| Commit message (Collapse) | Author | Age |
|
|
|
|
| |
--
MOS_MIGRATED_REVID=134823021
|
|
|
|
|
|
|
| |
It doesn't make sense to split a host Android rule's deps down fat APK settings.
--
MOS_MIGRATED_REVID=134814698
|
|
|
|
|
|
|
|
| |
Crash was triggered by overriding a @SkylarkCallable without repeating the
annotation in the subclass's method.
--
MOS_MIGRATED_REVID=134797463
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=134793032
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The BuildEventStreamer will listen for BuildEvents and stream them
to the provided transports. It also ensures events are properly chained:
for unsolicited events, it will add progress events to chain them and
at the end of a build it closes all announced but not produced events
as aborted.
--
Change-Id: I623b582657573fe1288821c96f084e0ab0bca4d4
Reviewed-on: https://bazel-review.googlesource.com/#/c/6275
MOS_MIGRATED_REVID=134787541
|
|
|
|
|
|
|
|
|
| |
Default is off (using ApkBuilderMain still).
Note that --use_singlejar_apkbuilder is incompatible with --apk_signing_method=legacy_v1 (but compatible with --apk_signing_method={v1,v2,v1_v2}).
--
MOS_MIGRATED_REVID=134783504
|
|
|
|
|
|
|
|
|
|
|
|
| |
This ensures that stale outputs from prior builds do not remain
to confuse the compiler.
Fixes #1778.
--
Change-Id: I31b5c3e7e5970cf45c3ff10144ddfc73540ef9af
Reviewed-on: https://bazel-review.googlesource.com/6250
MOS_MIGRATED_REVID=134780501
|
|
|
|
|
|
|
|
|
| |
(which results in single-character gRPC messages).
Fixes #1861.
--
MOS_MIGRATED_REVID=134775680
|
|
|
|
|
|
|
|
| |
Closes #1633.
--
Reviewed-on: https://github.com/bazelbuild/bazel/pull/1633
MOS_MIGRATED_REVID=134774446
|
|
|
|
|
|
|
| |
objc_library.
--
MOS_MIGRATED_REVID=134773363
|
|
|
|
|
|
|
|
|
|
|
|
| |
This restructures the way we set the default Spawn strategy so that each BlazeModule supplying a SpawnActionContext has an ActionContextConsumer that sets its own SpawnActionContext as the default, with the BazelRulesModule being put as the last module loaded in BazelMain, so that it can override that decision - it only does, if the user explicitly specifies a --spawn_strategy flag. IMHO this is a much saner approach than the older one.
So the flow is essentially this:
- StandaloneActionContextConsumer sets the default strategy to "standalone".
- SandboxActionContextConsumer sets the default strategy to "sandboxed", but only on Linux
- BazelRulesModule sets the default strategy to the value of the --spawn_strategy flag, if it is set.
--
MOS_MIGRATED_REVID=134770427
|
|
|
|
|
|
|
| |
Fixes #1863
--
MOS_MIGRATED_REVID=134767527
|
|
|
|
|
|
|
|
|
| |
simple synchronous cleanup.
Tested with bazel building itself that this does not result in a performance degradation.
--
MOS_MIGRATED_REVID=134766597
|
|
|
|
|
|
|
|
|
|
|
|
| |
Also:
- Move minimum os information into AppleConfiguration
- Update watchos and tvos minimum version defaulting behavior
such that, if unspecified in flags, the minimum OS is equal
to the SDK version.
- Expose Apple Platform to apple_common in skylark
--
MOS_MIGRATED_REVID=134712763
|
|
|
|
|
|
|
| |
RELNOTES: apple_static_library rule to create multi-architecture static archive files from Objc/C++/Swift dependencies on apple platforms
--
MOS_MIGRATED_REVID=134701599
|
|
|
|
|
|
|
|
| |
Tested with
bazel build --incremental_dexing_binary_types=monodex,multidex_unsharded,multidex_sharded -- //examples/android/java/bazel:hello_world
--
MOS_MIGRATED_REVID=134690103
|
|
|
|
|
|
|
| |
new AIDL files in relation to their location.
--
MOS_MIGRATED_REVID=134687843
|
|
|
|
|
|
|
|
|
|
| |
Profiling shows this shaves off 8% of getDynamicConfigurations' CPU time.
And brings down AttributeAndLabel instantiation from 30% of that time to 24%.
Over a simple cc_binary, this reduces the number of Objects.hash calls
by 96%.
