prebuilt_apex has been disabled when the device is configured for
flattened APEXes, sanitized, instrumented, or built unbundled. However,
apex_set which is another type of APEX prebuilt wasn't disabled for the
same conditions.
This change fixes the discripency. apex_set modules are also force
disabled when the prebuilts are not expected.
Bug: 161316762
Bug: 160933444
Test: OVERRIDE_TARGET_FLATTEN_APEX=true m
The built image has only the flattened APEXes and the device boots
Change-Id: I6c90dfb28d565861a473a1bdce93269ec370601d
Set ANDROID_LINT_CHECK to a comma-separated list of lint issues
that should be checked instead of the defaults. This will disable
all lint checks and enable only the given list.
Set ANDROID_LINT_CHECK_EXTRA_MODULES to a list of modules that provide
lint checks that should added to all modules when ANDROID_LINT_CHECK
is specified.
Bug: 153485543
Test: m ANDROID_LINT_CHECK=JavaKotlinApiUsedByModule ANDROID_LINT_CHECK_EXTRA_MODULES=JavaKotlinApiFinder TARGET_BUILD_APPS=Gallery2 lint-check dist
Change-Id: Ifdf9bf972b8550104315b0f5e98b34ad699dcb67
Break apart test helpers a bit to make it possible to enable
LinuxBionic in a single test, and add LinuxBionic support to
cc.GatherRequiredDepsForTest.
Test: m nothing
Bug: 160349757
Change-Id: Iace1024c964cee2308c130c945daef9e46c18c66
The stdlib's xml apparently cares about ordering but also doesn't
preserve it. There are a number of different recommendations of other
ways to do this, but they all depend on packages we don't have
available (lxml or xml.dom.ext, in particular).
Test: pytest
Bug: None
Change-Id: I07108d8977c302404e7c90ca75a4bf7a1144750f
Imports weren't working in tests because the package had been created.
The Python "binaries" built by Soong don't seem to take their own
pkg_path into account, so I split the separate pieces of code here out
into their own packages.
Note that the ndk_api_coverage_parser tests do not actually pass
before or after this change (seems like it might be a
non-deterministic ordering issue in the attributes of the generated
output?), but they can at least be run now.
Test: pytest ndkstubgen
Test: pytest symbolfile
Test: pytest ndk_api_coverage_parser
Test: out/host/linux-x86/nativetest64/test_ndkstubgen/test_ndkstubgen
Test: out/host/linux-x86/nativetest64/test_symbolfile/test_symbolfile
Test: out/host/linux-x86/nativetest64/test_ndk_api_coverage_parser/test_ndk_api_coverage_parser
Bug: None
Change-Id: I2ac22f7ced7566e4808070f2f72fd04355846e0b
Pass -n to blueprint to make it use validation dependencies for tests,
which will let them run in parallel with soong_build.
Bug: 156428456
Test: m nothing
Change-Id: I706c14330543e2a7728e5dc385976900c67cb2f5
Add a rule to build a zip containing the lint reports from transitive
dependencies for apps, and pass it to Make.
Bug: 153485543
Test: m TARGET_BUILD_APPS=Gallery2 lint-check
Change-Id: I523c09016251377ff89d76084769be7401b95425
This behavior is non-hermetic, but matches the behavior of the C++
support. On Linux, using lld works, but on OSX, this fails because lld
does not support Apple's new .tbd files.
Bug: 155302034
Test: On both Linux and Mac hosts:
cd external/rust; mma
Change-Id: I0ad489113d720bdb9c3b7a67cce9d1e72266f428
We frequently miss putting in the per-test skip condition, and since
mac isn't tested in presubmit this leads to build breakages and time
consuming reverts. To avoid that this blanket disables all the SDK
tests on mac. It's not used there and we got test coverage in many
linux-based builds.
Bug: 145598135
Bug: 161315642
Test: `go test -v ./sdk` in build/soong on mac and linux
Change-Id: I2aea92fef2c0f8c2742396fe36610501dc5a6f0f
rustc expects Rust dylibs to have a .dylib extension on darwin.
Bug: 155302034
Test: rust_library_host outputs have correct extension on darwin.
Change-Id: I17b0d7eaf0f42f8186185e3f14c3a8e992c74fde
This relands https://r.android.com/1356826 without changes. It got
reverted as part of b/161315642 but is not related to that issue.
Test: m nothing
Test: `m` with prebuilt SDKs in the manifest
Bug: 151303681
Bug: 161315642
Change-Id: I7af892ae28a8208113e0801b1b8a8204edb61ec8
Currently, the build timestamp saved to the build_date_timestamp
field in the MetricsBase metrics protobuf usually comes from a text
file named build_date.txt. It may not be accurate for metrics purposes.
Instead, use the build start time from Soong main UI for a better
timestamp.
Bug: b/140638454
Test: Ran and checked the build_date_timestamp by running the printproto
command.
Change-Id: I4d893fbbf9830ba21911e56ae13bc5272be47ae6
Remove old TODO setting sdk_version to "current" as the associated bug
appears to have been fixed.
Test: m checkapi
Change-Id: I7737fbd17d2f54fc4578e04d07875b4b8d408ef9
From several dashboards, there are high spikes from several
builds where the real time of executing a part of the build
system took many many hours. This is due to using the wall
time clock when calculating the duration. For metrics, the
monotonic clock should be used for more accurate measurement.
Also, cleaned up the time.go file to use the proper proto
functions when marshalling perf protobuf and deleted
unnecessary functions.
Bug: b/143492984
Test: Wrote and ran unit test case.
Change-Id: I1540cfbf37915e3045ecedaa2595c1f9429bcd62
The property args is currently a string, but represents a collection of
arguments to metalava/doclava. This has resulted in variables being used
rather than default modules to share arguments.
Test: go test java_tests
Bug: 145644363
Change-Id: I670235e8f1cf19bfe8909049c4765637d0accd47
We put reproducible timestamps in zip files so that the artifacts
are consistent, but that leads to old timestamps in the output
directory if they are unzipped as part of the build. Use
unzip -DD when unzipping to update the timestamps.
Bug: 161015009
Test: touch -d 2020-01-01 ref; find $OUT/system -not -newer ref
Change-Id: I70407a627cb070e24be510faa6a774e3d9eae3a8
Use prebuilts of the annotations.zip and api-versions.xml files
when running lint in an unbundled build.
Test: m TARGET_BUILD_APPS=Gallery2 lint-check
Change-Id: Idacf3758a2769678a635941486183673e95b43f8
There was a special case in cc_prebuilt_binary only, which resulted in
prebuilt libraries getting both 32 and 64 bit variants even when their
sources only had one of them, and the other variant would be defunct
since it wouldn't get any prebuilt artefact.
This moves the handling of compile_multilib completely to the common
update code, so that SDK members don't need to deal with it. It doesn't
take SDK member defaults into account, which means a bit more
boilerplate in the snapshots, but it's simpler and less error prone
(different SDK member types have different defaults).
Bug: 151303681
Test: m nothing
Test: build/soong/scripts/build-aml-prebuilts.sh art-module-host-exports
Check that the generated Android.bp passes Soong.
Change-Id: Ib73444c6788ee1c78480bdb103aa2b8ae8f2c63c
