RS_FILES and AIDL_FILES were used for cleaning up the intermediate
src dir, but are no longer used after
If7529979de6fa62a651933a3a974f47b033851d6 and
I01feff7cc399ac5b88b83333a1ac86928d0a81e6.
Bug: 73885582
Test: m checkbuild
Change-Id: If2653f846b568fd2587c8e055ef820e17760f133
Merged-In: If2653f846b568fd2587c8e055ef820e17760f133
(cherry picked from commit c3e7074df1)
Sharding is not used in make any more, the large java modules
(libcore and framework) have moved to Soong. Keeping sharding
support complicates java compiling, so remove it.
Bug: 73885582
Test: m checkbuild
Change-Id: I2932f9e96bfc8f2ebf82095b16d6c0b3c51613f0
Merged-In: I2932f9e96bfc8f2ebf82095b16d6c0b3c51613f0
(cherry picked from commit 5ef6e79c21)
The aapt flags were being set in java_common.mk, which is used by
various entry points. host_dalvik_java_library.mk and
host_java_library.mk, and java_library.mk don't support aapt, only
package_internal.mk and static_java_library.mk do. Move the aapt
flags into aapt_flags.mk, and include it from package_internal.mk
and static_java_library.mk.
Bug: 73885582
Test: no change to build-${TARGET_PRODUCT}.ninja
Change-Id: I124393846d37b9bbc941272cce4274121ac235ef
Merged-In: I124393846d37b9bbc941272cce4274121ac235ef
(cherry picked from commit b6013af085)
PRODUCT_AAPT_FLAGS was removed by Iacc914114616b5bd19d9a1011802f4f9bca9bc19,
remove one more that it missed.
Bug: 73885582
Test: m checkbuild
Change-Id: I31ea4954f5af5af37f13210925ed5d622625d71b
Merged-In: I31ea4954f5af5af37f13210925ed5d622625d71b
(cherry picked from commit 8dad6c62d4)
Dumping files in the *_intermediates/src causes incremental build
problems, because we don't know when to delete them if they are
no longer generated. Switch to outputting a srcjar instead, and
pass it in to the turbine and javac compiles.
Bug: 73885582
Test: m checkbuild
Change-Id: Ibef28507522339d930c09426d0c98372eb864cb6
Merged-In: Ibef28507522339d930c09426d0c98372eb864cb6
(cherry picked from commit 2d1cddd567)
Move R.txt from $(intermediates.COMMON)/src to $(intermediates.COMMON)
in preparation for moving aapt/aapt2 java files out of src. Also
make aapt2 generate R.txt, and make it an implicit output of the rule
and and input to the aar rule.
Bug: 73885582
Test: m out/target/common/obj/JAVA_LIBRARIES/SettingsLib_intermediates/javalib.aar
Test: m checkbuild
Change-Id: Ia858de226355965b69794c002f841333df16a612
Merged-In: Ia858de226355965b69794c002f841333df16a612
(cherry picked from commit 7c24f8ec85)
aapt2.mk will export a srcjar that will be compiled by java.mk, so
it needs to be evaluated first. Also move java_renderscript.mk
before aapt2.mk so that the generated resoures are available.
Bug: 73885582
Test: m checkbuild
Test: rm -rf out/target/common/obj/APPS/RsHelloCompute_intermediates &&
m out/target/common/obj/APPS/RsHelloCompute_intermediates/src/R.stamp
Change-Id: Ide3050ec993a945f7077b75e952c3b3532306ede
Merged-In: Ide3050ec993a945f7077b75e952c3b3532306ede
(cherry picked from commit 8528eabbba)
renderscript_target_api is set by java_renderscript.mk, it doesn't
need to be computed again in static_java_library.mk and
package_internal.mk.
Bug: 73885582
Test: no change to ninja file
Change-Id: Ib6e00be425c385f4911def19ed5042a740f2c79b
Merged-In: Ib6e00be425c385f4911def19ed5042a740f2c79b
(cherry picked from commit 55cce23360)
Move java renderscript support to java_renderscript.mk in
preparation to moving it before aapt. Filter *.rs files
out of LOCAL_SRC_FILES and forbid them in java_common.mk
so that using *.rs files is an error if java_renderscript.mk
was not included.
