| Age | Commit message (Collapse) | Author |
|---|
|
Bug: 377723687
Test: Unit tests and compare the ninja and mk files generated.
Change-Id: I957c09bba8fc047cb8959461294d8879486185a6
|
|
* changes:
Convert classLoaderContextForUsesLibDeps to use ModuleProxy.
Convert checkSdkVersions to use ModuleProxy.
|
|
|
|
Android lint considers code "test" code depending on if the --test flag
is passed. We pass it by default for *_test module types, but also allow
the user to control it via the "lint: { test: true }" property. However,
the module types were overriding the user-defined test property if
the user-defined one was supplied via a defaults module. Split the
test property into two so that modules can use a separate lower priority
one from the user-controlled one.
Fixes: 358643466
Test: m nothing --no-skip-soong-tests
Change-Id: I1b1ef7a73ca9f413aa29e0c6025134fc52dc7caf
|
|
Bug: 377723687
Test: Unit tests and compare the ninja and mk files generated.
Change-Id: I5d7150feb8dbb84ac4e01059e76f451fbfc5f7eb
|
|
Bug: 377723687
Test: Unit tests and compare the ninja and mk files generated.
Change-Id: Ie937a236f3c286fb5b2cd439614f09f7a85e884f
|
|
In Android Studio R8 is able to use bundled proguard
rules found both in AAR archives (under /proguard.txt)
and JAR archives (under /META-INF/proguard/*)
Soong currently only handles the _easier_ AAR format
and ignores the possibility to have rules in JAR archives.
To support those rules we use the newly introduced
ExtractR8Rules cli tool in R8 to extract all the embedded
rules in the imported JAR for later propagation.
Bug: 377144587
Test: m, including okhttp-4.12.0.jar now doesn't create R8 warnings.
Change-Id: I10f9bbf76bc8daec1ec470fb493e399d1beffc8a
|
|
headerJar and implementationJar have the same classes and methods, but
using implemetnationJar ensures that the jar is present in the
intermediates directory after an `m $module` run.
Bug: 384956156
Test: m android-support-fragment-nodeps out/soong/module_bp_java_deps.json
Test: cat out/soong/module_bp_java_deps.json | jq
'"android-support-fragment-nodeps"'
Test: verified that the jar reported by the above command exists
Test: m nothing --no-skip-soong-tests
Change-Id: Ibd70c455b6bf12958b8f55d6c46ff2553a2f9d84
|
|
Prepare for calling DepIsInSameApex from the apex transition mutator
by splitting all the implementations in two, one that is called on the
outgoing module and only takes the depTag, and one that is called on the
incoming module and only takes the depTag.
apexBundle.depVisitor was passing the child into android.IsDepInSameApex
for both the parent and child paramters. The parent field was only
used to find the type on which to call DepIsInSameApex, so this
effectively used the child's implementation of DepIsInSameApex. That
used to be necessary when the parent and child were of different
module types, as the parent module type may not have been aware
of the rules for the child module type, but is no longer necessary
with split outgoing and incoming DepIsInSameApex.
Bug: 372543712
Test: all soong tests pass
Change-Id: If7c81ec3f7b1ea69d77e9ad7694e238820194e59
|
|
Bug: 377723687
Test: Unit tests and compare the ninja and mk files generated.
Change-Id: I221f4ea7479d512d49cfb40a8cb06bba62ec2c7a
|
|
Aconfig files can be provided to java_api_library in two ways:
1. By specifying the aconfig_declarations module names in the
aconfig_declarations property
2. By specifying the java_aconfig_library (or its static rdeps) module
names in the libs/static_libs property
1. was behaving as expected for java_api_library, but 2. was not
behaving as expected and java_api_library is erroneously not collecting
the aconfig files from the libs/static_libs dependencies. This change
fixes such broken behavior.
Test: CI
Bug: 379807097
Change-Id: I1d6d1fa4eae32b4f28e68c88e4e67b2ff7ae7805
|
|
A rule that uses echo -e may execute in a dash shell instead of bash
when running on the build servers. The dash echo builtin is posix
compliant, and does not treat -e as an option[1]. Replace all uses
of echo -e with echo if it does not need escape sequences, or printf
if it does.