--
MOS_MIGRATED_REVID=134687748
|
|
|
|
|
|
|
| |
The new provider gathers actions generated by any Skylark-based RuleConfiguredTarget, so long as the rule definition has _skylark_test=True set. For the moment this flag is under the user's control, but the intention is that it will be set by a test runner.
--
MOS_MIGRATED_REVID=134687396
|
|
|
|
|
|
|
|
| |
Usually an OutputStream will do. Forgo the extra layer of indirection and
stream directly to the output.
--
MOS_MIGRATED_REVID=134682243
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
AttributeAndLabel instantiations. This shaves 25% off the method's execution time in --experimental_dynamic_configs=notrim mode.
This isn't a crucial optimization, since getDynamicConfigurations is already fast. Profiling:
$ bazel shutdown; bazel build --nobuild //testapp:cc
--experimental_dynamic_configs=notrim
shows getDynamicConfigurations takes 1.3% of the analysis phase's CPU time before this change and 1.0% after. Before this change, AttributeAndLabel instantiation took 44% of the method's CPU time. After: 30%.
Profiling done with JProfiler.
--
MOS_MIGRATED_REVID=134677107
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
To organize a clean build event stream, certain infrastructural events
are necessary.
- If the build is aborted prematurely, certain expected events will not happen;
hence we have to close those leafs with an event indicating that the actual event
did not happen, as the build was aborted.
- Occasionally, we update on progress of the build, using special events; those are
also used to chain failure events that come before the build events for which they
are root cause are reported.
Add those.
--
Change-Id: Ibf86a313eb678822754b5e1f1e9d7f1a2cbeaaef
Reviewed-on: https://bazel-review.googlesource.com/#/c/6274
MOS_MIGRATED_REVID=134667915
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=134652924
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=134651940
|
|
|
|
|
|
|
|
|
|
| |
This will allow a precise, but still human-readable, transcript of the
build-event protocol buffers generated.
--
Change-Id: I6284521f652b04ab6bf54e5eeed7b92f2d7e2245
Reviewed-on: https://bazel-review.googlesource.com/#/c/6273
MOS_MIGRATED_REVID=134637366
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bazel's dep configuration creation logic.
This essentially implements the following TODOs:
https://github.com/bazelbuild/bazel/blob/bc6045dcc8fa33d4241d231138020ac4bdecc14f/src/main/java/com/google/devtools/build/lib/skyframe/ConfiguredTargetFunction.java#L599
https://github.com/bazelbuild/bazel/blob/bc6045dcc8fa33d4241d231138020ac4bdecc14f/src/test/java/com/google/devtools/build/lib/skyframe/ConfigurationsForTargetsTest.java#L42
--
MOS_MIGRATED_REVID=134607049
|
|
|
|
|
|
|
| |
shared naive BFS implementation. Also implement RegexFilterExpression#parEval.
--
MOS_MIGRATED_REVID=134598046
|
|
|
|
|
|
|
|
| |
* When set, this flag will cause any objc_library or objc_binary rule with
Swift sources to error out. The native functionality is deprecated.
--
MOS_MIGRATED_REVID=134583636
|
|
|
|
|
|
|
|
|
|
|
| |
Instead have SkyQueryEnvironment#evaluateQuery be responsible for handling cleanup of its internal ForkJoinPool.
In addition to being a more sensible way of organizing the code (imo, it makes sense for SkyQueryEnvironment to clean up after itself), this fixes several issues:
(i) If query evaluation is interrupted, the AbstractBlazeQueryEnvironment#close call at the end of the try-with-resources statement in QueryCommand would be blocking and non-urgent. N.B. This was not an issue with the current ForkJoinPool usage, but was an issue with the old ThreadPoolExecutor usage.
(ii) Because of how the code in QueryCommand was structured, OutputFormatterCallback#close would happen _before_ the AbstractBlazeQueryEnvironment#close call. If query evaluation is interrupted, threads executing query tasks may be invoking the callback after the callback had been shut down!
--
MOS_MIGRATED_REVID=134573395
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=134566270
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=134540479
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=134537085
|
|
|
|
|
|
|
|
|
|
|
| |
My previous change carefully checked that the file was a symlink before
removing it and added a test with local repositories... and it of course isn't
a symlink with downloaded repositories and crashes.
Fixes #1697.
--
MOS_MIGRATED_REVID=134536130
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Update `select` statements in BUILD files with the
new config_setting.
This is a first step on a long path that leads to
us being able to compile bazel on Windows with
--cpu=x64_windows_msvc. Needless to say, we're not
there yet.
