So that they're set properly during dumpvars / product loading.
SHELL is already set in common/core.mk, so it's not necessary to be in
core/main.mk as well (which quickly loads core/config.mk, which loads
common/core.mk).
Test: get_build_var TOP
Test: get_build_var TOPDIR
Change-Id: I8179ac32628b74ccf29851dc53bf83cfb4b280fd
- Update the conversion from LSDUMP_PATHS to lsdump_paths.txt.
LSDUMP_PATHS is a list of "tag:path". Each element is converted to
"tag: path\n" in lsdump_paths.txt.
- Update the check for redundant ABI dumps.
prebuilts/abi-dumps/ndk contains the ABI dumps for NDK.
prebuilts/abi-dumps/vndk contains those for LLNDK, VNDK-core, VNDK-SP,
except NDK.
Test: make findlsdumps
Bug: 133176785
Change-Id: I90559e6c68a7f901a318cd8d5125e9a7401ac22c
Now that ONE_SHOT_MAKEFILE no longer exists, we don't have to rely on
the filesystem to store this informtion.
This removes ~16.7k files from our build graph
(aosp-master/aosp_arm64-eng), though only about 600 of them were being
used in a normal build.
Test: treehugger
Change-Id: I3ac12f5ea7f11d25064109a0599bc5be1976fba5
Now that mm/ONE_SHOT_MAKEFILE have been removed, we can expect to know
about all of our dependencies at the end of the build.
This removes 19k nodes from our build graph (aosp-master
aosp_arm64-eng), though in a default build, only 3k of those are used.
Test: ALLOW_MISSING_DEPENDENCIES=true, then trigger a missing dependency
Test: treehugger
Test: create link_type files, then apply CleanSpec.mk, ensure they're removed
Change-Id: I9506331e4a9911d2f26e59a2f72a97aef1644073
The build servers are no longer using this. It's hard to search for
others using it -- I've cleaned them up as I've run across them, but
we'll just have to see who breaks once this goes in.
This was the last user of makeparallel, so we can remove it.
Test: make
Test: treehugger
Change-Id: If6df3f1a67d6a6df36afaa4b07cd88a48a364fe1
I've switched all the build server configs to using TARGET_PRODUCT /
TARGET_BUILD_VARIANT and explicit goals instead.
Remove tools/check_builds.sh which relied on this, but hasn't been
touched in a long time.
Test: m PRODUCT-test
Test: treehugger
Change-Id: If5f8c714567b33aeb38223c176ca24ea649eb57d
apart.
Also make it more clear what the problem may be when the "find" check fails,
to aid troubleshooting.
Test: m
Change-Id: Iacdc4734f20eeaf3fb0d29e418df21f5089cecbc
This will allow us to unhardcode relative paths for native-bridge
binaries in soong.
Test: make PRODUCT-cf_x86_phone-userdebug
Change-Id: I72b33ee610c92d41e1370fcaa1512529c5d7c88c
This location allows the agent to be attached to arbitrary processes.
It only makes sense to include libdumpcoverage.so on coverage builds,
as these are the only builds that have any information to be dumped.
Test: manual, used examples in README (see other CL) to test whether
it works on a userdebug_coverage build on cuttlefish
Change-Id: Ib2fece1b41a3b5d16c8a2a444c5486137e475fda
Strip bitness information for PRODUCT_PACKAGES and
PRODUCT_HOST_PACKAGES before checking against ALL_MODULES.
Also correct spelling: nonexistant -> nonexistent
Bug:
If PRODUCT_ENFORCE_PACKAGES_EXIST was set, the build system would
complain about nonexistent packages for e.g. HALs where the bitness was
specified using a suffix, such as android.hardware.audio@4.0-impl:32
Test:
Add android.hardware.audio@4.0-impl:32 to PRODUCT_PACKAGES, specify
PRODUCT_ENFORCE_PACKAGES_EXIST, ensure no whilelist is set.
"make systemimage vendorimage" -> No error.
Signed-off-by: Felix <google@ix5.org>
Change-Id: Id59460a19320aa9437b8805411a0d97fa6432633
All users have been removed.