Bug: 73885582
Test: m checkbuild
Change-Id: Ib37ce89daee58b025dec98f733c5a08149641afd
Merged-In: Ib37ce89daee58b025dec98f733c5a08149641afd
(cherry picked from commit 71822d6fb9)
Replace RenderScript.stamp with a srcjar, with a resource zip as
an implicit output. This makes the renderscript rules self contained,
which will make it easier to move them before aapt.
Bug: 73885582
Test: m checkbuild
Change-Id: I12d33e52019aebac6ea33271939228e4690a4173
Merged-In: I12d33e52019aebac6ea33271939228e4690a4173
(cherry picked from commit e00e2fae52)
Old droiddoc will install the *-docs.zip & API files to the dist.
And also remove INTERNAL_PLATFORM_*API_FILE variables from config.mk and
get rid of the dependencies on api-stubs, system-api-stubs, test-api-stubs
timestamp.
Test: m -j api-stubs-docs
Bug: b/70351683
Change-Id: I08923bb5f62d2f55961d9a2865723d7b3fae7206
Merged-In: I08923bb5f62d2f55961d9a2865723d7b3fae7206
This CL handles a path that uses system-root-image on non-A/B device.
For this path, we can't generate recovery-from-boot patch with imgdiff,
because boot/recovery images contain different number of entries (only
recovery iamge has ramdisk image now).
Using BOARD_USES_FULL_RECOVERY_IMAGE can work around the issue, at the
cost of extra size. Compared to carrying full recovery image, this CL
saves the cost of the kernel size, by putting a patch that's roughly the
size of the recovery ramdisk.
The applypatch executable already detects and handles a bsdiff patch
automatically. No change required to that end.
Note that it won't further reduce the patch size by handling that
ramdisk entry specially, because (a) that's the only difference between
the two images; and (b) there's no corresponding data in boot image to
be diff'd against.
Bug: 72731506
Test: `m dist` with aosp_angler-userdebug. Check the device can install
recovery image successfully (after intentionally corrupting the
recovery image).
Test: Build aosp_angler-userdebug with BOARD_BUILD_SYSTEM_ROOT_IMAGE set.
Verify the generated patch files.
Test: Run validate_target_files.py with the target_files.zips.
Change-Id: I69c06f51ba8c39ae059c5e9a6872a9f10600cf17
Merged-In: I69c06f51ba8c39ae059c5e9a6872a9f10600cf17
(cherry picked from commit 6d5d623987)
if BOARD_BUILD_SYSTEM_ROOT_IMAGE != true: # case A
- BOOT/RAMDISK corresponds to the / under normal boot, with matching
fs_config in META/boot_filesystem_config.txt.
- RECOVERY/RAMDISK corresponds to the / under recovery, with fs_config
in META/recovery_filesystem_config.txt.
else:
if BOARD_USES_RECOVERY_AS_BOOT == true: # case B
- ROOT/ corresponds to the / under normal boot, with fs_config in
META/root_filesystem_config.txt.
- BOOT/RAMDISK corresponds to the / under recovery, with fs_config
in META/boot_filesystem_config.txt.
else: # case C
- ROOT/ corresponds to the / under normal boot, with fs_config in
META/root_filesystem_config.txt.
- RECOVERY/RAMDISK corresponds to the / under recovery, with fs_config
in META/recovery_filesystem_config.txt.
- BOOT/RAMDISK doesn't exist.
This CL fixes case C, where we shouldn't try to generate
'META/boot_filesystem_config.txt' for BOOT/RAMDISK. It wouldn't be fatal
without this fix, but would wrongly scan the current directory and
include a large fs_config output into target-files.zip.
