| Commit message (Collapse) | Author | Age |
... | |
|
|
|
|
|
|
|
|
|
|
| |
Instead, refactor the code to use TargetPatternPhaseValue exclusively. This
removes the need to convert from TargetPatternPhaseValue to LoadingResult, and
prepares for interleaving.
It also reduces the number of Skyframe calls which may speed up null builds a
bit, as a followup for https://github.com/bazelbuild/bazel/commit/1067310e18cb9ac203110726de0be53bdc403cea.
PiperOrigin-RevId: 205989338
|
|
|
|
|
|
|
| |
aren't used anymore.
RELNOTES: None.
PiperOrigin-RevId: 205984908
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For downloading output files / directories we trigger all
downloads concurrently and asynchronously in the background
and after that wait for all downloads to finish. However, if
a download failed we did not wait for the remaining downloads
to finish but immediately started deleting partial downloads
and continued with local execution of the action.
That leads to two interesting bugs:
* The cleanup procedure races with the downloads that are still
in progress. As it tries to delete files and directories, new
files and directories are created and that will often
lead to "Directory not empty" errors as seen in #5047.
* The clean up procedure does not detect the race, succeeds and
subsequent local execution fails because not all files have
been deleted.
The solution is to always wait for all downloads to complete
before entering the cleanup routine. Ideally we would also
cancel all outstanding downloads, however, that's not as
straightfoward as it seems. That is, the j.u.c.Future API does
not provide a way to cancel a computation and also wait for
that computation actually having determinated. So we'd need
to introduce a separate mechanism to cancel downloads.
RELNOTES: None
PiperOrigin-RevId: 205980446
|
|
|
|
|
|
|
|
| |
This makes the tests run much faster.
Fixes #5443
RELNOTES: None.
PiperOrigin-RevId: 205980410
|
|
|
|
| |
PiperOrigin-RevId: 205876673
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Fixed duplicate derived inputs bug. Test is in diffbase.
RELNOTES[INC]: If the same artifact is generated by two distinct but identical actions, and a downstream action has both those actions' outputs in its inputs, the artifact will now appear twice in the downstream action's inputs. If this causes problems in Skylark actions, you can use the uniquify=True argument in Args.add_args.
PiperOrigin-RevId: 205863806
|
|
|
|
|
|
|
|
| |
RELNOTES: java_common.compile creates the native headers jar accesible via JavaInfo.outputs.native_headers.
Closes #5662.
PiperOrigin-RevId: 205832180
|
|
|
|
|
|
|
|
|
|
|
| |
Also simplify LoadingPhaseCompleteEvent, and SkyframeExecutor, and remove
LoadingCallback, which is unnecessary now that we only have a single
implementation (previously LoadingPhaseRunner).
This also removes some of the excessive Skyframe calls introduced by
https://github.com/bazelbuild/bazel/commit/1067310e18cb9ac203110726de0be53bdc403cea, and prepares for interleaving target pattern eval and loading.
PiperOrigin-RevId: 205813197
|
|
|
|
|
|
|
| |
Each FileSystem instance has a digest function, but the getDigest and getHashDigest functions also accepted their own custom parameter functions. We only support 1 hash per filesystem instance, these parameters are redundant.
RELNOTES: None.
PiperOrigin-RevId: 205758571
|
|
|
|
|
|
|
| |
in a very specific window of time inbetween enqueueing one top-level node for evaluation and checking if another top-level node is done. See the added unit test for details.
RELNOTES: None
PiperOrigin-RevId: 205718683
|
|
|
|
|
|
|
|
|
| |
disabling relying on CROSSTOOL file in order to select the cc_toolchain label
CROSSTOOL file should not have any influence over selection of the cc_toolchain label. Ultimately the information that CROSSTOOL offers will be rerouted through an attribute of cc_toolchain.
RELNOTES: None.
PiperOrigin-RevId: 205651369
|
|
|
|
|
|
|
|
| |
Default value is true, and behavior related to //tools/defaults package is not
changed. If set it to false, then in-memory Dfaultpacked will not be created.
