aboutsummaryrefslogtreecommitdiffhomepage
path: root/src/main/java/com/google/devtools/build/lib/skyframe/SkyframeExecutor.java
Commit message (Collapse)AuthorAge
* Support tree artifacts in target complete.Gravatar tomlu2018-08-15
| | | | | RELNOTES: None PiperOrigin-RevId: 208861642
* Filter out events from analysis when constructing execution-phase values in ↵Gravatar janakr2018-08-14
| | | | | | | | | | Skyframe. This avoids some unnecessary iteration over already-emitted events that can show up in profiles, and allows us to store execution-phase values a bit more compactly, since we don't need to carry around wrapper objects and nested sets everywhere. This crucially depends on the fact that we can't build a target in the execution phase without first having analyzed it in a separate Skyframe call. Skyframe normally propagates all events/posts up the graph because it must be able to emit them if a user requests a node that only transitively depends on the node that emitted an event. However, because we do analysis in a separate Skyframe call, any warnings/posts associated with the analysis nodes will be emitted then, and we don't need to propagate them into execution. PiperOrigin-RevId: 208767078
* ActionAnalysisMetadata exposes whether loose headers are present. ↵Gravatar shahan2018-08-12
| | | | | | TransitiveVersionTable uses futures. PiperOrigin-RevId: 208402465
* When no new configured targets have been analyzed, only check for artifact ↵Gravatar janakr2018-08-03
| | | | | | | | | | conflicts if the current set of configured targets is not a subset of the largest set of configured targets that have been checked for conflicts. Also rework the flow between SkyframeBuildView and SkyframeActionExecutor to remove the SkyframeExecutor middleman. Also reword the error message in case of an ArtifactPrefixConflictException, since a clean should no longer be necessary. PiperOrigin-RevId: 207322139
* Add support to completion function to create a path resolver from its ↵Gravatar tomlu2018-08-03
| | | | | | | looked-up artifact values. RELNOTES: None PiperOrigin-RevId: 207295716
* Add a flag to evaluate the top level transitions in SkyframeGravatar ulfjack2018-08-02
| | | | | | | | | | | | | | | | | | | | | | | | This adds a new PrepareAnalysisPhaseFunction, which started out as a copy of some existing code from SkyframeExecutor, BuildView, AnalysisPhaseRunner, AnalysisUtils, and ConfigurationResolver, which was then modified to work inside Skyframe. Most of our tests already work with the new code, except for some of the tests related to configuration trimming in combination with dependency cycles. The reason for this is that we can only recover from dependency cycles at the end of a Skyframe invocation, but never inside a Skyframe invocation. The new code therefore cannot return partial results like the old code. This seems to make null builds a bit faster. In my testing, I saw null build times for a single test target go from ~50ms to ~40ms. This is probably due to slightly better caching - it seems that computing the configuration transitions and top-level targets is non-negligible, even if there's only a single top-level configuration for a single top-level target. This might be an even bigger win if there are a lot of top-level targets and configurations. PiperOrigin-RevId: 207083192
* Only detect artifact conflicts between targets in current build.Gravatar tomlu2018-08-01
| | | | | | | Previously, two successive builds of targets A and B would fail if the two sets had conflicting actions. Now, only fail such builds if the targets from the current build has action conflicts. RELNOTES: None PiperOrigin-RevId: 206974011
* Refactor setupPackageCache to get all BuildConfiguration.Options references ↵Gravatar juliexxia2018-08-01
| | | | | | | | out of it. Setting up the package cache is necessary for proper loading. In the future, BuildConfiguration.Option options will be targets that need to be loaded in order to be parsed. Thus, we will need to be able to do loading before BuildConfiguration.Options are parsed. PiperOrigin-RevId: 206964473
* Allows subclasses of SkyframeExecutor to enable analysis during execution.Gravatar shahan2018-07-30
| | | | PiperOrigin-RevId: 206568465
* Target pattern eval: use the same number of threads as loadingGravatar ulfjack2018-07-26
