R8 rules produce a depfile generated by R8. R8 sees the withres-withoutdex
jar as an input, and so adds it to the depfile. If the jar is deleted
after running R8 then ninja will always consider the rule dirty, as the
input file listed in the depfile is missing.
Fixes: 303064127
Test: m Calendar && m Calendar
Test: m framework-minus-apex && m framework-minus-apex
Change-Id: I5cbd780b56fc131b58598d6e569a8a78b7fe9395
There are many issues in the docs, so allow specifying a baseline
file that allows existing lint errors in doclava.
Bug: 217734059
Test: m docs
Change-Id: I8997e47a497e1366247e8354be492805f300a16e
This reverts commit 1180919dda.
Test: go test ./java && m TARGET_PRODUCT=sdk TESTING_TARGET_RELEASE_NEXT=true nothing and inspect ninja command for generating stubs and verify the flag is included && m TARGET_PRODUCT=sdk TARGET_RELEASE=trunk_food nothing and inspect ninja command for generating stubs and verify the flag is not included
Bug: 299570421
Change-Id: I4967376c0236bad729398af80fa59b48dbab5f21
The linter output can be referenced in other modules for further
processing.
Bug: 302400088
Test: m nothing
Change-Id: I1809b70d8d98ced5d8f3d5f068c6639c782ebebc
Delete files created within the d8 and r8 rules that are not considered
output files by ninja. This should reduce disk space usage without
affecting incremental build performance.
Bug: 302573555
Test: m checkbuild
Change-Id: Ifea86888c2f385415735ade7d80cf315bc4d057e
The Android lint --exitcode parameter can be skipped using
ANDROID_LINT_SUPPRESS_EXIT_CODE. Expose a similar attribute to Java
modules to ignore the exit code.
This is useful for integration testing. It is possible to build
libraries that report an error when linted. Othewise, such module would
break the checkbuild target.
Bug: 302400088
Test: m nothing
Change-Id: I372c42184f40d25753a688c380c0c63e9758ca00
This makes it possible for Android.bp files to depends on the srcjar.
Bug: 151360309
Test: m platform-bootclasspath.srcjar
Change-Id: Id42cc3cff89c084b55c7fb0726ad84ecd4f50f58
This output contains all the transitive sources that are contained in
the bootclasspath.
It's currently limited to source-built bootclasspath components. Future
work will make this deal with sdk_library prebuilts.
Bug: 151360309
Test: unit test, as well as manual inspection of platform-bootclasspath
Change-Id: Ie05d8125e19736d8b4b9ebafb70b88a0a40069d5
This new entry in the JavaInfoProvider lists all the transitive source
files contained within the library. That is, the source files of the
module and all its static dependencies.
Bug: 151360309
Test: unit test in java_go + some manual testing
Change-Id: I7fe3035b9e46774095c0e9196cd77fa1027adf6d
android_library does not accept deps when there are no srcs because
there is no compilation happening. The libs of android_library module
are unnecessary as deps on the android_library since they aren't
being propagated to any dependencies.
So we can drop deps here.
Test: the updated test case and CI
Bug: 302290600
Change-Id: Ib43a1c273a5f3b3e0aa6f29f4e8569ad3487e451
APIs annotated with @FlaggedApi should not be included in the artifact
when building sdk target products in the "next" release configuration.
This change adds such logic by passing additional flag to metalava in
droidstubs.
The flag does not need to be passed to metalava invocation done in
java_api_library, as java_api_library generates stubs using api
signature files (i.e. *-current.txt files), and they will not contain
apis marked @FlaggedApi. The metalava invocation in droidstubs is
responsible for removing such apis.
Test: go test ./java && m TARGET_PRODUCT=sdk TESTING_TARGET_RELEASE_NEXT=true nothing and inspect ninja command for generating stubs and verify the flag is included
Bug: 299570421
Change-Id: Ia4b699b6e3ff6324f050eecc9ff5b622fdc04621
java_library rules with javac_shard_size set split the sources into
shards to invoke javac multiple times, but were using a single javac
invocation for all srcjars. For fraemwork-minus-apex, this srcjar
shard was the long pole at 15.7 seconds, containing 266 srcjars with
1542 java files with a total of 614593 lines.
Use a rough approximation of 5 sources per srcjar to determine the
number of shards to split the srcjars into based on javac_shard_size.
This results in splitting the srcjars for frameworks-minus-apex into
8 shards, with the longest taking 10.5 seconds to compile.
The longest shard contains most of the aidl srcjars, which have been
generated by sharded groups of 50 aidl files and have a much higher
average number of sources per srcjar (a mean and median of 27). A
future improvement could be to shard those separately assuming a
higher number of sources per srcjar.
Bug: 302033097
Test: USE_RBE=false m frameworks-minus-apex
Change-Id: I85e740c7fcf5651cf18c0cdc90ab8c6ee39cb47b
This should be no-op, as the underlying mutator has not changed yet.
Some other refactoring is required and done in this CL:
- Delete some old, dead ApiBp2build code
- Fix casting to TopDownMutator when it's not necessary
This change is required to prepare for allowlist v2 work, as only
BottomUp mutators can AddDependency.
Bug: 285631638
Test: m nothing
Test: presubmits
Change-Id: I5212a5f5634cc13056195783e6df37ff8eb000da
