| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
|
|
|
| |
There's a realpath(3)-like Windows API function
called GetFinalPathNameByHandle{A,W}.
This commit removes JunctionResolver, implements
RealPath using GetFinalPathNameByHandleW, and
replaces JunctionResolver usages with RealPath.
Change-Id: I69efd29c139e957a93496cd75a2db351ed4453e1
PiperOrigin-RevId: 186598195
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Note that this dumps the current state in skyframe (which may contain more nodes than you're interested in):
- bazel build --nobuild //interesting:targets
- bazel dump --action_graph=/path/to/file
- printproto --proto2 --raw_protocol_buffer --message=action_graph.ActionGraphContainer --multiline --proto=third_party/bazel/src/main/protobuf/action_graph.proto /path/to/file
We'll add filtering options in a later CL.
RELNOTES[NEW]: Add option to dump the action graph to a file: 'bazel dump --action_graph=/path/to/file'.
PiperOrigin-RevId: 186597930
|
|
|
|
|
|
|
|
|
|
|
| |
Linkstamps were not re-built when only volatile data changed, i.e. when we
modified cc_binary source, linkstamp was not recompiled so we got old
timestamps. The proper behavior is to recompile linkstamp whenever any input to
cc_binary linking action changes. And the implementation in this cl solves this
by adding all linking inputs as inputs to linkstamp compile action.
RELNOTES: None.
PiperOrigin-RevId: 186595143
|
|
|
|
|
|
|
| |
Fixes #3507.
RELNOTES: None.
PiperOrigin-RevId: 186595099
|
|
|
|
|
|
|
|
|
|
| |
While rules that want to be called from different repositories may not
make any assumptions about locations of their dependencies, they still
may embed path obtained via expansion of $(location ...) into their files.
Add tests verifying that this behavior is preserved.
Change-Id: Id71588b4c128c15926d9190db507c9085fd17653
PiperOrigin-RevId: 186591069
|
|
|
|
|
|
|
|
|
|
|
| |
When --run_under is used, the run-under wrapper will be injected in front of the normal test command line. So, it's incorrect to use the beginning of the command line to infer the test name. However, we can always get the underlying test executable with the TEST_BINARY environmental variable.
Rename TEST_NAME to EXE to avoid future confusion.
Fixes https://github.com/bazelbuild/bazel/issues/4588.
Change-Id: I6fd05cca5e5441c13ee16290c4028ec84adec983
PiperOrigin-RevId: 186590311
|
|
|
|
|
|
|
| |
Methods that return NestedSet shouldn't claim they may return duplicates or their return value will have any sort of meaningful order.
Change-Id: I98f92ec6ddbf4d35fea47b12bcd072bb1b0f1e43
PiperOrigin-RevId: 186585490
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously, when inputs is a depset, we'd:
* Expand it to a collection, then
* Expand it again and add it as a flat list to the input builder
This would be square on both performance and analysis time memory.
Note that we _still_ expand the depset once to try to see if any input is a tool. This will have to be fixed in a future CL.
RELNOTES: None
PiperOrigin-RevId: 186566243
|
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 186565673
|
| |
|
|
|
|
|
|
| |
ErrorInfoManager.
PiperOrigin-RevId: 186524034
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Superceded by https://github.com/bazelbuild/bazel/commit/e734c479956df7a675c61f531d769609d3af3e5d
*** Original change description ***
JavaBuilder accepts a new --testonly argument to mark compilations of
test code. Error Prone will use these to distinguish between test and
production code, e.g. for enforcing @VisibleForTesting.
PiperOrigin-RevId: 186520809
|
|
|
|
|
|
|
|
| |
This bug was invisible to out tests, because our RemoteWorker will work regardless of whether the output was listed as a file or as a directory.
TESTED=same tests
RELNOTES: None
PiperOrigin-RevId: 186518904
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 186513938
|
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 186511953
|
|
|
|
|
|
|
|
|
| |
to check the completeness of the deps of aar_import targets.