[1] https://wiki.ubuntu.com/DashAsBinSh#echo
Bug: 378931009
Test: manual
Change-Id: I2974f59ea10a613bc3bd36e3591fd81f38a1769a
|
|
[retry reverted 45dca5c8cc538392edbde4ee1077fc86509b2e38 after rebase]
Implemented test_module_config for sh_tests.
There are some shell tests that have 130 cc_binaries and they get
duplicated in the vts .zip. test_module_config allows deduping for
lowmem and hwasan use cases freeing a 1 G in the zip.
Moved some of the knowledge about constructing the make rules back to
the provider.
I haven't tried this for cc_test or rust_test, but I think it will be
similar to sh_test if we ever need it.
There is currently only one usecase for using test_module_config with
sh_test (and none for cc_ or rust_)
Also added "extra_test_configs" to sh_test
Sample bp file.
test_module_config {
name: "vts_ltp_tmc",
base: "vts_ltp_test_x86_64",
test_suites: [
"general-tests",
"vts"
],
options: [
{name: "is-low-mem", value: "false"},
{name: "is-hwasan", value: "true"},
],
}
Test: go test ./
Test: m vts_ltp_test_x86_64_hwasan
Test: atest vts_ltp_test_x86_64_hwasan
Test: m vts # inspect the zip for symlinks
Test: CtsAppTestCases_android_server_am
Test: atest -c -v vts_ltp_test_x86_64:vts_ltp_test_x86_64_hwa
Change-Id: I208797f97e3448bcff7434630adfd7476fc275f9
|
|
|
|
This reverts commit 45dca5c8cc538392edbde4ee1077fc86509b2e38.
Reason for revert: DroidMonitor: Potential culprit for http://b/377964351 - 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: Idd5e4b0d04b3c04dc1cfc4541615fc5f42a85c3e
|
|
|
|
Implemented test_module_config for sh_tests.
There are some shell tests that have 130 cc_binaries and they get
duplicated in the vts .zip. test_module_config allows deduping for
lowmem and hwasan use cases freeing a 1 G in the zip.
Moved some of the knowledge about constructing the make rules back to
the provider.
I haven't tried this for cc_test or rust_test, but I think it will be
similar to sh_test if we ever need it.
There is currently only one usecase for using test_module_config with
sh_test (and none for cc_ or rust_)
Also added "extra_test_configs" to sh_test
Sample bp file.
test_module_config {
name: "vts_ltp_tmc",
base: "vts_ltp_test_x86_64",
test_suites: [
"general-tests",
"vts"
],
options: [
{name: "is-low-mem", value: "false"},
{name: "is-hwasan", value: "true"},
],
}
Test: go test ./
Test: m vts_ltp_test_x86_64_hwasan
Test: atest vts_ltp_test_x86_64_hwasan
Test: m vts # inspect the zip for symlinks
Test: CtsAppTestCases_android_server_am
Test: atest -c -v vts_ltp_test_x86_64:vts_ltp_test_x86_64_hwa
Change-Id: I5a36d798fd21180d653d09e54f40449a0ebcbd29
|
|
|
|
the apex" into main
|
|
As part of removing some of the complexity in Soong around using
stub vs. implementations for shared library dependencies a syntax
will be added to Soong to allow explicitly selecting stubs vs.
implementation. To avoid incorrect use, add a verification pass
on apexes that ensure that all transitive implementation libraries
used to link native libraries or binaries in the apex are
themselves in the apex.
Bug: 372543712
Test: TestApexVerifyNativeImplementationLibs
Flag: EXEMPT host only
Change-Id: I4aeaca00a359ce97e8f9efd2d8bffb8f9d2dc0df
|
|
Bug: 342332820
Test: TH
Change-Id: I3d440c79231cc470013610c8b3ec95a0b4a0cb6a
|
|
Invoking HideFromMake today would also make the module uninstallable.