| | | | | | | | Instead of using the default thread pool size of 200, use the number set for the loading phase. This is in preparation for interleaving the loading and target pattern eval phases. PiperOrigin-RevId: 206172915
* Support optional repository verificationGravatar Klaus Aehlig2018-07-25
| | | | | | | | | | | | | | | | Add an option to provide a file with a resolved value, that will be used to verify that the repositories mentioned in this file produce a correct directory tree. RELNOTES: newly added options --experimental_repository_hash_file and --experimental_verify_repository_rules allow to verify for repositories the directory generated against pre-recorded hashes. See documentation for those options. Work towards #5660. Change-Id: I2d8becb188d0fa51e890fb8f6139f321cca14b7b PiperOrigin-RevId: 206016792
* Remove LoadingResultGravatar ulfjack2018-07-25
| | | | | | | | | | | Instead, refactor the code to use TargetPatternPhaseValue exclusively. This removes the need to convert from TargetPatternPhaseValue to LoadingResult, and prepares for interleaving. It also reduces the number of Skyframe calls which may speed up null builds a bit, as a followup for https://github.com/bazelbuild/bazel/commit/1067310e18cb9ac203110726de0be53bdc403cea. PiperOrigin-RevId: 205989338
* Configured targets register created source artifacts with Skyframe.Gravatar shahan2018-07-24
| | | | PiperOrigin-RevId: 205876673
* Automated rollback of commit f309ad3be36363070e87eef0ee04b12f4956d601.Gravatar janakr2018-07-24
| | | | | | | | | | *** Reason for rollback *** Fixed duplicate derived inputs bug. Test is in diffbase. RELNOTES[INC]: If the same artifact is generated by two distinct but identical actions, and a downstream action has both those actions' outputs in its inputs, the artifact will now appear twice in the downstream action's inputs. If this causes problems in Skylark actions, you can use the uniquify=True argument in Args.add_args. PiperOrigin-RevId: 205863806
* Move LoadingPhaseCompleteEvent posting to TargetPatternPhaseFunctionGravatar ulfjack2018-07-24
| | | | | | | | | | | Also simplify LoadingPhaseCompleteEvent, and SkyframeExecutor, and remove LoadingCallback, which is unnecessary now that we only have a single implementation (previously LoadingPhaseRunner). This also removes some of the excessive Skyframe calls introduced by https://github.com/bazelbuild/bazel/commit/1067310e18cb9ac203110726de0be53bdc403cea, and prepares for interleaving target pattern eval and loading. PiperOrigin-RevId: 205813197
* Introduce option flag experimental_enable_tools_defaults_package.Gravatar dbabkin2018-07-23
| | | | | | | | Default value is true, and behavior related to //tools/defaults package is not changed. If set it to false, then in-memory Dfaultpacked will not be created. RELNOTES:none PiperOrigin-RevId: 205643628
* Automated rollback of commit d491bf10f42e213292382c98a1dc439537f00f43.Gravatar janakr2018-07-17
| | | | | | | | | | | | | | | | | | *** Reason for rollback *** Still bugs lurking. See linked bug. *** Original change description *** Automated rollback of commit eb587075b0d6ffab1cf9e69ede1b7e547905e547. *** Reason for rollback *** Depot has been fixed. RELNOTES[INC]: If the same artifact is generated by two distinct but identical actions, and a downstream action has both those actions' outputs in its inputs, the artifact will now appear twice in the downstream action's inputs. If this causes problems in Skylark actions, you can use the uniquify=True argument in Args.add_args. PiperOrigin-RevId: 204997569
* Automated rollback of commit eb587075b0d6ffab1cf9e69ede1b7e547905e547.Gravatar janakr2018-07-16
| | | | | | | | | | *** Reason for rollback *** Depot has been fixed. RELNOTES[INC]: If the same artifact is generated by two distinct but identical actions, and a downstream action has both those actions' outputs in its inputs, the artifact will now appear twice in the downstream action's inputs. If this causes problems in Skylark actions, you can use the uniquify=True argument in Args.add_args. PiperOrigin-RevId: 204827477
* Remove ToolchainContextException and make exception handling more robust.Gravatar John Cater2018-07-10
| | | | | Change-Id: Iefeab0de85541b963c6cc262339e73b79f630f8b PiperOrigin-RevId: 203952823
