The functionality of m* envsetup.sh commands has been implemented in
soong_ui. The bash functionality of m* commands can be replaced by
soong's modules-in* build commands.
Fixes: b/130049705
Test: * Loaded the envsetup.sh on a new environment
* Executed each m* command.
* Verified with "type <build command>" to make sure
the function generation is correct.
* Ran on Darwin environment.
Change-Id: I8792e4f1f1727aba2ca6ee4bdfe657f6ac42302a
Python 2 and 3 behave differently when calling ZipFile.writestr() with
zinfo.external_attr being 0. Python 3 uses `0o600 << 16` as the value
for such a case (since
18ee29d0b8),
which seems to make more sense. Otherwise the entry will end up with
0o000 as the permission bits. This CL updates common.ZipWriteStr to
follow the logic in Python 3, in order to get consistent behavior
between using the two versions.
Bug: 131631303
Test: `python -m unittest test_common.CommonZipTest`
Test: `python3 -m unittest test_common.CommonZipTest`
Change-Id: If8429855d922ef1ad76591f703215a0ce5089f0f
Previously it was using regular dict.
Test: python -m unittest test_common.DynamicPartitionsDifferenceTest
Change-Id: If108a4512aeaf9d3c8775c030cad6e44342b9d3d
Pure GSI build targets has no vendor partition, such as
aosp_$arch_ab and gsi_$arch. The system properties defined by
PRODUCT_PROPERTY_OVERRIDES will be in /system/build.prop.
The patch defined a fake BOARD_VENDORIMAGE_FILE_SYSTEM_TYPE to
let these system properties flow to vendor and won't pollute the
system.img.
The bug also move some properties to /product/build.prop.
Bug: 135508595
Bug: 131162245
Bug: 134781120
Test: check the /system/build.prop do not have "ro.carrier=unknown"
Test: adb remount on GSI Q on P
Change-Id: Ib200d66cf98fea572c26338e058bce29eb5e0cd7
Currently, GSI does not include apex support and TARGET_FLATTEN_APEX is true.
This will cause issues for vendors with apex support (e.g. Cuttlefish).
This change set ro.apex.updatable to false in product to override that sets
in vendor.
Bug: 135411972
Bug: 134673003
Test: $ lunch aosp_x86-userdebug; m -j; emulator
$ lunch aosp_cf_x86_phone-userdebug
# Replace system.img in super.img with GSI
# The resulted CF could boot and browse the web successfully.
Change-Id: I08fd7a1b254aac276926329e064c35b714764936
Instead of just depending on the single image file, depend on the entire
image that gets loaded -- the .oat / .vdex files, and the modules that
are part of the image that have their own files.
This is necessary for RBE, where the rule only gets the files that it
depends upon.
Test: treehugger
Test: build a system image with RBE
Change-Id: I93a580b42cf4ef1824ee1397e551c065f083348a
While searching for some fs_config related symbols, I ran across this
being out of date, so update it.
The new documentation defers to the help section of the underlying
script more to hopefully avoid this in the future.
Test: n/a
Change-Id: I94893cbb29b8d7e81963a12e3f8316ac3f41b7ae
Previously, setting PRODUCT_BUILD_SUPER_PARTITION to false for a partial
build (with PRODUCT_USE_DYNAMIC_PARTITIONS == true) would fail to
include necessary keys in misc_info.txt that are required when merging
two partial builds to create a dynamic-partition-enabled mixed build.
This change ensures these necessary keys are included even when
PRODUCT_BUILD_SUPER_PARTITION is false. Setting
PRODUCT_BUILD_SUPER_PARTITION to false causes partial builds to skip
building super.img and super_empty.img, instead relying on these images
to come from the final merged build.
Bug: 134764140
Test: Building & booting a dynamic-partition-enabled mixed build, and
inspecting partial builds' logs / out folder to ensure that
super.img/super_empty.img were not created.
