+best.length]) {
panic(fmt.Sprintf("match mismatch: %v != %v", src[best.s:best.s+best.length], src[best.offset:best.offset+best.length]))
}
}
// We have a match, we can store the forward value
s = best.s
if best.rep > 0 {
var seq seq
seq.matchLen = uint32(best.length - zstdMinMatch)
addLiterals(&seq, best.s)
// Repeat. If bit 4 is set, this is a non-lit repeat.
seq.offset = uint32(best.rep & 3)
if debugSequences {
println("repeat sequence", seq, "next s:", best.s, "off:", best.s-best.offset)
}
blk.sequences = append(blk.sequences, seq)
// Index old s + 1 -> s - 1
s = best.s + best.length
nextEmit = s
// Index skipped...
end := s
if s > sLimit+4 {
end = sLimit + 4
}
off := index0 + e.cur
for index0 < end {
cv0 := load6432(src, index0)
h0 := hashLen(cv0, bestLongTableBits, bestLongLen)
h1 := hashLen(cv0, bestShortTableBits, bestShortLen)
e.longTable[h0] = prevEntry{offset: off, prev: e.longTable[h0].offset}
e.table[h1] = prevEntry{offset: off, prev: e.table[h1].offset}
off++
index0++
}
switch best.rep {
case 2, 4 | 1:
offset1, offset2 = offset2, offset1
case 3, 4 | 2:
offset1, offset2, offset3 = offset3, offset1, offset2
case 4 | 3:
offset1, offset2, offset3 = offset1-1, offset1, offset2
}
if s >= sLimit {
if debugEncoder {
println("repeat ended", s, best.length)
}
break encodeLoop
}
continue
}
// A 4-byte match has been found. Update recent offsets.
// We'll later see if more than 4 bytes.
t := best.offset
offset1, offset2, offset3 = s-t, offset1, offset2
if debugAsserts && s <= t {
panic(fmt.Sprintf("s (%d) <= t (%d)", s, t))
}
if debugAsserts && int(offset1) > len(src) {
panic("invalid offset")
}
// Write our sequence
var seq seq
l := best.length
seq.litLen = uint32(s - nextEmit)
seq.matchLen = uint32(l - zstdMinMatch)
if seq.litLen > 0 {
blk.literals = append(blk.literals, src[nextEmit:s]...)
}
seq.offset = uint32(s-t) + 3
s += l
if debugSequences {
println("sequence", seq, "next s:", s)
}
blk.sequences = append(blk.sequences, seq)
nextEmit = s
// Index old s + 1 -> s - 1 or sLimit
end := s
if s > sLimit-4 {
end = sLimit - 4
}
off := index0 + e.cur
for index0 < end {
cv0 := load6432(src, index0)
h0 := hashLen(cv0, bestLongTableBits, bestLongLen)
h1 := hashLen(cv0, bestShortTableBits, bestShortLen)
e.longTable[h0] = prevEntry{offset: off, prev: e.longTable[h0].offset}
e.table[h1] = prevEntry{offset: off, prev: e.table[h1].offset}
index0++
off++
}
if s >= sLimit {
break encodeLoop
}
}
if int(nextEmit) < len(src) {
blk.literals = append(blk.literals, src[nextEmit:]...)
blk.extraLits = len(src) - int(nextEmit)
}
blk.recentOffsets[0] = uint32(offset1)
blk.recentOffsets[1] = uint32(offset2)
blk.recentOffsets[2] = uint32(offset3)
if debugEncoder {
println("returning, recent offsets:", blk.recentOffsets, "extra literals:", blk.extraLits)
}
}
// EncodeNoHist will encode a block with no history and no following blocks.
// Most notable difference is that src will not be copied for history and
// we do not need to check for max match length.
func (e *bestFastEncoder) EncodeNoHist(blk *blockEnc, src []byte) {
e.ensureHist(len(src))
e.Encode(blk, src)
}
// Reset will reset and set a dictionary if not nil
func (e *bestFastEncoder) Reset(d *dict, singleBlock bool) {
e.resetBase(d, singleBlock)
if d == nil {
return
}
// Init or copy dict table
if len(e.dictTable) != len(e.table) || d.id != e.lastDictID {
if len(e.dictTable) != len(e.table) {
e.dictTable = make([]prevEntry, len(e.table))
}
end := int32(len(d.content)) - 8 + e.maxMatchOff
for i := e.maxMatchOff; i < end; i += 4 {
const hashLog = bestShortTableBits
cv := load6432(d.content, i-e.maxMatchOff)
nextHash := hashLen(cv, hashLog, bestShortLen) // 0 -> 4
nextHash1 := hashLen(cv>>8, hashLog, bestShortLen) // 1 -> 5
nextHash2 := hashLen(cv>>16, hashLog, bestShortLen) // 2 -> 6
nextHash3 := hashLen(cv>>24, hashLog, bestShortLen) // 3 -> 7
e.dictTable[nextHash] = prevEntry{
prev: e.dictTable[nextHash].offset,
offset: i,
}
e.dictTable[nextHash1] = prevEntry{
prev: e.dictTable[nextHash1].offset,
offset: i + 1,
}
e.dictTable[nextHash2] = prevEntry{
prev: e.dictTable[nextHash2].offset,
offset: i + 2,
}
e.dictTable[nextHash3] = prevEntry{
prev: e.dictTable[nextHash3].offset,
offset: i + 3,
}
}
e.lastDictID = d.id
}
// Init or copy dict table
if len(e.dictLongTable) != len(e.longTable) || d.id != e.lastDictID {
if len(e.dictLongTable) != len(e.longTable) {
e.dictLongTable = make([]prevEntry, len(e.longTable))
}
if len(d.content) >= 8 {
cv := load6432(d.content, 0)
h := hashLen(cv, bestLongTableBits, bestLongLen)
e.dictLongTable[h] = prevEntry{
offset: e.maxMatchOff,
prev: e.dictLongTable[h].offset,
}
end := int32(len(d.content)) - 8 + e.maxMatchOff
off := 8 // First to read
for i := e.maxMatchOff + 1; i < end; i++ {
cv = cv>>8 | (uint64(d.content[off]) << 56)
h := hashLen(cv, bestLongTableBits, bestLongLen)
e.dictLongTable[h] = prevEntry{
offset: i,
prev: e.dictLongTable[h].offset,
}
off++
}
}
e.lastDictID = d.id
}
// Reset table to initial state
copy(e.longTable[:], e.dictLongTable)
e.cur = e.maxMatchOff
// Reset table to initial state
copy(e.table[:], e.dictTable)
}

1252
vendor/github.com/klauspost/compress/zstd/enc_better.go generated vendored Normal file
View File

File diff suppressed because it is too large Load Diff

1123
vendor/github.com/klauspost/compress/zstd/enc_dfast.go generated vendored Normal file
View File

File diff suppressed because it is too large Load Diff

891
vendor/github.com/klauspost/compress/zstd/enc_fast.go generated vendored Normal file
View File

