This change disallows Java modules in the vendor partition to use System
SDK that is newer than API level 34; 34 is the latest allowed.
Background 1: with Trunk Stable, the system/vendor interface is released
at Q2 whereas the system/app interface is released at Q3. In other
words, at Q2, the APIs which will be added to the system SDK at Q3 are
not available. Since the system/vendor interface (which is fronzen at
Q2) is what the modules in the vendor partition will be building
against, they can't and shouldn't use those new APIs that will be added
in the future (Q3). Using those APIs is risky because there's a chance
that those APIs get removed or changed between Q2 and Q3. For example,
2024 Q2 is technically still Android U, not Android V.
Background 2: The use of Java APIs in the vendor partition had many
issues. Most significantly, those "vendor" Java apps are categorized as
part of the system partition because all Java app processes require
access to platform internal libraries that are prohibited to vendor
processes. Furthermore, since the Project Treble, the vendor partition
was re-purposed to a partition to host SoC-dependent bits - usually
HALs. Implementing HALs in Java has never been officially supported and
has had many loop holes.
We'd like to use both background 1 and 2 as a chance to disallow any
Java code in the vendor partition. However, since there are already some
Java modules in the partition, we can't suddenly ban it. The deprecation
will be made gradually, and this CL is the start.
Note that sdk_version: "current" or "system_current" is automatically
overridden into 34 or system_34. This is to prevent sudden breakage of
vendor modules that have been targetting the latest (i.e. current) API
level. They will however fail if they use APIs newer than API level 34.
Bug: 314011075
Test: m blueprint_tests
Change-Id: I59f5ac15ce9ac2ff7cc89e9c110169359077c37c
hiddenapi processing require boot jars from apexes to determine the full
set of methods available at runtime.
When building with prebuilts, this comes via
java_import/java_sdk_library_import, which acts as a hook for
prebuilt_apex/apex_set. If we have multiple apexes in the tree, this
hook becomes 1:many. This CL prepares platform_bootclasspath to select the right
deapexerd .jar files when mutliple prebuilts exist.
Implementation details
- Create a dependency edge from platform_bootclasspath to
all_apex_contributions (DepsMutator)
- For every boot jar, query all_apex_contributions to get the path to
dexjar file (GenerateAndroidBuildActions)
Some other important details
- This CL does not drop the old mechanism to get the dex file (i.e. by
creating a dep on java_library). Once all mainline
modules have been flagged using apex_contributions, the old mechanism
will be dropped
- This CL has a functional change when building with source apexes. At
ToT, the unecoded hiddenapi dex jar is used for package check and
generating the monolithic stub file. After this change, the hiddenapi
encoded file will be used for these operations.
This should be fine since the
package and dex signature do not change across the encoded and
unencoded dex file. In fact, we already have a split today. When
building with prebuilts, we use the encoded dex file. When building
with source, we use the unecoded dex file.
Test: Added a unit test
Test: Manual testing in internal described below
- lunch cf_x86_64_phone-next-userdebug
- flagged com.google.android.adservices using apex_contributions
- aninja -t commands out/soong/hiddenapi/hiddenapi-stubs-flags.txt # no
diff before and after
Bug: 308790777
Change-Id: I72c70f0ae1b587679203ea254c9c12a48e7aa782
This is needed to be compatible with JDK 21+ which does not allow
other platforms (apart from the ones listed in the modules
plugin) when linking a system image.
Note the value of target platform doesn't matter to us, we just
choose a closest one to replace "android".
Bug: 313924276
Test: m EXPERIMENTAL_USE_OPENJDK21_TOOLCHAIN=true
Change-Id: Id805f31089fdbd78bf4db06c89ae391f25563448
hiddenapi and dexpreopt require boot and system server jars from apexes.
When building with prebuilts, this comes via
java_import/java_sdk_library_import, which acts as a hook for
prebuilt_apex/apex_set. If we have multiple apexes in the tree, this
hook becomes 1:many. This CL prepares dex_bootjars to select the right
deapexerd .jar files when mutliple prebuilts exist.
Implementation details
- Update prebuilt module types (prebuilt_apex/apex_set) and source
apexes to set a map of
library name to dex jar path on host.
- dex_bootjars will access the path of the .dex jar on host via the
provider. These then
copied/installed to the right locations.