RELNOTES:none
PiperOrigin-RevId: 205643628
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 205635805
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With invalid contents in the repository cache, silence the IOException
on RepositoryCache::get and re-download an artifact when attempting to
short-circuit that operation. The repository cache can easily get into
this state when a build is interrupted while downloading into the non-
atomic repository cache destination.
Possible solution to #5390
Closes #5392.
PiperOrigin-RevId: 205634761
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Instead, make ActionMetadataHandler implement the MetadataProvider interface.
This fixes an issue where an action that runs two spawns where one depends on
an output of the other was unable to get the metadata for the intermediate
output.
We don't currently have actions that do this, but we will have in a future
change (which will also implicitly act as a regression test).
PiperOrigin-RevId: 205629237
|
|
|
|
|
|
|
|
| |
It was missing the baseline coverage files, if any.
This is safe even if unknown commit is rolled back.
PiperOrigin-RevId: 205626149
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Update the Flutter rules AndroidSdkInfo provider to FlutterAndroidSdkInfo. AndroidSdkInfo should be unique in the repo now.
*** Original change description ***
Automated rollback of commit 4d10250291a813302de64151be3b22d57e94749d.
*** Reason for rollback ***
AndroidSdkInfo is already being used by the Flutter rules.
*** Original change description ***
Expose AndroidSdkProvider to Skylark (as AndroidSdkInfo).
RELNOTES: None.
PiperOrigin-RevId: 205431461
|
|
|
|
|
|
|
|
| |
contain a subset of the current :rpc3 dependencies.
This will allow the rpc3 server implementation to depend on real java_proto_libraries without a circular dependency.
PiperOrigin-RevId: 205420268
|
|
|
|
|
|
|
| |
RELNOTES[INC]:Labels in C++ rules' linkopts attribute are not expanded anymore
unless they are wrapped, e.g: $(location //foo:bar)
PiperOrigin-RevId: 205385711
|
| |
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 205288166
|
|
|
|
|
|
|
| |
Working towards #4570.
RELNOTES:none
PiperOrigin-RevId: 205274676
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previous behavior was to put flags coming from Bazel option --linkopt into
legacy_link_flags. They should be in user_link_flags instead (together with
flags coming from linkopts rule attribute). This cl introduces
--experimental_linkopts_in_user_link_flags option that flips the behavior.
There is another incompatible change. Previously cc_common.create_link_variables() included flags from --linkopt, with the flag flipped it doesn't anymore. I believe
--linkopt flags shouldn't be there by default because:
* We don't tie the API with the specifics of C++ rules/options, enabling theoretical use with other languages (objc)
* Users are free to use ctx.fragments.cpp to access C++ options and add them explicitly (https://github.com/bazelbuild/bazel/issues/5602)
* New behavior maintains the symmetry with --copt and user_compile_flags
RELNOTES: None.
PiperOrigin-RevId: 205274272
|
|
|
|
| |
PiperOrigin-RevId: 205236169
|
|
|
|
|
|
| |
encountered in a Fileset as a regular file.
PiperOrigin-RevId: 205152271
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
AndroidSdkInfo is already being used by the Flutter rules.
*** Original change description ***
Expose AndroidSdkProvider to Skylark (as AndroidSdkInfo).
RELNOTES: None.
PiperOrigin-RevId: 205142459
|
|
|
|
|
| |
RELNOTES: Fixed compatibility with aar_import when using aapt2. AAPT2 is now supported for Android app builds without resource shrinking. To use it, pass the `--android_aapt=aapt2` flag or define android_binary.aapt_version=aapt2.
PiperOrigin-RevId: 205136160
|
|
|
|
|
| |
RELNOTES: none.
PiperOrigin-RevId: 205100703
|
|
|
|
|
|
|
|
| |
FilesetEntryMap for the aspect's RuleContext.
