| Commit message (Collapse) | Author | Age |
... | |
|
|
|
|
|
|
|
|
| |
This prevents the case of a rule acquiring the null configuration in case the alias points to an input file. Which, in turn, makes "bazel build" work for these targets. The reason it breaks is that TargetCompletionValue instances are created from the associated ConfiguredTarget in SkyframeExecutor#buildArtifacts(), which means that if the configurations do not match, TargetCompletionFunction requests a different ConfiguredTargetValue than it was created from.
Fixes #1297.
--
MOS_MIGRATED_REVID=122973526
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122889821
|
|
|
|
|
|
|
|
|
| |
fragment was requested in the NullPointerException message.
This helps diagnose dynamic configuration errors.
--
MOS_MIGRATED_REVID=122855814
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122846987
|
|
|
|
|
|
|
|
|
| |
RuleClass is deserialized as part of a Skylark rule, the Environment is currently not present, but it is needed to detect changes to the rule.
Also precompute and store the Environment's hash code, and do a drive-by clean-up of a bunch of warnings in the Environment code.
--
MOS_MIGRATED_REVID=122838588
|
|
|
|
|
|
|
| |
allow for the exposure of values that are not typed as skylark primatives.
--
MOS_MIGRATED_REVID=122838580
|
|
|
|
|
|
|
|
|
|
|
|
| |
This change allows starting remote worker with an embedded hazelcast server that listens to
a specific port. This allows test to run reliably.
Fixes #1271.
--
Change-Id: I0bb36af5837f83cad3d76b8acb50f89cd599ee87
Reviewed-on: https://bazel-review.googlesource.com/c/3650/
MOS_MIGRATED_REVID=122829898
|
|
|
|
|
|
|
|
|
|
|
|
| |
This required a few assorted changes:
- The FDO build stamp is not special-cased anymore, it is treated as a preprocessor define like any other.
- When compiling a .pcm file, use interfaceContext instead of the regular context when setting up the build variables. This is a bit more consistent and would be a good cause for a future bug.
This is a retry of commit 7841a6ab100fc35a67600f1ce1a70d293c51350e, which made some bold changes to LIPO that didn't work out well.
--
MOS_MIGRATED_REVID=122829825
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Without this, old repositories are still cached across the change to
give a name in the WORKSPACE file, which results in annoying warnings.
This should have been included in
a7fea8a1467a89adde75e67bc99bf0e040f71fe9.
Fixes #1166.
--
Change-Id: I831b842dc07b4ca7925f003b68f4838d083b1166
Reviewed-on: https://bazel-review.googlesource.com/c/3660/
MOS_MIGRATED_REVID=122829553
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122825908
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122824226
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
breaks []
*** Original change description ***
Move the command line arguments for C++ preprocessor defines to a feature.
This required a few assorted changes:
- The LIPO compilation context is not merged anymore. Include paths for auxiliary files are apparently taken from the profile files, so it is not necessary.
- The FDO build stamp is not special-cased anymore, it is treated as a preprocessor define like any other.
- When compiling a .pcm file, use interfaceContext instead of the regular context when setting up the build variables. Th...
***
--
MOS_MIGRATED_REVID=122823591
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When a progress message has to be shortened, as it does not fit in a
line in the progress bar, add a new first attempt: if the message the
path implicit to the label, only shorten that path within the message
(if that gets short enough, leaving a reasonable part of the path);
usually, the additional information is more useful than having a longer
part of the path present.
While there, also fix incorrect length computation in a different case
of message shortening.
--
Change-Id: Ied80e03cace1b249fc0f4e11bce41f2b4207b6ad
Reviewed-on: https://bazel-review.googlesource.com/#/c/3670
MOS_MIGRATED_REVID=122818198
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122772973
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122772865
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122769355
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122755869
|
|
|
|
|
|
|
|
|
|
| |
to support cases like src/com/google/java, where the root is just src.
Also check for nested /src/ directories. This allows "blaze analyze"
to understand the classes provided by such "double src" rules and
match them up to java imports.
--
MOS_MIGRATED_REVID=122748028
|
|
|
|
|
|
|
| |
CLEANUP=<fixing documentation>
--
MOS_MIGRATED_REVID=122743083
|
|
|
|
|
|
|
| |
This fixed some fallout from commit 7894c18dbaf237a1c02d76beabe6ca54faf5039a. I also audited all the sites that change introduced ImmutableMap.Builder at.
--
MOS_MIGRATED_REVID=122738945
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Rollback of commit e37c55eccbd4516b2db7aaf58ef95209dfad3ed4 is complete so rolling this semantically unrelated change forward again
*** Original change description ***
Automated [] rollback of commit b8946eabd60a199a66a1892701d52d9801c7fb1a.
--
MOS_MIGRATED_REVID=122736257
|
|
|
|
|
|
|
| |
flags and environment variables.
--
MOS_MIGRATED_REVID=122735641
|
|
|
|
|
|
|
| |
ParameterFileWriteAction: Add support to write out CommandLine with tree artifact expansions.