This CL does not drop the old mechanism to get the dex file (i.e. by
creating a dep on java_library). Once all mainline
modules have been flagged using apex_contributions, the old mechanism
will be dropped
Bug: 308790457
Test: git_master-art-host:art-gtest https://android-build.corp.google.com/builds/abtd/run/L21500030000926533
Test: git_main:art_standalone_dexpreopt_tests https://android-build.corp.google.com/builds/abtd/run/L99000030000891212
Test: Added a unit test that checks that the right .jar is selected
when multiple prebuilts exists
Change-Id: I6ef94135b9303a35135810930af4b641df13a583
Since OtherModuleHasProvider will be deprecated soon, replace it with OtherModuleProvider
Test: go build ./java
Change-Id: I79736e44d85bc6e8c97f08ebf783b40533a3e6ae
Writing raw files as rules in the ninja file unnecessarily bloats
the ninja file. Write files immediately to disk instead to files
based on the hash of the contents, and then emit ninja rules to
copy the files into place during the build. Delete obsolete files
in a singleton at the end of analysis.
Bug: 306029038
Test: Run: m libc_musl_version.h
touch build/soong/Android.bp
m libc_musl_version.h
libc_musl_version.h/genrule.sbox.textproto is not recopied.
Test: Run: lunch aosp_cf_x86_64_phone-userdebug
m libc_musl_version.h
lunch aosp_x86_64-userdebug
m libc_musl_version.h
lunch aosp_cf_x86_64_phone-userdebug
m libc_musl_version.h
libc_musl_version.h/genrule.sbox.textproto is recopied but restat prevents rerunning the genrule.
Test: Run: touch out/soong/raw-aosp_cf_x86_64_phone/00/foo
touch build/soong/Android.bp
m nothing
out/soong/raw-aosp_cf_x86_64_phone/00/foo is removed.
Change-Id: I172869c4d49565504794c051e2e8c1f7cf46486e
Generating boot image requires a .prof file provided by the ART apex.
When building with prebuilts, this comes via the
prebuilt_bootclasspath_fragment module, which acts as a shim for
prebuilt_apex/apex_set. If we have multiple prebuilt apexes in the tree,
this shim becomes 1:many. This CL prepares dex_bootjars to select the
right .prof file when multiple prebuilts exist.
Implementation details
- Update deps mutator of dex_bootjars to create a dep on
all_apex_contributions. The latter contains information about which
apex is selected in a specific release configuration. dex_bootjars
will create a dependency on the selected apex in a postdeps phase
mutator.
- All apex module types (apex, prebuilt_apex and apex_set) will set a
provider that contains info about the location of the .prof file on
host
- dex_bootjars will access the provider of the selected apex to get the
location of the .prof file
This CL does not drop the old mechanism to get the .prof file (i.e. by
creating a dep on {prebuilt_}bootclasspath_fragment). Once all mainline
modules have been flagged using apex_contributions, the old mechanism
will be dropped
Bug: 308790457
Test: Added a unit test that checks that the right .prof is selected
when multiple prebuilts exists
Change-Id: I40fdb21416c46bed32f6ff187ce5153711ec2c69
Previously, the path to built files in the sandbox would be
out/.intermediates/... instead of out/soong/.intermediates/....
After this cl, it will be out/soong/.intermediates/.... This makes
it more consistent with the non-sandboxed paths, which is easier for
developers.
CtsApkVerityTestDebugFiles is a genrule that's used to find the paths
to other modules. Developers were expected to build it, then copy all
the files listed in its output file to a temporary directory. Those
paths would be wrong before this change.
Bug: 307824623
Test: ./build/soong/tests/genrule_sandbox_test.py CtsApkVerityTestDebugFiles
Change-Id: Iadf9e3240a2c828567e46b6b02cc14004f30e8bf
System property was removed in
https://r8-review.git.corp.google.com/c/r8/+/82140, as the
system property com.android.tools.r8.emitRecordAnnotationsInDex
is the only system property for not desugaring records.
Test: Existing
Fixes: b/316501817
Change-Id: I057a18c9e02a99365e910d3128890fd481e93541
FindDeapexerProviderForModule raises an exception if multiple apexes in
the tree has an export dep on the java module. In prepartation to
support multiple prebuilts, move this error check out of
FindDeapexerProviderForModule and into rdeps. i.e. raise an exception
only if an rdep calls DexJarBuildPath
- This should be a no-op for now.