The current logic in createFilesetEntryMap() uses the rule of the RuleContext.Builder to determine whether to create a filesetEntryMap or not, because of this, applying an aspect to a generated file of a fileset (e.g. 'out' attribute of a Fileset) causes createFilesetEntryMap to create the map even though the associated target is, for instance, a generated file. As a consequence, applying any aspect to a generated file of a fileset causes a NullPointerException.
PiperOrigin-RevId: 205065778
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 205060182
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
[Full NDK changelog](https://github.com/android-ndk/ndk/wiki/Changelog-r17)
Changes related to Bazel's NDK toolchain/CROSSTOOL generation:
- Removed support for targeting mips, mips64, armeabi.
- Bumped bundled clang version to 6.0.2.
- The default STL is now libc++. If Bazel detects that r17 is being used, it
aliases `@androidndk//:default_toolchain` to `@androidndk//:toolchain-libcpp`,
instead of `@androidndk//:toolchain-gnu-libstdcpp`.
- Removed tests that build against armeabi, mips, mips64.
- Stop passing `-fno-integrated-as` into the compiler by default.
TODO:
- [x] Java tests
- [x] ~Support for wrap.sh to use with ASANs~ Implementing in follow up PR.
- [ ] Fix or disable remote caching with android_ndk_integration_test https://github.com/bazelbuild/bazel/issues/4663
Fixes https://github.com/bazelbuild/bazel/issues/4742
RELNOTES: Added support for Android NDK r17. The default STL is now `libc++`, and support for targeting `mips`, `mips64` and `ARMv5` (`armeabi`) has been removed.
Closes #5371.
Change-Id: Id1ded004c3a80ea62f307746dc3ad6f633f2df2e
PiperOrigin-RevId: 205013758
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Still bugs lurking. See linked bug.
*** Original change description ***
Automated rollback of commit eb587075b0d6ffab1cf9e69ede1b7e547905e547.
*** Reason for rollback ***
Depot has been fixed.
RELNOTES[INC]: If the same artifact is generated by two distinct but identical actions, and a downstream action has both those actions' outputs in its inputs, the artifact will now appear twice in the downstream action's inputs. If this causes problems in Skylark actions, you can use the uniquify=True argument in Args.add_args.
PiperOrigin-RevId: 204997569
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 204983634
|
|
|
|
|
| |
RELNOTES: None
PiperOrigin-RevId: 204974291
|
|
|
|
|
| |
Change-Id: I13645199a21991b1458264f9d9ac2db2736066b2
PiperOrigin-RevId: 204963740
|
|
|
|
|
| |
Change-Id: I2195a3f7d60e9835bc00d6f151ea91271fdccf6f
PiperOrigin-RevId: 204954233
|
|
|
|
|
|
|
|
|
|
|
|
| |
This cl adds following flags:
* --experimental_disable_legacy_cc_compilation_api
* --experimental_disable_legacy_cc_linking_api
* --experimental_disable_linking_mode_flags
* --experimental_disable_compilation_mode_flags
* --experimental_disable_legacy_crosstool_fields
RELNOTES: None
PiperOrigin-RevId: 204924599
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
range used to use MutableList which would eagerly allocate an array list with
all range elements, which is not efficient for very large ranges or when only
a small number of its elements are used.
This implementation uses a constant amount of RAM and computes a value for
each requested index.
For the following Skylark snippet:
```
def check_content(t):
if t == []:
return t
return False
def modulo(n):
return n % 797
N = 10000000
[check_content(i) for i in range(N)]
[check_content(i) for i in range(N)]
[modulo(i) for i in range(N)]
[modulo(i) for i in range(N)]
```
the total runtime goes from
```
$ time bazel-bin/src/main/java/com/google/devtools/skylark/Skylark test.bzl
bazel-bin/src/main/java/com/google/devtools/skylark/Skylark test.bzl 93.09s user 1.67s system 316% cpu 29.930 total
```
to
```
$ time bazel-bin/src/main/java/com/google/devtools/skylark/Skylark test.bzl
bazel-bin/src/main/java/com/google/devtools/skylark/Skylark test.bzl 31.45s user 0.86s system 179% cpu 17.974 total
```
which reflects the reduced system time (fewer allocations) and performance.
