Prevent problem for some users using PAR file built by android's
build system which cause by the different site package of python
installed by user themselves.
Bug: 120469915
Test: 1. m atest
atest --help
2. apply aosp/861470
full build
Change-Id: I72c83f9397f0e88eecfcd32806f1746520299447
Soong doesn't wrap everything in bash, but relies on the ninja default
of /bin/sh. We should probably improve that in the future, but for now,
just fix the build.
On Mac, /bin/sh and /bin/bash are both bash, but /bin/sh implies `set -o
posix`, and ignores the `-n` flag.
Test: treehugger
Test: build on a mac
Change-Id: Icf41b1c5a1ce9eb0a56e39e68c433fc80b53620f
Instead of doing these string operations in the C++ code at every
startup of an embedded par, trim and convert the entry point file to a
python module path at build time.
Test: m
Change-Id: I04a6459a80c84a704b8a25a144ab1e19622d6cc3
The zip file format does support scripts/programs prefixed to the
archive, though many of the offsets are supposed to start from the
beginning of the file. Some tools (python and zipinfo) are able to
read zip files with arbitrary prefixes, but others (libziparchive and
zipdetails) fail.
So pass the file to prefix onto the zip file into merge_zips so that we
can set the offsets from the real start of the file.
Test: m sepolicy_tests (runs the embedded python interpreter)
Test: zipinfo out/host/linux-x86/bin/sepolicy_tests
Test: zipdetails out/host/linux-x86/bin/sepolicy_tests
Change-Id: If73d4c2465581f7de5aa47959284ecf2059df091
Otherwise, python will complain the variable has not been initialized,
in case ExtractRunFiles throws an exception.
Test: m
Change-Id: I3d9e3cda5e558436ee4182c44e4345151c031e41
The root cause is we didn't check if the optionalpath is valid or not,
the registerbuildation function directly invoke the path var and
cause the invalid memory error. We just return if the launcher doesn't
exist.
The ctx.VisitDirectDepsWithTag() also handles allowmissingdependency so
we are ok if launcher doesn't exist.
Test: N/A
Bug: b/116698229, b/67510844
Change-Id: I40941079a64d7797ab879fc5edaa29e835b493a0
VisitDepsDepthFirst is almost never correct, as it can't query
dependency tags of multiple dependencies between the same two modules.
Replace VisitDepsDepthFirst with WalkDeps in sanitize.go and
python.go. Also verify the dependency tag before continuing to
recurse to ensure that they don't recurse through genrules and into
unrelated modules.
Test: m checkbuild
Change-Id: I2f7560126f56b51a40ec39dfbdcc18b5891489f7
The problem came from Python libraries doesn't know the information that
we enabled embedded_launcher (only Python binary knows about that). And
we can't simply remove runfiles dir for Python libraries since host
Python mode need this.
Bug: b/80441699
Test: m perf_profo_flames
Change-Id: I73ffc4d7504f95a708ae7cca47bc6c15a673aa31
This reverts commit 09f4540d66.
Fixes the raw call to net.Listen in the tests to go through the listen()
helper and use the long socket path fallbacks.
Removes the use of timeouts from the tests -- the behaviors being tested
did not rely on timeouts, so removing them will reduce the flakiness if
the build is heavily loading the machine at the same time the test is
running.
Also fixes some potential nil pointer dereferences.
Test: OUT_DIR=<really long> m blueprint_tools
Test: `while .../soong-ui-build-paths/test/test; do sleep 0.01; done` with a build running
Change-Id: I16d44be7517bc415f1c808284088f4ba40df3bfa
This reverts commit c59a92cb1a.
Reason for revert: tests are broken with long OUT_DIRs
They're directly calling net.Listen, and not using the fallback
for long socket names.
Change-Id: Id14cbd499fd9b36c6926b7552d3554340cb0916c
This reverts commit 96c957ae20.
Fixes issues on some machines where the socket in TMPDIR ended up with a
unix domain socket pathname over 107 characters long, which Go will
reject due to underlying limitations in the system calls. If this
happens, we'll fall back to opening the directory, then using
/proc/self/fd/#/<file>, or manually creating a similar symlink in /tmp.
