| Commit message (Collapse) | Author | Age |
|
|
|
| |
PiperOrigin-RevId: 198074986
|
|
|
|
|
| |
RELNOTES: None
PiperOrigin-RevId: 198071932
|
|
|
|
|
|
|
|
|
|
| |
This is a rollforward of CL/197725926 after depot fix.
NEW: Additional deletions of unused providers (no constructor call references).
NEW[last rollforward]: Allow java_* rules to depend on proto_libraries via runtime_deps and exports. This should avoid the breakage that caused the original rollback. The edges are no-ops and could be removed.
PiperOrigin-RevId: 198060345
|
|
|
|
|
|
|
|
| |
This is part of a chain of CLs that will pull initialization of
CcCompilationContext from CcCompilationHelper.
RELNOTES:none
PiperOrigin-RevId: 198060027
|
|
|
|
|
|
|
|
| |
Fixes #5263
RELNOTES: execution strategies line no longer handles differently the case
where all processes have the same strategy.
PiperOrigin-RevId: 198057496
|
|
|
|
| |
PiperOrigin-RevId: 198053509
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Crashes lots of targets in nightly blaze-2018.05.24-1:
PiperOrigin-RevId: 198049395
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
https://github.com/bazelbuild/bazel/commit/56d1b1c3122a3d1ec111baab339631b8c42c2c31 changed the signature of SandboxfsProcess#mount() to take a
PathFragment instead of a Path, and this broke the build of the test above.
Given that the test is marked as manual, this was not caught by CI.
Further, the test was failing to link in the test runner class, so it
would not execute properly. The missing dependency was incorrectly removed
by unknown commit because it had been specified as a deps instead of a
runtime_deps.
RELNOTES: None.
PiperOrigin-RevId: 198048748
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This expands on https://github.com/bazelbuild/bazel/commit/33f08e75fa4e7480f8d46b8305ce03a171adefa0 to ensure list-like behavior of range for the following operations
- str/repr/type
- equality and ordering
- append / augmented assignment
- concatenation
- all of the above with slices
This is a prerequisite to both optimizing range() while preserving its current API, as well as changing its API via an incompatible change flag.
RELNOTES: None
PiperOrigin-RevId: 198046941
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Crashes lots of targets in nightly blaze-2018.05.24-1:
PiperOrigin-RevId: 198041861
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 198036365
|
|
|
|
|
| |
RELNOTES:none
PiperOrigin-RevId: 198025150
|
|
|
|
|
|
|
| |
Instead,defer to ActionFS regarding the choice about delegation to top-level Filesystem.
RELNOTES: None
PiperOrigin-RevId: 197960416
|
|
|
|
|
|
|
| |
Also refactor away SkylarkModule.Resolver.
RELNOTES: None
PiperOrigin-RevId: 197955164
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 197949354
|
|
|
|
|
|
|
|
|
| |
well-defined.
Moving forward, if a class not already annotated with @SkylarkModule has SkylarkModule supertypes A and B, then it must be the case that A is a type of B, or B is a type of A. The skylark type of a given class is dictated by the *most specific* superclass annotated with @SkylarkModule.
RELNOTES: None.
PiperOrigin-RevId: 197946898
|
|
|
|
| |
PiperOrigin-RevId: 197946783
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's more efficient to use properly sized array builders to avoid having
a need to dynamically adjust the size of the underlying array. The result
is improved performance (no need to copy elements to newly created array)
and better memory efficiency - no need to re-allocate and no extra space
is wasted.
Closes #5241.
PiperOrigin-RevId: 197946723
|
|
|
|
|
|
|
| |
This is useful before we optimize range() to return a lazy object that appears like a list to the user.
RELNOTES: None
PiperOrigin-RevId: 197944773
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 197941930
|
|
|
|
|
|
|
| |
with generic parameters
RELNOTES: None.
PiperOrigin-RevId: 197932265
|
|
|
|
| |
PiperOrigin-RevId: 197924560
|
|
|
|
|
| |
RELNOTES:none.
PiperOrigin-RevId: 197923013
|
|
|
|
|
|
| |
diagnosing of bloat.
PiperOrigin-RevId: 197915327
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 197915097
|
|
|
|
|
|
|
| |
and [AndroidResourcesInfo].r_txt to Skylark.
RELNOTES: none
PiperOrigin-RevId: 197902129
|
|
|
|
|
|
| |
should not exist.
PiperOrigin-RevId: 197895718
|
|
|
|
|
|
| |
Part of https://docs.google.com/document/d/1_UJKmAQ9EE8i3Pl0il3YLTYr-Q9EKYYyLatt2zohfyM/edit#
PiperOrigin-RevId: 197890685
|
|
|
|
|
|
|
|
|
|
|
|
| |
A bug in the lexer ignored indentation on the first line of a file. This now
causes an error.
Also, remove the COMMENT token from the lexer. Comments are now accessed separately. This will allow further optimizations in the lexer. It also aligns the code a bit more with the Go implementation.
RELNOTES[INC]:
Indentation on the first line of a file was previously ignored. This is now fixed.
