Commit Graph

2178 Commits

Author SHA1 Message Date
cmasone-attic dc07bca3d4 Merge pull request #1555 from cmasone-attic/dbcleanup
JS: make sure we call close() on Database most of the time
2016-05-20 08:18:53 -07:00
Mike Gray ade84de2da implementing map/set/orderedSequence diff (#1521)
resolving a portion of issue #1075
2016-05-20 09:20:01 -04:00
Chris Masone 917a4d8564 JS: make sure we call close() on Database most of the time
Other than DatasetSpec::value(), this should close all Database
instances that we create. I'm not sure how to deal with that one
case, though.
2016-05-19 16:49:54 -07:00
Rafael Weinstein 0e87d9f409 goPath (#1546)
Implement Go Path
2016-05-19 13:56:18 -07:00
Erik Arvidsson cd64ee5047 Update clients to 32.2 (#1545)
This updates the clients to use Foo instead of NomsFoo
2016-05-19 11:38:00 -07:00
Rafael Weinstein 4389605e47 Struct diff (#1548)
StructDiff
2016-05-19 11:09:26 -07:00
Mike Gray b911e7bc08 updating to go 1.6 (#1549) 2016-05-19 13:55:55 -04:00
Erik Arvidsson f653a20511 Make Splore Great Again! (#1541) 2016-05-18 17:57:57 -07:00
Ben Kalman 7f56776a8b Add dist back to .flowconfig (#1544) 2016-05-18 17:51:56 -07:00
Ben Kalman 34449f0b2a Delete sequence_chunker_test.go 2016-05-18 17:39:45 -07:00
Ben Kalman 65f4ae358c Make sequence chunker function return (meta tuple, collection) pair (#1540)
Both Go and JS. Currently they return (any, value), but it's useful to
have the more specific return type.
2016-05-18 16:01:13 -07:00
Erik Arvidsson 778acf8454 Make OrderedPutCache throw in browsers (#1539)
This is a stop gap to ensure that the browser version does not depend
on nodejs modules.

Fixes #1531
2016-05-18 15:28:09 -07:00
Erik Arvidsson d67d1fb924 Make Map use Array<[K, V]> instead of Array<K | V> (#1538) 2016-05-18 15:16:08 -07:00
Rafael Weinstein 4d47b9cf18 jsPath (#1527) 2016-05-18 14:35:26 -07:00
Erik Arvidsson 6e23e8f0e7 Update clients/js to v31 (#1532)
Note: This does not update splore or bar-chart since the SDK is now
broken in browsers.
2016-05-18 13:35:21 -07:00
Rafael Weinstein 7b439b6b52 Add rule to require a trailing newline & to disallow trailing whitespace (#1533) 2016-05-18 11:40:13 -07:00
Aaron Boodman 810e098349 Introduce clients/js/flickr (#1517)
Introduce clients/js/flickr

This version of the Flickr importer does its work in two stages:

1. Import the raw data
2. Parse out photo information

This is the type of thing I think will be common in archival use
cases for Noms: just get everything now. Worry about parsing it
later.
2016-05-18 11:10:10 -07:00
Ben Kalman 0585010274 Return io.Writer errors from WriteEncodedValueWithTags instead of panic (#1526)
A common case for io.Writer errors is during noms-log or noms-show when
piped through head, i.e. "noms log ldb:/path/to:noms | head".

noms-log handles this itself by recovering from the panic, but rather
than porting this logic to noms-show, just make it so that you don't
need to recover in the first place.

This is similar to how errors from fmt.Println etc don't panic.
Non-write errors (like crashes) will continue to panic.
2016-05-18 10:37:35 -07:00
Erik Arvidsson 643c12172b Noms data structure names (#1525)
* Change to default exports

* Rename NomsX to X in (Blob, List, Map, Set)

* Increment version
2016-05-18 09:26:36 -07:00
Erik Arvidsson 17a78061c0 Struct subtype (#1524)
* Subtype checking for structs

Towards #1491
2016-05-18 09:06:53 -07:00
cmasone-attic cefb014d15 Go: In httpBatchStore, order outgoing Chunks by ref-height (#1519)
* Go: In httpBatchStore, order outgoing Chunks by ref-height

Rather than sending chunks to the server in the order that callers put
them into an httpBatchStore, instead order them by ref-height at the
time that Flush() is called.  Making this change requires that
ref-height for a given chunk be passed in through the SchedulePut()
API, because Chunks (by design) don't carry much metadata about
themselves.

Toward #1510
2016-05-18 08:33:49 -07:00
Erik Arvidsson df1c91843d JS: No need to compute the union type in decode (#1523)
* JS: No need to compute the union type in decode

Instead trust that the serialization is correct

Towards #1491
2016-05-17 15:39:18 -07:00
Erik Arvidsson 653aa41776 Go: Dynamic types for Data structure API (#1516) 2016-05-17 15:02:53 -07:00
cmasone-attic ff6f63835a Merge pull request #1522 from cmasone-attic/issue1414
JS: add http idle timeout
2016-05-17 14:29:28 -07:00
Chris Masone 41d5ff8063 JS: add http idle timeout
Node's http.ClientRequest supports setting an idle-timeout, so
just go ahead and do that. Setting to 2 minutes to match Go.