--
MOS_MIGRATED_REVID=122734422
|
|
|
|
|
|
|
| |
ConfiguredTarget.
--
MOS_MIGRATED_REVID=122729392
|
|
|
|
|
|
|
|
|
|
|
| |
After the initial loading phase, further packages are loaded during
analysis. Show in the experimental UI the total number of packages
loaded during analysis phase to indicate at least some progress.
--
Change-Id: I345c6f806591e70e4397fc3e3365718dde95d7d2
Reviewed-on: https://bazel-review.googlesource.com/#/c/3645
MOS_MIGRATED_REVID=122728719
|
|
|
|
|
|
|
| |
RELNOTES: objc_framework now supports dynamic frameworks.
--
MOS_MIGRATED_REVID=122728042
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122727524
|
|
|
|
|
|
|
| |
Fixes #1290.
--
MOS_MIGRATED_REVID=122726071
|
|
|
|
|
|
|
|
|
|
|
| |
This required a few assorted changes:
- The LIPO compilation context is not merged anymore. Include paths for auxiliary files are apparently taken from the profile files, so it is not necessary.
- The FDO build stamp is not special-cased anymore, it is treated as a preprocessor define like any other.
- When compiling a .pcm file, use interfaceContext instead of the regular context when setting up the build variables. This is a bit more consistent and would be a good cause for a future bug.
--
MOS_MIGRATED_REVID=122725179
|
|
|
|
|
|
|
|
|
|
|
| |
Even during analysis, more packages will be loaded. However, those are
not tracked by the LoadingProgressReceiver. Therefore, provide a reference
to the AtomicInteger counting the total number of packages loaded already.
--
Change-Id: Ia546a2fec4afe6fd1bd69448b14ba889c68ae12c
Reviewed-on: https://bazel-review.googlesource.com/#/c/3644
MOS_MIGRATED_REVID=122722688
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122721164
|
|
|
|
|
|
|
|
|
|
| |
- mark the addDependencies method as @VisibleForTesting, and don't round-trip
through PathFragment
- use Path.getRelative(String) instead of Path.getRelative(PathFragment), which
may do additional optimizations
--
MOS_MIGRATED_REVID=122720947
|
|
|
|
|
|
|
| |
to it.
--
MOS_MIGRATED_REVID=122718503
|
|
|
|
|
|
|
| |
Due to reusing an old Reporter instead of grabbing the current one for each build, verbose messages were lost and not printed to the console. Also adds a test for this feature.
--
MOS_MIGRATED_REVID=122639383
|
|
|
|
|
|
|
|
|
|
|
| |
This completes the introduction of aspect configuration fragment enforcement
for static configuration builds; as of this change, it is no longer possible to
fall back to the base rule's set of requested configuration fragments.
This sort of fallback may become possible later, likely in a more controlled way.
--
MOS_MIGRATED_REVID=122638152
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122635220
|
|
|
|
|
|
|
|
|
|
| |
When reporting on the progress state while loading packages, also report the
number of packages currently being loaded simultaneously.
--
Change-Id: I5e46c6f392e1e583b752560878f4c42c160f12d2
Reviewed-on: https://bazel-review.googlesource.com/#/c/3643
MOS_MIGRATED_REVID=122632188
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122627792
|
|
|
|
|
|
|
|
|
| |
Label#equals was a cpu hotspot revealed during profiling.
The additional field for the precomputed hash code has no memory cost since each Label instance was already wasting 4 bytes due to object alignment requirements.
--
MOS_MIGRATED_REVID=122626707
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122626656
|
|
|
|
|
|
|
| |
This is a resubmission of commit 45d48bf1fe7503acbbb0c095822b7f8f558881e8. It turns out that we also need -fPIC for *assembler* command line options, because some assembler sources are preprocessed and they can say "#ifdef __PIC__".
--
MOS_MIGRATED_REVID=122626234
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122620007
|
|
|
|
|
|
|
| |
RELNOTES: new_objc_provider can now set the USES_SWIFT flag.
--
MOS_MIGRATED_REVID=122566589
|
|
|
|
|
|
|
| |
postProcess().
--
MOS_MIGRATED_REVID=122542545
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=122542339
|
|
|
|
|
|
|
| |
large number of rbuildfile dependencies, we retrieve them in batches of at most BATCH_CALLBACK_SIZE.
--
MOS_MIGRATED_REVID=122540401
|
|
|
|
|
|
|
|
|
|
|
| |
* Do not include cc info for non-C rules
* Omit exported headers (unused on the IDE side)
This CL reduces aspect output size (and memory footprint) by ~50MB
for an example real-life project.
--
MOS_MIGRATED_REVID=122537339
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Breaks //[]/go:all .
*** Original change description ***
Turn the addition of -fPIC to the command line of PIC actions into a feature.
--
MOS_MIGRATED_REVID=122526799
|
|
|
|
|
|
|
|
| |
Typical output looks like:
Compiling Python my/package/module.py
--
MOS_MIGRATED_REVID=122522202
|
|
|
|
|
|
|
| |
Fixes #1228.
--
MOS_MIGRATED_REVID=122511655
|