PiperOrigin-RevId: 197889775
|
|
|
|
|
|
|
|
| |
Netty's HttpClientCodec always emits a LastHttpContent event and so we don't need to track the received bytes manually to know when we are done reading. The HttpBlobStore compares the hashes of the received bytes to give us confidence that what we received is correct.
Closes #5244.
PiperOrigin-RevId: 197887877
|
|
|
|
|
|
|
|
|
|
|
| |
The Android data Skylark API includes references to FileProvider. To move that
API to the skylarkbuildapi, we must first move this provider.
For more information about this migration, see
https://docs.google.com/document/d/1UDEpjP_qWQRYsPRvx7TOsdB8J4o5khfhzGcWplW7zzI/
RELNOTES: none
PiperOrigin-RevId: 197882296
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 197881012
|
|
|
|
|
|
|
|
|
|
|
|
| |
This changeset introduces tests for the `shallow_since` functionality that was introduced in 6496b1d1d25025f33406b1eaf9949cab126f08bd.
We test that if the user specifies a tag as well as a `shallow_since` that it fails since they are incompatible. We also verify that specifying a commit before the `shallow_since` date causes the clone to fail (albeit currently with a non-useful error message; perhaps that could be improved at a later date).
It also documents more clearly how to use the feature by explaining the format of the 'since' string.
Closes #5183.
PiperOrigin-RevId: 197875727
|
|
|
|
|
|
|
|
|
|
|
|
| |
...and point --experimental_distdir there, so that offline builds
are again possible out of the distribution archive.
Related #5175.
Fixes #5202.
To be cherry-picked for #5056.
Change-Id: I634296e9d83e4e18ed966b42f35acc63061259d9
PiperOrigin-RevId: 197866998
|
|
|
|
|
|
|
|
|
|
| |
If ${JAVABASE} is a relative path prefix it with ${PWD}
but if it's an absolute path just use that i.e. when
specified via --javabase.
Closes #5238.
PiperOrigin-RevId: 197863163
|
|
|
|
|
|
|
|
| |
Split the AnalysisFailureEvent into a legacy and a new variant; always post an
AnalysisFailureEvent, even if the legacy variant is not posted - this gives us
some coverage for loading failures.
PiperOrigin-RevId: 197862257
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
'srcs') as textually included headers.
A private header is never exported to a different module and thus cannot be
used by a different module. We never need to compile a module just because one
of it's private headers is used from somewhere else. There must be a different
cc_library (and thus module) that exports the header.
At the same time, being compiled into a module doesn't provide benefits for the
compilation of that cc_library itself, because a module created for a
cc_library A is never used when compiling any of A's srcs.
PiperOrigin-RevId: 197846388
|
|
|
|
|
|
| |
construction. Also restrict visibility of some test-only OptionsDiff methods, and remove a server-specific part of the ConfiguredTargetKey #toString representation.
PiperOrigin-RevId: 197830577
|
|
|
|
|
|
| |
Part of https://docs.google.com/document/d/1_UJKmAQ9EE8i3Pl0il3YLTYr-Q9EKYYyLatt2zohfyM/edit#
PiperOrigin-RevId: 197800831
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 197799961
|
|
|
|
|
| |
RELNOTES: --batch is deprecated
PiperOrigin-RevId: 197780675
|
|
|
|
|
|
| |
This is going away with the removal of C++ LIPO support.
PiperOrigin-RevId: 197779736
|
|
|
|
|
|
| |
This results in less special logic in the implementation and a simpler API.
PiperOrigin-RevId: 197772283
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I've pulled out the API for separate review. It includes all
hooks from blaze/skylark used by the debugger.
Debuggable thread contexts are currently declared in 3 places:
- BuildFileAST (top-level evaluation of BUILD files)
- SkylarkRuleConfiguredTargetUtil (rules)
- SkylarkAspectFactory (aspects)
The purpose of declaring these contexts is so that the debugger
can track currently-active threads (and stop tracking them when
the task is completed).
Details of the actual debugging server are in unknown commit.
PiperOrigin-RevId: 197770547
|
|
|
|
|
|
|
| |
This unclashes with the incoming ConfigurationTransition.apply
method described in https://docs.google.com/document/d/1_UJKmAQ9EE8i3Pl0il3YLTYr-Q9EKYYyLatt2zohfyM/edit#heading=h.96gongkwg852.
PiperOrigin-RevId: 197769784
|
|
|
|
|
|
|
|
| |
using C++ modules (yet) and thus we are paying some of the cost such as
building transitive modules and a taller critical path without reaping any
benefits.
PiperOrigin-RevId: 197768129
|
|
|
|
|
|
|
|
|
| |
With this change, all classes in com.google.testing.junit.runner.util will be loaded from the system classloader for robolectric tests.
This fixes a problem where TestPropertyExporter and TestIntegrationsRunnerIntegration in the com.google.testing.junit.runner.util package were being loaded twice: once by the test runner into the system classloader, and again into the robolectric's classloader during test execution. Since there were two separate instances, the test properties exported during robolectric's testing lifecycle were disconnected from the test runner which writes out the properties.
RELNOTES: none
PiperOrigin-RevId: 197761010
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 197744356
|
|
|
|
|
|
|
| |
build API to a global environment.
RELNOTES: None.
PiperOrigin-RevId: 197742427
|