Test: treehugger
Change-Id: I2cfe720bef88cc5485e90746d85a7e0a5d068ef7
Merged-In: I2cfe720bef88cc5485e90746d85a7e0a5d068ef7
arm64 guest libraries are located in lib64/arm64
Bug: http://b/77159578
Test: lunch cf_x86_64_phone-userdebug in internal master; mm -j48 dist
Change-Id: I9f1f9426cbaed0a269126cb242e1a6df9249b167
Running 'm samplecode' fails with
"build/make/core/main.mk:1834: error: overriding commands for target
`out/target/common/samples/libsimplejni.so', previously defined at
build/make/core/main.mk:1834"
This proves that noone has been using 'samplecode' target at least since
the build started generating both 32- and 64-bit shared libraries by default,
because development/samples/SimpleJNI/jni/Android.mk defines libsimplejni.so
as 'sample', and the deleted code in core/main.mk causes duplicate rule in
such case.
Test: treehugger
Change-Id: I38a4c478fcb0e60793129fe2513a8cdc6d2c79fb
When ONE_SHOT_MAKEFILE is set, the targets of LOCAL_*_REQUIRED_MODULES
may not exist, so skip the error message.
Fixes: 131241991
Test: mmm cts/tests/signature/intent-check
Change-Id: I95c4c901610746854179af8486c8c4767dd18864
They should only be in the Runtime APEX, c.f. discussion on
http://r.android.com/957159.
Test: build/soong/build_test.bash through Forrest
Bug: 124293228
Change-Id: I33e97a56e67c3f4734d14f38e0f1597b697a2604
Bug: 132197773
Test: property present in vendor/build.prop
Currently, the property resides in product partition which gets
overwritten by the GSI image. Moving it to vendor will ensure
that the property is set even when a GSI image is flashed on the device.
Change-Id: I4d4a3c473194e15ba124a121f89bbb3ec1a73d19
Primarily for use with build_test during presubmit.
Moved the check to main.mk to avoid nonobvious dependency on
product_target_FILES from main.mk to Makefile.
Test: m systemimage
Test: build/soong/build_test.bash
Test: m out/target/product/taimen/system/lib/libdexfile.so; m systemimage;
check that "find"-based check fails
Test: Add a system library to APEX_MODULE_LIBS; m systemimage;
check that "product_target_FILES"-based check fails;
env DISABLE_APEX_LIBS_ABSENCE_CHECK=true m systemimage;
check that it works
Bug: 124293228
Change-Id: I7a93cc0ac4480404a3d56c75eb43ae499e8a3a05
If set to true,
- super.img is built from images in $PRODUCT_OUT
- super.img is built to $PRODUCT_OUT
- super.img is built when 'make'.
'make dist' still builds super image from images in target files to
intermediates directory.
This flag is useful for virtual devices, but isn't intended to be set for
actual devices. For actual devices, userspace fastboot should flash
super_empty.img and individual user images separately.
Test: on cuttlefish (flag is set),
make # generates $OUT/super.img
make superimage # generates $OUT/super.img
Test: on real non-retrofit DAP device (flag is not set)
make # no super.img is generated
make superimage # generates $OUT/obj/PACKAGING/.../super.img
make dist # generates out/dist/super.img
Bug: 113175337
Change-Id: Ieb81e4fbb663bb4b69e9962c3fa9f16b03aeb907
The two new debugging images adds additional files based on
boot.img and ramdisk.img/ramdisk-recovery.img, respectively.
File /force_debuggable is to trigger special logic in /init to load an
userdebug version of sepolicy and an additional property file from this
ramdisk to allow adb root, if the device is unlocked.
It's intentional to skip signing for boot-debug.img, as it can
only be used if the device is unlocked, where verification error
is allowed.
Those debugging images allows adb root on user build
system.img, vendor.img, product.img, etc. This can facilitate more
automated testings on user builds and is helpful to narrow down the
delta between what's being tested v.s. what's being shipped.
Bug: 126493225
Test: `make dist`, checks both boot-debug.img and ramdisk-debug.img
are in $OUT/ and out/dist.
Test: `make dist`, checks installed-files-ramdisk-debug.{json,txt} are
in out/dist.
Test: `system/core/mkbootimg/unpack_bootimg.py --boot_img $OUT/boot-debug.img`,
checks the extracted out/ramdisk is as expected
Test: Run `gunzip -c ramdisk | cpio -idm` for the ramdisk extracted from
$OUT/boot-debug.img and $OUT/boot.img, respectively.
Then compare the root dirs of both, e.g.,
`diff -rq --no-dereference ./ramdisk ./ramdisk-debug`
Test: `make ramdisk_debug-nodeps` and `make bootimage_debug-nodeps`
Change-Id: I30137c3caef91805d9143d404e5e4d06c0fccc30
Core platform API violation reporting is disabled by default and can be
enabled by setting the persist.debug.dalvik.vm.core_platform_api_policy
property. Set it to "just-warn" for non-user builds and leave disabled
on user builds.