Bug: 72731506
Test: `lunch aosp_bullhead-userdebug` and `m dist`. (case A)
Test: `lunch aosp_marlin-userdebug` and `m dist`. (case B)
Test: Define 'BOARD_BUILD_SYSTEM_ROOT_IMAGE := true' for angler. `m
dist` and check the generated target-files.zip. (case C)
Change-Id: I5582ce8cca464d535af0718be0fd8e65791bd6c2
Merged-In: I5582ce8cca464d535af0718be0fd8e65791bd6c2
(cherry picked from commit da01b211f8)
Tradefed depends on Loganalysis so it should also be part
of cts. Luckily there was no overlap (cts never called a
part of TF that use loganalysis) so it never failed
with classNotFound before.
Test: build cts
Bug: 75979908
Change-Id: I65edf4d1a46c20875e31a61ddfcd343dd7c639aa
Merged-In: I2542ecd077149f7165aff120d67054943df20b54
Tradefed depends on Loganalysis so it should also be part
of cts. Luckily there was no overlap (cts never called a
part of TF that use loganalysis) so it never failed
with classNotFound before.
Test: build cts
Bug: 75979908
Change-Id: I2542ecd077149f7165aff120d67054943df20b54
If HOST_CROSS_OUT_TESTCASES is not defined, then a module that
builds for windows and has a test config will attempt to install
the config file to /<module>/<module>.config.
Fixes:
[ 54% 99/183] Copy: /simpleperf_unit_test/x86/simpleperf_unit_test.exeninja: error: mkdir(/simpleperf_unit_test): Permission denied
Bug: 75971265
Test: m simpleperf_unit_test
Change-Id: I07dcecbce9c710b29fec5fbdedc462be3245b578
Merged-In: I07dcecbce9c710b29fec5fbdedc462be3245b578
(cherry picked from commit e887ccbf12)
This CL handles a path that uses system-root-image on non-A/B device.
For this path, we can't generate recovery-from-boot patch with imgdiff,
because boot/recovery images contain different number of entries (only
recovery iamge has ramdisk image now).
Using BOARD_USES_FULL_RECOVERY_IMAGE can work around the issue, at the
cost of extra size. Compared to carrying full recovery image, this CL
saves the cost of the kernel size, by putting a patch that's roughly the
size of the recovery ramdisk.
The applypatch executable already detects and handles a bsdiff patch
automatically. No change required to that end.
Note that it won't further reduce the patch size by handling that
ramdisk entry specially, because (a) that's the only difference between
the two images; and (b) there's no corresponding data in boot image to
be diff'd against.
Bug: 72731506
Test: `m dist` with aosp_angler-userdebug. Check the device can install
recovery image successfully (after intentionally corrupting the
recovery image).
Test: Build aosp_angler-userdebug with BOARD_BUILD_SYSTEM_ROOT_IMAGE set.
Verify the generated patch files.
Test: Run validate_target_files.py with the target_files.zips.
Change-Id: I69c06f51ba8c39ae059c5e9a6872a9f10600cf17
If HOST_CROSS_OUT_TESTCASES is not defined, then a module that
builds for windows and has a test config will attempt to install
the config file to /<module>/<module>.config.
Fixes:
[ 54% 99/183] Copy: /simpleperf_unit_test/x86/simpleperf_unit_test.exeninja: error: mkdir(/simpleperf_unit_test): Permission denied
Test: m simpleperf_unit_test
Change-Id: I07dcecbce9c710b29fec5fbdedc462be3245b578
For APCT auto-generated configuration, the template will
include the test-suite-tag that will allow it to be
picked up by the current ATP gcl config.
Test: none
Bug: 74440293
Change-Id: I66e9f2f75e43a9371cf8b55130cd49a962753167
Merged-In: Ia82e8287f212f4daec97523394517bfbbe7a86d3
This time around we're really deprecating the 32-bit
binder interface (v7), even for 32-bit devices.
Upgrading devices can keep using it.
Bug: 69775513
Test: manual makefile prodding
Change-Id: I3850418ca267e72593f5f12b940e5382b29b024a
Merged-In: I3850418ca267e72593f5f12b940e5382b29b024a
Use zipsync instead of extract_srcjars.sh.
Also fixes an issue where unzipping a resource zip would
ignore anything that wasn't a *.java file.
Bug: 73885582
Test: m checkbuild
Change-Id: I1c25f65afa196614879071c8655866febebac185
Some javac and turbine compiles have no source files but sources
from srcjars. Continue with compilation if the java source list
is empty but the srcjar source list is not.