@@ -0,0 +1,891 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"fmt"
)
const (
tableBits = 15 // Bits used in the table
tableSize = 1 << tableBits // Size of the table
tableShardCnt = 1 << (tableBits - dictShardBits) // Number of shards in the table
tableShardSize = tableSize / tableShardCnt // Size of an individual shard
tableFastHashLen = 6
tableMask = tableSize - 1 // Mask for table indices. Redundant, but can eliminate bounds checks.
maxMatchLength = 131074
)
type tableEntry struct {
val uint32
offset int32
}
type fastEncoder struct {
fastBase
table [tableSize]tableEntry
}
type fastEncoderDict struct {
fastEncoder
dictTable []tableEntry
tableShardDirty [tableShardCnt]bool
allDirty bool
}
// Encode mimmics functionality in zstd_fast.c
func (e *fastEncoder) Encode(blk *blockEnc, src []byte) {
const (
inputMargin = 8
minNonLiteralBlockSize = 1 + 1 + inputMargin
)
// Protect against e.cur wraparound.
for e.cur >= e.bufferReset-int32(len(e.hist)) {
if len(e.hist) == 0 {
for i := range e.table[:] {
e.table[i] = tableEntry{}
}
e.cur = e.maxMatchOff
break
}
// Shift down everything in the table that isn't already too far away.
minOff := e.cur + int32(len(e.hist)) - e.maxMatchOff
for i := range e.table[:] {
v := e.table[i].offset
if v < minOff {
v = 0
} else {
v = v - e.cur + e.maxMatchOff
}
e.table[i].offset = v
}
e.cur = e.maxMatchOff
break
}
s := e.addBlock(src)
blk.size = len(src)
if len(src) < minNonLiteralBlockSize {
blk.extraLits = len(src)
blk.literals = blk.literals[:len(src)]
copy(blk.literals, src)
return
}
// Override src
src = e.hist
sLimit := int32(len(src)) - inputMargin
// stepSize is the number of bytes to skip on every main loop iteration.
// It should be >= 2.
const stepSize = 2
// TEMPLATE
const hashLog = tableBits
// seems global, but would be nice to tweak.
const kSearchStrength = 6
// nextEmit is where in src the next emitLiteral should start from.
nextEmit := s
cv := load6432(src, s)
// Relative offsets
offset1 := int32(blk.recentOffsets[0])
offset2 := int32(blk.recentOffsets[1])
addLiterals := func(s *seq, until int32) {
if until == nextEmit {
return
}
blk.literals = append(blk.literals, src[nextEmit:until]...)
s.litLen = uint32(until - nextEmit)
}
if debugEncoder {
println("recent offsets:", blk.recentOffsets)
}
encodeLoop:
for {
// t will contain the match offset when we find one.
// When existing the search loop, we have already checked 4 bytes.
var t int32
// We will not use repeat offsets across blocks.
// By not using them for the first 3 matches
canRepeat := len(blk.sequences) > 2
for {
if debugAsserts && canRepeat && offset1 == 0 {
panic("offset0 was 0")
}
nextHash := hashLen(cv, hashLog, tableFastHashLen)
nextHash2 := hashLen(cv>>8, hashLog, tableFastHashLen)
candidate := e.table[nextHash]
candidate2 := e.table[nextHash2]
repIndex := s - offset1 + 2
e.table[nextHash] = tableEntry{offset: s + e.cur, val: uint32(cv)}
e.table[nextHash2] = tableEntry{offset: s + e.cur + 1, val: uint32(cv >> 8)}
if canRepeat && repIndex >= 0 && load3232(src, repIndex) == uint32(cv>>16) {
// Consider history as well.
var seq seq
length := 4 + e.matchlen(s+6, repIndex+4, src)
seq.matchLen = uint32(length - zstdMinMatch)
// We might be able to match backwards.
// Extend as long as we can.
start := s + 2
// We end the search early, so we don't risk 0 literals
// and have to do special offset treatment.
startLimit := nextEmit + 1
sMin := s - e.maxMatchOff
if sMin < 0 {
sMin = 0
}
for repIndex > sMin && start > startLimit && src[repIndex-1] == src[start-1] && seq.matchLen < maxMatchLength-zstdMinMatch {
repIndex--
start--
seq.matchLen++
}
addLiterals(&seq, start)
// rep 0
seq.offset = 1
if debugSequences {
println("repeat sequence", seq, "next s:", s)
}
blk.sequences = append(blk.sequences, seq)
s += length + 2
nextEmit = s
if s >= sLimit {
if debugEncoder {
println("repeat ended", s, length)
}
break encodeLoop
}
cv = load6432(src, s)
continue
}
coffset0 := s - (candidate.offset - e.cur)
coffset1 := s - (candidate2.offset - e.cur) + 1
if coffset0 < e.maxMatchOff && uint32(cv) == candidate.val {
// found a regular match
t = candidate.offset - e.cur
if debugAsserts && s <= t {
panic(fmt.Sprintf("s (%d) <= t (%d)", s, t))
}
if debugAsserts && s-t > e.maxMatchOff {
panic("s - t >e.maxMatchOff")
}
break
}
if coffset1 < e.maxMatchOff && uint32(cv>>8) == candidate2.val {
// found a regular match
t = candidate2.offset - e.cur
s++
if debugAsserts && s <= t {
panic(fmt.Sprintf("s (%d) <= t (%d)", s, t))
}
if debugAsserts && s-t > e.maxMatchOff {
panic("s - t >e.maxMatchOff")
}
if debugAsserts && t < 0 {
panic("t<0")
}
break
}
s += stepSize + ((s - nextEmit) >> (kSearchStrength - 1))
if s >= sLimit {
break encodeLoop
}
cv = load6432(src, s)
}
// A 4-byte match has been found. We'll later see if more than 4 bytes.
offset2 = offset1
offset1 = s - t
if debugAsserts && s <= t {
panic(fmt.Sprintf("s (%d) <= t (%d)", s, t))
}
if debugAsserts && canRepeat && int(offset1) > len(src) {
panic("invalid offset")
}
// Extend the 4-byte match as long as possible.
l := e.matchlen(s+4, t+4, src) + 4
// Extend backwards
tMin := s - e.maxMatchOff
if tMin < 0 {
tMin = 0
}
for t > tMin && s > nextEmit && src[t-1] == src[s-1] && l < maxMatchLength {
s--
t--
l++
}
// Write our sequence.
var seq seq
seq.litLen = uint32(s - nextEmit)
seq.matchLen = uint32(l - zstdMinMatch)
if seq.litLen > 0 {
blk.literals = append(blk.literals, src[nextEmit:s]...)
}
// Don't use repeat offsets
seq.offset = uint32(s-t) + 3
s += l
if debugSequences {
println("sequence", seq, "next s:", s)
}
blk.sequences = append(blk.sequences, seq)
nextEmit = s
if s >= sLimit {
break encodeLoop
}
cv = load6432(src, s)
// Check offset 2
if o2 := s - offset2; canRepeat && load3232(src, o2) == uint32(cv) {
// We have at least 4 byte match.
// No need to check backwards. We come straight from a match
l := 4 + e.matchlen(s+4, o2+4, src)
// Store this, since we have it.
nextHash := hashLen(cv, hashLog, tableFastHashLen)
e.table[nextHash] = tableEntry{offset: s + e.cur, val: uint32(cv)}
seq.matchLen = uint32(l) - zstdMinMatch
seq.litLen = 0
// Since litlen is always 0, this is offset 1.
seq.offset = 1
s += l
nextEmit = s
if debugSequences {
println("sequence", seq, "next s:", s)
}
blk.sequences = append(blk.sequences, seq)
// Swap offset 1 and 2.
offset1, offset2 = offset2, offset1
if s >= sLimit {
break encodeLoop
}
// Prepare next loop.
cv = load6432(src, s)
}
}
if int(nextEmit) < len(src) {
blk.literals = append(blk.literals, src[nextEmit:]...)
blk.extraLits = len(src) - int(nextEmit)
}
blk.recentOffsets[0] = uint32(offset1)
blk.recentOffsets[1] = uint32(offset2)
if debugEncoder {
println("returning, recent offsets:", blk.recentOffsets, "extra literals:", blk.extraLits)
}
}
// EncodeNoHist will encode a block with no history and no following blocks.
// Most notable difference is that src will not be copied for history and
// we do not need to check for max match length.
func (e *fastEncoder) EncodeNoHist(blk *blockEnc, src []byte) {
const (
inputMargin = 8
minNonLiteralBlockSize = 1 + 1 + inputMargin
)
if debugEncoder {
if len(src) > maxCompressedBlockSize {
panic("src too big")
}
}
// Protect against e.cur wraparound.
if e.cur >= e.bufferReset {
for i := range e.table[:] {
e.table[i] = tableEntry{}
}
e.cur = e.maxMatchOff
}
s := int32(0)
blk.size = len(src)
if len(src) < minNonLiteralBlockSize {
blk.extraLits = len(src)
blk.literals = blk.literals[:len(src)]
copy(blk.literals, src)
return
}
sLimit := int32(len(src)) - inputMargin
// stepSize is the number of bytes to skip on every main loop iteration.
// It should be >= 2.
const stepSize = 2
// TEMPLATE
const hashLog = tableBits
// seems global, but would be nice to tweak.
const kSearchStrength = 6
// nextEmit is where in src the next emitLiteral should start from.
nextEmit := s
cv := load6432(src, s)
// Relative offsets
offset1 := int32(blk.recentOffsets[0])
offset2 := int32(blk.recentOffsets[1])
addLiterals := func(s *seq, until int32) {
if until == nextEmit {
return
}
blk.literals = append(blk.literals, src[nextEmit:until]...)
s.litLen = uint32(until - nextEmit)
}
if debugEncoder {
println("recent offsets:", blk.recentOffsets)
}
encodeLoop:
for {
// t will contain the match offset when we find one.
// When existing the search loop, we have already checked 4 bytes.
var t int32
// We will not use repeat offsets across blocks.
// By not using them for the first 3 matches
for {
nextHash := hashLen(cv, hashLog, tableFastHashLen)
nextHash2 := hashLen(cv>>8, hashLog, tableFastHashLen)
candidate := e.table[nextHash]
candidate2 := e.table[nextHash2]
repIndex := s - offset1 + 2
e.table[nextHash] = tableEntry{offset: s + e.cur, val: uint32(cv)}
e.table[nextHash2] = tableEntry{offset: s + e.cur + 1, val: uint32(cv >> 8)}
if len(blk.sequences) > 2 && load3232(src, repIndex) == uint32(cv>>16) {
// Consider history as well.
var seq seq
length := 4 + e.matchlen(s+6, repIndex+4, src)
seq.matchLen = uint32(length - zstdMinMatch)
// We might be able to match backwards.
// Extend as long as we can.
start := s + 2
// We end the search early, so we don't risk 0 literals
// and have to do special offset treatment.
startLimit := nextEmit + 1
sMin := s - e.maxMatchOff
if sMin < 0 {
sMin = 0
}
for repIndex > sMin && start > startLimit && src[repIndex-1] == src[start-1] {
repIndex--
start--
seq.matchLen++
}
addLiterals(&seq, start)
// rep 0
seq.offset = 1
if debugSequences {
println("repeat sequence", seq, "next s:", s)
}
blk.sequences = append(blk.sequences, seq)
s += length + 2
nextEmit = s
if s >= sLimit {
if debugEncoder {
println("repeat ended", s, length)
}
break encodeLoop
}
cv = load6432(src, s)
continue
}
coffset0 := s - (candidate.offset - e.cur)
coffset1 := s - (candidate2.offset - e.cur) + 1
if coffset0 < e.maxMatchOff && uint32(cv) == candidate.val {
// found a regular match
t = candidate.offset - e.cur
if debugAsserts && s <= t {
panic(fmt.Sprintf("s (%d) <= t (%d)", s, t))
}
if debugAsserts && s-t > e.maxMatchOff {
panic("s - t >e.maxMatchOff")
}
if debugAsserts && t < 0 {
panic(fmt.Sprintf("t (%d) < 0, candidate.offset: %d, e.cur: %d, coffset0: %d, e.maxMatchOff: %d", t, candidate.offset, e.cur, coffset0, e.maxMatchOff))
}
break
}
if coffset1 < e.maxMatchOff && uint32(cv>>8) == candidate2.val {
// found a regular match
t = candidate2.offset - e.cur
s++
if debugAsserts && s <= t {
panic(fmt.Sprintf("s (%d) <= t (%d)", s, t))
}
if debugAsserts && s-t > e.maxMatchOff {