Fixes #1414
2016-05-17 14:11:34 -07:00
cmasone-attic 590bc69cbc Go: replace httpBatchStore overall timeout with idle-timeout (#1518)
The old timeout mechanism started counting when the client
began trying to connect to the server. This meant that sending
a large batch of chunks could cause the request to fail before
all the data even made it to the server. What we actually want
is to timeout when the connection has been idle for some amount
of time. This means that long server processing times can still
cause clients to time out, but that's probably as it should be.

Towards #1414
2016-05-17 13:27:09 -07:00
Erik Arvidsson d125446955 Clear dist before publish (#1480)
This is so that we do not keep old files around.

Fixes #1380
2016-05-16 19:13:44 -07:00
Ben Kalman 2b671664a1 Don't descend into leaf nodes during SomeChunksP (#1506) 2016-05-16 17:48:57 -07:00
Dan Willhite 998ad17ad0 Merge pull request #1512 from willhite/work
Dataspec improvements
2016-05-16 15:52:40 -07:00
Dan Willhite 06b54eb523 Dataspec improvements 2016-05-16 15:27:53 -07:00
cmasone-attic e0eb7bc297 JS: back BatchStore's put-cache with tingodb (#1479)
* JS: back BatchStore's put-cache with tingodb

tingodb is a pure-JS, file-backed mongodb-api-compatible database.
Before this patch, BatchStore was caching all pending chunks in memory,
which seems unlikely to scale. Putting them in a tingodb spills them to
disk as needed, though we can still iterate them in put-order when it's
time to send them.

Fixes #1349
2016-05-16 14:38:25 -07:00
Aaron Boodman ae12160f57 Add progress to csv-importer (#1507)
Add progress to csv-importer
2016-05-16 12:03:39 -07:00
Aaron Boodman 07e747ef4c Add lost clients/js/url_fetch (#1509)
This was accidentally removed in the clients dir reshuffle a few days back.
2016-05-16 11:01:19 -07:00
Erik Arvidsson 5db7fe8285 JS: Update the collection/struct API to compute the type. (#1502)
When you create or modify a data structure we now compute the type
based on the type of the values in the data structure.

Towards #1491
2016-05-16 10:58:50 -07:00
Aaron Boodman e4540aae9f Add a little description heading to all the noms commands (#1508)
Add a little description heading to all the noms commands
2016-05-16 10:56:26 -07:00
Ben Kalman 5caad77f81 js/walk: allow callback to return boolean in addition to Promise (#1511) 2016-05-16 10:52:31 -07:00
Aaron Boodman 49cfe3ddc3 Update README.md 2016-05-15 18:57:07 -07:00
Aaron Boodman 65839a4a62 Update cli-tour.md 2016-05-14 14:01:32 -07:00
mnm678 576bb34017 Merge pull request #1505 from mnm678/noms-serve
Introduced noms-serve command
2016-05-13 17:14:26 -07:00
Marina Moore b3f647d8f6 Introduced noms-serve command 2016-05-13 17:08:59 -07:00
Mike Gray 54e4c18806 primitives.go split into bool.go and number.go (#1503)
* primitives.go split into bool.go and number.go, interfaces implemented marked, a few .gitignore files updated
* removing old binary names
2016-05-13 19:13:19 -04:00
Aaron Boodman a339168e08 Make sure all our client apps are compiled in production mode. (#1501)
This doesn't seem to make fs fast for some reason, but that might
be unrelated.
2016-05-13 16:02:00 -07:00
Aaron Boodman ec3b3b6bb9 Optimize clients/js/hash-test-rig (#1487)
Massively improve performance of buzhash in JavaScript.

Because of 'let' statements, v8 was not optimizing our code.
2016-05-13 15:04:43 -07:00
mnm678 c380823ac6 Merge pull request #1467 from mnm678/noms-shove
Introduce noms-sync command
2016-05-13 14:35:57 -07:00
Marina Moore 0d3cdf0aeb Introduced noms-sync and deleted shove 2016-05-13 14:22:36 -07:00
Erik Arvidsson 2883609860 Update makeUnionType type to flatten and remove duplicates (#1492)
makeUnionType now flattens and removes duplicate types. With this
change makeUnionType might no longer actually create a union
type in the case there is only one type in the union.

Towards #1491
2016-05-13 09:21:03 -07:00
Dan Willhite 7c119ba447 Merge pull request #1494 from willhite/work
Add stdout-is-tty argument to all noms command line utils
2016-05-12 17:41:57 -07:00
Dan Willhite cd6bb22b64 Add stdout-is-tty argument to all noms command line utils 2016-05-12 17:34:58 -07:00
Dan Willhite ffb312d7b3 Merge pull request #1447 from willhite/work
Provisional noms-log command
2016-05-12 17:09:12 -07:00
Dan Willhite ee7fd768c8 Provisional noms-log command 2016-05-12 17:04:16 -07:00