This was a safe assumption since the packaging sytem was in build/make -
if a module is not visible to make, it is also uninstallable.
With the advent of soong packaging system, this might not always be
true. This CL disentangles SkipInstall from HideFromMake. This CL also
sets SkipInstall on modules where the installation rules should not be
emitted, e.g. to prevent duplicate installation rules between platform
and apex variants of libraries
Test: m nothing --no-skip-soong-tests
Test: no diff in aosp_cf_x86_64_phone's system file_list.txt
Change-Id: I80cdd60d2ebdba22fd37e748eb00242cc412bda1
|
|
|
|
This reverts commit a8bf946a22d5e7022c5fd2c748da226f0ee0a16e.
Reason for revert: Replaced with aosp/3322568
Change-Id: Ided7e261d50fbc3e41d4d731ae86b69258b2e7a2
|
|
Test: m --no-skip-soong-tests nothing
Bug: 292141694
Flag: EXEMPT host test change only
Change-Id: Ide365dce9ff55ab07d304b10d660317c83f827fc
|
|
There are some host tests in internal that rely on the vendor variation
of other modules, that they include in their `data` property.
Add explicit properties for these dependencies so that they don't rely
on the 1-variant fallback.
Bug: 372091092
Test: m nothing
Change-Id: I51a50db2a06fdd3f885ac21f3ad3ce15386feb53
|
|
* changes:
Update DepSet references
Move DepSet to blueprint
Convert DepSet to a wrapper around a pointer
|
|
Update all references to depset to use blueprint/depset, and to use
DepSet instead of *DepSet.
Bug: 375276086
Test: all soong tests pass
Flag: EXEMPT refactor
Change-Id: I59a7836d0975366ddc336225fb770ac7e6e0c8ea
|
|
The testing package defined test_spec and code_metadata module types,
but they were never used and the owner left google.
Bug: 372091092
Test: Presubmits
Change-Id: I2d34ce173ec46483719a1d333180937b3da1fa04
|
|
Some java host tests add data dependencies on
`compile_multilib: "first_prefer32"` modules. In order to match that
with the requested variations, add a device_first_prefer32_data
property.
Bug: 372091092
Test: m nothing
Change-Id: I54d056cc864512b22ad3da882c807052344ac1ca
|
|
|
|
These new properties are essentially methods to specify "outgoing
transitions" in blueprint files. There are lots of host tests
that want to include apps built for device in their data, so they
need a property that adds dependencies based on the device variants
instead of copying the same host variants.
After this cl is submitted, I'll do an LSC to update all the usages
that are relying on the 1-variant fallback to use these properties
instead.
Bug: 372091092
Test: m nothing --no-skip-soong-tests
Change-Id: I45b8fb024da120ad61606e3a21de86e4392be2a4
|
|
Requested for the mk2bp conversion.
Fixes: 373772769
Test: m nothing --no-skip-soong-tests
Change-Id: I9a36e73f20c73e25b555fb3f0a074b30a548ae50
|
|
`none` is not a soong module, and can be skipped.
Test: go test ./java
Test: verified that `none` no longer appears in deps of
`android-non-updatable.stubs.from-text`
Bug: 373837953
Change-Id: I3bf60b762a52f1b236dde8dd020aa291697939b2
|
|
* changes:
Remove MutatorHandle.Parallel()
Fix TestFinalDepsPhase for parallel mutator
|
|
All mutators are parallel by default, remove the explicit Parallel()
calls.
Bug: 372540665
Test: all soong tests pass
Flag: EXEMPT refactor
Change-Id: I41e3a71bd13c75b7fceae91b1c4cfe678ab62df2
|
|
Previously, java binaries had both common and the first arch variants.
This was origionally added in aosp/556843. The reasoning there was so
that a java binary could be used as the dependency of a genrule.
However, now with transition mutators, we can make the incoming
transition redirect the arch variant to the common variant if the module
only has a common variant. This allows genrules to depend on the common
arch variant easily. That change is also included in this cl.