* In SkyframeExecutor#invalidateConfigurationCollection, actually delete the ↵Gravatar janakr2018-07-03
| | | | | | nodes, instead of invalidating them. Invalidating isn't sound, because we're depending on externally injected state to be re-read during evaluation. Also rename the method to resetConfigurationCollectionForTesting since it's only used in tests. PiperOrigin-RevId: 203172895
* Add some debugging info to ArtifactFactory, and allow tests to access the ↵Gravatar janakr2018-06-22
| | | | | | ArtifactResolverSupplier in SkyframeExecutor. PiperOrigin-RevId: 201705857
* Automated rollback of commit 45b308a62f42c2c0bcfe79dcd4046c4025a31059.Gravatar janakr2018-06-20
| | | | | | | | | | | | | | | | | | | | | | | | *** Reason for rollback *** Breaks Javascript compilation. There's currently no way to iterate over a depset of Artifacts and deduplicate by identical paths in Skylark. This means that actions that want to do something once per Artifact in a depset (add a flag to the command line with the path of the Artifact for instance) will have duplicate entries if there are multiple Artifacts with the same path. This is not a true automated rollback, since I tried to make this as minimal as possible, to avoid merge conflicts and keep some of the benefits of the rolled back CL. In particular, the tests that were changed in the original CL to give artifacts their correct owners did not need to be changed back, since the owners are still correct. Moreover, this effectively contains rollbacks of unknown commit and https://github.com/bazelbuild/bazel/commit/39d6f89644107a8f7c080c4f4aec8527c1a73954, but keeps test coverage from those CLs as well. Comments added to CL thread where not a clean rollback (there are plenty of files not changed at all: ActionArtifactCycleReporter is the main wart, since it can still handle SkyKeys with Artifact as the argument, but Artifact is no longer the argument to any SkyKey). RELNOTES: None *** Original change description *** Make Artifact#equals take the owner into account for derived artifacts. Derived artifacts' owners are important because they are used to determine the artifact's generating action. Source artifacts' owners are not used in this way, so I left them alone. This allows us to get rid of most uses of ArtifactSkyKey. We may be able to delete it entirely in a follow-up. PiperOrigin-RevId: 201464780
* Add new BuildMetrics event to BEP.Gravatar tomlu2018-06-19
| | | | | | | To start we add just a single metric, the number of actions constructed in the current build. RELNOTES: None PiperOrigin-RevId: 201248490
* Remove support for --discard_actions_after_execution.Gravatar tomlu2018-06-18
| | | | | | | The memory savings from this flag are not worth the complexity, and it interferes with action restarting. RELNOTES: Remove support for --discard_actions_after_execution. PiperOrigin-RevId: 201077905
* Refactoring: uses OutputService for ActionFileSystem injectionGravatar shahan2018-06-18
| | | | | | Moves more Action-oriented from skyframe package, which has become very unwieldy, to action package. This is needed to avoid circular dependencies caused when build-base is needed for skyframe. PiperOrigin-RevId: 200996982
* Remove LIPO supportGravatar hlopko2018-06-15
| | | | | RELNOTES: Support for LIPO has been fully removed. PiperOrigin-RevId: 200724578
* Add a bit more profiler coverageGravatar ulfjack2018-06-15
| | | | | | | | | | | | | | | | | | The intent is that the main thread has ~zero gaps in the profile (though there may still be small gaps due to the time between one try block and an immediately subsequent try block). We need to be careful not to wrap markPhase calls (or methods that call markPhase) in try blocks for the profiler - the Profiler requires that all markPhase calls happen at top level, and throws an exception if not. This should not have any performance impact - all of these are once per build, or at most once per module per build, and we don't expect a very large number of modules (and if we see an increasing number, we need to change the module API to not have to call every single module, but only those that are actually interested in certain events, maybe with an EventBus-based setup). PiperOrigin-RevId: 200712677
