| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
| |
unwrapped ObjcProvider." after skylark rule fixes.
RELNOTES: None.
PiperOrigin-RevId: 163121697
|
|
|
|
|
|
|
| |
Fixes https://github.com/bazelbuild/continuous-integration/issues/103.
Change-Id: Icbf62514baf0f581ca2bc5c115da1f3633ef7828
PiperOrigin-RevId: 163117605
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Generally speaking, it should be as easy as possible to create custom
rule classes in tests. Aside from the tests that *need* this, this also
helps disentangle traditional tests from the implementation logic
of predefined rules.
This nudges us further forward toward the paradigm of Bazel as "build
execution engine" with all rule implementations (and tests) in
Skylark.
PiperOrigin-RevId: 163114414
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 163114020
|
|
|
|
|
|
|
| |
The getMetadata call implicitly adds files to the metadata handler if they weren't known before,
which may include statting the file.
PiperOrigin-RevId: 163106987
|
|
|
|
|
|
|
| |
Also fix minor bug where a[b] += c would evaluate b twice.
RELNOTES: None
PiperOrigin-RevId: 163103618
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Breaks targets in the depot. http://b/64022432
*** Original change description ***
Expose to Skylark Strings instead of PathFragments from CppConfiguration.
Fixes #2931.
RELNOTES: None.
PiperOrigin-RevId: 163100552
|
|
|
|
|
|
|
|
|
| |
Currently, allowable resource directories are defined in two places:
LocalResourceContainer, and android_ide_common (used, for example, by
ResourceFilter). Instead, keep them defined only in android_ide_common, a more authoritative source that we already rely on anyway.
RELNOTES: none
PiperOrigin-RevId: 163100497
|
|
|
|
|
|
| |
CPU constraints.
PiperOrigin-RevId: 163100038
|
|
|
|
|
|
|
|
| |
Documentation was not being generated for the Apple "platform" module because it collided with a configuration fragment of the same name.
This should be a harmless change because the underlying module name should not be used by anyone; the name isn't exported as a top-level type and would only be seen in debug output and in the string returned by the type() function.
PiperOrigin-RevId: 163079704
|
|
|
|
|
|
|
| |
RecursivePackageProvider dealing with the concept of "excluded directories".
RELNOTES: None
PiperOrigin-RevId: 163074794
|
|
|
|
|
|
|
| |
Also, a few small cleanups of some duplicate test methods.
RELNOTES: None
PiperOrigin-RevId: 163066349
|
|
|
|
|
|
|
|
|
|
|
| |
finished extracting the FDO profile.
Also fix a truly embarrassing infinite recursion bug introduced by Yours Truly in unknown commit .
This avoids a failure mode where, when two profiles at the same path are used in two builds close one after the other, the file handle would get erroneously re-used.
RELNOTES: None.
PiperOrigin-RevId: 163063976
|
|
|
|
|
|
| |
Also fix two warnings about missing generic types.
PiperOrigin-RevId: 163062670
|
|
|
|
|
|
|
|
|
| |
This and further changes may contain minor modifications to BUILD files that
don't serve any apparent purpose. The reason for these changes is that we're
switching from checked-in BUILD files to generated BUILD files, and there may
be small differences between these files.
PiperOrigin-RevId: 163062395
|
|
|
|
|
|
|
| |
Fixes https://github.com/bazelbuild/bazel/issues/1906
RELNOTES: none
PiperOrigin-RevId: 163059900
|
|
|
|
|
|
| |
TESTED=ran the test in a loop and with runs_per_test=100
RELNOTES: fixes #3421
PiperOrigin-RevId: 163059760
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Broke bazel_apple_rules
*** Original change description ***
Make all WithLegacySkylarkName providers declared providers.
RELNOTES: None
PiperOrigin-RevId: 163054821
|
|
|
|
| |
PiperOrigin-RevId: 163053841
|
|
|
|
|
|
| |
Closes #3444.
PiperOrigin-RevId: 163047183
|
|
|
|
|
|
|
| |
Almost all implementations simply return this, all of which can be removed
now.
PiperOrigin-RevId: 163046912
|
|
|
|
|
| |
RELNOTES: None
PiperOrigin-RevId: 163042362
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
the configuration.
This is part of the crusade to kill package loading in configuration creation.
AppleConfiguration still retains the version information for the following reasons:
- Some Skylark rules access it that way. The plan is to create an xcode_config_alias() rule that can be an implicit dependency of Skylark rules and to export the version information to Skylark from XcodeVersionsProvider.
- The minimum OS version is part of the name of the output directory. The plan for that is simply to remove it. Hopefully we can get away with it.
- They allow us to add assertions to make sure the new mechanism works like the old one.
RELNOTES: None.
PiperOrigin-RevId: 163038724
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a method to test if a path is /dev/null (or
case-insensitive "NUL" on Windows), and use it in
blaze::MakeAbsolute.
In this commit:
- implement blaze_util::IsDevNull with POSIX and
Windows semantics + add tests
- blaze::MakeAbsolute calls blaze::ConvertPath
on its input to convert MSYS paths on Windows
- blaze_util::GetCwdW (thus GetCwd) always returns
a lowercase path so that it is deterministic
- blaze_util::GetCurrentDrive returns lowercase
letter to be consisent with blaze::ConvertPath,
which also returns a lowercase path
Fixes https://github.com/bazelbuild/bazel/issues/3440
Change-Id: I3af5ba0a033d542fe64a676d67f27472298d1089
PiperOrigin-RevId: 163038503
|
|
|
|
|
|
|
| |
on Apple rules.
RELNOTES: None.
PiperOrigin-RevId: 163035922
|
|
|
|
|
|
|
|
|
|
|
|
| |
SKIP_KOKORO: Kokoro is out of quota.