Bug: 372091092
Test: m nothing --no-skip-soong-tests
Change-Id: Iea612d050bff608d661f81566884653239015213
|
|
This adds jni_libs support back to device variants, which was removed in
https://r.android.com/3230584. These will be installed by the two packaging
systems in different ways
1. Kati: By being listed in `LOCAL_REQUIRED_MODULES` of the
autogenerated Android.mk
2. Soong: Via the`jniInstallTag`, whose `InstallDepNeeded` value is true
This CL also makes it an error for a native library to be listed in
`required` of a `java_binary(_host)`.
Test: m nothing --no-skip-soong-tests
Test: Verified that there are no diff in
system_intermediates/file_list.txt (device variants)
Test: m installclean && m <host_java_bin> ; verified that the jni
library is installed in out/host/linux-x86/lib64 (host variants)
Bug: 370110572
Change-Id: I6aebc648191ab39f2da15a5f2807d46abab22cb0
|
|
|
|
Bug: 323382414
Bug: 370563864
Test: m nothing --no-skip-soong-tests
Change-Id: I11d8350d338e44d27537c6e5529034ba8a38265c
|
|
|
|
|
|
The enforce_strict_updatability_linting and apex_strict_updatability_lint
are some of the last top down mutators, and removing them will help
with incremental analysis. Both mutators are used to propagate a flag
to transitive java dependencies that causes them to add extra checks to
their lint rules to require that baselines not include any skipped
NewApi checks.
Instead of modifying dependencies to check the baselines, propagate the
baselines up to the modules that are requesting the checks, and perform
the checks on all transitive baselines there.
Bug: 367784740
Test: TestJavaLintStrictUpdatabilityLinting
Test: TestApexStrictUpdtabilityLint
Flag: EXEMPT refactor
Change-Id: Ief2e3b26d745da61f13e621d635a5879d9c56779
|
|
|
|
Instead, provide the information of the source/prebuilt java_sdk_library
to the rdeps via the SdkLibraryInfoProvider.
Test: m nothing --no-skip-soong-tests && diff ninja file
Bug: 348040422
Change-Id: If6cd3cd260a8ce8dccead7f302840cabf68a9fae
|
|
Bug: b/348040422
Test: lint_test.go
Flag: EXEMPT refactor
Change-Id: I420bee2c7056a3f4acee3af5955f79c504ea61d9
|
|
* changes:
Remove internal deapexer module
Remove deapex support from java_*_import
|
|
prebuilt_apex/apex_set currently supports its deapexed contents to be
returned via a sibling java_import/java_sdk_import module. This is
necesssary for
1. dexpreopt/hiddenapi processing
2. usage as shared library
(1) is no longer necessary. this information is provided by the
top-level prebuilt apexes now
(2) is no longer possible since `exported_java_libs` has been removed in
https://r.android.com/3272110
This CL uses a hack for java_sdk_library_improt. Even though (1) is
provided by the top-level apex, there are still some places where
dexpreopt/hiddenapi processing visits the import modules. This CL uses
a bogus path to make analysis work. If this bogus path gets used, there
will be an error during ninja execution
Test: go test ./apex
Test: in internal lunch cf_x86_64_phone-next-userdebug (uses mainline
prebuilts)
Test: verified that file_list.txt is same
Bug: 368337090
Change-Id: I0ea2327f648f0fc60e337b232f7265e140772ffd
|
|
modules"
These are all remaining usages of "magic", which are being removed with
this change. The follow up change will remove the "magic".
Test: m nothing --no-skip-soong-tests
Bug: 366069293
Change-Id: If432cdc6621ac00564e14acbaefca53db1723686
|
|
This change allows the impl lib of java_sdk_library to call AddContext(...) instead of AddContextMap(...) and ensure that the behavior between the top level java_sdk_library and the impl lib is identical when creating the class loader context.
Test: CI && ABTD
Bug: 366572133
Change-Id: Ia0bc7558ca737fcd3ad4ed62b3190aa567aa103f
|