These entries are proprietary, which are defined in vendor/google/build
in git_main, but vendor projects do not exist in
aosp-main-future-without-vendor branch. This change adds entries that
are supposed to be included in the future aosp release.
Ignore-AOSP-First: Intended aosp-main-future only change
Test: m nothing
Bug: 338660802
Change-Id: If48f56757e069a467a8ae0466820f27f97b011a8
This reverts commit bb678f82d6.
Reason for revert: Likely culprit for b/362670437 - verifying through ABTD before revert submission. This is part of the standard investigation process, and does not mean your CL will be reverted.
Change-Id: I21a9fbc79618fce48f5713fdb9856f3eb751ec45
This reverts commit bb678f82d6.
Reason for revert: Droidmonitor created revert due to b/362639013. Will be verifying through ABTD before submission.
Change-Id: I1dcb71672eeb44a17cf176d3731bcaf9aaf81edf
Bugs: me
Test: clear error message, for instance:
error: frameworks/native/libs/binder/ndk/Android.bp:252:1: module "libbinder_ndk.ndk" variant "android_x86_64_silvermont_sdk_shared_29": OS mismatch between "libbinder_ndk.ndk" (android) and "libbinder_headers_platform_shared" (linux_glibc)
Change-Id: Idfa6d27daeb89056781c93a893e4080e49b3fcb0
The external code that depends on this field has been changed to use the
new interface, this field can be removed now.
Bug: 358425833
Test: Manually verified all the generated ninja and mk files and CI.
Change-Id: I12cb6e22c091bc90d3a3cbbf554e81b8dc27e264
The IDEInfo() methods read properties. To make those properties
configurable, we need a context to evaluate them with.
Bug: 362579941
Test: m nothing --no-skip-soong-tests
Change-Id: I26d4b7084439b3006e50b02277298f74a929e1aa
In order to keep the changes minimal, call the old IdeInfo method right
after GenerateAndroidBuildActions and then set the result of that
into a provider.
Test: Diff'd ninja files and saw no difference.
Change-Id: Ic706e7dbcb7c6e3d2a381ddae383eb87e105f385
AddFarVariationDependencies runs into issues in partner workspaces which
do not contain the sources of google mainline prebuilts, since the
prebuilt would be renamed in a prior prebuilt_rename mutator. To account
for this rename, add `prebuilt_` prefix in
`addDependenciesOntoSelectedBootImageApexes` to ensure that the prebuilt
apex does not get elided
Test: m nothing --no-skip-soong-tests
Test: in v-aml-prebuilt dev,
aninja -t query out/soong/dexpreopt_arm64/dex_bootjars_input/okhttp.jar | head -n3
out/soong/dexpreopt_arm64/dex_bootjars_input/okhttp.jar:
input: g.android.Cp
out/soong/.intermediates/vendor/partner_modules/ArtPrebuilt/prebuilt_com.google.android.art_compressed.deapexer/android_common/deapexer/javalib/okhttp.jar
(uses the prebuilt deapexed boot jar, and no the source apex boot jar)
Bug: 362163831
Change-Id: Ie94a36911ab40ac09a470efe505fd00ef3a5bf9c
* changes:
Skip `-Wl,--version-script` on libclang_rt.* stubs (temp hack)
Special case the stubgen args for bionic libs
Add stub generation support to cc_prebuilt_library
This allows us to get most of the benefits of genrule sandboxing,
and defer deciding what to do about the build number until later.
Bug: 307824623
Test: m aidl_camera_build_version
Change-Id: Ib3f4dd6d270ac56d068593c95ee40db44962b845
https://r.android.com/3083813 added support for overriding
min_sdk_version, and creating separate variants of transitive
dependencies of override apexes. This causes issues when there is a skew
between apex boot jars of source and prebuilt mainline modules. For
mediaprovider google apex, we now have _two_ source variants of
mediaprovider bcp fragment module.
This CL uses ApexInfo.BaseApexName to skip the check on both variants of
bootclasspath/systemserverclasspath fragments.
This CL also removes the prefix based check to determine if prebuilts
are active. If RELEASE_APEX_CONTRIBUTIONS for a specific mainline module
is set, it will be assumed that prebuilts for that mainline module are
active.
Test: go test ./apex ./java
Bug: 361771739
Change-Id: I01f7ef437df8f10b74ffa397ba461ff4f1f9bed6
