This way config.mk no longer needs to check which java is in PATH and
fix it. It'll be consistent for all build steps under soong_ui.
Also unify handling of ANDROID_JAVA_HOME / JAVA_HOME with
OVERRIDE_ANDROID_JAVA_HOME / EXPERIMENTAL_USE_OPENJDK9.
Test: m nothing
Test: build/soong/soong_ui.bash --make-mode nothing (w/o envsetup.sh)
Test: aosp_arm ninja files are the same before/after
Test: before/after ninja files match with OVERRIDE_ANDROID_JAVA_HOME
Test: before/after ninja files match with EXPERIMENTAL_USE_OPENJDK9
Change-Id: Icdb65093d9c346524074de239a4f895e4230a24d
Strip debug info in javac when PRODUCT_MINIMIZE_JAVA_DEBUG_INFO
is set.
Test: m with PRODUCT_MINIMIZE_JAVA_DEBUG_INFO=true
Change-Id: I167e742662801291c516bf1ff826486560d22147
Also adapt to the renaming of d8 to d8-compat-dx.
Test: USE_D8=true m -j16 PRODUCT-aosp_arm64-eng showcommands core-oj
Change-Id: Ic7aaa6a5da34a6a5706c20f53922a83cb1de1100
This is done by implementing the HostToolPath interface for java
Binary objects.
Currently, in order for a java_binary_host to be used in this way the
java_host_binary must include:
compile_multilib: "first"
Otherwise the genrule is unable to find the correct variant.
Test: ./build_test.bash
Test: Manual
Bug: 68397812
Change-Id: I6849488dc13fa8c383df69d00f62ad815ec17876
The min-sdk-version issue is solved.
Same goes for the ART layout gtests failure found on the art host bots.
Bug: 67569208
Test: m checkbuild / tests
Change-Id: I88c7f5e9ea5c5c5d1b13404ba5db1479f9f7f2f7
kotlinc expects -classpath, not --classpath.
Also add a test that uses only .kt files to exercise the code when
there are no java sources.
Test: java_test.go
Change-Id: Ifa5a007b460b40ea2188d0907570fbdca6c48da7
Source jars were not working as designed because javac will only
compile files from the -sourcepath if there are references to them
starting from files on the command line. Switch to extracting
the source jars into a directory and passing a list of the files
to javac.
Test: m checkbuild
Change-Id: I9f7d824f8538d081b2f5ad64ae3cbfd0e96213af
OpenJDK9 system modules are not build modules, which means they
can't be built with m core-system-modules. Add a phony target.
Test: m core-system-modules
Change-Id: If6d512ff7a009b49743fb25cbb566935ec1c0153
ALLOW_MISSING_DEPENDENCIES=true can cause missing dependencies to be
hidden from GenerateAndroidBuildActions, which is triggering the
len(jars) == 0 check in compileJavaHeader. For the
ALLOW_MISSING_DEPENDENCIES=true case the check is unnecessary, as the
build rules will be replaced with runtime errors. On the off chance
the check would have been hit in a real build, allow merge_zips to
create an empty zip file if it is asked to.
Test: m ALLOW_MISSING_DEPENDENCIES=true
Change-Id: I3171f921d51229ddf38a1a647d32566658c673fa
Also adds checks that the dependencies are android.Modules and
are not disabled.
Test: m checkbuild
Change-Id: I05e945f38915d49cd3c0ab72a86576949bc7eff2
Now that android.ModuleContext does not include blueprint.ModuleContext
we can rename android.ModuleContext.ModuleBuild to
android.ModuleContext.Build without colliding with
blueprint.ModuleContext.Build. Leave ModuleBuild as a wrapper around
Build for now to avoid having to update all the users outside
build/soong simultaneously.
Test: m checkbuild
Change-Id: I18eb8cc04faf002049a11d9aac97e9732ff5d638
The source for dx may not be available in PDK builds, use the
prebuilt one from prebuilts/build-tools instead.
Bug: 67663308
Test: m TARGET_BUILD_PDK=true
Change-Id: I9090b5190539f901fc05264a472133c12d4ea2a1
Copy the logic from cc.go to fail when a module depends on a disabled
module. A future change will refactor this to remove the duplication
and cover all other module types.