panic("s - t >e.maxMatchOff")
}
if debugAsserts && t < 0 {
panic("t<0")
}
break
}
s += stepSize + ((s - nextEmit) >> (kSearchStrength - 1))
if s >= sLimit {
break encodeLoop
}
cv = load6432(src, s)
}
// A 4-byte match has been found. We'll later see if more than 4 bytes.
offset2 = offset1
offset1 = s - t
if debugAsserts && s <= t {
panic(fmt.Sprintf("s (%d) <= t (%d)", s, t))
}
if debugAsserts && t < 0 {
panic(fmt.Sprintf("t (%d) < 0 ", t))
}
// Extend the 4-byte match as long as possible.
l := e.matchlen(s+4, t+4, src) + 4
// Extend backwards
tMin := s - e.maxMatchOff
if tMin < 0 {
tMin = 0
}
for t > tMin && s > nextEmit && src[t-1] == src[s-1] {
s--
t--
l++
}
// Write our sequence.
var seq seq
seq.litLen = uint32(s - nextEmit)
seq.matchLen = uint32(l - zstdMinMatch)
if seq.litLen > 0 {
blk.literals = append(blk.literals, src[nextEmit:s]...)
}
// Don't use repeat offsets
seq.offset = uint32(s-t) + 3
s += l
if debugSequences {
println("sequence", seq, "next s:", s)
}
blk.sequences = append(blk.sequences, seq)
nextEmit = s
if s >= sLimit {
break encodeLoop
}
cv = load6432(src, s)
// Check offset 2
if o2 := s - offset2; len(blk.sequences) > 2 && load3232(src, o2) == uint32(cv) {
// We have at least 4 byte match.
// No need to check backwards. We come straight from a match
l := 4 + e.matchlen(s+4, o2+4, src)
// Store this, since we have it.
nextHash := hashLen(cv, hashLog, tableFastHashLen)
e.table[nextHash] = tableEntry{offset: s + e.cur, val: uint32(cv)}
seq.matchLen = uint32(l) - zstdMinMatch
seq.litLen = 0
// Since litlen is always 0, this is offset 1.
seq.offset = 1
s += l
nextEmit = s
if debugSequences {
println("sequence", seq, "next s:", s)
}
blk.sequences = append(blk.sequences, seq)
// Swap offset 1 and 2.
offset1, offset2 = offset2, offset1
if s >= sLimit {
break encodeLoop
}
// Prepare next loop.
cv = load6432(src, s)
}
}
if int(nextEmit) < len(src) {
blk.literals = append(blk.literals, src[nextEmit:]...)
blk.extraLits = len(src) - int(nextEmit)
}
if debugEncoder {
println("returning, recent offsets:", blk.recentOffsets, "extra literals:", blk.extraLits)
}
// We do not store history, so we must offset e.cur to avoid false matches for next user.
if e.cur < e.bufferReset {
e.cur += int32(len(src))
}
}
// Encode will encode the content, with a dictionary if initialized for it.
func (e *fastEncoderDict) Encode(blk *blockEnc, src []byte) {
const (
inputMargin = 8
minNonLiteralBlockSize = 1 + 1 + inputMargin
)
if e.allDirty || len(src) > 32<<10 {
e.fastEncoder.Encode(blk, src)
e.allDirty = true
return
}
// Protect against e.cur wraparound.
for e.cur >= e.bufferReset-int32(len(e.hist)) {
if len(e.hist) == 0 {
e.table = [tableSize]tableEntry{}
e.cur = e.maxMatchOff
break
}
// Shift down everything in the table that isn't already too far away.
minOff := e.cur + int32(len(e.hist)) - e.maxMatchOff
for i := range e.table[:] {
v := e.table[i].offset
if v < minOff {
v = 0
} else {
v = v - e.cur + e.maxMatchOff
}
e.table[i].offset = v
}
e.cur = e.maxMatchOff
break
}
s := e.addBlock(src)
blk.size = len(src)
if len(src) < minNonLiteralBlockSize {
blk.extraLits = len(src)
blk.literals = blk.literals[:len(src)]
copy(blk.literals, src)
return
}
// Override src
src = e.hist
sLimit := int32(len(src)) - inputMargin
// stepSize is the number of bytes to skip on every main loop iteration.
// It should be >= 2.
const stepSize = 2
// TEMPLATE
const hashLog = tableBits
// seems global, but would be nice to tweak.
const kSearchStrength = 7
// nextEmit is where in src the next emitLiteral should start from.
nextEmit := s
cv := load6432(src, s)
// Relative offsets
offset1 := int32(blk.recentOffsets[0])
offset2 := int32(blk.recentOffsets[1])
addLiterals := func(s *seq, until int32) {
if until == nextEmit {
return
}
blk.literals = append(blk.literals, src[nextEmit:until]...)
s.litLen = uint32(until - nextEmit)
}
if debugEncoder {
println("recent offsets:", blk.recentOffsets)
}
encodeLoop:
for {
// t will contain the match offset when we find one.
// When existing the search loop, we have already checked 4 bytes.
var t int32
// We will not use repeat offsets across blocks.
// By not using them for the first 3 matches
canRepeat := len(blk.sequences) > 2
for {
if debugAsserts && canRepeat && offset1 == 0 {
panic("offset0 was 0")
}
nextHash := hashLen(cv, hashLog, tableFastHashLen)
nextHash2 := hashLen(cv>>8, hashLog, tableFastHashLen)
candidate := e.table[nextHash]
candidate2 := e.table[nextHash2]
repIndex := s - offset1 + 2
e.table[nextHash] = tableEntry{offset: s + e.cur, val: uint32(cv)}
e.markShardDirty(nextHash)
e.table[nextHash2] = tableEntry{offset: s + e.cur + 1, val: uint32(cv >> 8)}
e.markShardDirty(nextHash2)
if canRepeat && repIndex >= 0 && load3232(src, repIndex) == uint32(cv>>16) {
// Consider history as well.
var seq seq
length := 4 + e.matchlen(s+6, repIndex+4, src)
seq.matchLen = uint32(length - zstdMinMatch)
// We might be able to match backwards.
// Extend as long as we can.
start := s + 2
// We end the search early, so we don't risk 0 literals
// and have to do special offset treatment.
startLimit := nextEmit + 1
sMin := s - e.maxMatchOff
if sMin < 0 {
sMin = 0
}
for repIndex > sMin && start > startLimit && src[repIndex-1] == src[start-1] && seq.matchLen < maxMatchLength-zstdMinMatch {
repIndex--
start--
seq.matchLen++
}
addLiterals(&seq, start)
// rep 0
seq.offset = 1
if debugSequences {
println("repeat sequence", seq, "next s:", s)
}
blk.sequences = append(blk.sequences, seq)
s += length + 2
nextEmit = s
if s >= sLimit {
if debugEncoder {
println("repeat ended", s, length)
}
break encodeLoop
}
cv = load6432(src, s)
continue
}
coffset0 := s - (candidate.offset - e.cur)
coffset1 := s - (candidate2.offset - e.cur) + 1
if coffset0 < e.maxMatchOff && uint32(cv) == candidate.val {
// found a regular match
t = candidate.offset - e.cur
if debugAsserts && s <= t {
panic(fmt.Sprintf("s (%d) <= t (%d)", s, t))
}
if debugAsserts && s-t > e.maxMatchOff {
panic("s - t >e.maxMatchOff")
}
break
}
if coffset1 < e.maxMatchOff && uint32(cv>>8) == candidate2.val {
// found a regular match
t = candidate2.offset - e.cur
s++
if debugAsserts && s <= t {
panic(fmt.Sprintf("s (%d) <= t (%d)", s, t))
}
if debugAsserts && s-t > e.maxMatchOff {
panic("s - t >e.maxMatchOff")
}
if debugAsserts && t < 0 {
panic("t<0")
}
break
}
s += stepSize + ((s - nextEmit) >> (kSearchStrength - 1))
if s >= sLimit {
break encodeLoop
}
cv = load6432(src, s)
}
// A 4-byte match has been found. We'll later see if more than 4 bytes.
offset2 = offset1
offset1 = s - t
if debugAsserts && s <= t {
panic(fmt.Sprintf("s (%d) <= t (%d)", s, t))
}
if debugAsserts && canRepeat && int(offset1) > len(src) {
panic("invalid offset")
}
// Extend the 4-byte match as long as possible.
l := e.matchlen(s+4, t+4, src) + 4
// Extend backwards
tMin := s - e.maxMatchOff
if tMin < 0 {
tMin = 0
}
for t > tMin && s > nextEmit && src[t-1] == src[s-1] && l < maxMatchLength {
s--
t--
l++
}
// Write our sequence.
var seq seq
seq.litLen = uint32(s - nextEmit)
seq.matchLen = uint32(l - zstdMinMatch)
if seq.litLen > 0 {
blk.literals = append(blk.literals, src[nextEmit:s]...)
}
// Don't use repeat offsets
seq.offset = uint32(s-t) + 3
s += l
if debugSequences {
println("sequence", seq, "next s:", s)
}
blk.sequences = append(blk.sequences, seq)
nextEmit = s
if s >= sLimit {
break encodeLoop
}
cv = load6432(src, s)
// Check offset 2
if o2 := s - offset2; canRepeat && load3232(src, o2) == uint32(cv) {
// We have at least 4 byte match.
// No need to check backwards. We come straight from a match
l := 4 + e.matchlen(s+4, o2+4, src)
// Store this, since we have it.
nextHash := hashLen(cv, hashLog, tableFastHashLen)
e.table[nextHash] = tableEntry{offset: s + e.cur, val: uint32(cv)}
e.markShardDirty(nextHash)
seq.matchLen = uint32(l) - zstdMinMatch
seq.litLen = 0
// Since litlen is always 0, this is offset 1.
seq.offset = 1
s += l
nextEmit = s
if debugSequences {
println("sequence", seq, "next s:", s)
}
blk.sequences = append(blk.sequences, seq)
// Swap offset 1 and 2.
offset1, offset2 = offset2, offset1
if s >= sLimit {
break encodeLoop
}
// Prepare next loop.
cv = load6432(src, s)
}
}
if int(nextEmit) < len(src) {
blk.literals = append(blk.literals, src[nextEmit:]...)
blk.extraLits = len(src) - int(nextEmit)
}
blk.recentOffsets[0] = uint32(offset1)
blk.recentOffsets[1] = uint32(offset2)
if debugEncoder {
println("returning, recent offsets:", blk.recentOffsets, "extra literals:", blk.extraLits)
}
}
// ResetDict will reset and set a dictionary if not nil
func (e *fastEncoder) Reset(d *dict, singleBlock bool) {
e.resetBase(d, singleBlock)
if d != nil {
panic("fastEncoder: Reset with dict")
}
}
// ResetDict will reset and set a dictionary if not nil
func (e *fastEncoderDict) Reset(d *dict, singleBlock bool) {
e.resetBase(d, singleBlock)
if d == nil {
return
}
// Init or copy dict table
if len(e.dictTable) != len(e.table) || d.id != e.lastDictID {
if len(e.dictTable) != len(e.table) {
e.dictTable = make([]tableEntry, len(e.table))
}
if true {
end := e.maxMatchOff + int32(len(d.content)) - 8
for i := e.maxMatchOff; i < end; i += 2 {
const hashLog = tableBits
cv := load6432(d.content, i-e.maxMatchOff)
nextHash := hashLen(cv, hashLog, tableFastHashLen) // 0 -> 6
nextHash1 := hashLen(cv>>8, hashLog, tableFastHashLen) // 1 -> 7
e.dictTable[nextHash] = tableEntry{
val: uint32(cv),
offset: i,
}
e.dictTable[nextHash1] = tableEntry{
val: uint32(cv >> 8),
offset: i + 1,
}
}
}
e.lastDictID = d.id
e.allDirty = true
}
e.cur = e.maxMatchOff
dirtyShardCnt := 0
if !e.allDirty {
for i := range e.tableShardDirty {
if e.tableShardDirty[i] {
dirtyShardCnt++
}
}
}
const shardCnt = tableShardCnt
const shardSize = tableShardSize
if e.allDirty || dirtyShardCnt > shardCnt*4/6 {
//copy(e.table[:], e.dictTable)
e.table = *(*[tableSize]tableEntry)(e.dictTable)
for i := range e.tableShardDirty {
e.tableShardDirty[i] = false
}
e.allDirty = false
return
}
for i := range e.tableShardDirty {
if !e.tableShardDirty[i] {
continue
}
//copy(e.table[i*shardSize:(i+1)*shardSize], e.dictTable[i*shardSize:(i+1)*shardSize])
*(*[shardSize]tableEntry)(e.table[i*shardSize:]) = *(*[shardSize]tableEntry)(e.dictTable[i*shardSize:])
e.tableShardDirty[i] = false
}
e.allDirty = false
}
func (e *fastEncoderDict) markAllShardsDirty() {
e.allDirty = true
}
func (e *fastEncoderDict) markShardDirty(entryNum uint32) {
e.tableShardDirty[entryNum/tableShardSize] = true
}