Closes #5240.
PiperOrigin-RevId: 204918577
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Depot has been fixed.
RELNOTES[INC]: If the same artifact is generated by two distinct but identical actions, and a downstream action has both those actions' outputs in its inputs, the artifact will now appear twice in the downstream action's inputs. If this causes problems in Skylark actions, you can use the uniquify=True argument in Args.add_args.
PiperOrigin-RevId: 204827477
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 204801047
|
|
|
|
|
|
|
| |
parameters of the skylark attr module.
RELNOTES: None.
PiperOrigin-RevId: 204797954
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Tulsi uses BEP json output in its UI, to simulate Bazel terminal outptut. This
means we have to promptly flush the stream. It's sufficient to do flushing at
the granularity of whole events, not any specific count of bytes, since whole
events are what's being consumed by the user.
To balance IO throughput and interactivity, let's flush at a regular
sub-second interval.
(The alternative solution of using a stream with smaller buffer could still
end up with small-sized event descriptions buffered arbitrarily long; and
abandoning buffering altogether would be suboptimal for throughput when
writing a long sequence of small build events.)
RELNOTES: None.
PiperOrigin-RevId: 204790794
|
|
|
|
|
|
|
| |
Temporary solution finding the SkylarkCallable fields using reflection,
mirroring the existing handling of ClassObject.
PiperOrigin-RevId: 204779295
|
|
|
|
|
|
|
|
|
|
| |
referencing the main repository using its name caused bazel to treat it as a separate external repository.
Closes #5586.
Fixes #3115.
RELNOTES: None
PiperOrigin-RevId: 204752150
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
non-hermetic events that happen as part of repository rules).
Defining representation for Execute events for workspace logging.
In the future:
- Add more events
- Allowing to specify log file rather than dumping to INFO
- Log levels, full or alerts only
RELNOTES: None
PiperOrigin-RevId: 204748436
|
|
|
|
|
|
|
|
|
|
|
| |
This was providing runtime libraries. This provider is redundant because the
same libraries are in CcLinkParamsStore.
Note to sheriff: Shouldn't break anything. If any, breakages would most likely be in Go test rules. I will look at the nightly tomorrow anyway. The breakages in the linked TGP re-run are not caused by this change, later re-runs have unrelated changes.
TESTED=[]
RELNOTES:none
PiperOrigin-RevId: 204738136
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- CcLinkingInfo has a getter for CcLinkParams.
- Artifacts and linkopts of CcLinkParams are accessible.
- CcCompilationInfo constructor now accepts defines and include dirs.
- CcCompilation now has getters for headers, defines and include dirs.
CcCompiilationInfo(
headers=depset([Artifacts]),
defines=depset([Strings]),
include_dirs=depet([Strings])
CcCompilationInfo.headers
CcCompilationInfo.defines
CcCompilationInfo.include_dirs
--
cc_common.merge([CcLinkingInfos]) returns CcLinkingInfo
--
CcLinkingInfo.static_shared_params
CcLinkingInfo.static_no_shared_params
CcLinkingInfo.no_static_shared_params
CcLinkingInfo.no_static_no_shared_params
--
CcLinkParams.libraries
CcLinkParams.linkopts
CcLinkParams.dynamic_libraries_for_runtime
RELNOTES:none
PiperOrigin-RevId: 204700779
|
|
|
|
|
|
|
| |
are difficult or impossible to verify by annotation processor.
RELNOTES: None.
PiperOrigin-RevId: 204540521
|
|
|
|
|
|
|
|
|
| |
files which define new provider types and instantiate Labels.
Also improve docstring usage in attr definitions by doing whitespace trimming.
RELNOTES: None.
PiperOrigin-RevId: 204537623
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 204505003
|