* Change WorkspaceStatusAction incrementality logic. We no longer manually ↵Gravatar janakr2018-06-12
| | | | | | | | invalidate the BUILD_INFO_KEY node on --workspace_status_command and related flag changes. Instead, the action has a supplier that allows it to retrieve the correct values at execution time. This does not sacrifice correctness because the action executes unconditionally on every build, so it will never have stale data. PiperOrigin-RevId: 200265375
* Remove LoadingPhaseRunner, inline the only implementationGravatar ulfjack2018-06-12
| | | | PiperOrigin-RevId: 200210007
* Split TargetPatternEvaluator into two interfacesGravatar ulfjack2018-06-12
| | | | | | | | | | | | | | | | | - the TargetPatternPreloader is still used for query in all its forms - the remaining TargetPatternEvaluator part is no longer used except in tests - also make both implementations stateless and pass the offset to the methods instead; note that they both modify the underlying skyframe graph, so there are side effects to the calls even if there's no direct state anymore The intent is to migrate the relevant tests to LoadingPhaseRunnerTest (which could also now be renamed since it's not doing a loading phase), and then delete the TargetPatternEvaluator interface. This depends on the previous commit that removed the last direct use of TPE from an internal command. PiperOrigin-RevId: 200198067
* Allow delegation to the underlying filesystem only for the source tree.Gravatar felly2018-06-11
| | | | | | | The on-disk execRoot/blaze-out is now off limits to ActionFS. RELNOTES: None PiperOrigin-RevId: 200080287
* Make Artifact#equals take the owner into account for derived artifacts.Gravatar janakr2018-06-08
| | | | | | | | Derived artifacts' owners are important because they are used to determine the artifact's generating action. Source artifacts' owners are not used in this way, so I left them alone. This allows us to get rid of most uses of ArtifactSkyKey. We may be able to delete it entirely in a follow-up. PiperOrigin-RevId: 199836436
* Remove the legacy loading phase runnerGravatar ulfjack2018-06-08
| | | | PiperOrigin-RevId: 199775400
* Deletes ActionInputFileCache.Gravatar shahan2018-06-07
| | | | PiperOrigin-RevId: 199732415
* Rewrite TargetPattern failure reportingGravatar ulfjack2018-06-07
| | | | | | | | | | | | | | Report failures in TargetPatternFunction, rather than in its callers. Since we can't distinguish between keep_going and nokeep_going modes, we otherwise end up double-reporting errors. In the particular case that's covered by the build_event_stream_test.sh, we end up reporting the same target pattern as both skipped and failed. Unfortunately, this means we cannot report whether the target pattern was skipped or failed, so the pattern_skipped event is now unused (if we agree that this is acceptable, I'll remove the corresponding infrastructure). PiperOrigin-RevId: 199593700
* Refactoring: makes the code simpler by deleting Metadata and InputArtifactData.Gravatar shahan2018-06-06
| | | | | | (minor) ActionFS now implements MetadataProvider.getInput PiperOrigin-RevId: 199575194
* Remove CachingPackageLocator interface from the PackageProvider hierarchyGravatar ulfjack2018-06-06
| | | | | | This simplifies the code quite a bit. PiperOrigin-RevId: 199427943
* Allow rules and targets to specify extra execution platform constraints.Gravatar John Cater2018-06-05
| | | | | | | | | | | | | | | | | | | | | | RuleClass.Builder now allows authors to specify whether a rule's targets can add additional constraints on the execution platform, and to declare additional constraints for all targets of that rule. Targets which support this now have an attribute, "exec_compatible_with", which supports specifying additional constraints that the execution platform used must match. This attribute is non-configurable. It will only affect execution platforms used during toolchain resolution. Part of #5217. Change-Id: Id2400dbf869a00aa2be3e3d2f085c2850cd6dc00 Closes #5227. Change-Id: If7d55f08f7f44bc7d7f6dfec86a3e6bcd68574b9 PiperOrigin-RevId: 199326255
* Process 'repo_mapping' attribute from WORKSPACE rules. 'repo_mapping' is a ↵Gravatar dannark2018-06-04
