| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
android_binary targets have their own R.java files (built from merging
dependencies and any resources that belong directly to the target). As such,
they don't need inherited R.java files at runtime. Taking these out makes for
smaller APKs and less inheritance from the target's dependencies.
Add a flag to control this behavior. Have it default to continue to include
R.class files as runtime dependencies so we can control rollout of this
behavior.
Add tests of android_binary to ensure the JAR is filtered out as appropriate,
and of android_robolectrictest to ensure that those tests still have access to
the JARs.
RELNOTES: none
PiperOrigin-RevId: 153177074
|
|
|
|
|
|
| |
`DefaultInfo in target` should always return `true` for any target.
PiperOrigin-RevId: 153166669
|
|
|
|
| |
PiperOrigin-RevId: 153161442
|
|
|
|
| |
PiperOrigin-RevId: 153140874
|
|
|
|
| |
PiperOrigin-RevId: 153096906
|
|
|
|
|
|
|
|
|
| |
In this way, all indices are reported with 1 being the smallest
possible. Also, the numbers fit better with the file names generated
for log files etc.
Change-Id: I7671e5a79dd47c3e3afac16108acaeacdf018fc5
PiperOrigin-RevId: 153080339
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 153058374
|
|
|
|
|
| |
RELNOTES: Use action_config in crosstool for static library archiving, remove ar_flag.
PiperOrigin-RevId: 153046587
|
|
|
|
|
|
|
|
|
| |
For transports that are purely local (like the ones writing to a local file),
it sometimes can be useful to skip path conversion and use the local paths directly.
Support this for the text and binary format file transports.
Change-Id: I2ac2e187ebb11ff82c4e1ddf4881ea54f9d4205d
PiperOrigin-RevId: 153044267
|
|
|
|
|
|
| |
any extra defines.
PiperOrigin-RevId: 153039146
|
|
|
|
|
|
| |
used.
PiperOrigin-RevId: 153020672
|
|
|
|
|
|
|
|
|
|
|
|
| |
This filegroup is used by the Android testing infrastructure to launch QEMU2
emulators.
This filegroup will only build correctly on Linux.
Tracking issue: https://github.com/bazelbuild/bazel/issues/903
RELNOTES: None
PiperOrigin-RevId: 153003267
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
fa97703c1edf ("allow skylark implicit output callbacks to use the rule
name") fixed a limitation of the skylark implicit outputs by adding a
special case for "name" to the implicit outputs code. Later,
015e5954157a ("Remove special handling of name attribute") fixed the
general problem of "name" being a special case in the attribute
map. Therefore, we can remove my original fix. We may also excise an
older special case in the implicit outputs templating code.
Change-Id: I606c9decd98a8df492d2359abe540d3263f99fe1
PiperOrigin-RevId: 152974774
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 152965874
|
|
|
|
|
|
|
| |
The memory regression was introduced in https://github.com/bazelbuild/bazel/commit/360fb4d9a1e2c44154b17aeb866e07bac2dd1b5b , now default providers
are optimized and are built only on demand for all types of targets.
PiperOrigin-RevId: 152957220
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Before this change, attributes with select() were completely skipped.
This doesn't attempt to merge "trivial" selects, e.g. even though:
attr = [":foo"] + select({"//conditions:default": [":bar"]})
always resolves to:
attr = [":foo", ":bar"]
this change still produces:
attr = [":foo"] + [":bar"]
We could merge these in a future change, if desired. But it's not
even clear that's desired. There's conceptual value in keeping the
lists separate since that's how they were originally written. That
gives users a cue to maybe refactor their rules.
RELNOTES[NEW]: "query --output=build" now includes select()s
PiperOrigin-RevId: 152956939
|
|
|
|
|
|
|
|
|
| |
Before this change Bazel was adding this implicitly. But that wasn't
working properly with --incremental_dexing. And that implementation is a
lot messier and bug-prone vs. just including it as a standard dependency.
Issue: #2694
PiperOrigin-RevId: 152956638
|
|
|
|
|
|
|
|
| |
This logger makes it easier to parse log statements and is now enabled for Bazel's java.log.
RELNOTES[INC]: Bazel now prints logs in single lines to java.log
PiperOrigin-RevId: 152954337
|
|
|
|
|
|
|
| |
This issues a configuration transition on the dependencies of the rule, causing deps to be compiled with the appropriate minimum os version.
RELNOTES: None.
PiperOrigin-RevId: 152952168
|
|
|
|
|
|
|
|
|
|
|
| |
This is the first actual user of the config_feature_flag rule, able to actually
set its value (and thus give a point to using it!)
This feature is not fully launched yet, but it is usable in any build
containing this change.
RELNOTES: None.
PiperOrigin-RevId: 152948153
|
|
|
|
|
| |
RELNOTES: None
PiperOrigin-RevId: 152947523
|
|
|
|
|
|
| |
from an apple_binary split transition context. Refactors ProtobufSupport to not rely on ruleContext for dependencies.
