Permissions of apex variants of shared java sdk libraries should use the
base apex name (com.android.foo) and not the override apex name
(com.mycompany.android.foo). The override apex name started being used
when separate variants were created for override apex transitive deps
(https://r.android.com/3083813)
Bug: 346486531
Test: m com.google.android.ipsec
Test: verified that
out/target/product/<product>/apex/com.google.android.ipsec/etc/permissions/android.net.ipsec.ike.xml
contains /apex/com.androiqd.ipsec/javalib/android.net.ipsec.ike.jar #
previously it contained com.google.android.ipsec
Test: https://android-build.corp.google.com/builds/abtd/run/L59900030004679761
Change-Id: I863170d78efa017bef5e4ece3b1471f74ba84b0c
flags_packages property was added to supports flagging the manifest
files. The listed names of `aconfig_declarations` modules are passed to
aapt2. However, this is currently scoped to the module level, and is not
propagated to the reverse dependencies. In other words, if the manifest
is flagged with `featureFlag` property, all of the reverse dependency of
the android_app/android_library should specify `flags_packages` property
in the bp module definition, leading to huge toil for the users.
In order to resolve such inconvenience, this change modifies the build
rules of android_app,android_library and runtime_resource_overlay such
that flags_packages of the static dependencies are collected.
Test: Patch ag/27816261 && m Settings --no-skip-soong-tests
Bug: 347289274
Change-Id: I4c3855541dd09cb72293515eb2626eaf4ae8c0df
Having a message name and field name that are the same causes C++
generated code to fail.
Bug: 347076012
Test: manual, TH
Change-Id: I198e92dc906d476881ef351e603ef2ea63ce5848
This is another place where we can trade performance for additional
binary size savings.
Test: presubmit
Change-Id: Icf9600ef2ca4092e4adccc0c26465b542b038f7b
Some of the getters were unused, which means that all the fields
could be made into local variables.
Test: m nothing --no-skip-soong-tests
Change-Id: Iea38ed7f9da952803f54194c0c5a8fda9b6007f5
In OutputFiles method, TestingModule is able to get its own output
files by reading its module base property. If the TestingModule
never updates its outputFiles property, it will fall back to use the OutputFileProducer interface.
Only empty string tag case is added in this CL since all the testing modules are getting the output files using empty string tag.
Test: CI
Bug: 339477385
Change-Id: I3009ae45d8d909653e3b6b222bced9ccabbaede7
This reverts commit ce320f86a0.
Reason for revert: <Dorid monitor: Likely culprit for b/347143262 - 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: I70eabe014acc6feb123f28ac487f7d7d39e9fb6b
Currently, when you do `m` repeatedly, it does a little bit of
rebuilding to copy release config files around. If we change
release-config to only rewrite the files if they've changed, we
get a proper "ninja: no work to do." message.
Bug: 346757289
Test: m repeatedly
Change-Id: I9c1f6d34ec20d14b684a0183c5ec457ea92440f9