619
vendor/github.com/klauspost/compress/zstd/encoder.go generated vendored Normal file
View File

@@ -0,0 +1,619 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"crypto/rand"
"fmt"
"io"
"math"
rdebug "runtime/debug"
"sync"
"github.com/klauspost/compress/zstd/internal/xxhash"
)
// Encoder provides encoding to Zstandard.
// An Encoder can be used for either compressing a stream via the
// io.WriteCloser interface supported by the Encoder or as multiple independent
// tasks via the EncodeAll function.
// Smaller encodes are encouraged to use the EncodeAll function.
// Use NewWriter to create a new instance.
type Encoder struct {
o encoderOptions
encoders chan encoder
state encoderState
init sync.Once
}
type encoder interface {
Encode(blk *blockEnc, src []byte)
EncodeNoHist(blk *blockEnc, src []byte)
Block() *blockEnc
CRC() *xxhash.Digest
AppendCRC([]byte) []byte
WindowSize(size int64) int32
UseBlock(*blockEnc)
Reset(d *dict, singleBlock bool)
}
type encoderState struct {
w io.Writer
filling []byte
current []byte
previous []byte
encoder encoder
writing *blockEnc
err error
writeErr error
nWritten int64
nInput int64
frameContentSize int64
headerWritten bool
eofWritten bool
fullFrameWritten bool
// This waitgroup indicates an encode is running.
wg sync.WaitGroup
// This waitgroup indicates we have a block encoding/writing.
wWg sync.WaitGroup
}
// NewWriter will create a new Zstandard encoder.
// If the encoder will be used for encoding blocks a nil writer can be used.
func NewWriter(w io.Writer, opts ...EOption) (*Encoder, error) {
initPredefined()
var e Encoder
e.o.setDefault()
for _, o := range opts {
err := o(&e.o)
if err != nil {
return nil, err
}
}
if w != nil {
e.Reset(w)
}
return &e, nil
}
func (e *Encoder) initialize() {
if e.o.concurrent == 0 {
e.o.setDefault()
}
e.encoders = make(chan encoder, e.o.concurrent)
for i := 0; i < e.o.concurrent; i++ {
enc := e.o.encoder()
e.encoders <- enc
}
}
// Reset will re-initialize the writer and new writes will encode to the supplied writer
// as a new, independent stream.
func (e *Encoder) Reset(w io.Writer) {
s := &e.state
s.wg.Wait()
s.wWg.Wait()
if cap(s.filling) == 0 {
s.filling = make([]byte, 0, e.o.blockSize)
}
if e.o.concurrent > 1 {
if cap(s.current) == 0 {
s.current = make([]byte, 0, e.o.blockSize)
}
if cap(s.previous) == 0 {
s.previous = make([]byte, 0, e.o.blockSize)
}
s.current = s.current[:0]
s.previous = s.previous[:0]
if s.writing == nil {
s.writing = &blockEnc{lowMem: e.o.lowMem}
s.writing.init()
}
s.writing.initNewEncode()
}
if s.encoder == nil {
s.encoder = e.o.encoder()
}
s.filling = s.filling[:0]
s.encoder.Reset(e.o.dict, false)
s.headerWritten = false
s.eofWritten = false
s.fullFrameWritten = false
s.w = w
s.err = nil
s.nWritten = 0
s.nInput = 0
s.writeErr = nil
s.frameContentSize = 0
}
// ResetContentSize will reset and set a content size for the next stream.
// If the bytes written does not match the size given an error will be returned
// when calling Close().
// This is removed when Reset is called.
// Sizes <= 0 results in no content size set.
func (e *Encoder) ResetContentSize(w io.Writer, size int64) {
e.Reset(w)
if size >= 0 {
e.state.frameContentSize = size
}
}
// Write data to the encoder.
// Input data will be buffered and as the buffer fills up
// content will be compressed and written to the output.
// When done writing, use Close to flush the remaining output
// and write CRC if requested.
func (e *Encoder) Write(p []byte) (n int, err error) {
s := &e.state
for len(p) > 0 {
if len(p)+len(s.filling) < e.o.blockSize {
if e.o.crc {
_, _ = s.encoder.CRC().Write(p)
}
s.filling = append(s.filling, p...)
return n + len(p), nil
}
add := p
if len(p)+len(s.filling) > e.o.blockSize {
add = add[:e.o.blockSize-len(s.filling)]
}
if e.o.crc {
_, _ = s.encoder.CRC().Write(add)
}
s.filling = append(s.filling, add...)
p = p[len(add):]
n += len(add)
if len(s.filling) < e.o.blockSize {
return n, nil
}
err := e.nextBlock(false)
if err != nil {
return n, err
}
if debugAsserts && len(s.filling) > 0 {
panic(len(s.filling))
}
}
return n, nil
}
// nextBlock will synchronize and start compressing input in e.state.filling.
// If an error has occurred during encoding it will be returned.
func (e *Encoder) nextBlock(final bool) error {
s := &e.state
// Wait for current block.
s.wg.Wait()
if s.err != nil {
return s.err
}
if len(s.filling) > e.o.blockSize {
return fmt.Errorf("block > maxStoreBlockSize")
}
if !s.headerWritten {
// If we have a single block encode, do a sync compression.
if final && len(s.filling) == 0 && !e.o.fullZero {
s.headerWritten = true
s.fullFrameWritten = true
s.eofWritten = true
return nil
}
if final && len(s.filling) > 0 {
s.current = e.EncodeAll(s.filling, s.current[:0])
var n2 int
n2, s.err = s.w.Write(s.current)
if s.err != nil {
return s.err
}
s.nWritten += int64(n2)
s.nInput += int64(len(s.filling))
s.current = s.current[:0]
s.filling = s.filling[:0]
s.headerWritten = true
s.fullFrameWritten = true
s.eofWritten = true
return nil
}
var tmp [maxHeaderSize]byte
fh := frameHeader{
ContentSize: uint64(s.frameContentSize),
WindowSize: uint32(s.encoder.WindowSize(s.frameContentSize)),
SingleSegment: false,
Checksum: e.o.crc,
DictID: e.o.dict.ID(),
}
dst := fh.appendTo(tmp[:0])
s.headerWritten = true
s.wWg.Wait()
var n2 int
n2, s.err = s.w.Write(dst)
if s.err != nil {
return s.err
}
s.nWritten += int64(n2)
}
if s.eofWritten {
// Ensure we only write it once.
final = false
}
if len(s.filling) == 0 {
// Final block, but no data.
if final {
enc := s.encoder
blk := enc.Block()
blk.reset(nil)
blk.last = true
blk.encodeRaw(nil)
s.wWg.Wait()
_, s.err = s.w.Write(blk.output)
s.nWritten += int64(len(blk.output))
s.eofWritten = true
}
return s.err
}
// SYNC:
if e.o.concurrent == 1 {
src := s.filling
s.nInput += int64(len(s.filling))
if debugEncoder {
println("Adding sync block,", len(src), "bytes, final:", final)
}
enc := s.encoder
blk := enc.Block()
blk.reset(nil)
enc.Encode(blk, src)
blk.last = final
if final {
s.eofWritten = true
}
s.err = blk.encode(src, e.o.noEntropy, !e.o.allLitEntropy)
if s.err != nil {
return s.err
}
_, s.err = s.w.Write(blk.output)
s.nWritten += int64(len(blk.output))
s.filling = s.filling[:0]
return s.err
}
// Move blocks forward.
s.filling, s.current, s.previous = s.previous[:0], s.filling, s.current
s.nInput += int64(len(s.current))
s.wg.Add(1)
go func(src []byte) {
if debugEncoder {
println("Adding block,", len(src), "bytes, final:", final)
}
defer func() {
if r := recover(); r != nil {
s.err = fmt.Errorf("panic while encoding: %v", r)
rdebug.PrintStack()
}
s.wg.Done()
}()
enc := s.encoder
blk := enc.Block()
enc.Encode(blk, src)
blk.last = final
if final {
s.eofWritten = true
}
// Wait for pending writes.
s.wWg.Wait()
if s.writeErr != nil {
s.err = s.writeErr
return
}
// Transfer encoders from previous write block.
blk.swapEncoders(s.writing)
// Transfer recent offsets to next.
enc.UseBlock(s.writing)
s.writing = blk
s.wWg.Add(1)
go func() {
defer func() {
if r := recover(); r != nil {
s.writeErr = fmt.Errorf("panic while encoding/writing: %v", r)
rdebug.PrintStack()
}
s.wWg.Done()
}()
s.writeErr = blk.encode(src, e.o.noEntropy, !e.o.allLitEntropy)
if s.writeErr != nil {
return
}
_, s.writeErr = s.w.Write(blk.output)
s.nWritten += int64(len(blk.output))
}()
}(s.current)
return nil
}
// ReadFrom reads data from r until EOF or error.
// The return value n is the number of bytes read.
// Any error except io.EOF encountered during the read is also returned.
//
// The Copy function uses ReaderFrom if available.
func (e *Encoder) ReadFrom(r io.Reader) (n int64, err error) {
if debugEncoder {
println("Using ReadFrom")
}
// Flush any current writes.
if len(e.state.filling) > 0 {
if err := e.nextBlock(false); err != nil {
return 0, err
}
}
e.state.filling = e.state.filling[:e.o.blockSize]
src := e.state.filling
for {
n2, err := r.Read(src)
if e.o.crc {
_, _ = e.state.encoder.CRC().Write(src[:n2])
}
// src is now the unfilled part...
src = src[n2:]
n += int64(n2)
switch err {
case io.EOF:
e.state.filling = e.state.filling[:len(e.state.filling)-len(src)]
if debugEncoder {
println("ReadFrom: got EOF final block:", len(e.state.filling))
}
return n, nil
case nil:
default:
if debugEncoder {
println("ReadFrom: got error:", err)
}
e.state.err = err
return n, err
}
if len(src) > 0 {
if debugEncoder {
println("ReadFrom: got space left in source:", len(src))
}
continue
}
err = e.nextBlock(false)
if err != nil {
return n, err
}
e.state.filling = e.state.filling[:e.o.blockSize]
src = e.state.filling
}
}
// Flush will send the currently written data to output
// and block until everything has been written.
// This should only be used on rare occasions where pushing the currently queued data is critical.
func (e *Encoder) Flush() error {
s := &e.state
if len(s.filling) > 0 {
err := e.nextBlock(false)
if err != nil {
return err
}
}
s.wg.Wait()
s.wWg.Wait()
if s.err != nil {
return s.err
}
return s.writeErr
}
// Close will flush the final output and close the stream.
// The function will block until everything has been written.
// The Encoder can still be re-used after calling this.
func (e *Encoder) Close() error {
s := &e.state
if s.encoder == nil {
return nil
}
err := e.nextBlock(true)
if err != nil {
return err
}
if s.frameContentSize > 0 {
if s.nInput != s.frameContentSize {
return fmt.Errorf("frame content size %d given, but %d bytes was written", s.frameContentSize, s.nInput)
}
}
if e.state.fullFrameWritten {
return s.err
}
s.wg.Wait()
s.wWg.Wait()
if s.err != nil {
return s.err
}
if s.writeErr != nil {
return s.writeErr
}
// Write CRC
if e.o.crc && s.err == nil {
// heap alloc.
var tmp [4]byte
_, s.err = s.w.Write(s.encoder.AppendCRC(tmp[:0]))
s.nWritten += 4
}
// Add padding with content from crypto/rand.Reader
if s.err == nil && e.o.pad > 0 {
add := calcSkippableFrame(s.nWritten, int64(e.o.pad))
frame, err := skippableFrame(s.filling[:0], add, rand.Reader)
if err != nil {
return err
}
_, s.err = s.w.Write(frame)
}
return s.err
}
// EncodeAll will encode all input in src and append it to dst.
// This function can be called concurrently, but each call will only run on a single goroutine.
// If empty input is given, nothing is returned, unless WithZeroFrames is specified.
// Encoded blocks can be concatenated and the result will be the combined input stream.
// Data compressed with EncodeAll can be decoded with the Decoder,
// using either a stream or DecodeAll.
func (e *Encoder) EncodeAll(src, dst []byte) []byte {
if len(src) == 0 {
if e.o.fullZero {
// Add frame header.
fh := frameHeader{
ContentSize: 0,
WindowSize: MinWindowSize,
SingleSegment: true,
// Adding a checksum would be a waste of space.
Checksum: false,
DictID: 0,
}
dst = fh.appendTo(dst)
// Write raw block as last one only.
var blk blockHeader
blk.setSize(0)
blk.setType(blockTypeRaw)
blk.setLast(true)
dst = blk.appendTo(dst)
}
return dst
}
e.init.Do(e.initialize)
enc := <-e.encoders
defer func() {
// Release encoder reference to last block.
// If a non-single block is needed the encoder will reset again.
e.encoders <- enc
}()
// Use single segments when above minimum window and below window size.
single := len(src) <= e.o.windowSize && len(src) > MinWindowSize
if e.o.single != nil {
single = *e.o.single
}
fh := frameHeader{
ContentSize: uint64(len(src)),
WindowSize: uint32(enc.WindowSize(int64(len(src)))),
SingleSegment: single,
Checksum: e.o.crc,
DictID: e.o.dict.ID(),
}
// If less than 1MB, allocate a buffer up front.
if len(dst) == 0 && cap(dst) == 0 && len(src) < 1<<20 && !e.o.lowMem {
dst = make([]byte, 0, len(src))
}
dst = fh.appendTo(dst)
// If we can do everything in one block, prefer that.
if len(src) <= e.o.blockSize {
enc.Reset(e.o.dict, true)
// Slightly faster with no history and everything in one block.
if e.o.crc {
_, _ = enc.CRC().Write(src)
}
blk := enc.Block()
blk.last = true
if e.o.dict == nil {
enc.EncodeNoHist(blk, src)
} else {
enc.Encode(blk, src)
}
// If we got the exact same number of literals as input,
// assume the literals cannot be compressed.
oldout := blk.output
// Output directly to dst
blk.output = dst
err := blk.encode(src, e.o.noEntropy, !e.o.allLitEntropy)
if err != nil {
panic(err)
}
dst = blk.output
blk.output = oldout
} else {
enc.Reset(e.o.dict, false)
blk := enc.Block()
for len(src) > 0 {
todo := src
if len(todo) > e.o.blockSize {
todo = todo[:e.o.blockSize]
}
src = src[len(todo):]
if e.o.crc {
_, _ = enc.CRC().Write(todo)
}
blk.pushOffsets()
enc.Encode(blk, todo)
if len(src) == 0 {
blk.last = true
}
err := blk.encode(todo, e.o.noEntropy, !e.o.allLitEntropy)
if err != nil {
panic(err)
}
dst = append(dst, blk.output...)
blk.reset(nil)
}
}
if e.o.crc {
dst = enc.AppendCRC(dst)
}
// Add padding with content from crypto/rand.Reader
if e.o.pad > 0 {
add := calcSkippableFrame(int64(len(dst)), int64(e.o.pad))
var err error
dst, err = skippableFrame(dst, add, rand.Reader)
if err != nil {
panic(err)
}
}
return dst
}
// MaxEncodedSize returns the expected maximum
// size of an encoded block or stream.
func (e *Encoder) MaxEncodedSize(size int) int {
frameHeader := 4 + 2 // magic + frame header & window descriptor
if e.o.dict != nil {
frameHeader += 4
}
// Frame content size:
if size < 256 {
frameHeader++
} else if size < 65536+256 {
frameHeader += 2
} else if size < math.MaxInt32 {
frameHeader += 4
} else {
frameHeader += 8
}
// Final crc
if e.o.crc {
frameHeader += 4
}
// Max overhead is 3 bytes/block.
// There cannot be 0 blocks.
blocks := (size + e.o.blockSize) / e.o.blockSize
// Combine, add padding.
maxSz := frameHeader + 3*blocks + size
if e.o.pad > 1 {
maxSz += calcSkippableFrame(int64(maxSz), int64(e.o.pad))
}
return maxSz
}