| | | | | | | | | | | | | | | | way to remap references to repositories within an external repository by another name. This CL only adds the mappings to the Package object, but it does not actually enable the reassignments. Example usage (in a WORKSPACE file): local_repository( name = ?a?, path = ?../a?, repo_mapping = {?@x? : ?@y?} ) This change also creates a new SkyKey which represents the mappings. This is to prevent all packages from depending on the external package, and instead depending just on the mappings. i.e. a change to the WORKSPACE file that does not touch the mappings shouldn't cause a reload of the package. RELNOTES:None PiperOrigin-RevId: 199187963
* Remove ConfiguredTarget from TargetCompletionValue, since it is no longer ↵Gravatar janakr2018-06-04
| | | | | | needed: we can get the ConfiguredTargetKey directly from the TargetCompletionKey. Since that was the only use of the actual value in EvaluationProgressReceiver#evaluated, remove it, instead just provide a boolean enum that gives whether or not evaluation succeeded. PiperOrigin-RevId: 199178047
* Rename PatchTransition.apply to PatchTransiton.patch.Gravatar gregce2018-05-23
| | | | | | | This unclashes with the incoming ConfigurationTransition.apply method described in https://docs.google.com/document/d/1_UJKmAQ9EE8i3Pl0il3YLTYr-Q9EKYYyLatt2zohfyM/edit#heading=h.96gongkwg852. PiperOrigin-RevId: 197769784
* Optimizes performance of ActionFS staging and eliminates ActionFS updates.Gravatar shahan2018-05-15
| | | | | | | | | Extracts a class, InputArtifactData to hold the input data instead of using a raw map. This provides the flexibility needed to support both ActionFS and existing code so ActionFS does not need to rekey the input data. Uses the smaller, getDeclaredIncludeSrcs instead of getAllowedDerivedInputs when possible for staging optional inputs in ActionFS. PiperOrigin-RevId: 196736703
* Potentially allow children of a dirty node to be missing from the graph. ↵Gravatar janakr2018-05-15
| | | | | | Also pass the GraphInconsistencyReciever into SkyframeExecutor as a parameter. PiperOrigin-RevId: 196716642
* Move --track_incremental_state and --keep_state_after_build to ↵Gravatar felly2018-05-15
| | | | | | CommonCommandOptions to better facilitate certain kinds of benchmarking. PiperOrigin-RevId: 196695342
* Ease the configuration-related restrictions of several more tests.Gravatar mstaib2018-05-14
| | | | | | | | | Use update() and other methods which properly respect the trimming transition over those which don't. Avoid using the target configuration if the configuration actually used by a target is available. RELNOTES: None. PiperOrigin-RevId: 196588405
* Adds an action-scoped filesystem.Gravatar shahan2018-05-09
| | | | PiperOrigin-RevId: 195973862
* Clean up code that directly imports nested classes like Builder, Entry, etc.Gravatar jcater2018-05-01
| | | | PiperOrigin-RevId: 194998981
* Allow SkyFunctions to return a sentinel value indicating that all of a ↵Gravatar janakr2018-04-30
| | | | | | | | | | node's in-progress data should be forgotten, and its evaluation should be restarted from scratch, as if it were freshly created/dirtied. To guard against this happening unexpectedly, any such events are passed to a GraphInconsistencyReceiver, which can verify that the SkyFunction is behaving properly. This is the first change in a series to permit action rewinding when it is discovered that a previously generated input file is no longer available. When an action detects that one of its inputs is unusable, it can return this sentinel value, causing it to be re-evaluated from scratch. Follow-up changes will make the node corresponding to the input, and the node corresponding to the action that generated the input, dirty when this happens, causing the upstream action to be re-run, regenerating the desired input. Currently works for builds that do not keep edges, although follow-ups may make this possible for all builds. PiperOrigin-RevId: 194863097
* Widen visibility for SkyframeExecutor.setSkylarkSemanticsGravatar mschaller2018-04-27
| | | | | | | | Allows subclasses to use different strategies for injecting Skylark semantics. RELNOTES: None. PiperOrigin-RevId: 194546199
* Fix skyframe target pattern evaluator thread usageGravatar ulfjack2018-04-26
| | | | PiperOrigin-RevId: 194387085
generated by cgit v1.2.3 (git 2.39.1) at 2024-06-20 01:53:05 +0000