Bug: 73885582
Test: m checkbuild
Change-Id: Ica412de2dd1b1bfc12c01c3eafca884220d42771
The latest version of turbine requires a -- after the javac
options, and uses multiple --bootclasspath arguments instead
of a javac-style colon separated list.
Test: m javac-check
Bug: 74339924
Change-Id: I3f694ac375eab61954d4fe63c1166a18314bf75f
* changes:
Allow aapt2 rules to take zipped resource files
Add LOCAL_SRCJARS support
Remove unused .fs file suppport
Remove unused file lists
Remove sharding support from make
Move aapt flags from java_common.mk to aapt_flags.mk
Allow aapt2 rules to take zipped resource files similar to the
way the javac rules take srcjars. For now, unzip the files using
extract_srcjars.sh, and then use aapt2 compile --dir. Eventually
it could be done with aapt2 compile --zip.
Bug: 73885582
Bug: 74574557
Test: m checkbuild
Change-Id: I97ea3ff305e7a9ef2a908b4cf0bca903af62c6b1
Fixed bugs which was discovered in the last attempt.
Bug: 74508478
Test: m -j PRODUCT-taimen-userdebug
Test: m -j PRODUCT-elfin-userdebug
Test: Check that some .vdex files built have R8 generated code
(cherry picked from commit a3b796d998)
Change-Id: Ibe08137cd6db94b1fc487267e169d3f5ee598846
Merged-In: I2a8e0e5cad1d698ae73ba152e34a08a496112d7d
Allow java modules to take jars containing .java files as an input.
For turbine, pass them directly to turbine. For javac, extract them
to a directory and pass a list of the files to javac.
Bug: 73885582
Test: m checkbuild
Change-Id: Ib4351a5b4b165a55e5fb02ee2680466db368983a
RS_FILES and AIDL_FILES were used for cleaning up the intermediate
src dir, but are no longer used after
If7529979de6fa62a651933a3a974f47b033851d6 and
I01feff7cc399ac5b88b83333a1ac86928d0a81e6.
Bug: 73885582
Test: m checkbuild
Change-Id: If2653f846b568fd2587c8e055ef820e17760f133
Sharding is not used in make any more, the large java modules
(libcore and framework) have moved to Soong. Keeping sharding
support complicates java compiling, so remove it.
Bug: 73885582
Test: m checkbuild
Change-Id: I2932f9e96bfc8f2ebf82095b16d6c0b3c51613f0
The aapt flags were being set in java_common.mk, which is used by
various entry points. host_dalvik_java_library.mk and
host_java_library.mk, and java_library.mk don't support aapt, only
package_internal.mk and static_java_library.mk do. Move the aapt
flags into aapt_flags.mk, and include it from package_internal.mk
and static_java_library.mk.
Bug: 73885582
Test: no change to build-${TARGET_PRODUCT}.ninja
Change-Id: I124393846d37b9bbc941272cce4274121ac235ef
Setting LOCAL_AAPT_NAMESPACES := true enables
namespaces for an app and requires any library resources
to be fully qualified by their package.
Test: manual
Bug: 73885582
Change-Id: Ia6bfc82ee8b8d6335f8e7c2abe94ce36467a012d
Merged-In: Ia6bfc82ee8b8d6335f8e7c2abe94ce36467a012d
Not all native tests have the "tests" tag so adjust the logic to be:
NATIVE_TESTS||NATIVE_BENCHMARKS||(APPS&&tags==tests)
Bug: 74339384
Test: Rebuilt module-info.json, verified native tests now have
their suite set to null-suite and have auto_test_config=True
Double checked a frameworks App unittest as well.
Change-Id: Idea00adcebfc1a78b0c6d8ade40feddb1e90ed83
Merged-In: Idea00adcebfc1a78b0c6d8ade40feddb1e90ed83
For APCT auto-generated configuration, the template will
include the test-suite-tag that will allow it to be
picked up automatically.
