| Commit message (Collapse) | Author | Age |
|
|
|
|
| |
--
MOS_MIGRATED_REVID=89131807
|
|
|
|
|
|
|
|
| |
In the future, we will have linkstamp flags be generated by the crosstool
feature configuration, just like normal compile actions.
--
MOS_MIGRATED_REVID=89131127
|
|
|
|
|
|
|
|
|
|
|
|
| |
Removing ActionMetadata#getInputCount instead of relying on all
implementations of ActionMetadata keeping getInputCount and
getInputs in sync. getInputCount is only used by a few tests
and has no applications in the main code at the moment, so it
is trivial to remove before we may come to rely on it more or its
semantics get more complicated.
--
MOS_MIGRATED_REVID=89130009
|
|
|
|
|
|
|
| |
RELNOTES: Referencing a path outside the execution root is now an error.
--
MOS_MIGRATED_REVID=89129910
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=89123900
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=89123292
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=89123209
|
|
|
|
|
|
|
| |
../lib/rules package. Unused code is removed. Anchor for lists is fixed in the doc.
--
MOS_MIGRATED_REVID=89117189
|
|
|
|
|
|
|
| |
We don't need a provisioning profile if we're building for a simulator
--
MOS_MIGRATED_REVID=89098211
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=89095609
|
|
|
|
|
|
|
|
| |
Move castList, castMap, toMap utilities away from SkylarkFunction to SkylarkType
where the cast utility already has been moved.
--
MOS_MIGRATED_REVID=89042835
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=89038414
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Reflection, and saves ~15% of initialization time in []RuleClassProvider.create().
Before change:
Over 14 samples,
average = 976
median = 969.9
After change:
Over 14 samples,
average = 811.5
median = 813.9
--
MOS_MIGRATED_REVID=89036261
|
|
|
|
|
|
|
|
|
|
|
| |
The build mode ('dbg', 'opt', 'fastbuild') is specified as a feature. The
crosstool can use that feature to switch between different flag configurations.
As one of these is always provided, the crosstool can hang default
configurations off of those features (via implies).
--
MOS_MIGRATED_REVID=89035231
|
|
|
|
|
|
|
|
|
| |
unsupported features for a rule.
Previously this would lead to a NPE later on.
--
MOS_MIGRATED_REVID=89033249
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=89032399
|
|
|
|
|
|
|
|
|
|
|
| |
Without this change we create a xcode target for each combination (target, architecture)
which causes the xcode project generation to blow up: Xcode transparently handles
architectures. Eventually we want to match Xcode's behaviour but for now we export just
a single architecture's xcode targets which should allow developers to work in xcode as
before while preserving multi-architecture compilation for IPAs in blaze.
--
MOS_MIGRATED_REVID=88973190
|
|
|
|
|
|
|
|
|
|
|
| |
is known that the action is a cache hit.
This adds momentary memory overhead when checking the action cache, but should prevent a host of potential errors.
Note that this cl assumes that an action that discovers its inputs does *not* take the names of its inputs into account when calculating its key, which is already stated as part of the javadoc of Action#getKey.
--
MOS_MIGRATED_REVID=88971626
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=88964412
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of repeatedly scanning all transitively required include scannables, do
one pass from all include scannables; this prevents us revisiting the common
transitive closure of the include scannables.
Additionally, only scan command line includes relatively to a main source file,
not relatively to all source files. This is better than what we had before, but
it's still not exactly right - we should actually scan the command line
includes relatively to the module map file ([]
Brings include scanning times for large TUs with modules down from ~60 seconds
to < 2 seconds.
--
MOS_MIGRATED_REVID=88963159
|
|
|
|
|
|
|
|
|
|
|
| |
Before, external jars would stomp on each other and the last jar loaded would
"win" (all others would be overwritten and the build would fail). This changes
symlink forest path generation to generate unique paths under
<exec-root>/external/ for jars, so multiple external jar dependencies can be
used.
--
MOS_MIGRATED_REVID=88952191
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Moves pattern resolving logic from TargetPatternFunction.Resolver to
a top level class. Adds a layer of abstraction to the Resolver
implementation enabling it to be backed by either an Environment or
a Graph, for use in SkyFunction evaluation or on-the-fly evaluation,
respectively. Finally, SkyQueryEnvironment#preloadOrThrow now checks
to see if each target pattern exists in the graph, and any that
don't will be resolved on-the-fly.
--
MOS_MIGRATED_REVID=88861201
|
|
|
|
|
|
|
|
| |
Have actoolzip abs-ify the partial info plist path because otherwise it picks
seemingly random directories to which to output the info plist files.
--
MOS_MIGRATED_REVID=88851706
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=88789823
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=88742425
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=88739509
|
|
|
|
|
|
|
| |
RELNOTES: Allow ios_application in xctest_app attribute.
--
MOS_MIGRATED_REVID=88730548
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Broke one of our integration tests.
*** Original change description ***
Declare dependencies on discovered inputs before execution instead of after.
As a side effect, we no longer restart ActionExecutionFunction after the action has executed unless the action ran locally (and therefore didn't discover its inputs beforehand). This also means that we no longer need to store an out-of-Skyframe cache for discovered includes except when checking the action cache. Since my suspicion is that the out-of-Skyframe cache will have a minimal performance impact once it is just...