Also fixes some issues on Mac where os.Executable returns the symlink
instead of the underlying file, sending a message over a unix domain
socket will block if the reader isn't reading, and sandboxing was
preventing us from running `ps`.
Test: m blueprint_tools
Test: m blueprint_tools on mac
Change-Id: Ib19ccfe10cb0a79f1476fb1d5cd20ed0495be367
This reverts commit bfb0aed2dd.
Reason for revert: <LOCAL_REQUIRED_MODULES overrides the "required:" property in Android.bp>
Test: m -j sepolicy_tests
Bug: b/80140289
Change-Id: I4e51007a7d4dda69cbf4f31253e9a56d3f5e35c7
As I mentioned in b/79751992 #4, when launcher loads __main__ as entry point,
the module dict got set using absolute pkg path which
was runfiles.google.protobuf which caused the type checking failed
(expected: google.protobuf for type()). After then, starting with non entrypoint
python files, the importing module dict would be set with from...import...
So remove dir:runfiles/ from zip section so that during entrypoint Python file
importing the proto files using
from...import..., the type checking can be passed.
Bug: b/79751992
Test: m clean && m out/host/linux-x86/bin/perf_proto_stack;
And execute it:
usage: perf_proto_stack [-h] [--syms SYMS] [--json-out JSON_OUT]
[--print-samples] [--skip-kernel-syms]
[--print-pid-histogram] [--print-sym-histogram]
[--print-dso-histogram] [--parallel PARALLEL]
file [file ...]
perf_proto_stack: error: too few arguments
Change-Id: I95a4a9ee95ce7f90d893ea0529c384d7b944abe4
protoc dependency files, at least for C++ outputs, uses the form of:
a/b.c \
a/b.h: <dep1> <dep2>...
Ninja will fail the command when it parses a dep file and there's more
than one output file (even though it doesn't care what the output file
name is). So this tool will parse the original file, and output a
version with only a single output file.
Bug: 67329638
Test: NINJA_ARGS="-t deps ...pb.c" m
Test: NINJA_ARGS="-t deps ...srcjar" m
Test: NINJA_ARGS="-t deps ...srcszip" m
Test: Run dep_fixer across all of taimen's dep files, no failures.
Test: Run dep_fixer against the processed files, no changes.
Test: Run androidmk across all of our Android.mk files, inspect the diffs
Change-Id: I4263b7d5faea37285afa6b24dedf5964aa7d19dc
Python protobuf std libs will be wrapped in final binary/test par file.
Bug: b/70568913
Test: manually create real examples.
Change-Id: I7376ec9175f3e03d1adbd20858a7f74e826387ad
Each Python module will generate a zip file containing source & data
files. The Python binary will collect all its dependencies and use
merge_zips to merge each zip file to create a final .par file.
Test: m -j checkbuild && real examples:
Bug: b/70568913
Change-Id: I9ff232d461d33e1c06026e7dcb5b124bf02c3ce5
This reverts commit 606e9de344.
Reason for revert: <try to fix the broken build yesterday>
Change-Id: I2963b9af63c7c7398159e5e9a1e448266e1c81d5
Test: unittest
Add a method on ModuleContext and TopDownMutatorContext to visit
direct dependencies that have a given dependency tag.
Test: m checkbuild
Change-Id: Ib875563091dcae6b7282b3e3427d0eb07d8c8af5
This reverts commit 0a32e5936b.
Reason for revert: <libsqlite was missing for darwin_x86_64 before>
Change-Id: I2e13e849a503a705ffad425df292380f2f73954e
Add python_test to generate executables for both host and device side.
And change python_library_host to python_library.
installation for python device test will be under
target/product/generic_arm[64]/data/nativetest[64].
Bug: b/69114465
Test: m clean && m -j
Device Test:
adb root; adb push nan_devicetest /data;
adb shell; ./nan_devicetest
print(sys.path)
['/data/nan_devicetest/runfiles', '/data/nan_devicetest', '/data/nan_devicetest/internal', '/data/nan_devicetest/internal/stdlib']
Change-Id: If8317070a3aa1b6dab2e84b8df2d037f495d7247
Also adds checks that the dependencies are android.Modules and
are not disabled.
