* Current default is not using lld.
* When USE_CLANG_LLD or LOCAL_USE_CLANG_LLD is true or 1,
* Use *GLOBAL_LLDFLAGS instead of *GLOBAL_LDFLAGS.
GLOBAL_LLDFLAGS should call lld and with correct lld flags.
* set my_pack_module_relocations to false.
Bug: 73768157
Test: make checkbuild
Change-Id: I3e63cf8ae0865d01d2bc1f36e9304f4a5d092cb8
Vendor apks should be able to reference native vendor libraries.
Vendor apks have worked around this by building without
LOCAL_SDK_VERSION, which then allowed to use all libs. However,
since vendor apks now needs to be built with SystemSDK, that
workaround can no longer be used.
Bug: 76398918
Test: BOARD_SYSTEMSDK_VERSION=P m -j
Change-Id: Idb13d5db71f4dfd542658483b6a24e7ece18ce26
libclang_rt libs are prebuilt libs that have different names by the
architecture they are built.
Bug: 77931086
Bug: 77816590
Bug: 67002788
Test: 'm check-vndk-list' for various architectures.
Change-Id: Iacb3979b6e5df7e9ba8073470aab23c603b4db55
This change allows removing some vendor properties from
(vendor|system/vendor)/build.prop file based on a blacklist.
For WearOS Unified Builds, which can change the product name depending on
the chosen locale, we use runtime-generated value for ro.build.fingerprint,
but since the ro.vendor.build.fingerprint cannot be generated the same way,
we always hit a "Mismatched fingerprints" error.
Bug: 71555551
Test: manual
Change-Id: Ifad793187e930a28fbf9325b03468c7ea86076b7
Currently the build system will automatically attribute a NOTICE
file with the target of $(BUILD_PHONY_PACKAGE). This shouldn't
be the case.
Disable notice file inclusion for fake targets so that the
/fake_packages/blah_blah-timestamp paths don't show up in
NOTICE.xml.gz.
Bug: 77910458
Test: NOTICE files are not attributed to fake targets.
Change-Id: Ia942cac41b750efbd5a23d896d85ac0820ee8b4e
We've had these warnings for a while, but we've still got a lot of
warnings that people have been trained to ignore. Lets just silence them
for now.
The Merged-In is just a random change from AOSP that isn't in pi-dev, to
stop this from being merged to master, where I'm just fixing the
warnings.
Bug: 77883700
Test: m <fewer warnings>
Change-Id: I00d8173a527760d62dca5837bcc86d5f03b87e83
Merged-In: I78ca65e6f0c81f09e7da848eda797b3a8f97a521
There is an extra ) in the implicit output path for R.txt, which causes
the rule to rerun every time because of a missing output file. There
is already an implicit output for R.txt on line 180 (which is why
the incorrect path didn't cause an immediate "No rule to generate R.txt"
error), so just remove the incorrect one.
Bug: 77244156
Test: m checkbuild && m checkbuild
Change-Id: Id960ee211b89a9a5f5104cdcac23bc3124742145
Merged-In: Id960ee211b89a9a5f5104cdcac23bc3124742145
(cherry picked from commit ac48f41ba5)
Instead of passing the entire contents of
$(COMPATIBILITY.$(suite).FILES) to eval, which may keep that string
around, delay the evaluation of that, and the new files until inside the
eval.
This saves ~2.8GB: 7.4GB -> 4.6GB of ckati max resident memory for a
relatively small internal build. It also saves ~10% of the makefile
loading time (81 -> 73 seconds).
Bug: 77879502
Test: build-aosp_arm.ninja is identical
Change-Id: If45a4796f1bbf6d67dff388ea877a6115a4e06f4
Merged-In: If45a4796f1bbf6d67dff388ea877a6115a4e06f4
(cherry picked from commit a4155e8512)
There is an extra ) in the implicit output path for R.txt, which causes
the rule to rerun every time because of a missing output file. There
is already an implicit output for R.txt on line 180 (which is why
the incorrect path didn't cause an immediate "No rule to generate R.txt"
error), so just remove the incorrect one.
Bug: 77244156
Test: m checkbuild && m checkbuild
Change-Id: Id960ee211b89a9a5f5104cdcac23bc3124742145
Chih-Hung Hsieh