Test: none
Bug: 74440293
Change-Id: Ia82e8287f212f4daec97523394517bfbbe7a86d3
There isn't an Address Sanitizer prebuilt of xz, so always use the same
xz prebuilt.
Bug: 36130900
Test: SANITIZE_HOST=address m nothing; check ninja file
Test: m libc
Change-Id: Iba3599d4289ad5afb573dbec5ac8fc463668b1c6
This reverts commit 86a596c711.
Reverting as that commit breaks ART tests' results (see
b/73804944).
(cherry picked from commit 209b96e152)
Test: m test-art-host
Bug: 73804944
Bug: 26275726
Change-Id: I06af4b170d845d3e55e044c28ec22d4b3ca3dede
All apps in master comply with this now, but new offenders
are being added daily. This will ensure compliance going
forward.
Test: build/soong/build_test.bash
Bug: 73535841
Exempt-From-Owner-Approval: cp from master
Change-Id: If28049750a9787bf455f72402e624dbddf842253
Merged-In: Id3fbb502d26d91b64480fec74e241ba77fa0faff
I plan on turning the error on for APPS in AOSP soon, and in preparation for
that I'm introducing a finer granularity of warning/error control.
Also add an almost-empty whitelist, which will likely need to be expanded
in the future.
Bug: 73535841
Test: make
Exempt-From-Owner-Approval: cp from aosp and master
Change-Id: I2fc6700a504b7af50aa7bde727047bc56b167937
Merged-In: I13ebe3ead2d19aa797bcc39a7bbccdb55b9c7d1c
Merged-In: I87c968b2e8314300b155483bbb7ce5e169fe8f0c
Pre-extracting the zip files are more sustainable for git, so that
objects can be shared if unchanged, rather than the 900MB zip file
changing on every build.
This also has the advantage that we could put an Android.bp file inside
the PDK, and Soong would just pick it up.
Bug: 68767391
Test: Build mini_arm64 PDK with platform.zip
Test: Build mini_arm64 PDK with extracted zip file
Change-Id: I16db030a731aea55b69c1d6e2260dbd70b167544
Merged-In: I16db030a731aea55b69c1d6e2260dbd70b167544
(cherry picked from commit 0663f685c2)
Modules from Soong do not have dependency specified (via
LOCAL_JAVA_LIBRARIES). So, link-type check 'from' them doesn't make
sense. Resetting my_warn_types and my_allowed_types
Bug: 69899800
Test: m -j checkbuild on aosp_walleye, aosp_sailfish
Test: build/soong/build_test.bash --dist
Merged-In: I2a409e64f388bd8cb9336be589bf6af33e48a6f4
Change-Id: I2a409e64f388bd8cb9336be589bf6af33e48a6f4
(cherry picked from commit 98a4f52c28)
This reverts commit 49b67af95e.
Reason for revert: Missing a bunch of fixes from master.
Test: make
Bug: 73535841
Exempt-From-Owner-Approval: fixes the build
Change-Id: Ide0998bad5876aa409d1bb0f8331e6b097c3c8a2
Merged-In: Id3fbb502d26d91b64480fec74e241ba77fa0faff
(cherry picked from commit 455f2c0223)
Pre-extracting the zip files are more sustainable for git, so that
objects can be shared if unchanged, rather than the 900MB zip file
changing on every build.
This also has the advantage that we could put an Android.bp file inside
the PDK, and Soong would just pick it up.
Bug: 68767391
Test: Build mini_arm64 PDK with platform.zip
Test: Build mini_arm64 PDK with extracted zip file
Change-Id: I16db030a731aea55b69c1d6e2260dbd70b167544
This reverts commit 49b67af95e.
Reason for revert: Missing a bunch of fixes from master.
Test: make
Bug: 73535841
Exempt-From-Owner-Approval: fixes the build
Change-Id: Ide0998bad5876aa409d1bb0f8331e6b097c3c8a2
Merged-In: Id3fbb502d26d91b64480fec74e241ba77fa0faff
All apps in master comply with this now, but new offenders
are being added daily. This will ensure compliance going
forward.
Test: build/soong/build_test.bash
Bug: 73535841
Change-Id: If28049750a9787bf455f72402e624dbddf842253
All apps in master comply with this now, but new offenders
are being added daily. This will ensure compliance going
forward.