Change-Id: I99431a9a342e9b0617510e250597f3024ef39322
BOARD_SUPER_PARTITION_ERROR_LIMIT can be set by OEMs to cause a
build failure when the sum of sizes of logical partitions exceeds the same.
It is set to BOARD_SUPER_PARTITION_SIZE by default.
Bug: 133329143
Test: mmm -j32
Change-Id: Ief85c9ab3e3b9d071cd22b70e13b5c64bb784b9c
Bug: 134509111
Test: add "require_root: true" to init_benchmarks and libpower_test
build the modules, confirm the extra target preparer is added in the
test config.
Change-Id: I2fdae79d45fd1e5866ee94d1f0e59df106be2a87
Current information is not enough if the prebuilt modules are depend on
AAR libs Therefore, AIDEGen has to collect the classes.jar of the
prebuilt modules from build system.
Build module-info.json without this change:
Build time: 1m55.001s
File size: 14,918,354 Bytes
Build module-info.json with this change:
Build time: 1m56.292s
File size: 15,577,031 Bytes
Bug: 132768299
Test: 1. Checkout the internal master branch
2. Patch this CL
3. m -j out/target/product/generic_x86_64/module-info.json
4. Open the module-info.json
5. For verifying soong_java_prebuilt.mk, check the classes.jar of
module prebuilt-google-play-* exists.
e.g.
"prebuilt-google-play-...": {
...
"classes_jar": [
"out/target/common/obj/JAVA_LIBRARIES/
prebuilt-google-..._intermediates/classes.jar"
]
}
5. For verifying java_prebuilt_internal.mk, check the classes.jar
of module ink exists.
e.g.
"ink": {
...
"classes_jar": [
"out/target/common/obj/JAVA_LIBRARIES/ink_intermediates/
classes.jar"
]
}
Change-Id: I09518c92260db47d2686493fa13951f316159d13
* changes:
Fix dependency on BOARD_AVB_SYSTEM_KEY_PATH
Fix some java library dependencies
Depend on NDK libraries
Call clean-path in more places
Add missing dependecy on SOONG_ZIP
Add dependencies on the toolchain (clang, etc)
When an Android.mk specifies a local path that is outside the current
directory, it uses '..', since we prepend $(LOCAL_PATH)/ to the given
path.
This path ('a/b/../c') gets inserted into command lines and
dependencies. For dependencies, when it gets to Ninja, Ninja calls
CanonicalizePath and removes those references ('a/c'), but the command
line is preserved.
So we've got a command line that references 'a/b', but no dependency on
it. Usually that's not a big problem, but it's an issue if we're trying
to only expose dependencies to the rule, like with RBE only sending the
dependencies to the remote workers. So use our 'clean-path' macro to
collapse the '..' references in Kati so that they don't show up in the
command line either.
Test: treehugger
Test: build a system image with RBE
Change-Id: I0706faa2ac15e9c12c720f43c7bb8171c5efd97c
Historically, we've relied on the path to the clang compiler changing in
order to trigger rebuilds instead of direct dependencies on the compiler
itself. That's somewhat more reliable, since the actual $(CLANG) file is
often a script that doesn't actually change during every update.
In Soong, we've added the dependency anyway, so do it here as well. This
makes it easier for my RBE experiments to identify when we intend to use
clang and to send the entire compiler.
Test: treehugger
Test: build a system image with RBE, find that every command that needed
the compiler gets it.
Change-Id: I3838b5d77884394a8c0f7a1ba133142102602084
JVM occasionally (0.5% time) crashes on entering a method with SIGSEGV.
It might be due to excessive load on the host machine. Reduce the number
of the compiler threads to 6, and do not start all the GC threads
immediately. If JVM crashes, copy full diagnostics to stderr.
Bug: 132766811
Test: treehugger
Change-Id: Ic1b09e0f2cbcf4c22214eaaeae464899b2b99ef7
These were found when trying to run remotely on RBE with only the
sources depended upon available for each rule.
Bug: 130111713
Test: treehugger
Change-Id: Id763f8fc7dfbe60445f98604db3422147165f537