Followup cl of unknown commit
RELNOTES: None.
PiperOrigin-RevId: 186503514
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 186501019
|
|
|
|
|
|
| |
cannot finish building until top registers a dep on it, and we can only know that by listening to the addReverseDep call.
PiperOrigin-RevId: 186499232
|
|
|
|
|
| |
RELNOTES: None
PiperOrigin-RevId: 186494122
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 186470386
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 186456188
|
|
|
|
|
|
|
|
| |
The old TransitiveInfoProvider is deprecated. Providers used from Skylark
should use NativeInfo as specified in[]
RELNOTES:none
PiperOrigin-RevId: 186447814
|
|
|
|
|
|
|
| |
This is in preparation for migrating to the new way of specifying providers as described in[]
RELNOTES:none
PiperOrigin-RevId: 186436462
|
|
|
|
| |
PiperOrigin-RevId: 186427907
|
|
|
|
|
|
|
|
| |
invokespecial invocations for core interfaces.
RELNOTES: None.
PiperOrigin-RevId: 186404206
|
|
|
|
| |
PiperOrigin-RevId: 186397865
|
|
|
|
|
|
| |
to select a codec.
PiperOrigin-RevId: 186378153
|
|
|
|
|
|
| |
the information we need. It is a waste of CPU time otherwise as we will just lose the state. Note that we don't pass the env directly into the processDeps function so this wasn't a skyframe restart optimization. Also, none of the state recorded in the processDeps for the direct label deps is used in computing the aspect deps, as we immediately call targetAndErrorIfAny#getTarget.
PiperOrigin-RevId: 186373573
|
|
|
|
|
|
|
|
| |
corresponding Proguard-style -keep rules
RELNOTES: None.
PiperOrigin-RevId: 186372769
|
|
|
|
|
|
|
|
|
| |
regex@attempts, similarly to --runs_per_test.
Also re-arrange some option converters to be closer to their options.
RELNOTES: flaky_test_attempts supports the regex@attempts syntax, like runs_per_test.
PiperOrigin-RevId: 186358437
|
|
|
|
|
|
|
|
|
|
|
|
| |
* AutoCodec now delegates to the registry.
* Adds getSuperclass logic for resolving a codec.
* Small cleanups for classes that break the registry.
TODO after this change:
* Explicit CODEC definitions are no longer needed and existing ones should be cleaned up.
* POLYMORPHIC is no longer be needed and should be cleaned up.
PiperOrigin-RevId: 186351845
|
|
|
|
|
|
| |
CharSequence fields.
PiperOrigin-RevId: 186323424
|
|
|
|
|
|
| |
...instead of passing it everywhere so we can rewrite the local env.
PiperOrigin-RevId: 186316859
|
|
|
|
|
|
| |
AbstractAction.
PiperOrigin-RevId: 186316435
|
|
|
|
| |
PiperOrigin-RevId: 186314781
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since the skylark implementation of http_archive provides all
the features available in the native http_archive and a few
more (patching), there is no need to have to rule implementations
doing the same thing. Hence deprecate the native one as part
of our plan of moving more functionality to skylark.
RELNOTES: The native http_archive rule is deprecated. Use the
Skylark version available via load("@bazel_tools//tools/build_defs/repo:http.bzl", "http_archive")
instead.
Change-Id: I107c2f25f5a37c67f56b4362e7c9d9ade8428c16
PiperOrigin-RevId: 186314624
|
|
|
|
|
|
|
|
|
|
| |
Add a test demonstrating how a repository can provide a rule with
an implicit dependency on a target in its own repository in a way
that works both, locally and remotely without the need to know
the name of the repository the rule resides in.
Change-Id: I2446be1ba16a382576800f8141da17331e64f8b1
PiperOrigin-RevId: 186314462
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
https://github.com/bazelbuild/bazel/commit/3904ac33a983fd8faebba1b52bcac5a3ff942029 (https://github.com/bazelbuild/bazel/commit/3904ac33a983fd8faebba1b52bcac5a3ff942029). Fix #4625 by running the test process in a sub-shell.