Test: vendor/google/build/build_test.bash --dist
Bug: 73535841
Exempt-From-Owner-Approval: reviewed in master/aosp, but this blocks aosp
Change-Id: Id3fbb502d26d91b64480fec74e241ba77fa0faff
Merged-In: Id3fbb502d26d91b64480fec74e241ba77fa0faff
This time around we're really deprecating the 32-bit
binder interface (v7), even for 32-bit devices.
Upgrading devices can keep using it.
Bug: 69775513
Test: manual makefile prodding
Change-Id: I3850418ca267e72593f5f12b940e5382b29b024a
Modules from Soong do not have dependency specified (via
LOCAL_JAVA_LIBRARIES). So, link-type check 'from' them doesn't make
sense. Resetting my_warn_types and my_allowed_types
Bug: 69899800
Test: m -j checkbuild on aosp_walleye, aosp_sailfish
Test: build/soong/build_test.bash --dist
Change-Id: I2a409e64f388bd8cb9336be589bf6af33e48a6f4
This prevents cases where system_server is running in interpreter
only mode.
Removed unused flag in product.mk
(cherry-picked from commit 27f4287406)
Bug: 74209329
Test: WITH_DEXPREOPT=false make
Merged-In: I4ab3afed95a5baf77d0cd089dafaa18bcc1913e5
Change-Id: I4ab3afed95a5baf77d0cd089dafaa18bcc1913e5
Not all native tests have the "tests" tag so adjust the logic to be:
NATIVE_TESTS||NATIVE_BENCHMARKS||(APPS&&tags==tests)
Bug: 74339384
Test: Rebuilt module-info.json, verified native tests now have
their suite set to null-suite and have auto_test_config=True
Double checked a frameworks App unittest as well.
Change-Id: Idea00adcebfc1a78b0c6d8ade40feddb1e90ed83
This prevents cases where system_server is running in interpreter
only mode.
Removed unused flag in product.mk
Bug: 74209329
Test: WITH_DEXPREOPT=false make
Change-Id: I4ab3afed95a5baf77d0cd089dafaa18bcc1913e5
It has been allowed that a java module built with Android SDK or System
SDK to link against other java module built with broader API surface.
For example, an app that is building with SDK (LOCAL_SDK_VERSION :=
current or <number>) can link libs like telephony-common or bouncycastle
which are built without SDK but in fact exposing private APIs.
From now on, this is no longer allowed because it prevents the app from
being unbundled. In general, a Java module A cannot be linked to Java
module B if B is built with broader API surface than A.
Bug: 69899800
Test: m -j checkbuild on walleye, sailfish, and crosshatch
Test: m -j ANDROID_BUILDSPEC=vendor/google/build/app_build_spec.mk
Test: ./vendor/google/build/build_test.bash --dist
Change-Id: Ibfdb1a6777f4e0606927d834d56f808639eb91f0
PRODUCT_AAPT_FLAGS was removed by Iacc914114616b5bd19d9a1011802f4f9bca9bc19,
remove one more that it missed.
Test: m checkbuild
Change-Id: I31ea4954f5af5af37f13210925ed5d622625d71b
Silence warnings when merging header jars that were missed by
Ifb67dbf7734b5f53941d110f578c4fe642e36005.
Bug: 69316739
Test: m checkbuild
Change-Id: I5cfea62e46d99f9b315a589d8915bf62307220aa
framework_res_package_export is now always either the prebuilt
android.jar from an SDK or a package-export.apk generated by
Soong, so the dependency on R.stamp is never useful. Remove
all assignments to framework_res_package_export_deps, and replace
usages with frameworks_res_package_export.
Test: m checkbuild
Change-Id: If484e9eb08061cb0ed0697755f13db71d741aaab
This reverts commit 86a596c711.
Reverting as that commit breaks ART tests' results (see
b/73804944).
