| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
| |
are created, as opposed to when they are requested from the ParallelEvaluator. That delay can lead to large memory spikes and churn.
--
MOS_MIGRATED_REVID=116224565
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Breaks on Java 1.7
See http://ci.bazel.io/job/Bazel/385/.
*** Original change description ***
Canonicalize the name of the target itself for java indirect dep errors.
Otherwise it prints the wrong message when the target which is missing
a dependency is in a repository other than the default one.
--
MOS_MIGRATED_REVID=116164887
|
|
|
|
|
|
|
| |
Fixed bazel github issue #978
--
MOS_MIGRATED_REVID=116164610
|
|
|
|
|
|
|
| |
--
Change-Id: I7f9edbef4dd94b5d37aabb1a17333224af3615c6
Reviewed-on: https://bazel-review.googlesource.com/#/c/3042
MOS_MIGRATED_REVID=116163936
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=116146417
|
|
|
|
|
|
|
|
| |
This is a temporary workaround to enable the use
of a cc_configure() rules to auto-configure C++ toolchain.
--
MOS_MIGRATED_REVID=116140726
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Manually break lines in the progress bar (and only in the progress bar!)
that are longer than the terminal width. In this way, we avoid any
confusion about when the terminal advances a line; in particular, when
the terminal is wider than we believe, we do not erase back into the
scrollback buffer.
--
Change-Id: I19f21bdbd6e6af4665d1e712b921f1a15c7d901c
Reviewed-on: https://bazel-review.googlesource.com/#/c/3041
MOS_MIGRATED_REVID=116138487
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=116138214
|
|
|
|
|
|
|
|
|
|
| |
Otherwise it prints the wrong message when the target which is missing
a dependency is in a repository other than the default one.
--
Change-Id: I3762e6be911b79a133a2959fce3a8bf34d183232
Reviewed-on: https://bazel-review.googlesource.com/#/c/2782/
MOS_MIGRATED_REVID=116128663
|
|
|
|
|
|
|
|
|
|
|
|
| |
The production JvmConfigurationLoader never invents Jvm labels, and this
is only used for the sanity check that all implicit labels can be derived
from explicit labels. Unfortunately, our tests often use the
MockJvmFactory, which happily invents Jvm labels.
This is a step towards making the sanity check run unconditionally.
--
MOS_MIGRATED_REVID=116125095
|
|
|
|
|
|
|
| |
cpp, java, jvm) to the documentation.
--
MOS_MIGRATED_REVID=116092594
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=116076100
|
|
|
|
|
|
|
|
|
| |
that labels and label_lists are actually of type Target, and adding more cross linking to various documentation to make it easier to peruse.
Tested all of the links to make sure they work.
--
MOS_MIGRATED_REVID=116076080
|
|
|
|
|
|
|
|
|
| |
transformation 'rdeps(<sky_query_environment_universe_scope>, T, depth)' -> 'allrdeps(T, depth)'.
SkyQueryEnvironment can evaluate such allrdeps queries much more efficiently since it doesn't need to bother filtering out targets outside of universe, meaning it doesn't need to have all targets in the universe in memory at the same time.
--
MOS_MIGRATED_REVID=116075008
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=116058328
|
|
|
|
|
|
|
|
|
| |
second time. If the first call is notifying listeners that the command is complete, the second call can overtake it (since it skips notifying listeners) and kill the JVM prematurely.
This change does open up the possibility that the CommandCompleteEvent triggers a hang or blocking process that prevents us from exiting, but I'm not aware of any known cases of that.
--
MOS_MIGRATED_REVID=116053085
|
|
|
|
|
|
|
|
|
|
|
| |
For print(fragmentCollection)
you get:
target: [ 'apple', 'cpp', 'java', 'jvm' ]
instead of
com.google.devtools.build.lib.analysis.config.FragmentCollection@402dc7f2
--
MOS_MIGRATED_REVID=116045778
|
|
|
|
|
|
|
|
| |
The only input to FdoSupport.prepareToBuild is now the exec root, and
changes to that invalidate the entire server anyway.
--
MOS_MIGRATED_REVID=116027192
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=116020389
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=116017697
|
|
|
|
|
|
|
|
|
| |
...so that it can be used by other event handlers as well.
--
Change-Id: Ic7b37f9041aab63611d05bdb5f18e72bb0c8f1d9
Reviewed-on: https://bazel-review.googlesource.com/#/c/3040
MOS_MIGRATED_REVID=116012068
|
|
|
|
|
|
|
|
|
|
|
|
| |
In order for Aspects to support dynamic configuration, they need to have two
configurations: one to instantiate the Aspect with, containing all the fragment
dependencies of the Aspect itself, and one to create the ConfiguredTargetValue.key
with, containing only the dependencies of the Rule. This expands AspectKey to
have a second configuration, although it currently does not populate that key with
anything different.
--
MOS_MIGRATED_REVID=115997454
|
|
|
|
|
|
|
|
| |
If we don't do that, the host configuration still uses the default
value, despite the invocation policy we apply internally.
--
MOS_MIGRATED_REVID=115986998
|
|
|
|
|
|
|
|
| |
Migrates --java_langtools, --javabuilder_top, --singlejar_top, --genclass_top,
and --ijar_top to java_toolchain.{javac,javabuilder,singlejar,genclass,ijar}
--
MOS_MIGRATED_REVID=115949619
|
|
|
|
|
|
|
|
|
|
| |
* Add support for rules with capital letters, such as Fileset
* Add support for static page headings with periods.
* Run the expander on the HTML doc for the name attribute.
* Add functions to initial rule index.