339
vendor/github.com/klauspost/compress/zstd/encoder_options.go generated vendored Normal file
View File

@@ -0,0 +1,339 @@
package zstd
import (
"errors"
"fmt"
"math"
"math/bits"
"runtime"
"strings"
)
// EOption is an option for creating a encoder.
type EOption func(*encoderOptions) error
// options retains accumulated state of multiple options.
type encoderOptions struct {
concurrent int
level EncoderLevel
single *bool
pad int
blockSize int
windowSize int
crc bool
fullZero bool
noEntropy bool
allLitEntropy bool
customWindow bool
customALEntropy bool
customBlockSize bool
lowMem bool
dict *dict
}
func (o *encoderOptions) setDefault() {
*o = encoderOptions{
concurrent: runtime.GOMAXPROCS(0),
crc: true,
single: nil,
blockSize: maxCompressedBlockSize,
windowSize: 8 << 20,
level: SpeedDefault,
allLitEntropy: false,
lowMem: false,
}
}
// encoder returns an encoder with the selected options.
func (o encoderOptions) encoder() encoder {
switch o.level {
case SpeedFastest:
if o.dict != nil {
return &fastEncoderDict{fastEncoder: fastEncoder{fastBase: fastBase{maxMatchOff: int32(o.windowSize), bufferReset: math.MaxInt32 - int32(o.windowSize*2), lowMem: o.lowMem}}}
}
return &fastEncoder{fastBase: fastBase{maxMatchOff: int32(o.windowSize), bufferReset: math.MaxInt32 - int32(o.windowSize*2), lowMem: o.lowMem}}
case SpeedDefault:
if o.dict != nil {
return &doubleFastEncoderDict{fastEncoderDict: fastEncoderDict{fastEncoder: fastEncoder{fastBase: fastBase{maxMatchOff: int32(o.windowSize), bufferReset: math.MaxInt32 - int32(o.windowSize*2), lowMem: o.lowMem}}}}
}
return &doubleFastEncoder{fastEncoder: fastEncoder{fastBase: fastBase{maxMatchOff: int32(o.windowSize), bufferReset: math.MaxInt32 - int32(o.windowSize*2), lowMem: o.lowMem}}}
case SpeedBetterCompression:
if o.dict != nil {
return &betterFastEncoderDict{betterFastEncoder: betterFastEncoder{fastBase: fastBase{maxMatchOff: int32(o.windowSize), bufferReset: math.MaxInt32 - int32(o.windowSize*2), lowMem: o.lowMem}}}
}
return &betterFastEncoder{fastBase: fastBase{maxMatchOff: int32(o.windowSize), bufferReset: math.MaxInt32 - int32(o.windowSize*2), lowMem: o.lowMem}}
case SpeedBestCompression:
return &bestFastEncoder{fastBase: fastBase{maxMatchOff: int32(o.windowSize), bufferReset: math.MaxInt32 - int32(o.windowSize*2), lowMem: o.lowMem}}
}
panic("unknown compression level")
}
// WithEncoderCRC will add CRC value to output.
// Output will be 4 bytes larger.
func WithEncoderCRC(b bool) EOption {
return func(o *encoderOptions) error { o.crc = b; return nil }
}
// WithEncoderConcurrency will set the concurrency,
// meaning the maximum number of encoders to run concurrently.
// The value supplied must be at least 1.
// For streams, setting a value of 1 will disable async compression.
// By default this will be set to GOMAXPROCS.
func WithEncoderConcurrency(n int) EOption {
return func(o *encoderOptions) error {
if n <= 0 {
return fmt.Errorf("concurrency must be at least 1")
}
o.concurrent = n
return nil
}
}
// WithWindowSize will set the maximum allowed back-reference distance.
// The value must be a power of two between MinWindowSize and MaxWindowSize.
// A larger value will enable better compression but allocate more memory and,
// for above-default values, take considerably longer.
// The default value is determined by the compression level and max 8MB.
func WithWindowSize(n int) EOption {
return func(o *encoderOptions) error {
switch {
case n < MinWindowSize:
return fmt.Errorf("window size must be at least %d", MinWindowSize)
case n > MaxWindowSize:
return fmt.Errorf("window size must be at most %d", MaxWindowSize)
case (n & (n - 1)) != 0:
return errors.New("window size must be a power of 2")
}
o.windowSize = n
o.customWindow = true
if o.blockSize > o.windowSize {
o.blockSize = o.windowSize
o.customBlockSize = true
}
return nil
}
}
// WithEncoderPadding will add padding to all output so the size will be a multiple of n.
// This can be used to obfuscate the exact output size or make blocks of a certain size.
// The contents will be a skippable frame, so it will be invisible by the decoder.
// n must be > 0 and <= 1GB, 1<<30 bytes.
// The padded area will be filled with data from crypto/rand.Reader.
// If `EncodeAll` is used with data already in the destination, the total size will be multiple of this.
func WithEncoderPadding(n int) EOption {
return func(o *encoderOptions) error {
if n <= 0 {
return fmt.Errorf("padding must be at least 1")
}
// No need to waste our time.
if n == 1 {
n = 0
}
if n > 1<<30 {
return fmt.Errorf("padding must less than 1GB (1<<30 bytes) ")
}
o.pad = n
return nil
}
}
// EncoderLevel predefines encoder compression levels.
// Only use the constants made available, since the actual mapping
// of these values are very likely to change and your compression could change
// unpredictably when upgrading the library.
type EncoderLevel int
const (
speedNotSet EncoderLevel = iota
// SpeedFastest will choose the fastest reasonable compression.
// This is roughly equivalent to the fastest Zstandard mode.
SpeedFastest
// SpeedDefault is the default "pretty fast" compression option.
// This is roughly equivalent to the default Zstandard mode (level 3).
SpeedDefault
// SpeedBetterCompression will yield better compression than the default.
// Currently it is about zstd level 7-8 with ~ 2x-3x the default CPU usage.
// By using this, notice that CPU usage may go up in the future.
SpeedBetterCompression
// SpeedBestCompression will choose the best available compression option.
// This will offer the best compression no matter the CPU cost.
SpeedBestCompression
// speedLast should be kept as the last actual compression option.
// The is not for external usage, but is used to keep track of the valid options.
speedLast
)
// EncoderLevelFromString will convert a string representation of an encoding level back
// to a compression level. The compare is not case sensitive.
// If the string wasn't recognized, (false, SpeedDefault) will be returned.
func EncoderLevelFromString(s string) (bool, EncoderLevel) {
for l := speedNotSet + 1; l < speedLast; l++ {
if strings.EqualFold(s, l.String()) {
return true, l
}
}
return false, SpeedDefault
}
// EncoderLevelFromZstd will return an encoder level that closest matches the compression
// ratio of a specific zstd compression level.
// Many input values will provide the same compression level.
func EncoderLevelFromZstd(level int) EncoderLevel {
switch {
case level < 3:
return SpeedFastest
case level >= 3 && level < 6:
return SpeedDefault
case level >= 6 && level < 10:
return SpeedBetterCompression
default:
return SpeedBestCompression
}
}
// String provides a string representation of the compression level.
func (e EncoderLevel) String() string {
switch e {
case SpeedFastest:
return "fastest"
case SpeedDefault:
return "default"
case SpeedBetterCompression:
return "better"
case SpeedBestCompression:
return "best"
default:
return "invalid"
}
}
// WithEncoderLevel specifies a predefined compression level.
func WithEncoderLevel(l EncoderLevel) EOption {
return func(o *encoderOptions) error {
switch {
case l <= speedNotSet || l >= speedLast:
return fmt.Errorf("unknown encoder level")
}
o.level = l
if !o.customWindow {
switch o.level {
case SpeedFastest:
o.windowSize = 4 << 20
if !o.customBlockSize {
o.blockSize = 1 << 16
}
case SpeedDefault:
o.windowSize = 8 << 20
case SpeedBetterCompression:
o.windowSize = 8 << 20
case SpeedBestCompression:
o.windowSize = 8 << 20
}
}
if !o.customALEntropy {
o.allLitEntropy = l > SpeedDefault
}
return nil
}
}
// WithZeroFrames will encode 0 length input as full frames.
// This can be needed for compatibility with zstandard usage,
// but is not needed for this package.
func WithZeroFrames(b bool) EOption {
return func(o *encoderOptions) error {
o.fullZero = b
return nil
}
}
// WithAllLitEntropyCompression will apply entropy compression if no matches are found.
// Disabling this will skip incompressible data faster, but in cases with no matches but
// skewed character distribution compression is lost.
// Default value depends on the compression level selected.
func WithAllLitEntropyCompression(b bool) EOption {
return func(o *encoderOptions) error {
o.customALEntropy = true
o.allLitEntropy = b
return nil
}
}
// WithNoEntropyCompression will always skip entropy compression of literals.
// This can be useful if content has matches, but unlikely to benefit from entropy
// compression. Usually the slight speed improvement is not worth enabling this.
func WithNoEntropyCompression(b bool) EOption {
return func(o *encoderOptions) error {
o.noEntropy = b
return nil
}
}
// WithSingleSegment will set the "single segment" flag when EncodeAll is used.
// If this flag is set, data must be regenerated within a single continuous memory segment.
// In this case, Window_Descriptor byte is skipped, but Frame_Content_Size is necessarily present.
// As a consequence, the decoder must allocate a memory segment of size equal or larger than size of your content.
// In order to preserve the decoder from unreasonable memory requirements,
// a decoder is allowed to reject a compressed frame which requests a memory size beyond decoder's authorized range.
// For broader compatibility, decoders are recommended to support memory sizes of at least 8 MB.
// This is only a recommendation, each decoder is free to support higher or lower limits, depending on local limitations.
// If this is not specified, block encodes will automatically choose this based on the input size and the window size.
// This setting has no effect on streamed encodes.
func WithSingleSegment(b bool) EOption {
return func(o *encoderOptions) error {
o.single = &b
return nil
}
}
// WithLowerEncoderMem will trade in some memory cases trade less memory usage for
// slower encoding speed.
// This will not change the window size which is the primary function for reducing
// memory usage. See WithWindowSize.
func WithLowerEncoderMem(b bool) EOption {
return func(o *encoderOptions) error {
o.lowMem = b
return nil
}
}
// WithEncoderDict allows to register a dictionary that will be used for the encode.
//
// The slice dict must be in the [dictionary format] produced by
// "zstd --train" from the Zstandard reference implementation.
//
// The encoder *may* choose to use no dictionary instead for certain payloads.
//
// [dictionary format]: https://github.com/facebook/zstd/blob/dev/doc/zstd_compression_format.md#dictionary-format
func WithEncoderDict(dict []byte) EOption {
return func(o *encoderOptions) error {
d, err := loadDict(dict)
if err != nil {
return err
}
o.dict = d
return nil
}
}
// WithEncoderDictRaw registers a dictionary that may be used by the encoder.
//
// The slice content may contain arbitrary data. It will be used as an initial
// history.
func WithEncoderDictRaw(id uint32, content []byte) EOption {
return func(o *encoderOptions) error {
if bits.UintSize > 32 && uint(len(content)) > dictMaxLength {
return fmt.Errorf("dictionary of size %d > 2GiB too large", len(content))
}
o.dict = &dict{id: id, content: content, offsets: [3]int{1, 4, 8}}
return nil
}
}