Tested: on Linux, Darwin, Windows/MSYS
--
MOS_MIGRATED_REVID=134534613
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Before this change, a non-trivial real-world build was measured to have about 20% analysis time overhead. After, that's theoretically down to 2.4%. But that's only the analysis phase, so the impact on full builds is smaller. And the impact on analysis-cached builds is zero. And practical tests show no obvious difference (JProfiler is probably overstating the impact since it excludes known heavyweight methods).
The improvements, in short:
- Optimize a sanity check that expects each <Attribute, Label> pair to only have one transition. This alone was over half the original performance penalty.
- Simplify logic for null configuration deps (of which there are many, e.g.: all source files)
- Skip a check for required fragments not available in the configuration. This is irrelevant for notrim mode.
There are still some places we could optimize. Dependency.withNullConfiguration in particular takes measurable time (I think from being constructed all the time and in its hashCode calls). But this doesn't seem pressing given the new numbers.
--
MOS_MIGRATED_REVID=134533452
|
|
|
|
|
|
|
|
|
| |
empty, writable directory at a specified path when running actions. (Supported on Linux only for now.)
RELNOTES: Added a new flag --sandbox_tmpfs_path, which asks the sandbox to mount an empty, writable directory at a specified path when running actions. (Supported on Linux only for now.)
--
MOS_MIGRATED_REVID=134526345
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=134523222
|
|
|
|
|
|
|
|
|
|
|
| |
Bazel in the will provide a machine-readable stream of important build
events. These interfaces set up the framework and expectations about
the produced events and the entities distributing those events.
--
Change-Id: If2c3b2e11c31b0136b57eadeef2d2f8f8fe5e2e7
Reviewed-on: https://bazel-review.googlesource.com/#/c/6272
MOS_MIGRATED_REVID=134522369
|
|
|
|
|
|
|
|
|
| |
command line labels" sanity check.
It's only used for --libc_top and makes configuration creation much more heavyweight that it would otherwise be.
--
MOS_MIGRATED_REVID=134513058
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
of a SkylarkCallable.
CL already reviewed in commit 5972bee6ebfa53cf165befab9fa7962e19d5f620.
Rolled back due to casting error in Java 1.7. This is fixed now:
https://github.com/bazelbuild/bazel/issues/1832
Old:
matchingMethod =
new Pair<>(
method, argumentListConversionResult.getArguments());
New:
matchingMethod =
new Pair<MethodDescriptor, List<Object>>(
method, argumentListConversionResult.getArguments());
--
MOS_MIGRATED_REVID=134503884
|
|
|
|
|
|
|
|
| |
It's possible that there are multiple outstanding interrupts. When this
happens avoid interleaving their stack traces.
--
MOS_MIGRATED_REVID=134477540
|
|
|
|
|
|
|
| |
artifacts from AndroidSemantics.
--
MOS_MIGRATED_REVID=134475699
|
|
|
|
|
|
|
|
|
| |
action sees a different shared action that is in the midst of being an action cache hit.
Concretely, suppose that the user builds A, then requests C and D. C depends on A, D depends on B, which is a shared action with A. If B executes at just the right time, as C is finishing execution, C can think that it must add itself to B's execution path, which is incorrect.
--
MOS_MIGRATED_REVID=134475095
|
|
|
|
|
|
|
| |
in android_* targets (behind existing flag).
--
MOS_MIGRATED_REVID=134453743
|
|
|
|
|
|
|
| |
false.
--
MOS_MIGRATED_REVID=134452391
|
|
|
|
|
|
|
|
|
| |
This entails adding a read-locking mechanism to Mutability contexts.
RELNOTES[INC]: Updating list/dicts while they are being looped over is not allowed. Use an explicit copy if needed ("for x in list(mylist):").
--
MOS_MIGRATED_REVID=134442701
|
|
|
|
|
|
|
|
|
|
| |
Of course, the sources of the buildevent protos should be visible by
the including srcs targets.
--
Change-Id: I76f5155990dd767d5eb46ad7150980589050217e
Reviewed-on: https://bazel-review.googlesource.com/#/c/6290
MOS_MIGRATED_REVID=134414797
|
|
|
|
|
|
|
|
|
|
| |
EvaluationProgressReceiver objects have two common naming schemes
currently, and calling them invalidationReceiver is misleading, so to
make the naming convention standard, all object names are based on
"progressReceiver."
--
MOS_MIGRATED_REVID=134411011
|
|
|
|
|
|
|
| |
protos from linking into the final binary.
--
MOS_MIGRATED_REVID=134409481
|
|
|
|
|
|
|
|
|
| |
PackageCacheOptions as a bundle.
This will allow me to thread a new flag through more easily in the future.
--
MOS_MIGRATED_REVID=134406676
|