Test: m test-art-host
Bug: 73804944
Bug: 26275726
Change-Id: I06af4b170d845d3e55e044c28ec22d4b3ca3dede
Adds Make support for -fsanitze-minimal-runtime when using
the integer overflow sanitizers. This makes the crashes due to these
sanitizers less mysterious. This also cleans up the handling of the
integer_overflow flag.
Bug: 64091660
Test: Compiled and checked the generated compiler commands
Test: Checked program that overflows for the abort reason
Test: Checked integer_overflow flag still emits overflow checks
Change-Id: I11012ed0cbbf51935f549a08bd17109b5ce6f330
if BOARD_BUILD_SYSTEM_ROOT_IMAGE != true: # case A
- BOOT/RAMDISK corresponds to the / under normal boot, with matching
fs_config in META/boot_filesystem_config.txt.
- RECOVERY/RAMDISK corresponds to the / under recovery, with fs_config
in META/recovery_filesystem_config.txt.
else:
if BOARD_USES_RECOVERY_AS_BOOT == true: # case B
- ROOT/ corresponds to the / under normal boot, with fs_config in
META/root_filesystem_config.txt.
- BOOT/RAMDISK corresponds to the / under recovery, with fs_config
in META/boot_filesystem_config.txt.
else: # case C
- ROOT/ corresponds to the / under normal boot, with fs_config in
META/root_filesystem_config.txt.
- RECOVERY/RAMDISK corresponds to the / under recovery, with fs_config
in META/recovery_filesystem_config.txt.
- BOOT/RAMDISK doesn't exist.
This CL fixes case C, where we shouldn't try to generate
'META/boot_filesystem_config.txt' for BOOT/RAMDISK. It wouldn't be fatal
without this fix, but would wrongly scan the current directory and
include a large fs_config output into target-files.zip.
Bug: 72731506
Test: `lunch aosp_bullhead-userdebug` and `m dist`. (case A)
Test: `lunch aosp_marlin-userdebug` and `m dist`. (case B)
Test: Define 'BOARD_BUILD_SYSTEM_ROOT_IMAGE := true' for angler. `m
dist` and check the generated target-files.zip. (case C)
Change-Id: I5582ce8cca464d535af0718be0fd8e65791bd6c2
Typical failure case for boot image dex2oat is an inconsistent boot
classpath left over from incomplete build dependencies. Give advice
to run a top-level build.
Bug: 73749543
Test: m
Change-Id: I81c4ce9d02b2b360fe867b594d0a2b21c763e473
Factor out ANDROID_LOG_TAGS for boot image compilation. Do not use
the setting when ART_BOOT_IMAGE_EXTRA_ARGS is set.
Bug: 73749543
Test: m
Test: ART_BOOT_IMAGE_EXTRA_ARGS="--runtime-arg -verbose:verifier" m art-boot-image
Change-Id: Ia599381991f74f243fee966184715b0172742e78
I plan on turning the error on for APPS in AOSP soon, and in preparation for
that I'm introducing a finer granularity of warning/error control.
Also add an almost-empty whitelist, which will likely need to be expanded
in the future.
Bug: 73535841
Test: make
Change-Id: I2fc6700a504b7af50aa7bde727047bc56b167937
$(OUT_DIR)/target/product/$(TARGET_DEVICE)/lsdump_paths.txt will contain all
.lsdump paths relative to $(ANDROID_BUILD_TOP). This helps faster lookup while
running scripts to generate reference dumps.
Test: m -j findlsdumps for aosp_arm64_ab.
$OUT_DIR/lsdump_paths/generic_arm64_ab/paths.txt has paths to lsdump files
generated for the build.
Test: m -j findlsdumps for aosp_arm_ab.
$OUT_DIR/target/product/generic_arm_ab/lsump_paths.txt has paths to
lsdump files generated for the build.
Change-Id: Iab1640f57bf9d0af5e88e6dda64a610fedcbe87e
Bug: 72552006
Test: Make a module with no source files,
run `m -j nothing`, notice that the error tells
which module has no sources
Change-Id: Ib169e7b3cb86d840a3acd644e42cd1f9f65e1304
If there's a prebuilt with LOCAL_MULTILIB := true, but only a single
LOCAL_SRC_FILES entry, we end up with a weird build error where `cp`
is trying to copy the local directory. Exit early with an error in this
case.