413
vendor/github.com/klauspost/compress/zstd/framedec.go generated vendored Normal file
View File

@@ -0,0 +1,413 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"encoding/binary"
"encoding/hex"
"errors"
"io"
"github.com/klauspost/compress/zstd/internal/xxhash"
)
type frameDec struct {
o decoderOptions
crc *xxhash.Digest
WindowSize uint64
// Frame history passed between blocks
history history
rawInput byteBuffer
// Byte buffer that can be reused for small input blocks.
bBuf byteBuf
FrameContentSize uint64
DictionaryID uint32
HasCheckSum bool
SingleSegment bool
}
const (
// MinWindowSize is the minimum Window Size, which is 1 KB.
MinWindowSize = 1 << 10
// MaxWindowSize is the maximum encoder window size
// and the default decoder maximum window size.
MaxWindowSize = 1 << 29
)
const (
frameMagic = "\x28\xb5\x2f\xfd"
skippableFrameMagic = "\x2a\x4d\x18"
)
func newFrameDec(o decoderOptions) *frameDec {
if o.maxWindowSize > o.maxDecodedSize {
o.maxWindowSize = o.maxDecodedSize
}
d := frameDec{
o: o,
}
return &d
}
// reset will read the frame header and prepare for block decoding.
// If nothing can be read from the input, io.EOF will be returned.
// Any other error indicated that the stream contained data, but
// there was a problem.
func (d *frameDec) reset(br byteBuffer) error {
d.HasCheckSum = false
d.WindowSize = 0
var signature [4]byte
for {
var err error
// Check if we can read more...
b, err := br.readSmall(1)
switch err {
case io.EOF, io.ErrUnexpectedEOF:
return io.EOF
case nil:
signature[0] = b[0]
default:
return err
}
// Read the rest, don't allow io.ErrUnexpectedEOF
b, err = br.readSmall(3)
switch err {
case io.EOF:
return io.EOF
case nil:
copy(signature[1:], b)
default:
return err
}
if string(signature[1:4]) != skippableFrameMagic || signature[0]&0xf0 != 0x50 {
if debugDecoder {
println("Not skippable", hex.EncodeToString(signature[:]), hex.EncodeToString([]byte(skippableFrameMagic)))
}
// Break if not skippable frame.
break
}
// Read size to skip
b, err = br.readSmall(4)
if err != nil {
if debugDecoder {
println("Reading Frame Size", err)
}
return err
}
n := uint32(b[0]) | (uint32(b[1]) << 8) | (uint32(b[2]) << 16) | (uint32(b[3]) << 24)
println("Skipping frame with", n, "bytes.")
err = br.skipN(int64(n))
if err != nil {
if debugDecoder {
println("Reading discarded frame", err)
}
return err
}
}
if string(signature[:]) != frameMagic {
if debugDecoder {
println("Got magic numbers: ", signature, "want:", []byte(frameMagic))
}
return ErrMagicMismatch
}
// Read Frame_Header_Descriptor
fhd, err := br.readByte()
if err != nil {
if debugDecoder {
println("Reading Frame_Header_Descriptor", err)
}
return err
}
d.SingleSegment = fhd&(1<<5) != 0
if fhd&(1<<3) != 0 {
return errors.New("reserved bit set on frame header")
}
// Read Window_Descriptor
// https://github.com/facebook/zstd/blob/dev/doc/zstd_compression_format.md#window_descriptor
d.WindowSize = 0
if !d.SingleSegment {
wd, err := br.readByte()
if err != nil {
if debugDecoder {
println("Reading Window_Descriptor", err)
}
return err
}
printf("raw: %x, mantissa: %d, exponent: %d\n", wd, wd&7, wd>>3)
windowLog := 10 + (wd >> 3)
windowBase := uint64(1) << windowLog
windowAdd := (windowBase / 8) * uint64(wd&0x7)
d.WindowSize = windowBase + windowAdd
}
// Read Dictionary_ID
// https://github.com/facebook/zstd/blob/dev/doc/zstd_compression_format.md#dictionary_id
d.DictionaryID = 0
if size := fhd & 3; size != 0 {
if size == 3 {
size = 4
}
b, err := br.readSmall(int(size))
if err != nil {
println("Reading Dictionary_ID", err)
return err
}
var id uint32
switch len(b) {
case 1:
id = uint32(b[0])
case 2:
id = uint32(b[0]) | (uint32(b[1]) << 8)
case 4:
id = uint32(b[0]) | (uint32(b[1]) << 8) | (uint32(b[2]) << 16) | (uint32(b[3]) << 24)
}
if debugDecoder {
println("Dict size", size, "ID:", id)
}
d.DictionaryID = id
}
// Read Frame_Content_Size
// https://github.com/facebook/zstd/blob/dev/doc/zstd_compression_format.md#frame_content_size
var fcsSize int
v := fhd >> 6
switch v {
case 0:
if d.SingleSegment {
fcsSize = 1
}
default:
fcsSize = 1 << v
}
d.FrameContentSize = fcsUnknown
if fcsSize > 0 {
b, err := br.readSmall(fcsSize)
if err != nil {
println("Reading Frame content", err)
return err
}
switch len(b) {
case 1:
d.FrameContentSize = uint64(b[0])
case 2:
// When FCS_Field_Size is 2, the offset of 256 is added.
d.FrameContentSize = uint64(b[0]) | (uint64(b[1]) << 8) + 256
case 4:
d.FrameContentSize = uint64(b[0]) | (uint64(b[1]) << 8) | (uint64(b[2]) << 16) | (uint64(b[3]) << 24)
case 8:
d1 := uint32(b[0]) | (uint32(b[1]) << 8) | (uint32(b[2]) << 16) | (uint32(b[3]) << 24)
d2 := uint32(b[4]) | (uint32(b[5]) << 8) | (uint32(b[6]) << 16) | (uint32(b[7]) << 24)
d.FrameContentSize = uint64(d1) | (uint64(d2) << 32)
}
if debugDecoder {
println("Read FCS:", d.FrameContentSize)
}
}
// Move this to shared.
d.HasCheckSum = fhd&(1<<2) != 0
if d.HasCheckSum {
if d.crc == nil {
d.crc = xxhash.New()
}
d.crc.Reset()
}
if d.WindowSize > d.o.maxWindowSize {
if debugDecoder {
printf("window size %d > max %d\n", d.WindowSize, d.o.maxWindowSize)
}
return ErrWindowSizeExceeded
}
if d.WindowSize == 0 && d.SingleSegment {
// We may not need window in this case.
d.WindowSize = d.FrameContentSize
if d.WindowSize < MinWindowSize {
d.WindowSize = MinWindowSize
}
if d.WindowSize > d.o.maxDecodedSize {
if debugDecoder {
printf("window size %d > max %d\n", d.WindowSize, d.o.maxWindowSize)
}
return ErrDecoderSizeExceeded
}
}
// The minimum Window_Size is 1 KB.
if d.WindowSize < MinWindowSize {
if debugDecoder {
println("got window size: ", d.WindowSize)
}
return ErrWindowSizeTooSmall
}
d.history.windowSize = int(d.WindowSize)
if !d.o.lowMem || d.history.windowSize < maxBlockSize {
// Alloc 2x window size if not low-mem, or window size below 2MB.
d.history.allocFrameBuffer = d.history.windowSize * 2
} else {
if d.o.lowMem {
// Alloc with 1MB extra.
d.history.allocFrameBuffer = d.history.windowSize + maxBlockSize/2
} else {
// Alloc with 2MB extra.
d.history.allocFrameBuffer = d.history.windowSize + maxBlockSize
}
}
if debugDecoder {
println("Frame: Dict:", d.DictionaryID, "FrameContentSize:", d.FrameContentSize, "singleseg:", d.SingleSegment, "window:", d.WindowSize, "crc:", d.HasCheckSum)
}
// history contains input - maybe we do something
d.rawInput = br
return nil
}
// next will start decoding the next block from stream.
func (d *frameDec) next(block *blockDec) error {
if debugDecoder {
println("decoding new block")
}
err := block.reset(d.rawInput, d.WindowSize)
if err != nil {
println("block error:", err)
// Signal the frame decoder we have a problem.
block.sendErr(err)
return err
}
return nil
}
// checkCRC will check the checksum, assuming the frame has one.
// Will return ErrCRCMismatch if crc check failed, otherwise nil.
func (d *frameDec) checkCRC() error {
// We can overwrite upper tmp now
buf, err := d.rawInput.readSmall(4)
if err != nil {
println("CRC missing?", err)
return err
}
want := binary.LittleEndian.Uint32(buf[:4])
got := uint32(d.crc.Sum64())
if got != want {
if debugDecoder {
printf("CRC check failed: got %08x, want %08x\n", got, want)
}
return ErrCRCMismatch
}
if debugDecoder {
printf("CRC ok %08x\n", got)
}
return nil
}
// consumeCRC skips over the checksum, assuming the frame has one.
func (d *frameDec) consumeCRC() error {
_, err := d.rawInput.readSmall(4)
if err != nil {
println("CRC missing?", err)
}
return err
}
// runDecoder will run the decoder for the remainder of the frame.
func (d *frameDec) runDecoder(dst []byte, dec *blockDec) ([]byte, error) {
saved := d.history.b
// We use the history for output to avoid copying it.
d.history.b = dst
d.history.ignoreBuffer = len(dst)
// Store input length, so we only check new data.
crcStart := len(dst)
d.history.decoders.maxSyncLen = 0
if d.o.limitToCap {
d.history.decoders.maxSyncLen = uint64(cap(dst) - len(dst))
}
if d.FrameContentSize != fcsUnknown {
if !d.o.limitToCap || d.FrameContentSize+uint64(len(dst)) < d.history.decoders.maxSyncLen {
d.history.decoders.maxSyncLen = d.FrameContentSize + uint64(len(dst))
}
if d.history.decoders.maxSyncLen > d.o.maxDecodedSize {
if debugDecoder {
println("maxSyncLen:", d.history.decoders.maxSyncLen, "> maxDecodedSize:", d.o.maxDecodedSize)
}
return dst, ErrDecoderSizeExceeded
}
if debugDecoder {
println("maxSyncLen:", d.history.decoders.maxSyncLen)
}
if !d.o.limitToCap && uint64(cap(dst)) < d.history.decoders.maxSyncLen {
// Alloc for output
dst2 := make([]byte, len(dst), d.history.decoders.maxSyncLen+compressedBlockOverAlloc)
copy(dst2, dst)
dst = dst2
}
}
var err error
for {
err = dec.reset(d.rawInput, d.WindowSize)
if err != nil {
break
}
if debugDecoder {
println("next block:", dec)
}
err = dec.decodeBuf(&d.history)
if err != nil {
break
}
if uint64(len(d.history.b)-crcStart) > d.o.maxDecodedSize {
println("runDecoder: maxDecodedSize exceeded", uint64(len(d.history.b)-crcStart), ">", d.o.maxDecodedSize)
err = ErrDecoderSizeExceeded
break
}
if d.o.limitToCap && len(d.history.b) > cap(dst) {
println("runDecoder: cap exceeded", uint64(len(d.history.b)), ">", cap(dst))
err = ErrDecoderSizeExceeded
break
}
if uint64(len(d.history.b)-crcStart) > d.FrameContentSize {
println("runDecoder: FrameContentSize exceeded", uint64(len(d.history.b)-crcStart), ">", d.FrameContentSize)
err = ErrFrameSizeExceeded
break
}
if dec.Last {
break
}
if debugDecoder {
println("runDecoder: FrameContentSize", uint64(len(d.history.b)-crcStart), "<=", d.FrameContentSize)
}
}
dst = d.history.b
if err == nil {
if d.FrameContentSize != fcsUnknown && uint64(len(d.history.b)-crcStart) != d.FrameContentSize {
err = ErrFrameSizeMismatch
} else if d.HasCheckSum {
if d.o.ignoreChecksum {
err = d.consumeCRC()
} else {
d.crc.Write(dst[crcStart:])
err = d.checkCRC()
}
}
}
d.history.b = saved
return dst, err
}