- In the short-term future, a java import module will be allowed to have
multiple deapexers. An error will be raised if anyone actually tries
to depend on the dexjar
- In the long-term future, this function will be removed. All processing
will be done at the prebuilt apex level and not at the prebuilt java
library level
Since this check now happens in the moduleCtx of rdeps, add some
additional props to unit tests to ensure that it does not exit early on
unrelated validation checks (e.g. hidden_api prop is not set)
Test: go test ./apex ./java
Bug: 308790457
Change-Id: I3323d993c1ea8f43305834cae8e65b6fe41dfefd
dexpreopt of apex system server from prebuilts involves three soong
modules
1. prebuilt_apex / apex_set
2. an internal deapexer module created by the prebuilt apex
3. java_import/java_sdk_library
(3) acts as a shim for the deapexer to set the dexjar extracted from the
prebuilt apex. This methodolody requires a 1:1 correspondence across the
three modules
This breaks down when we have multiple versions of the same prebuilt
apex in the tree. In preparation for this, move the dexpreopt
processing from (3) to (1). Each prebuilt_apex will create the necessary
rules for dexpreopting the jars deapexed from itself. In the future,
apex_contributions will be used to pick which service-foo.{odex|.vdex} to
install depending on which prebuilt apex is selected.
Implementation details
- Embed dexpreopter in prebuiltApex structs so that this module type can
register the dexpreopt rules. Since a single apex can have multiple
system server jars, this also requires creating an additional scope in
dexpreopt.go to prevent name collisions
- Add the dexpreopt modules as required in initApexFilesForAndroidMk
- Add the depreopt modules to androidMk in AndroidMkEntries. Drop the
equivalent from java_import and java_sdk_library_import
Bug: 308790457
Test: existing soong unit tests
Test: lunch cf_x86_64_phone-next-userdebug && m out/target/product/vsoc_x86_64/system/apex/com.google.android.adservices.apex
Test: Verified that the above command installs
/out/target/product/vsoc_x86_64/system/framework/oat/x86_64/apex@com.android.adservices@javalib@service-adservices.jar@classes.{odex|vdex} and the equivalent files of service-sdksandbox
Test: presubmits
Change-Id: I01cea8956d2857fb864b415e73d3d2686d069b5e
Convert all of the callers of ModuleProvider/ModuleHasProvider to use the
type-safe android.SingletonModuleProvider API.
Bug: 316410648
Test: builds
Change-Id: I6f11638546b64749e451cebbf33140248dc1d193
Convert all of the callers of OtherModuleProvider/OtherModuleHasProvider
to use the type-safe android.OtherModuleProvider API.
Bug: 316410648
Test: builds
Change-Id: Id77f514d68761a262d9ea830a601dbed804bbbe5
Convert all of the callers of Provider/HasProvider to use the type-safe
android.ModuleProvider API.
Bug: 316410648
Test: builds
Change-Id: I73479de1625fa2865b6c73444cd477e50d56dc5a
Using generics for the providers API allows a type to be associated
with a ProviderKey, resulting in a type-safe API without that doesn't
require runtime type assertions by every caller.
Unfortunately, Go does not allow generic types in methods, only in
functions [1]. This prevents a type-safe API on ModuleContext, and
requires moving the API to be functions that take a ModuleContext as
a parameter.
This CL creates the new API, but doesn't convert all of the callers.
[1] https://go.googlesource.com/proposal/+/refs/heads/master/design/43651-type-parameters.md#no-parameterized-methods)
Bug: 316410648
Test: builds
Change-Id: I3e30d68b966b730efd968166a38a25cc144bd6de
TestAarImportProducesJniPackages was attempting to assert that an
aar_import module always had a JniPackageProvider, but was instead
asserting that the return type was always a JniPackageInfo. Since
ModuleProvider returned the zero value of JniPackageInfo even when
there was no provider that was always true.
Fix aar_import to always set JniPackageProvider.
Test: TestAarImportProducesJniPackages
Change-Id: Idaf106e14373490d4eb807892e174aaaf094bbcf
Convert all of the callers to NewProvider and NewMutatorProvider
to use a generic type parameter instead of an example object.
Bug: 316410648
Test: builds
Change-Id: Ic9cdafc87336e26730d3fd596df05de0e7267542
Convert all of the callers of SetProvider to use the type-safe
android.SetProvider API.
Bug: 316410648
Test: builds
Change-Id: If58f4b5355264ddab2045bc3591a4eac19cd58fc
