Since these bazel labels are removed from depsets anyways,
there is little reason to convert they to ninja build statements
Test: m nothing (with bazel af426041)
Bug: 301638491
Change-Id: Ie920477231d147d0b5b7dbcd1c59ed9985a80abb
Revert submission 2651299-bazel_sandwich
Reason for revert: b/293883239, checking if this breaks the build.
Reverted changes: /q/submissionid:2651299-bazel_sandwich
Change-Id: I5f1d4bb13d21e6599f5c353dcaba2375f5ec234d
The "bazel sandwich" is a mechanism for bazel to depend on make/soong
outputs. The name comes from the fact that bazel is now at the top
and bottom of the build graph. This is intended to allow us to work
on converting the partition builds to bazel while not all of the
dependencies of the partition have been converted.
It works by adding the bazel_sandwich_import_file rule, which emits a
dangling symlink that starts with bazel_sandwich:, and includes
information that the aquery handler in soong reads. The aquery handler
rewrites the symlink so that it points to a file generated by
make/soong, and adds a ninja dependency from the symlink to the file
it's targeting.
This allows us to depend on make-built files from bazel, but notably
it doesn't allow us to depend on analysis-time information from make.
This shouldn't be a problem for the partitions, but limits the use of
the bazel sandwich to similar, less complicated types of builds.
go/roboleaf-bazel-sandwich
Bug: 265127181
Test: m bazel_sandwich
Change-Id: Ic41bae7be0b55f251d04a6a95f846c50ce897adc
Creation of build statements is largely parallelizable because each
action is independent apart from updates/reads to
depsetHashToArtifactPathsCache. Locally resulted in build statements
taking ~.45 seconds on staging mode to ~.02 seconds
Test: CI
Change-Id: Iab00c8394a9eab17353f71230885ff0870e17f24
This prevents unnecessary copies of data from the query by using a
pointer to the proto or its contents
Test: go test soong tests
Test: m nothing
Change-Id: I5962b27fe31f7103d2af6cef159f11747cff3ed9
Instead of using a sentinel file that was created under bazel_workspace
prune them out. The sentinel file was being created under bazel_workspace,
which would be deleted and thus rendered useless from the perspective of
not retriggering a build.
Test: `touch Android.bp` then `NINJA_ARGS="-d explain" m --bazel-mode-staging adbd_test`
Bug: 265155778
Change-Id: If5d8f0ea7f3b8828fda0646faafd1a621f0cb27c
This change constitutes a number of fixes which cause mixed builds to
have deterministic ninja file output:
1. Depsets are identified based on a hash of their contents instead of
an arbitrary ID integer from Bazel
2. Depset definitions in the ninja file are sorted by the above hashes
3. BuildStatements (action information from Bazel's aquery) are sorted
by their contents
Test: Ran `USE_BAZEL_ANALYSIS=1 m nothing` three times and verified the
md5sum of out/soong/build.ninja was identical all three runs.
Test: mixed_droid
Change-Id: Iffdf6cc62c31d76fbbfa78726827497516171f4f
Each depset now corresponds to a phony rule which depends on other
depsets or on full paths; thus, bazel's depset structure is preserved in
the form of phony rules of name bazel_depset_{id}.
Previously, flattening and recopying large lists of file path strings
was quite inefficient. This was particularly evident as we enumerated
hundreds of clang headers for each cc compile action.
This reduces soong_build analysis time by about 30% for mixed builds.
It also reduces ninja file size by ~750MB.
Fixes: 229405615
Test: Unit tests, manually verified metrics, mixed_droid CI
Change-Id: I78df152ac1488ae0c6807afdde4b4ad5e6d26287
Bazel run shell actions begin `bin/bash -c <command>`, without escaping,
the command can be treated as an argument to `bin/bash`.
Test: build/bazel/mixed_droid.sh
Change-Id: I423cb393da2e6ac97448ec77b2596f12670dfd31
hardlinks are incompatible with sandboxing
relative links are incopatible with Soong's use of `cp -d`
Test: build/bazel/ci/mixed_libc.sh
Change-Id: I8c776cda6a27c680c51466d9a7af1b499f2f566d
This allows support of generated hdrs / srcs in mixed builds.
Test: Manually verified that libc_bionic_ndk passes compilation (failed
previously due to missing generated heaer)
Test: bp2build and mixed_libc CI scripts
Test: New aquery test
Change-Id: I88e359a4bd9eba383c207d5cf812272725ff0a3d
In Bazel aquery responses, Bazel represents action inputs by preserving
the form of the depset data structure which contains them.
(https://docs.bazel.build/versions/master/skylark/lib/depset.html)
Thus, Soong's aquery handler must appropriately "flatten" this
data structure in cases where the depset consists of multiple levels.
Test: m nothing
Test: lunch aosp_flame && USE_BAZEL_ANALYSIS=1 m libc
Change-Id: I2ebacac1deea7538eb34ad1975594caae71091a2
Also introduce test suite for aquery handling
Test: m nothing
Test: lunch aosp_flame && USE_BAZEL_ANALYSIS=1 m libc
Change-Id: I2493d42782099ea0b575968fca38bce6f0d59015
