Since commit b6a5382217 (Ninja: depend on language module information
files directly, 2023-02-10, v3.27.0-rc1~502^2), the return value of
`cmCommonTargetGenerator::GetLinkedTargetDirectories` must account for
linked object libraries because they may provide modules (#25112).
These were added by commit b665966933 (cmComputeLinkInformation: track
OBJECT library dependencies, 2023-07-22, v3.27.1~5^2). However, targets
named by `$<TARGET_OBJECTS:...>` sources are also needed (#25365).
The latter were added by commit 22da18b995 (Fortran: Restore support for
TARGET_OBJECTS providing modules, 2023-10-27, v3.28.0-rc4~9^2) and
commit 035302b7e3 (cmComputeLinkDepends: also copy the target from
object link items, 2023-10-27, v3.28.0-rc4~9^2~2). However, their
approach added link entries not actually specified by projects. It also
incorrectly re-used `cmComputeLinkDepends::AddLinkObject` for object
library targets when it is meant for their individual object files.
These problems caused additional regressions (#25417). Revert the
implementation parts of those commits and leave behind an assertion and
comment to help avoid the mistake in the future. Instead, track targets
named by `$<TARGET_OBJECTS:...>` sources with a dedicated member.
Issue: #25112
Issue: #25365Fixes: #25417
Co-authored-by: Ben Boeckel <ben.boeckel@kitware.com>
Continue b665966933 (cmComputeLinkInformation: track OBJECT library
dependencies, 2023-07-22) which added explicitly listed `OBJECT`
libraries to the list of targets which the collator needs to consider.
Now also consider targets which provide objects directly to the target
via a `$<TARGET_OBJECT>` source lists.
Also add tests which use target objects directly and through an
`INTERFACE` library with target objects in its own sources.
Fixes: #25365
Since refactoring in commit 0a5aeaf302 (cmCTestRunTest: Consolidate test
timeout selection logic, 2023-05-04, v3.27.0-rc1~120^2) we accidentally
truncate `--timeout` values to ctest's default `TimeOut`. Fix the
logic to prefer the flag whenever the `TIMEOUT` property is not set.
In combination with the prior refactoring, this also fixes a bug that
caused `--timeout` values of 10000000 seconds or more to be ignored.
Fixes: #23979
Update logic added by commit 099934e313 (Add generator expression
support to static code analysis hooks, 2023-03-24, v3.27.0-rc1~261^2),
and preserved by commit 993dde925f (TargetGenerator: Factor out
generation of code check rules, 2023-05-12, v3.27.0-rc1~84^2~2), to
handle a generator expression that evaluates to the empty string.
If `<LANG>_CPPCHECK`, `<LANG>_CPPLINT`, `<LANG>_CLANG_TIDY`, or
`<LANG>_INCLUDE_WHAT_YOU_USE` are empty after evaluating generator
expressions, do not run their lints.
Fixes: #25265
Since commit 74b1d6caf3 (cmComputeLinkInformation: compute link info for
module-using targets, 2023-09-05, v3.27.5~7^2) we accidentally try to
compute link information for custom targets if they have Fortran
sources. For module dependencies, we only need to consider target types
that can compile.
Fixes: #25252
Previously an internal error was raised which ended up causing an
internal exception to be thrown. This is a typo situation that should
fall into an explicit error.
Fixes: #25207
In commit c5c3aff1f5 (Autogen: Add INTERFACE_AUTOMOC_MACRO_NAMES target
property, 2023-04-03, v3.27.0-rc1~197^2) we forgot to mark the property
as a usage requirement. This is needed for efficient evaluation over
the target dependency closure.
Fixes: #25238
Without `CudaLink` fields, MSBuild tries to perform a device link step
for object libraries with CUDA_SEPARABLE_COMPILATION on them even when
not needed. This commonly was encountered when generating PTX or
OPTIX-IR output.
After CDash PR 1519 introduced HTTP status codes greater than 200 for
various error cases, CMake commit b7c871f745 (ctest: Update ctest_submit
for CDash behavior change, 2023-07-24, v3.27.1~3^2) modified CTest's
submit handler to check the status returned by CDash and throw an error
when this status is not equal to 200.
That change had the unintended side effect of causing CTest submissions
to fail when uploading results to a URL that returns a redirect status
code (3xx). Fix this by configuring cURL to follow the redirect. The
status cURL reports to CTest is now 200 instead of 3xx when CDash is
located behind a redirect.