137
vendor/github.com/klauspost/compress/zstd/frameenc.go generated vendored Normal file
View File

@@ -0,0 +1,137 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"encoding/binary"
"fmt"
"io"
"math"
"math/bits"
)
type frameHeader struct {
ContentSize uint64
WindowSize uint32
SingleSegment bool
Checksum bool
DictID uint32
}
const maxHeaderSize = 14
func (f frameHeader) appendTo(dst []byte) []byte {
dst = append(dst, frameMagic...)
var fhd uint8
if f.Checksum {
fhd |= 1 << 2
}
if f.SingleSegment {
fhd |= 1 << 5
}
var dictIDContent []byte
if f.DictID > 0 {
var tmp [4]byte
if f.DictID < 256 {
fhd |= 1
tmp[0] = uint8(f.DictID)
dictIDContent = tmp[:1]
} else if f.DictID < 1<<16 {
fhd |= 2
binary.LittleEndian.PutUint16(tmp[:2], uint16(f.DictID))
dictIDContent = tmp[:2]
} else {
fhd |= 3
binary.LittleEndian.PutUint32(tmp[:4], f.DictID)
dictIDContent = tmp[:4]
}
}
var fcs uint8
if f.ContentSize >= 256 {
fcs++
}
if f.ContentSize >= 65536+256 {
fcs++
}
if f.ContentSize >= 0xffffffff {
fcs++
}
fhd |= fcs << 6
dst = append(dst, fhd)
if !f.SingleSegment {
const winLogMin = 10
windowLog := (bits.Len32(f.WindowSize-1) - winLogMin) << 3
dst = append(dst, uint8(windowLog))
}
if f.DictID > 0 {
dst = append(dst, dictIDContent...)
}
switch fcs {
case 0:
if f.SingleSegment {
dst = append(dst, uint8(f.ContentSize))
}
// Unless SingleSegment is set, framessizes < 256 are not stored.
case 1:
f.ContentSize -= 256
dst = append(dst, uint8(f.ContentSize), uint8(f.ContentSize>>8))
case 2:
dst = append(dst, uint8(f.ContentSize), uint8(f.ContentSize>>8), uint8(f.ContentSize>>16), uint8(f.ContentSize>>24))
case 3:
dst = append(dst, uint8(f.ContentSize), uint8(f.ContentSize>>8), uint8(f.ContentSize>>16), uint8(f.ContentSize>>24),
uint8(f.ContentSize>>32), uint8(f.ContentSize>>40), uint8(f.ContentSize>>48), uint8(f.ContentSize>>56))
default:
panic("invalid fcs")
}
return dst
}
const skippableFrameHeader = 4 + 4
// calcSkippableFrame will return a total size to be added for written
// to be divisible by multiple.
// The value will always be > skippableFrameHeader.
// The function will panic if written < 0 or wantMultiple <= 0.
func calcSkippableFrame(written, wantMultiple int64) int {
if wantMultiple <= 0 {
panic("wantMultiple <= 0")
}
if written < 0 {
panic("written < 0")
}
leftOver := written % wantMultiple
if leftOver == 0 {
return 0
}
toAdd := wantMultiple - leftOver
for toAdd < skippableFrameHeader {
toAdd += wantMultiple
}
return int(toAdd)
}
// skippableFrame will add a skippable frame with a total size of bytes.
// total should be >= skippableFrameHeader and < math.MaxUint32.
func skippableFrame(dst []byte, total int, r io.Reader) ([]byte, error) {
if total == 0 {
return dst, nil
}
if total < skippableFrameHeader {
return dst, fmt.Errorf("requested skippable frame (%d) < 8", total)
}
if int64(total) > math.MaxUint32 {
return dst, fmt.Errorf("requested skippable frame (%d) > max uint32", total)
}
dst = append(dst, 0x50, 0x2a, 0x4d, 0x18)
f := uint32(total - skippableFrameHeader)
dst = append(dst, uint8(f), uint8(f>>8), uint8(f>>16), uint8(f>>24))
start := len(dst)
dst = append(dst, make([]byte, f)...)
_, err := io.ReadFull(r, dst[start:])
return dst, err
}

307
vendor/github.com/klauspost/compress/zstd/fse_decoder.go generated vendored Normal file
View File

@@ -0,0 +1,307 @@
// Copyright 2019+ Klaus Post. All rights reserved.
// License information can be found in the LICENSE file.
// Based on work by Yann Collet, released under BSD License.
package zstd
import (
"encoding/binary"
"errors"
"fmt"
"io"
)
const (
tablelogAbsoluteMax = 9
)
const (
/*!MEMORY_USAGE :
* Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; etc.)
* Increasing memory usage improves compression ratio
* Reduced memory usage can improve speed, due to cache effect
* Recommended max value is 14, for 16KB, which nicely fits into Intel x86 L1 cache */
maxMemoryUsage = tablelogAbsoluteMax + 2
maxTableLog = maxMemoryUsage - 2
maxTablesize = 1 << maxTableLog
maxTableMask = (1 << maxTableLog) - 1
minTablelog = 5
maxSymbolValue = 255
)
// fseDecoder provides temporary storage for compression and decompression.
type fseDecoder struct {
dt [maxTablesize]decSymbol // Decompression table.
symbolLen uint16 // Length of active part of the symbol table.
actualTableLog uint8 // Selected tablelog.
maxBits uint8 // Maximum number of additional bits
// used for table creation to avoid allocations.
stateTable [256]uint16
norm [maxSymbolValue + 1]int16
preDefined bool
}
// tableStep returns the next table index.
func tableStep(tableSize uint32) uint32 {
return (tableSize >> 1) + (tableSize >> 3) + 3
}
// readNCount will read the symbol distribution so decoding tables can be constructed.
func (s *fseDecoder) readNCount(b *byteReader, maxSymbol uint16) error {
var (
charnum uint16
previous0 bool
)
if b.remain() < 4 {
return errors.New("input too small")
}
bitStream := b.Uint32NC()
nbBits := uint((bitStream & 0xF) + minTablelog) // extract tableLog
if nbBits > tablelogAbsoluteMax {
println("Invalid tablelog:", nbBits)
return errors.New("tableLog too large")
}
bitStream >>= 4
bitCount := uint(4)
s.actualTableLog = uint8(nbBits)
remaining := int32((1 << nbBits) + 1)
threshold := int32(1 << nbBits)
gotTotal := int32(0)
nbBits++
for remaining > 1 && charnum <= maxSymbol {
if previous0 {
//println("prev0")
n0 := charnum
for (bitStream & 0xFFFF) == 0xFFFF {
//println("24 x 0")
n0 += 24
if r := b.remain(); r > 5 {
b.advance(2)
// The check above should make sure we can read 32 bits
bitStream = b.Uint32NC() >> bitCount
} else {
// end of bit stream
bitStream >>= 16
bitCount += 16
}
}
//printf("bitstream: %d, 0b%b", bitStream&3, bitStream)
for (bitStream & 3) == 3 {
n0 += 3
bitStream >>= 2
bitCount += 2
}
n0 += uint16(bitStream & 3)
bitCount += 2
if n0 > maxSymbolValue {
return errors.New("maxSymbolValue too small")
}
//println("inserting ", n0-charnum, "zeroes from idx", charnum, "ending before", n0)
for charnum < n0 {
s.norm[uint8(charnum)] = 0
charnum++
}
if r := b.remain(); r >= 7 || r-int(bitCount>>3) >= 4 {
b.advance(bitCount >> 3)
bitCount &= 7
// The check above should make sure we can read 32 bits
bitStream = b.Uint32NC() >> bitCount
} else {
bitStream >>= 2
}
}
max := (2*threshold - 1) - remaining
var count int32
if int32(bitStream)&(threshold-1) < max {
count = int32(bitStream) & (threshold - 1)
if debugAsserts && nbBits < 1 {
panic("nbBits underflow")
}
bitCount += nbBits - 1
} else {
count = int32(bitStream) & (2*threshold - 1)
if count >= threshold {
count -= max
}
bitCount += nbBits
}
// extra accuracy
count--
if count < 0 {
// -1 means +1
remaining += count
gotTotal -= count
} else {
remaining -= count
gotTotal += count
}
s.norm[charnum&0xff] = int16(count)
charnum++
previous0 = count == 0
for remaining < threshold {
nbBits--
threshold >>= 1
}
if r := b.remain(); r >= 7 || r-int(bitCount>>3) >= 4 {
b.advance(bitCount >> 3)
bitCount &= 7
// The check above should make sure we can read 32 bits
bitStream = b.Uint32NC() >> (bitCount & 31)
} else {
bitCount -= (uint)(8 * (len(b.b) - 4 - b.off))
b.off = len(b.b) - 4
bitStream = b.Uint32() >> (bitCount & 31)
}
}
s.symbolLen = charnum
if s.symbolLen <= 1 {
return fmt.Errorf("symbolLen (%d) too small", s.symbolLen)
}
if s.symbolLen > maxSymbolValue+1 {
return fmt.Errorf("symbolLen (%d) too big", s.symbolLen)
}
if remaining != 1 {
return fmt.Errorf("corruption detected (remaining %d != 1)", remaining)
}
if bitCount > 32 {
return fmt.Errorf("corruption detected (bitCount %d > 32)", bitCount)
}
if gotTotal != 1<<s.actualTableLog {
return fmt.Errorf("corruption detected (total %d != %d)", gotTotal, 1<<s.actualTableLog)
}
b.advance((bitCount + 7) >> 3)
return s.buildDtable()
}
func (s *fseDecoder) mustReadFrom(r io.Reader) {
fatalErr := func(err error) {
if err != nil {
panic(err)
}
}
// dt [maxTablesize]decSymbol // Decompression table.
// symbolLen uint16 // Length of active part of the symbol table.
// actualTableLog uint8 // Selected tablelog.
// maxBits uint8 // Maximum number of additional bits
// // used for table creation to avoid allocations.
// stateTable [256]uint16
// norm [maxSymbolValue + 1]int16
// preDefined bool
fatalErr(binary.Read(r, binary.LittleEndian, &s.dt))
fatalErr(binary.Read(r, binary.LittleEndian, &s.symbolLen))
fatalErr(binary.Read(r, binary.LittleEndian, &s.actualTableLog))
fatalErr(binary.Read(r, binary.LittleEndian, &s.maxBits))
fatalErr(binary.Read(r, binary.LittleEndian, &s.stateTable))
fatalErr(binary.Read(r, binary.LittleEndian, &s.norm))
fatalErr(binary.Read(r, binary.LittleEndian, &s.preDefined))
}
// decSymbol contains information about a state entry,
// Including the state offset base, the output symbol and
// the number of bits to read for the low part of the destination state.
// Using a composite uint64 is faster than a struct with separate members.
type decSymbol uint64
func newDecSymbol(nbits, addBits uint8, newState uint16, baseline uint32) decSymbol {
return decSymbol(nbits) | (decSymbol(addBits) << 8) | (decSymbol(newState) << 16) | (decSymbol(baseline) << 32)
}
func (d decSymbol) nbBits() uint8 {
return uint8(d)
}
func (d decSymbol) addBits() uint8 {
return uint8(d >> 8)
}
func (d decSymbol) newState() uint16 {
return uint16(d >> 16)
}
func (d decSymbol) baselineInt() int {
return int(d >> 32)
}
func (d *decSymbol) setNBits(nBits uint8) {
const mask = 0xffffffffffffff00
*d = (*d & mask) | decSymbol(nBits)
}
func (d *decSymbol) setAddBits(addBits uint8) {
const mask = 0xffffffffffff00ff
*d = (*d & mask) | (decSymbol(addBits) << 8)
}
func (d *decSymbol) setNewState(state uint16) {
const mask = 0xffffffff0000ffff
*d = (*d & mask) | decSymbol(state)<<16
}
func (d *decSymbol) setExt(addBits uint8, baseline uint32) {
const mask = 0xffff00ff
*d = (*d & mask) | (decSymbol(addBits) << 8) | (decSymbol(baseline) << 32)
}
// decSymbolValue returns the transformed decSymbol for the given symbol.
func decSymbolValue(symb uint8, t []baseOffset) (decSymbol, error) {
if int(symb) >= len(t) {
return 0, fmt.Errorf("rle symbol %d >= max %d", symb, len(t))
}
lu := t[symb]
return newDecSymbol(0, lu.addBits, 0, lu.baseLine), nil
}
// setRLE will set the decoder til RLE mode.
func (s *fseDecoder) setRLE(symbol decSymbol) {
s.actualTableLog = 0
s.maxBits = symbol.addBits()
s.dt[0] = symbol
}
// transform will transform the decoder table into a table usable for
// decoding without having to apply the transformation while decoding.
// The state will contain the base value and the number of bits to read.
func (s *fseDecoder) transform(t []baseOffset) error {
tableSize := uint16(1 << s.actualTableLog)
s.maxBits = 0
for i, v := range s.dt[:tableSize] {
add := v.addBits()
if int(add) >= len(t) {
return fmt.Errorf("invalid decoding table entry %d, symbol %d >= max (%d)", i, v.addBits(), len(t))
}
lu := t[add]
if lu.addBits > s.maxBits {
s.maxBits = lu.addBits
}
v.setExt(lu.addBits, lu.baseLine)
s.dt[i] = v
}
return nil
}
type fseState struct {
dt []decSymbol
state decSymbol
}
// Initialize and decodeAsync first state and symbol.
func (s *fseState) init(br *bitReader, tableLog uint8, dt []decSymbol) {
s.dt = dt
br.fill()
s.state = dt[br.getBits(tableLog)]
}
// final returns the current state symbol without decoding the next.
func (s decSymbol) final() (int, uint8) {
return s.baselineInt(), s.addBits()
}