Test: builds, boots
Bug: 125701194
Change-Id: I2f4be42373de9fdbc71c3178de6d34e07809f13a
So that we actually respect different LOCAL_REQUIRED_MODULES for the
host and device versions instead of unioning them. That got particularly
problematic when LOCAL_SHARED_LIBRARIES is implicitly added to
LOCAL_REQUIRED_MODULES. We also used to walk through device-only modules
when filling out the list of required modules, which triggered even more
extra installations.
This also changes the requirements for PRODUCT_HOST_PACKAGES so that it
no longer accepts target-only phony modules (since we can now
differentiate them). They were all removed in previous patches.
Test: treehugger; diff resulting builds
Test: diff list of product_target_FILES and product_host_FILES
Change-Id: I2ed8950320d31f5693323ad8cef6ec5b6780b7d4
The standard target modules stopped working when restrictions in
soong were introduced to which binaries are allowed to be used during
builds.
Listing a very large amount of modules in columns does not make it
more readable and harder to work with in tools so just drop it.
Bug: 129800175
Test: make modules
Change-Id: I26040479a03916161fb5d072de1af640d8080c7f
This CL simplifies the PRODUCTS.$(INTERNAL_PRODUCT).X accesses of
product variables, and removes unnecessary stripping of them.
Replace: '\$\(PRODUCTS\.\$\(INTERNAL_PRODUCT\)\.([^\)]*)\)' with '$(\1)'
Replace: '\$\(strip\s*\$\(PRODUCT_([^\)]*)\)\)' with '$(PRODUCT_\1)'
A few minor manual tweaks.
Bug: 116769560
Test: presubmit
Change-Id: I70c54f1582e3cc780028535960147d99ebc2e0e1
With this change, all PRODUCT_ variables are treated the same
when it comes to stripping and assigning them to their final
variable name. In the past, all the PRODUCT variables needed
to be listed in two places to achieve this.
The documentation previously attached to the strip/assignment
is moved to the PRODUCT_ variable list in product.mk.
Also refactor some of the default value logic to cope with
the new automation.
Many places in the build system that currently refer to
$(PRODUCTS.$(INTERNAL_PRODUCT).X) can now be modified to
use $(X) directly.
Bug: 116769560
Test: verified noop on PRODUCT_ variables on all products in the tree
Change-Id: I5677c355e81359b1d3c0db2a2232941097a05047
This change changes auto-generated RROs from DEVICE_PACKAGE_OVERLAYS
to be generated in the vendor partition, as opposed to /product where
they were generated in the past.
Note that PRODUCT_PACKAGE_OVERLAYS continue generating RRO packages
to /product, which means that a single app can be overlayed from
different partitions. These RROs have been given module and package
names based on their location.
Bug: 127758779
Test: verify noop on presubmit targets
Change-Id: I5cee70e28e3969e67b2d83eaf25d9c6e3a11102d
When TARGET_VNDK_USE_CORE_VARIANT is set to true, the vendor variant of
VNDK libraries are by default not installed. Instead, the core variant
will be used by vendor binaries at runtime.
To ensure the core variant of VNDK libraries are installed, we also add
a flag LOCAL_VNDK_DEPEND_ON_CORE_VARIANT to indicate that the vendor
variant module depends on the core variant module. This flag should be
set by Soong for all VNDK libraries without the vendor variant
installed. When the flag is set, the vendor variant binary is also
compared against the core variant binary to ensure they are
functionally identical.
As we are merging the two variants for some libraries, we need a new
link type to denote a module is usable as both native:vndk and
native:platform. We add native:platform_vndk for this.
Bug: 119423884
Test: With the corresponding Soong change, build with
TARGET_VNDK_USE_CORE_VARIANT set to true.
Test: Add a dummy VNDK library and a dummy vendor binary that depends
on it. Build with no-vendor-variant VNDK and check the core
variant is installed.
Test: Add conditional compilation based on __ANDROID_VNDK__ in the
dummy VNDK library and check build fails.
Change-Id: I40000f2728e8193212113c1ee950e9d697f2d40d
Like If1f817d855cbe329b83caee9fdd68c2cce55f02b, but for
PRODUCT_PACKAGES, which is only enabled for a few builds.
Also share the ALLOW_MISSING_DEPENDENCIES check with the host
version.
Test: m product-graph
Change-Id: Iab55072e7d7c0fc9f4680cc515e139a5214dc3b4