Test: m checkbuild
Change-Id: I05e945f38915d49cd3c0ab72a86576949bc7eff2
Now that android.ModuleContext does not include blueprint.ModuleContext
we can rename android.ModuleContext.ModuleBuild to
android.ModuleContext.Build without colliding with
blueprint.ModuleContext.Build. Leave ModuleBuild as a wrapper around
Build for now to avoid having to update all the users outside
build/soong simultaneously.
Test: m checkbuild
Change-Id: I18eb8cc04faf002049a11d9aac97e9732ff5d638
Ubuntu has both aliases, but Mac only declares python2.7, not python2.
Test: `m sepolicy_tests` on internal master
Change-Id: Ia2b5d2d95195f307befbc1893a465752f14b8a7b
launcher.
For Python2, we bundle embedded launcher as bootstrapper within every
.par file. This feature is only enabled for linux_x86_64 for now. We
provide a user flag: hermetic_enabled within bp file. By default, Pyhon2
still use classic bootstrapping way to construct .par file and relys on
host interpreter. Once embedded_launcher is enabled, launcher will be
used to bootstrap .par file and execute user program.
For Python3, the launcher will be ready soon, and for now it still relys
on classic bootstrapping.
Test: Real example is used to test.
Bug: b/63018041
Change-Id: I28deba413d8ad3af407595e46f77d663e79a3705
Split InstallFileName into InstallExecutable that does chmod +x
after copying the file. Also remove InstallFile and rename
InstallFileName to InstallFile.
Test: m -j checkbuild
Change-Id: Id41ad4eafe521f6cd5d8cc250b7747ecb3da8dfc
It's never anything except nil, and it unnecessarily complicates
the implementations.
Test: m -j checkbuild
Change-Id: I3e3b7251f32ffa84dbdfd0448faf248c306ca808
It's never anything except nil, and it unnecessarily complicates
the implementations.
Test: m -j checkbuild
Change-Id: I8a117a86aa39aeb07d9d8d0686ef869c52784f19
Fix mutator registration for tests to allow different tests
in the same package to register different mutators.
Allow tests to track the resulting ModuleBuildParams objects
to use in assertions, and provide helpers for getting them.
For example:
config := android.TestConfig(buildDir)
ctx := android.NewTestContext()
ctx.RegisterModuleType(...)
ctx.MockFileSystem(...)
ctx.ParseBlueprintsFile("Android.bp")
ctx.PrepareBuildActions(config)
ctx.Register()
// Get the Inputs value passed to the javac rule for the foo module
inputs := ctx.ModuleForTests("foo".Rule("javac").Inputs
Test: java_test.go
Change-Id: I10c82967f5f3586d2c176f169906b571ed82fc73
Change module factories from returning a blueprint.Module and a list
of property structs to returning an android.Module, which holds the
list of property structs.
Test: build.ninja identical except for Factory: comment lines
Change-Id: Ica1d823f009db812c518f271a386fbff39c9766f
Added params to WriteFile rule that allow users to turn on the
specific 'echo' options, such as "enable the interpretation of backslash
escapes".
Bug: b/38261000
Test: tested with real data files.
Change-Id: I5145056b5c2522a3a72b9436055ec95d2fbd94c0
Descriptions currently look like:
[ 0% 4/29328] cc out-soong/.intermediates/external/clang/lib/Sema/libclangSema/android_arm_armv7-a-neon_denver_static_core/obj/external/clang/lib/Sema/SemaCodeComplete.o
This is not very helpful - most of the characters are used to show the
output path, which contains useful information like target architecture,
but also contains most of the path to the source files twice, and less
useful information like the exact variant name used by soong.
Make the descriptions look like:
[ 0% 3/29329] //external/clang/lib/Sema:libclangSema clang++ SemaTemplateInstantiate.cpp
This is //path/to/module:modulename tool relative/path/to/source/file
Test: builds, looks pretty
Change-Id: I3087aa7d4eb1860ef6239d77407b8b35445616d7
The base module handles all the common functionalites, such as version
compatibilty check, version variations split, source file format check,
source/data file duplicate check.
The library/binary module focuses on how to generate binary build actions,
such as setting up stub script, zipping, filling in __init__.py in
runfiles dir tree.
Bug: b/31676493
Test: go test under python package
Change-Id: I06608369f350f7195873d459e1c8d1bdb811e77e
Logan Chien