65
vendor/github.com/klauspost/compress/zstd/fse_decoder_amd64.go generated vendored Normal file
View File

@@ -0,0 +1,65 @@
//go:build amd64 && !appengine && !noasm && gc
// +build amd64,!appengine,!noasm,gc
package zstd
import (
"fmt"
)
type buildDtableAsmContext struct {
// inputs
stateTable *uint16
norm *int16
dt *uint64
// outputs --- set by the procedure in the case of error;
// for interpretation please see the error handling part below
errParam1 uint64
errParam2 uint64
}
// buildDtable_asm is an x86 assembly implementation of fseDecoder.buildDtable.
// Function returns non-zero exit code on error.
//
//go:noescape
func buildDtable_asm(s *fseDecoder, ctx *buildDtableAsmContext) int
// please keep in sync with _generate/gen_fse.go
const (
errorCorruptedNormalizedCounter = 1
errorNewStateTooBig = 2
errorNewStateNoBits = 3
)
// buildDtable will build the decoding table.
func (s *fseDecoder) buildDtable() error {
ctx := buildDtableAsmContext{
stateTable: &s.stateTable[0],
norm: &s.norm[0],
dt: (*uint64)(&s.dt[0]),
}
code := buildDtable_asm(s, &ctx)
if code != 0 {
switch code {
case errorCorruptedNormalizedCounter:
position := ctx.errParam1
return fmt.Errorf("corrupted input (position=%d, expected 0)", position)
case errorNewStateTooBig:
newState := decSymbol(ctx.errParam1)
size := ctx.errParam2
return fmt.Errorf("newState (%d) outside table size (%d)", newState, size)
case errorNewStateNoBits:
newState := decSymbol(ctx.errParam1)
oldState := decSymbol(ctx.errParam2)
return fmt.Errorf("newState (%d) == oldState (%d) and no bits", newState, oldState)
default:
return fmt.Errorf("buildDtable_asm returned unhandled nonzero code = %d", code)
}
}
return nil
}

126
vendor/github.com/klauspost/compress/zstd/fse_decoder_amd64.s generated vendored Normal file
View File

@@ -0,0 +1,126 @@
// Code generated by command: go run gen_fse.go -out ../fse_decoder_amd64.s -pkg=zstd. DO NOT EDIT.
//go:build !appengine && !noasm && gc && !noasm
// func buildDtable_asm(s *fseDecoder, ctx *buildDtableAsmContext) int
TEXT ·buildDtable_asm(SB), $0-24
MOVQ ctx+8(FP), CX
MOVQ s+0(FP), DI
// Load values
MOVBQZX 4098(DI), DX
XORQ AX, AX
BTSQ DX, AX
MOVQ (CX), BX
MOVQ 16(CX), SI
LEAQ -1(AX), R8
MOVQ 8(CX), CX
MOVWQZX 4096(DI), DI
// End load values
// Init, lay down lowprob symbols
XORQ R9, R9
JMP init_main_loop_condition
init_main_loop:
MOVWQSX (CX)(R9*2), R10
CMPW R10, $-1
JNE do_not_update_high_threshold
MOVB R9, 1(SI)(R8*8)
DECQ R8
MOVQ $0x0000000000000001, R10
do_not_update_high_threshold:
MOVW R10, (BX)(R9*2)
INCQ R9
init_main_loop_condition:
CMPQ R9, DI
JL init_main_loop
// Spread symbols
// Calculate table step
MOVQ AX, R9
SHRQ $0x01, R9
MOVQ AX, R10
SHRQ $0x03, R10
LEAQ 3(R9)(R10*1), R9
// Fill add bits values
LEAQ -1(AX), R10
XORQ R11, R11
XORQ R12, R12
JMP spread_main_loop_condition
spread_main_loop:
XORQ R13, R13
MOVWQSX (CX)(R12*2), R14
JMP spread_inner_loop_condition
spread_inner_loop:
MOVB R12, 1(SI)(R11*8)
adjust_position:
ADDQ R9, R11
ANDQ R10, R11
CMPQ R11, R8
JG adjust_position
INCQ R13
spread_inner_loop_condition:
CMPQ R13, R14
JL spread_inner_loop
INCQ R12
spread_main_loop_condition:
CMPQ R12, DI
JL spread_main_loop
TESTQ R11, R11
JZ spread_check_ok
MOVQ ctx+8(FP), AX
MOVQ R11, 24(AX)
MOVQ $+1, ret+16(FP)
RET
spread_check_ok:
// Build Decoding table
XORQ DI, DI
build_table_main_table:
MOVBQZX 1(SI)(DI*8), CX
MOVWQZX (BX)(CX*2), R8
LEAQ 1(R8), R9
MOVW R9, (BX)(CX*2)
MOVQ R8, R9
BSRQ R9, R9
MOVQ DX, CX
SUBQ R9, CX
SHLQ CL, R8
SUBQ AX, R8
MOVB CL, (SI)(DI*8)
MOVW R8, 2(SI)(DI*8)
CMPQ R8, AX
JLE build_table_check1_ok
MOVQ ctx+8(FP), CX
MOVQ R8, 24(CX)
MOVQ AX, 32(CX)
MOVQ $+2, ret+16(FP)
RET
build_table_check1_ok:
TESTB CL, CL
JNZ build_table_check2_ok
CMPW R8, DI
JNE build_table_check2_ok
MOVQ ctx+8(FP), AX
MOVQ R8, 24(AX)
MOVQ DI, 32(AX)
MOVQ $+3, ret+16(FP)
RET
build_table_check2_ok:
INCQ DI
CMPQ DI, AX
JL build_table_main_table
MOVQ $+0, ret+16(FP)
RET

73
vendor/github.com/klauspost/compress/zstd/fse_decoder_generic.go generated vendored Normal file
View File

@@ -0,0 +1,73 @@
//go:build !amd64 || appengine || !gc || noasm
// +build !amd64 appengine !gc noasm
package zstd
import (
"errors"
"fmt"
)
// buildDtable will build the decoding table.
func (s *fseDecoder) buildDtable() error {
tableSize := uint32(1 << s.actualTableLog)
highThreshold := tableSize - 1
symbolNext := s.stateTable[:256]
// Init, lay down lowprob symbols
{
for i, v := range s.norm[:s.symbolLen] {
if v == -1 {
s.dt[highThreshold].setAddBits(uint8(i))
highThreshold--
v = 1
}
symbolNext[i] = uint16(v)
}
}
// Spread symbols
{
tableMask := tableSize - 1
step := tableStep(tableSize)
position := uint32(0)
for ss, v := range s.norm[:s.symbolLen] {
for i := 0; i < int(v); i++ {
s.dt[position].setAddBits(uint8(ss))
for {
// lowprob area
position = (position + step) & tableMask
if position <= highThreshold {
break
}
}
}
}
if position != 0 {
// position must reach all cells once, otherwise normalizedCounter is incorrect
return errors.New("corrupted input (position != 0)")
}
}
// Build Decoding table
{
tableSize := uint16(1 << s.actualTableLog)
for u, v := range s.dt[:tableSize] {
symbol := v.addBits()
nextState := symbolNext[symbol]
symbolNext[symbol] = nextState + 1
nBits := s.actualTableLog - byte(highBits(uint32(nextState)))
s.dt[u&maxTableMask].setNBits(nBits)
newState := (nextState << nBits) - tableSize
if newState > tableSize {
return fmt.Errorf("newState (%d) outside table size (%d)", newState, tableSize)
}
if newState == uint16(u) && nBits == 0 {
// Seems weird that this is possible with nbits > 0.
return fmt.Errorf("newState (%d) == oldState (%d) and no bits", newState, u)
}
s.dt[u&maxTableMask].setNewState(newState)
}
}
return nil
}

Some files were not shown because too many files have changed in this diff Show More