So that we can more easily enable these new flags on several
modules.
It seems you can still apply a python_defaults to a python_library
even with this change, so I didn't break it out into a separate
python_binary_defaults.
Bug: 245583294
Test: m py_dont_import_folder_of_entrypoint_test && out/host/linux-x86/testcases/py_dont_import_folder_of_entrypoint_test/x86_64/py_dont_import_folder_of_entrypoint_test
Change-Id: Iecf9c1845df4630bafdea9957eb2450c15f8596b
bazel_path conversion would treat "./" as a package boundary
which would be incorrect
Bug: b/243072036
Test: unit test build_conversion_test.go and `m libc`
Change-Id: I0fc44d2fd7a4270442c8755219a11fd585244c5c
The python interpreter will by default add the directory of
the entrypoint script to the beginning of sys.path. This
can be disabled in python 3.11+ (which is not released yet)
using the PYTHON_SAFE_PATH environment variable or the -P flag.
As a workaround to have this behavior in older python versions,
we can make an __soong_entrypoint_redirector__.py file at the
root of the zip file that is the entrypoint, and then that
file will redirect to the real entrypoint.
This brings non-embedded-launcher python modules closer to
the embedded launcher version. The embedded launcher binaries
already act like this because they start at an __main__.py file
at the root of the zip file.
Bug: 245583294
Test: m py_dont_import_folder_of_entrypoint_test && out/host/linux-x86/nativetest64/py_dont_import_folder_of_entrypoint_test/py_dont_import_folder_of_entrypoint_test
Change-Id: I39aaf04fb19c3ba7f5c9d98220872d6d08abf736
The API finalization process for platform releases copies various files
from the apistubs/ dist directory into prebuilts/sdk/<n>/... Having
them use the same name in both places makes that process simpler. For
most modules the name of the file is derived from the name of the
module by appending a suffix but unfortunately, for some modules that
does not work. e.g. the conscrypt.txt file is produced by the
conscrypt.module.public.api module. The dist_stem property was added to
java_sdk_library to allow the stem name of the file to differ from the
module name.
The API finalization process for extension APIs does something similar
as it extracts various files from the snapshots and copies them into
the appropriate extension API specific directory in
prebuilts/sdk/extensions/<n>. Instead of copying files from the
apistubs/ dist directory (which is not built for trains) it copies them
from sdk snapshots that are built as part of the train.
Previously, the sdk snapshot used to derive the name of the files
within the snapshot from the name of the module, it ignored the
dist_stem property. This change causes it to use the dist_stem property
to make it consistent with the apistubs/ dist directory naming.
The file name is created in
sdkLibrarySdkMemberProperties.AddPropertyToSet() which does not have
access to the dist_stem property. So, it has to be supplied in the
sdkLibrarySdkMemberProperties instance in the new Stem property.
Bug: 248258460
Test: m nothing
BUILD_NUMBER=fixed packages/modules/common/build/mainline_modules_sdks.sh --build-release=latest
# Ran the previous command before and after and make sure that the
# conscrypt, art and icu sdk snapshots use the dist_stem value but
# none of the other snapshots are affected.
Change-Id: Ied52003de63dcdb86a252a39bb8781f85d51a6ff
Fields tagged with `sdk:"keep"` will keep the value even if it would
normally be cleared because it was common across a number of structs.
This will allow a module type to specify the same value across all
structs populated directly from variants and have it be copied into all
common property structs without clearing it from the variant specific
structs.
Bug: 248258460
Test: m nothing
Change-Id: I147d946b11fd8766a7d875d9206e8f5034f585d6
"ignore" is a better description of the effect that this tag has on the
optimization that is done as part of the sdk snapshot generation.
"keep" implies that something special is done with the field during the
optimization process but actually the optimization process completely
ignores the property, just as it does with unexported fields.
Bug: 248258460
Test: m nothing
Change-Id: Idfdfd87f927d641d4d301cbbde72b463c89155bf
When an app_set module is included in an APEX, we don't use its
`builtFile` which is the main APK, , but `PackedAdditionalOutputs()`
which is the zip file where all APK files (main + additional) files are
packed. We then unzip them into the corresponding directory inside the
APEX.
However, `buildFile` still has been used when constructing
canned_fs_config whose content should match with what are actually
included in the APEX. Fixing the bug by using
`PackedAdditionalOutputs()` instead.
Bug: 236299724
Test: m nothing (apex_test.go amended)
Test: follow https://buganizer.corp.google.com/issues/236299724#comment34
* mkdir for-236299724
* cd for-236299724
* repo init -u https://android.googlesource.com/platform/manifest -b
* master --partial-clone --clone-filter=blob:none
* repo sync -c -j8
* unzip -l ~/Downloads/mainline_t_2022_jun_t1004275.zip | cut -c31- | grep | xargs rm -f
* unzip ~/Downloads/mainline_t_2022_jun_t1004275.zip
* source build/envsetup.sh
* choosecombo 1 aosp_arm64 userdebug
* m com.android.extservices.gms
Change-Id: If9752c07748300dbb963568de4e879b041b0a206
When gensrcs's srcs prop is set with filegroup from external package and especially when gensrcs generates cpp headers, it's not trivial to know what the output paths and their correponding include paths look like.
In mixed build, there are a few bugs in gensrcs's include paths for cpp headers (b/248555356 and b/248554581) that we'll eventually need to fix. These tests serve as an documentation of the existing behavior when gensrcs generate cpp headers.
Test: go test
Bug: 248555356
Change-Id: I10168dd4229be8f110a31955d214ef792c8050de
stub_template_host.txt added all the top-level modules to the
PYTHONPATH, which isn't correct, and caused absl.logging to
override the built-in logging module.
Removing this also makes it more consistent with python binaries
built with embedded_launcher: true. embedded_launcher: true
binaries don't add the top-level modules.
Fixes: 245583294
Test: m py_dont_add_top_level_dirs_test && out/host/linux-x86/testcases/py_dont_add_top_level_dirs_test/x86_64/py_dont_add_top_level_dirs_test
Change-Id: Id3069565d2b2c4b2bda0ff5301e757a7b4201751
This adds a new flag to the optimize section to allow invoking the resource shrinker
Bug: 246217952
Test: m
Change-Id: I2e7851af1189db2a6adf6f9f9f444a1d7f3a8d60
Add a filter for cyclic references in required module names in bp2build.
Test: CI
Bug: 247985196
Bug: 225291837
Fixes: 234748998
Change-Id: Id1cfd0b2e5f217cdb01218ec3bbe11d5070cc9df
This is an improvement over directly referencing the prebuilt because
bp2build can now emit BUILD files that contain a dep to
//build/soong/zip/cmd:soong_zip even though we're not converting the
soong_zip module.
It helps to make progress despite not having go rules support to
convert/build soong_zip. This unblocks many modules, like genrules and
gensrcs that use soong_zip in tools. soong_zip doesn't change often, so
it's not risky to depend on the prebuilt instead of source version.
Bug: 194644518
Bug: 248222296
Test: bp2build CI
Change-Id: Idef0c27ec0fb43a9097d750ec72fba00644c9f2a