Apparently, nested background processes interfere with SIGINT handling in bash. I don't 100% understand why and how, but I do have a small bash script that demonstrates the problem: script A that spawns a background process, sends it a SIGINT, and verifies it was received. The script works, *unless* run in the background by a process B; this extra layer of backgrounding cause process A's logic to stop working. See experimental/users/olaola/shell/ for examples. See also https://stackoverflow.com/questions/48847722/nested-background-processes-and-sigint-handling
*** Original change description ***
Fixing test-setup.sh occasionally missing stdout/stderr, on systems where "tail --pid" is supported.
The solutions aren't mine, the new test was taken from Ola's unknown commit and the way to avoid race condition courtesy of sethkoehler@
Mitigates #4608 for compatible Linux systems.
TESTED=presubmits, manual shell tests on new bazel
RELNOTES: None
PiperOrigin-RevId: 186312008
|
|
|
|
|
|
|
|
| |
This option was removed in https://github.com/bazelbuild/bazel/commit/2ea4fa26281175c316651ec50784b820a9f409cf.
Closes #4509.
PiperOrigin-RevId: 186311923
|
|
|
|
| |
PiperOrigin-RevId: 186309181
|
|
|
|
|
|
| |
Closes #4646.
PiperOrigin-RevId: 186301842
|
|
|
|
| |
PiperOrigin-RevId: 186298405
|
|
|
|
|
| |
RELNOTES: None.
PiperOrigin-RevId: 186297559
|
|
|
|
|
|
|
|
|
| |
Remove the static kAsciiPropertyBits data from
strings.cc and replace its only reference with
simpler code.
Change-Id: I95fb12b7d4bca4589ae64603b63019ee6e12bfc2
PiperOrigin-RevId: 186295165
|
|
|
|
|
|
|
|
|
|
|
| |
This will make protoc see as direct dependencies the .proto files that were included using the proto_source_root flag.
Until now, Bazel passed to protoc the direct dependencies of a target as the path relative to the WORKSPACE, which made it fail when a shorter path, relative to the package was used.
Progress on #4544.
RELNOTES: None.
PiperOrigin-RevId: 186294997
|
|
|
|
|
|
|
|
|
| |
Since the http_archive inherits the native download_and_extract function,
the support of plain tar archives introduced in 4c8fa1bab507fa7f0a1cbeac0724751d9b574f89
applies here as well. Fix the doc strings accordingly.
Change-Id: Ica4037215abf7498c3a8ec9f21ac32b2e59e8966
PiperOrigin-RevId: 186293816
|
|
|
|
|
|
|
|
|
|
|
|
| |
We have two places in internal code that require an ActionInput for an embedded
tool. The embedded tool is currently not declared as an input to the
corresponding actions, which in turn causes the remote execution to stat the
file directly through SingleBuildFileCache. I'm working on removing
SingleBuildFileCache, so this needs to be solved. This is the quick-and-dirty
fix. A more principled fix might be to check in the relevant tools instead of
embedding them into Bazel, but it's also significantly more work.
PiperOrigin-RevId: 186292203
|
|
|
|
| |
PiperOrigin-RevId: 186290994
|
|
|
|
|
|
|
|
| |
So far, nobody uses it, but I want to start using this field soon.
TESTED=unit test
RELNOTES: None
PiperOrigin-RevId: 186290375
|
|
|
|
|
|
|
|
|
| |
- Make sure that we correctly resolve symlinks starting with ./ and ../. This
is currently failing, although it's working at head.
From my reading, https://github.com/bazelbuild/bazel/commit/a729b9b4c3d7844a7d44934bf3365f92633c0a60 changes PathFragment.getRelative to always
normalize the return value, which it wasn't doing before.
PiperOrigin-RevId: 186289431
|