***
--
MOS_MIGRATED_REVID=88724988
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
to keep the existing behavior when turning on C++ modules. The problem is that
a #include of a header gets internally translated into an import of the
header's (sub)module. Now, if all headers of a cc_library are in the same
module, this means, a single import of one of them makes the symbols of all of
them accessible. While this isn't inherently problematic, it breaks the
similarity between a modules and non-modules build.
This behavior is hidden behind a feature for now as I have just committed
necessary changes to Clang, which will need to be release to stable, before we
can flip everything.
Also, don't put a header as both normal and private header if it is present in both. (Or don't output two submodules with the new implementation).
--
MOS_MIGRATED_REVID=88723759
|
|
|
|
|
|
|
|
|
| |
root when used with protobuf or XML outputs.
genquery was non-deterministic accross machines / workspace location without that setting.
--
MOS_MIGRATED_REVID=88711033
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Using the new flag works only on ios_application and ios_extension targets,
resulting in their output binaries being "universal" with all given architectures
included.
Internally this is achieved by splitting the "binary" dependency (and its transitive
dependencies) of ios_{application,extension} to be built in one of the architectures
using split transitions. The results are then combined. Since an ios_application may
depend on ios_extension, we transition back to the top-level configuration when
encoutering such a dependency using the newly introduced TOP_LEVEL transition and then
split configurations again.
RELNOTES: --ios_multi_cpus allows building fat ios_applications.
--
MOS_MIGRATED_REVID=88555540
|
|
|
|
|
|
|
| |
the command line when running an objc_binary or ios_application.
--
MOS_MIGRATED_REVID=88544811
|
|
|
|
|
|
|
|
|
|
|
| |
As a side effect, we no longer restart ActionExecutionFunction after the action has executed unless the action ran locally (and therefore didn't discover its inputs beforehand). This also means that we no longer need to store an out-of-Skyframe cache for discovered includes except when checking the action cache. Since my suspicion is that the out-of-Skyframe cache will have a minimal performance impact once it is just being used for the action cache, I may delete it in a follow-up cl.
After this change, we will overapproximate the set of includes because we depend on all includes, rather than just the ones that the action was found to depend on after execution.
This is a prerequisite for Skyframe-native include scanning in that Skyframe-native include scanning will need to add at least as many Skyframe nodes and edges. If we end up punting on it, then we may want to revert this change. But for now I think it's worth having. I'll run some more numbers to see what the actual performance impact is.
--
MOS_MIGRATED_REVID=88492955
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We define unused artifacts as those that aren't consumed by any
action. This can be because an action produced more outputs than
a dependent action needed, or because it's a top level artifact
and we don't care about its contents, just that it was built
without issue. Actions may prevent outputs from being discarded
by declaring them as mandatory. This is particularly useful for
test outputs. The motivation behind this change is to reduce
storage overhead for things we can do without.
It is worth noting this change doesn't cover all cases. In particular
it has difficulty identifying *_binary artifacts as orphaned. This
is due to the insertion of a virtual runfiles artifact which depends
upon the rule's outputs.
--
MOS_MIGRATED_REVID=88467504
|
|
|
|
|
|
|
|
| |
ctx.var(x) is kept for compatibility and should be removed after next
release.
--
MOS_MIGRATED_REVID=88438195
|
|
|
|
|
|
|
| |
case of a missing synopsis (except for deprecated attributes).
--
MOS_MIGRATED_REVID=88436777
|
|
|
|
|
|
|
|
|
| |
given name is present.
Rule implementations are cleaned up.
--
MOS_MIGRATED_REVID=88430329
|
|
|
|
|
|
|
|
| |
This should remove the needs for the --cpu=darwin flag under darwin. Bazel build are now using the host cpu too.
This commit also fix the host cpu detection. The host cpu detection was too late and and some corner case was not really working with configurable attributes. It is now done as the default value for the --host_cpu flag.
--
MOS_MIGRATED_REVID=88427551
|
|
|
|
|
|
|
|
|
|
| |
BazelTestSuiteRule.
- "suites" is deprecated
- it has the exact same functionality as "tests"
--
MOS_MIGRATED_REVID=88426066
|
|
|
|
|
|
|
|
|
|
|
| |
We currently use .inc as file extension for textual includes, as per
the style guide. This is not feasible for third_party libraries, including
libc and lib(std)c++.
RELNOTES: Add textual_hdrs attribute to cc_library rules.
--
MOS_MIGRATED_REVID=88422306
|
|
|
|
|
|
|
|
|
| |
case of nested bundles.
RELNOTES: objc: Fixes a case where nested bundle would pick up parent bundle_id value
--
MOS_MIGRATED_REVID=88372062
|
|
|
|
|
|
|
|
|
| |
probably want to store these and may declare dependencies in advance so all metadata is present in the graph.
Long-term, we may be able to use the returned set to make input-discovering actions immutable, which would be nice.
--
MOS_MIGRATED_REVID=88348068
|
|
|
|
|
|
|
| |
RELNOTES: objc_* rules can use structured_resources
--
MOS_MIGRATED_REVID=88337360
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=88322388
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=88236380
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=88224368
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=88218457
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=88137421
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=87956686
|
|
|
|
|
|
|
| |
application for memory leaks.
--
MOS_MIGRATED_REVID=87839695
|