Test: m TARGET_BUILD_PDK=true doesn't panic
Bug: 67663308
Change-Id: Iab2b142cebdbd74df934e4733f0de83bd3334d86
aidlFlags was always looking for a src directory at the top level
instead of in the module's source directory.
Test: m checkbuild
Change-Id: Ie31baa21c8632c2bd9a6ebc42e6291417b1f31cd
Ninja doesn't support depfiles on a rule with multiple outputs.
Use a single output and put all the rest as implicit outputs.
Bug: 68057449
Test: java_test.go
Change-Id: Ia544493b1b3b51b185c865149d8f3e0eb3c57ee2
The android_*stubs_current modules don't yet exist, so fall back to
the prebuilt jars for now.
Test: m checkbuild
Change-Id: Iabd32b30954b3f4a6d9a779fde52a032b684807e
If sdk jars(android_stubs_current, etc) are compiled using soong java
modules, we have to filter them when running Java build with Turbine.
TODO: provide more unit-tests.
Test: m clean && m -j32; go test java_test
Change-Id: Iad7c241b0e8b0ca760950733f513124b56c84564
Make all the jar intermediates called modulename.jar, and put each
in a subdir for the tool that generated it. This will simplify
using the jars as inputs to genrules and resources.
Test: java_test.go
Change-Id: If00e16bd7df5a4ba24ebc4b68c9ccf2cfda49544
Modules may need to use jarjar on the device to move classes into
the android namespace, but not on the host.
Test: m checkbuild
Change-Id: I910ebbe20e45e98edecca0d7c4fb18e806bc3c6c
Allow java libraries to specify .kt sources, precompile them with
kotlin, and then pass them in the classpath to javac.
Bug: 65219535
Test: java_test.go
Change-Id: Ife22b6ef82ced9ec26a9e5392b2dadacbb16546f
srcFileLists was an ill-fated attempt to deal with generators that
produce a set of java sources that is not known ahead of time.
For example, the list of files produced by protoc depends on the
package statement in the .proto file. srcFileLists put the list
of generated files into a file, which was then passed to javac
using the @file syntax. This worked, but it was too easy to cause
missing dependencies, and will not work well in a future distributed
build environment.
Switch to putting generated sources into a jar, and then pass them
jar to javac using -sourcepath.
Test: m checkbuild
Change-Id: Iaab7a588a6c1239f7bf46e4f1b102b3ef517619b
Pass the output file name into the java.Transform* functions.
This consistently puts control of the filename into java.go,
which is often necessary to avoid collisions when the same
rule is used multiple times in a single module. It also
has the side-effect of removing the poorly named "stem"
parameters.
Test: java_test.go
Change-Id: I7bc1d1f3bfae6f9d2c92870e6df381817817aab4
Adds a java_system_modules module type that (when
EXPERIMENTAL_USE_OPENJDK9 is set to true) converts a list of
java library modules and prebuilt jars into system modules,
and plumbs the system modules through to the javac command
line.
Also exports the location of the system modules to make
variables, as well as the name of the default system module.
Test: TestClasspath in java_test.go, runs automatically as part of the build
Bug: 63986449
Change-Id: I27bd5d2010092422a27b69c91568e49010e02f40
Use a helper function to set up errorprone and javac compiles.
Test: m -j checkbuild
Test: m -j RUN_ERROR_PRONE=true
Change-Id: Icef3a5e1b359487eea3c3306d3d5763dab912b38
If hostdex: true is specified for a java library, create an extra
Make module that copies the dex jar to a module with a -hostdex
suffix in the host output directory.
Bug: 67600882
Test: m -j checkbuild
Change-Id: I859dfaabeefdca714b566de94e00f74e03c85939
Generated source files were being included twice, once manually
and once by ctx.ExpandSources. Remove the manual one.
Bug: 67364649
Test: TestGeneratedSources in later patch
Change-Id: Ia6760b01a34a96767a914bdcb911a39fa6683300
java_resource_dirs was using a file list file that was generated
at ninja time to get the list of files to include, which meant
there were no dependencies on the files to cause res.jar to get
rebuilt. Switch to using a glob at soong time instead. This
is substantially similar to a glob in java_resources, except that
java_resource_dirs strips the listed directories off the paths
that end up in the jar.
Test: TestResources in java_test.go
Change-Id: I4b9b38f7b6b38a013cbb4e211187e7282a6795c0