Bug: 73904572
Test: build-aosp_marlin.ninja is identical
Change-Id: Ie2821817c237087a96e87fb9602e430e0f86584a
Instead of relying on whatever version of xmllint is on the host system,
build and use the version in external/libxml2.
Test: diff build_aosp-marlin.ninja, expected changes.
Test: m $(xmllint targets in build_aosp-marlin.ninja)
Test: introduce xml error, build fails
Change-Id: I39579f06db3777e3b5c8dda7c7541c25a35887b2
Another attempt, waiting for b/73762903 to be mitigated
Change-Id: I496b1297f5dfe0737828ab01fc1b628f64a57304
Test: build, flash and cts/tests/signature/runSignatureTests.sh, manual
Bug: 18027885
Before this CL, the build could fail with error messages such as:
Error: out/target/common/obj/JAVA_LIBRARIES/core-oj_intermediates/classes.jar
contains class file jdk/internal/HotSpotIntrinsicCandidate.class, which is not
in the whitelist
This error message was only moderately helpful because it left a few
questions unanswered or misled ("Whitelist for what?", "Where does the
whitelist live?", "Is it a whitelist of class files or of packages?").
This CL clarifies that:
- it's a whitelist of packages allowed on the bootclasspath,
- where it lives (currently
build/make/core/tasks/check_boot_jars/package_whitelist.txt)
which makes the error message more actionable.
Test: manually checked that the error message now looks okay.
Bug: 17434570
Change-Id: I2f52a5e2eb532bc4945bedf9811de5857f67a9a3
<Two phase commits> Since internal master code has more places that use
BUILD_NUMBER (mostly in vendor/) than AOSP (conflict). We can't
deprecate BUILD_NUMBER directly. Therefore, we try to switch to
BUILD_NUMBER_FROM_FILE as much as possible at first. Then we will do
a one-off deprecation for BUILD_NUMBER in internal master next step.
Test: m -j
Bug: b/70351683
Change-Id: I14ffee7381933c9fde14c4bde8c0c14e45fe98bf
Added product property: PRODUCT_DEX_PREOPT_GENERATE_DM_FILES.
If this property is true, APKs compiled as verify will have the dex
files left compressed and the vdex put in a dm file. The vdex file
and oat files are not copied to system partition in this case.
Bug: 70934104
Test: manual
Change-Id: Ie137e14f14642b803a506162de6db8ac65a43f46
The only enforcement right now is that modules don't specify both
LOCAL_SDK_VERSION and LOCAL_PRIVATE_PLATFORM_APIS, but a warning
can be turned on by setting the JAVA_SDK_ENFORCEMENT_WARNING variable.
Bug: 73535841
Test: make w/ and wo/ flag.
Change-Id: I3cc720bf060e2bc789e10884ff888d527b91cb50
The only case where we expect to actually enable this behavior is for
modules that are just an unpacking of a prebuilt, because those are
the only cases for which we can't conveniently just change the source
Bug: 72552006
Test: ./build/soong/scripts/diff_build_graphs.sh \
--products=aosp_arm 'build/make:work^' \
'build/make:work'
Test: cherry-pick the following patch that depends on this one and
the rest of that patch's topic, and then do `m -j jeffapp`
Change-Id: I62e231900d6d87673ddf19c7fc2686fef2828590
Change Ibe4ceb1b224c5e8f3e1a8a59cfee89217765a901 caused kati to
regenerate on every build:
$(shell cat out/build_date.txt) was changed, regenerating...
Switch back to $(shell date +%s) which is special cased in kati
to not cause a regeneration. Everything besides DATE will still
use BUILD_DATETIME_FROM_FILE from soong_ui.
Once BUILD_NUMBER is moved to soong_ui $(DATE) is no longer
necessary and the whole thing can be removed.
Bug: 70351683
Bug: 73352417
Test: m && m doesn't regenerate
Test: m BUILD_DATETIME=1234 shows up in ro.build.version.incremental
Change-Id: I8444beb2b47a49095ef760083fc85f220821bffe
Colin Cross