--
MOS_MIGRATED_REVID=115925884
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=115921810
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=115920640
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=115912069
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=115902893
|
|
|
|
|
|
|
|
|
|
|
| |
*** Reason for rollback ***
Breaking Blaze: []
*** Original change description ***
--
MOS_MIGRATED_REVID=115886167
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=115882941
|
|
|
|
|
|
|
| |
subdirectory package lookup values. This should improve change pruning speed when we have to check a glob. It also keeps GlobFunction closer to the contract of Skyframe, because in order to avoid quadratic restarts, it wasn't checking for missing deps between getValue calls.
--
MOS_MIGRATED_REVID=115882309
|
|
|
|
|
|
|
|
|
|
| |
This change adds a to_json method on Skylark structs, paralleling the
existing to_proto built-in.
RELNOTES[NEW]: Adds a to_json method to Skylark structs, providing conversion to JSON format.
--
MOS_MIGRATED_REVID=115873229
|
|
|
|
|
|
|
| |
Another part of #976.
--
MOS_MIGRATED_REVID=115859513
|
|
|
|
|
|
|
|
| |
Fixes #587 by properly setting up the Proxy object for HttpDownloader and by setting JVM arguments for GitCloner.
--
Reviewed-on: https://github.com/bazelbuild/bazel/pull/795
MOS_MIGRATED_REVID=115857070
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=115852489
|
|
|
|
|
|
|
|
|
|
|
| |
the same artifact.
This fix is that we pick the random executable. Filed #987 for a better
fix (the current behavior just crashes the build randomly and the real
fix is involved).
--
MOS_MIGRATED_REVID=115835198
|
|
|
|
|
|
|
|
| |
This target include all non tests targets of Bazel to do integration tests of
bootstrapping.
--
MOS_MIGRATED_REVID=115830741
|
|
|
|
|
|
|
|
|
|
|
|
| |
In order to correctly erase the current progress bar, we
need to know how many lines it is long. Instead of trying
to compute it, enforce explicit breaks shorter than the terminal
width.
--
Change-Id: Ifac16b351e1390f553d0ceac2b647b1178b58d0b
Reviewed-on: https://bazel-review.googlesource.com/#/c/3024
MOS_MIGRATED_REVID=115829390
|
|
|
|
|
|
|
|
| |
This unify the way which() and os.environ() get their environment
and won't lead to weird issue.
--
MOS_MIGRATED_REVID=115829154
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In order to update Bazel's progress bar, the old one has to be removed
first; this requires knowledge about the number of lines it currently
uses. For small terminals, this requires us to take line breaks into
account. While this normally works well, there are situations where our
believe about the terminal width is smaller that the actual width of the
terminal, causing the deletion of too many lines. This wrapper class provides
a solution for this case by explicitly breaking lines at the given width.
--
Change-Id: I9d33e389730568ab8c15ee082594de9b35abb71c
Reviewed-on: https://bazel-review.googlesource.com/#/c/3023
MOS_MIGRATED_REVID=115827354
|
|
|
|
|
|
|
| |
RELNOTES:
--
MOS_MIGRATED_REVID=115804885
|
|
|
|
|
| |
--
MOS_MIGRATED_REVID=115800229
|
|
|
|
|
|
|
| |
While it works on Darwin, "ps -p <PID>" on Linux prints more information than just the PID, so the string magic there couldn't have worked. (I'm not exactly sure why we didn't notice this earlier...).
--
MOS_MIGRATED_REVID=115799799
|
|
|
|
|
|
|
|
|
|
| |
- Remove unneeded abstract override AbstractAction#getResourceConsumption()
- Remove comments justifying returning ResourceSet.ZERO from getResourceConsumption() - instead document it once on the base method.
- Replace ResourceSet.create(0, 0, 0) with ResourceSet.ZERO.
- Make sure no one ever returns null from getResourceConsumption() and then simplify the code in SkyframeActionExecutor to not check for it.
--
MOS_MIGRATED_REVID=115739250
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This makes very little sense, because we only do this when compiling
them within a rule. The preprocessed files are not shipped to other rules,
meaning we treat things differently depending on the dependency
structure. And, worst of all, this can lead to a bug in the aidl compiler,
which causes the preprocessed definition to take precedence over the input
file, which causes certain modifiers to be stripped away.
Also, Gradle doesn't do it, and that's proof enough that this is no longer
the way to go, if ever it was.
Unfortunately, this causes a problem: the preprocessing had an effect, in
that all preprocessed types are available without the use of imports. This
means that .aidl files which were previously legal before this change may
now be broken, if they relied on this behavior. But, those .aidl files are
actually not legal according to the .aidl specification. Unfortunate, but
the right thing to do.
This CL also updates the idl documentation. Which, let's face it, could
probably have used some improvement.
RELNOTES[INC]: .aidl files correctly require import statements for types
defined in the same package and the same android_library.
--
MOS_MIGRATED_REVID=115718918
|
|
|
|
|
|
|
|
|
|
|
| |
Currently, the Python launcher does not ensure that the `python` file it finds
in a search directory is an actual executable file. As a result, if it finds a
directory called `python`, it thinks that it has found the `python` binary.
Fixes #981
--
MOS_MIGRATED_REVID=115710677
|
|
|
|
|
|
|
|
|
|
|
| |
Map.
This makes it possible to request multiple implementations of the same ActionContext to be available via Executor#getContext().
Currently, specialized SpawnActionContexts like the sandbox or the worker strategy that might have to do a fallback each instantiate their own private copy of e.g. the StandaloneSpawnStrategy. With this change, they can instead get a global instance from the Executor.
--
MOS_MIGRATED_REVID=115705811
|
|
|
|
|
|
|
| |
SkyFunction
--
MOS_MIGRATED_REVID=115671161
|
|
|
|
|
|
|
| |
willExecuteRemotely(). Remove unused parameter "mnemonic".
--
MOS_MIGRATED_REVID=115666410
|