*** Reason for rollback ***
Causes memory regression, somehow: b/63934093
*** Original change description ***
PiperOrigin-RevId: 163023580
|
|
|
|
|
|
|
|
|
| |
This is a temporary flag that causes Extension objects to elide some built-in symbols, for the benefit of tests. Currently it only applies to the "native" builtin for Extensions created by SkylarkImportLookupFunction.
A future Skylark language change will have the effect of excluding all builtin symbols from all Extension objects anyway, so this flag will eventually be subsumed by that incompatible change.
RELNOTES: None
PiperOrigin-RevId: 163016504
|
|
|
|
|
| |
RELNOTES: none
PiperOrigin-RevId: 162997060
|
|
|
|
|
|
| |
to indicate unserializability, improve error message in Path when serialization fails, and add some test-only methods to SkyframeExecutor and PackageFactory.
PiperOrigin-RevId: 162993806
|
|
|
|
|
|
| |
necessary, the reset occurs before the command starts.
PiperOrigin-RevId: 162973194
|
|
|
|
|
|
|
| |
It now exists as a separate project at
https://github.com/bazelbuild/migration-tooling.
PiperOrigin-RevId: 162970509
|
|
|
|
|
|
|
| |
keywords.
Change-Id: I8421e9d2d8a5e6720d7b8d6de4417bee71c0fa68
PiperOrigin-RevId: 162966141
|
|
|
|
|
|
|
|
|
|
| |
the owner of the method reference have assignable relation. If yes, use the
bridge method.
This CL addresses the integration problem between Desugar and the lambda factory of JDK 9. The change in JDK 9 is here,http://hg.openjdk.java.net/jdk9/dev/jdk/rev/a3b3c7b6464d.
RELNOTES: None
PiperOrigin-RevId: 162965244
|
|
|
|
|
|
|
| |
Fixes TODO in AndroidRuleClasses.java.
RELNOTES: None
PiperOrigin-RevId: 162955052
|
|
|
|
| |
PiperOrigin-RevId: 162949753
|
|
|
|
|
| |
RELNOTES: None
PiperOrigin-RevId: 162945309
|
|
|
|
|
| |
RELNOTES: nocopts now also filter copts
PiperOrigin-RevId: 162944771
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Please look to page:
Note "Target" link.
Actual: https://docs.bazel.build/versions/master/skylark/lib/lib/Target.html
`404 - Page not found
$ bazel build :what-you-were-looking-for
...............
ERROR: no such page ':what-you-were-looking-for': BUILD file not found on package path.
INFO: Elapsed time: 0.567s`
Expected: https://docs.bazel.build/versions/master/skylark/lib/Target.html
Closes #3439.
PiperOrigin-RevId: 162923519
|
|
|
|
|
|
|
|
|
|
| |
(https://github.com/bazelbuild/bazel/blob/master/tools/cpp/lib_cc_configure.bzl#L106-L107)), forcing bazel to use the `SingleJar_deploy.jar` java implementation of SingleJar fixes this issue.
I suspect 32-bit Unix has the same issue, which could be fixed by adding the same configuration for `piii`, but I don't have a 32-bit x86 system to test it on. But if anyone is here by Googling the above error code, try replacing all the `arm` references in this pull request with `piii`!
Closes #3382.
PiperOrigin-RevId: 162923371
|
|
|
|
|
|
|
|
| |
Incorrect instruction for OSX (apt-get instead of brew)
Closes #3437.
PiperOrigin-RevId: 162923095
|
|
|
|
|
|
| |
Change the return type for RPC PublishBuildToolEventStream from OrderedBuildEvent to PublishBuildToolEventStreamRequest.
Change-Id: I8421e9d2d8a5e6720d7b8d6de4417bee71c0fa68
|
|
|
|
|
|
|
|
|
|
| |
When using LLVM LIPO is automatically mapped to ThinLTO.
While for LIPO, fully dynamic linking needs to be off for correctness issues,
there is no such restriction for ThinLTO.
RELNOTES[NEW]: Do not disable fully dynamic linking with ThinLTO when invoked via LIPO options.
PiperOrigin-RevId: 162918429
|
|
|
|
|
|
|
| |
Also change it to java.time.Duration, rather than Jodatime. Now that we're on
Java 8, we no longer need Jodatime.
PiperOrigin-RevId: 162917526
|
|
|
|
|
|
|
| |
It's bad to crash with an exception, let's show proper rule error instead.
RELNOTES: None.
PiperOrigin-RevId: 162916138
|
|
|
|
| |
PiperOrigin-RevId: 162915070
|
|
|
|
|
|
|
|
| |
After the ByteUploader changes, the Retrier no longer needs the onFailure(s) functions. Removing them will simplify both the code and the stack traces used for debugging problems.
TESTED=unit tests
RELNOTES: none
PiperOrigin-RevId: 162913762
|
|
|
|
|
|
|
| |
Fixes #3183.
Change-Id: Id208e0b723f0ffe8313cae4852dd77df43be477f
PiperOrigin-RevId: 162911817
|
|
|
|
|
|
|
|
|
| |
In this way, for each event there can only be two event payloads
for every event identifier: the expected one, and the Aborted event.
The type of failure will be reported in the AbortReason enum value.
Change-Id: Icce225ec156d1fc20c8facd18df13343712f775c
PiperOrigin-RevId: 162911732
|
|
|
|
|
|
| |
This removes a bunch of code duplication that I previously introduced.
PiperOrigin-RevId: 162909430
|
|
|
|
|
|
|
| |
Fixes #2931.
RELNOTES: None.
PiperOrigin-RevId: 162907348
|