linkageMutator removes srcs property of the shared variant of a lib in
order to reuse *.o files compiled for the static variant also to the
shared variant.
However, this causes problem when vendor-specific srcs are specified in
target: {vendor: {srcs: ["..."]}}. For example, let's assume
cc_library {
name: "libfoo",
srcs: ["foo.c"],
target: {
vendor: {
srcs: ["bar.c"],
},
},
}
Then,
static_vendor: inputs = foo.o, bar.o
shared_vendor: inputs = foo.o (from static_vendor), bar.o (from
static_vendor), bar.o
So, bar.o is included twice and this causes multiple symbol definition
error.
In order to handle the problem, vendor mutator is applied before the
linkage mutator and the vendor-specific srcs are squashed in the vendor
mutator.
Bug: 67731122
Test: build
Test: cc_test.go
Change-Id: I2a5390295dddfc41260e9b6f02746908cdf47228
Bug: http://b/65598278
Non-PGO builds (with this environment variable set) can be used to
measure the performance difference induced by PGO.
Test: Build PGO modules with ANDROID_PGO_PROFILE_USE set.
Change-Id: Ib23bad5208ac7f54894c7768d7532f53b6b91179
These methods only need the 'PgoProperties' struct. Avoid the extra
indirection by directly using this struct.
Test: Build modules with PGO property.
Change-Id: I1923ebde4b0d546810de8e696514d218b3a4f54b
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
When a lib is explicitly marked as `vendor_available: false`, then it
can't be directly depended by a vendor lib which is installed to /vendor
partition. This is to hide some VNDK libs (including llndk) from vendors
so that platform owners can have a freedom of modifying their ABI
without breaking vendors.
In addition, the list of the private libs are exported to the make world
as VNDK_PRIVATE_LIBRARIES.
Also, fixed a bug that allowed a vndk lib to link against to vendor
library (or vendor variant of a system lib) if the lib is prebuilt.
Bug: 64730695
Bug: 64994918
Test: Add `vendor_available: false` to libft2 and libcompiler_rt.
Add the libs to shared_libs property of a vendor library in soong
(i.e. libnbaio_mono). The build fails with the error message.
Change-Id: Iab575db96bb4f6739a592f3fa0a75124296bea0c
The crash we're following is likely a stack overflow. Use an alternate
stack to actually get the backtraces.
Also write out ulimit to reference what the stack size should be.
Bug: 36182021
Test: ulimit -s 768; build/soong/build_test.bash
Change-Id: I8826e5e53f38af552c8b80a1a8a490fde55f2d4b
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
Putting a genrule dependency after a glob in a srcs property
was causing the generated files to be treated as globbed files
and hitting an interface conversion panic.
Bug: 67364649
Test: TestGeneratedSources in later patch
Change-Id: I8a076c9998fdd07e53769922be433a793ca575a9
This can make it easier to add build/soong/build_test.sh to continuous
build branches to catch new breaks, but where some products are known to
be broken.
Test: build/soong/build_test.bash --skip-products aosp_arm,aosp_arm64
Change-Id: Ib240570b68957077047ca75b78658429cb8c6a49
Bug: http://b/65598278
The profile runtime depends on libgcc for some symbols (only under some
circumstances - armv5, ndk r14, static executables). Since Android
build passes -nostdlib and adds libgcc manually, the profile runtime
gets passed to the linker later than libgcc.
Instead, explicitly add the profile runtime to the linker command (and
pass one other flag added by the clang driver to the link).
Test: Build a library with profile instrumentation that otherwise fails
instrumented build.
Change-Id: I24b34cebd2c3bb6a540f8f4c465ace1be4eb90f3
Signed-off-by: Pirama Arumuga Nainar <pirama@google.com>
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
Honor the default resource exclusions in java_resources properties
including *.java.
Test: m -j checkbuild
Change-Id: Ia117a000680161b54c189758926ddb2068a2a2c3
Dex jars were getting their resources from the res.jar files of
their transitive static dependencies. This accidentally bypassed
jarjar on resources, since the jarjar pass only happened once the
resources jar was combined into the classpath jar. Switch to
getting the resources out of the classpath jar by merging it
with the dex jar while skipping *.class.
Test: m -j checkbuild
Test: compare ext.jar to one generated by make
Change-Id: I5f6f3da738dcb0af56ab9a1bd7174ed5359de2b2
Putting resources in dex jars is going to need to pull files from
classpath jars that do not match *.class.
Test: m -j checkbuild
Change-Id: Ia37b8b9387e5f5636769afc6937b9aa3c9ec1ced
This reverts commit 2370af0e23.
Reason for revert: New Build Breakage: aosp-master/aosp_arm64_ab-userdebug @ 4376965
Change-Id: Ibe4b819c4292457c454bf42e6d94fba3071ec04b
Libraries with no source files but static lib dependencies should
still be installable.
Test: m -j checkbuild
Change-Id: Ibc91aa3153241092ce86888dbf90e99b66069ff2
The partial hostdex support was causing the host variant of
java_library modules to depend on core-oj and core-libart,
which caused the tagsoup jar to use the wrong
java.lang.System.arraycopy signature. Remove the hostdex
code that was causing the problem, and add a test.
Test: java_test.go
Change-Id: I4f7b1f29c99aae328ba19b042538d9d35544aa43