PiperOrigin-RevId: 152939984
|
|
|
|
|
| |
Change-Id: I55fd5b2f38cad9d2b18d369ae51dbd93c2c92288
PiperOrigin-RevId: 152925304
|
|
|
|
| |
PiperOrigin-RevId: 152916191
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
files for all android system images installed in the SDK. For system images that are not installed, create genrule "poison pills" that print a helpful error message if you attempt to build them.
These filegroups are intended to be used as the system_image attribute of the android_device rule.
Example:
$ bazel build @android_test_support//tools/android/emulated_devices/pixel:android_25_x86
Extracting Bazel installation...
.
INFO: Found 1 target...
ERROR: /usr/local/google/home/ajmichael/.cache/bazel/_bazel_ajmichael/efa32264346ef1fc4fec492202b88b31/external/androidsdk/BUILD.bazel:56:1: Executing genrule @androidsdk//:emulator_images_android_25_x86 failed: Process exited with status 1 [sandboxed].
This rule requires that the Android SDK used by Bazel has the following system image installed: emulator_images_android_25_x86. Please install this system image through the Android SDK Manager and try again.
RELNOTES: None
PiperOrigin-RevId: 152889824
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 152882585
|
|
|
|
|
|
|
|
| |
tool during incremental dexing
RELNOTES: none
PiperOrigin-RevId: 152838197
|
|
|
|
|
|
|
| |
This follows our CamelCaseInfo naming conventions for providers.
RELNOTES: None.
PiperOrigin-RevId: 152832215
|
|
|
|
| |
PiperOrigin-RevId: 152810023
|
|
|
|
|
|
|
|
| |
accidentally regressed.
TESTED=local RemoteWorker without work_path
RELNOTES: n/a
PiperOrigin-RevId: 152806430
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 152804937
|
|
|
|
|
|
|
| |
This is an encore of https://github.com/bazelbuild/bazel/commit/6127358c1799d8d83cebbd499edac89f869df41b .
RELNOTES: None.
PiperOrigin-RevId: 152803621
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 152800780
|
|
|
|
| |
PiperOrigin-RevId: 152799488
|
|
|
|
|
|
|
|
|
| |
Output artifacts are now always reported as named sets. So the
field for reporting artifacts directly in no longer needed. Clean
it up.
Change-Id: I41efc66116d4a733088d4ee685ccd3fcf0a62d05
PiperOrigin-RevId: 152799435
|
|
|
|
|
|
|
| |
Default providers can now be used not only to return standard providers values
from a rule implementation function, but also to access these values provided
by other rules.
PiperOrigin-RevId: 152797193
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 152793682
|
|
|
|
|
|
|
|
| |
If aspect a3 sees aspect a2, and aspect a2 sees aspect a1, propagation
of the aspect list [a1,a2,a3] should not lose any aspects.
RELNOTES: None.
PiperOrigin-RevId: 152786900
|
|
|
|
|
| |
RELNOTES: none
PiperOrigin-RevId: 152783295
|
|
|
|
|
| |
RELNOTES: None
PiperOrigin-RevId: 152782181
|
|
|
|
| |
PiperOrigin-RevId: 152760365
|
|
|
|
|
|
|
|
|
|
| |
Previously, there was no configuration transition in the default case.
While this was "convenient" for testing, it makes these tests somewhat inconsistent,
and furthermore invalid once these rules switch to always using a configuration transition
with the inclusion of minimum OS as an attribute.
RELNOTES: None.
PiperOrigin-RevId: 152734307
|
|
|
|
| |
PiperOrigin-RevId: 152714468
|
|
|
|
|
|
|
| |
--j2objc_dead_cdoe_report accepts a label that can generate a dead code report as argument. If specified, Blaze should pass "--dead-code-report path/to/dead/code/report" to J2ObjC.
RELNOTES: None.
PiperOrigin-RevId: 152712841
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously multidex did not work in Bazel because it required a dependency on
zip, which we do not ship with Bazel. Then the
--experimental_android_use_singlejar_for_multidex flag was added to use a
different tool that is included with Bazel for multidex. This change sets that
flag to be the default.
Fixes https://github.com/bazelbuild/bazel/issues/1936
RELNOTES: android_binary multidex should now work without additional flags.
PiperOrigin-RevId: 152712301
|
|
|
|
|
| |
Change-Id: I407240708f4aacc89ec5c00bf3e8ff46a1d6d6d6
PiperOrigin-RevId: 152692981
|
|
|
|
|
|
|
|
| |
In this way, shared subsets of artifacts are only reported once,
even if occurring for many aspect-target pairs.
Change-Id: Ia300126f427af4a9cc630fbfca649760d8b72262
PiperOrigin-RevId: 152692099
|
|
|
|
|
|
|
| |
Only checking for GitHub right now, this could probably be expanded.
Fixes #2147
PiperOrigin-RevId: 152689610
|
|
|
|
|
|
|
|
| |
To avoid artifacts rolled up from other targets to be reported several
times.
Change-Id: I8a329f1c53ad3fcb37cc6602b906472dfce1a12f
PiperOrigin-RevId: 152688681
|
|
|
|
|
|
|
| |
GroupedListHelper
RELNOTES: None
PiperOrigin-RevId: 152687516
|