Fixes: #25159
Targets which contain C++ module or Fortran sources need to participate
in link information unconditionally regardless of whether they actually
have link artifacts or not.
Fixes: #25223
Extend commit 78bbd58545 (Source: Link libatomic when needed on Linux
32-bit ARM, 2023-07-27, v3.27.2~10^2) to check for libatomic on more
architectures.
Fixes: #25204
Update the comment added by commit 485256f2d0 (Unity: Suppress
clang-tidy 'bugprone-suspicious-include' warning, 2023-04-27,
v3.27.0-rc1~149^2) to work for C.
Fixes: #25184
The policy added by commit f90c8ab54e (VS: Select latest available
Windows SDK version by default, 2023-04-03, v3.27.0-rc1~206^2~1) applied
only when targeting Windows 10+. Apply it to older versions too.
Fixes: #25170
Issue: #16202
Make logic choosing between Windows 10 SDKs and the Windows 8.1 SDK
easier to follow by consolidating it in the VS 14 generator. The only
information we need from VS 15+ generators is whether the 8.1 SDK is
installed.
Refactoring in commit bff468c988 (cmFileCommand: Use cm::optional for
keyword argument presence, 2022-06-30, v3.25.0-rc1~512^2) accidentally
broke the check that the input argument is either `INPUT` or `CONTENT`.
The check is supposed to fail when arguments are passed in the wrong
order. For example:
file(GENERATE OUTPUT ...
TARGET <target>
CONTENT <content>)
Prior to this fix, the input method would be CONTENT, but because the
first parsed keyword is not `CONTENT`, `inputIsContent` would be false.
The first parsed keyword isn't INPUT either, so we would not continue
into the error condition. CMake would then try to handle this as an
input file, when there isn't one, resulting in uninitialized memory
usage and segfaults or corruption later on.
Fixes: #25169
Since commit df58dbb0e9 (VS: Add CMake input files to ZERO_CHECK,
2023-03-19, v3.27.0-rc1~157^2), projects that specify the same file both
as input to `configure_file` and as the `MAIN_DEPENDENCY` of a custom
command fail to configure. Revert the change pending further
investigation. Add a test case demonstrating the problem.
Issue: #24557Fixes: #25149
Fix a typo from commit e522f8ca06 (cmTarget: factor out properties
initialized for target which compile, 2023-02-01, v3.27.0-rc1~577^2~13),
that broke this property.
Also update the test suite to use this upgrade as some level of
coverage.
Fixes: #25123
open.cdash.org was recently upgraded in preparation for the release of
CDash v3.2.0. This upgrade brought a change in behavior where CDash now
responds with HTTP 400 (bad request) rather than HTTP 200 (OK) for MD5
checksum mismatches during submission time.
This commit removes our usage of CURLOPT_FAILONERROR in
cmCTestSubmitHandler.cxx This was necessary to pass along the contents
of the request body ("md5 mismatch") in cases where the web server
(CDash) responds with an error status (400).
Fixes: #25121
In commit b6a5382217 (Ninja: depend on language module information files
directly, 2023-02-10), introduced via !8197, language-specific module
information files (`CMakeFiles/<target>.dir/<lang>Modules.json`) files
were added as real dependencies to the dyndep collation steps.
Previously, the behavior was to inform the collator of all possible
targets and search for the files manually ignoring those which did not
exist with ordering enforced by depending on the linker output of all
dependent targets. This behavior could lead to stale information being
used (e.g., if a target stops providing any targets) and also did not
reliably build everything needed on rebuilds. Afterwards, the internal
computation changed the dependency from all possible targets to an exact
set of "these targets might have modules" query, however one that did
not include `OBJECT` libraries since do not have `LinkEntry` items
internally (their objects are instead treated as source files).
As a stopgap measure, track `OBJECT` libraries in a separate list and
query them explicitly when gathering targets which may have interesting
information. Future work can add `LinkEntry` items to represent these
targets once all `LinkEntry` consumers have been audited to make sure
they are not surprised by any `OBJECT` library entries.
Fixes: #25112
Inno Setup implicitly creates three installation types if none is
specified in the script. This causes some component features (e.g.
`REQUIRED`) to lose their functionality. Teach the generator to always
specify a "custom" installation type when using components.
Fixes: #25083
