aboutsummaryrefslogtreecommitdiffhomepage
path: root/src
Commit message (Collapse)AuthorAge
...
* Global cleanup change.Gravatar Eddie Aftandilian2017-01-20
| | | | | | -- PiperOrigin-RevId: 145059124 MOS_MIGRATED_REVID=145059124
* Refactor SkylarkNestedSet to not implement IterableGravatar Jon Brandvein2017-01-20
| | | | | | | | | | | | | | This is not intended to be a user-visible semantic change, aside from error messages. This is to help avoid unintentional flattening of depsets, and to narrow down the number of call sites where this can occur, to help us print warning/deprecation messages. EvalUtils#toIterable will now return an ImmutableList in place of SkylarkNestedSet. This should be ok since the caller shouldn't be relying on the result being a Skylark-safe type. Code that takes Iterable because it accepts either a list or set, can instead be changed to take Object and use EvalUtils#toIterableStrict for validation. Note that NestedSet still implements Iterable, so native code can still easily and accidentally flatten sets. -- PiperOrigin-RevId: 145044023 MOS_MIGRATED_REVID=145044023
* Fall back to the transitive classpath for top-level completion failuresGravatar Liam Miller-Cushon2017-01-20
| | | | | | | | | | These exceptions are typically handled in javac and result in a diagnostic, but it's possible for them to be thrown to the top-level by plugins or annotation processors. If that happens they will be printed to the error output stream. -- PiperOrigin-RevId: 145032097 MOS_MIGRATED_REVID=145032097
* Description redacted.Gravatar Carmi Grushko2017-01-20
| | | | | | -- PiperOrigin-RevId: 145030082 MOS_MIGRATED_REVID=145030082
* Filter out javac deferred diagnostic notes in JavaBuilderGravatar Liam Miller-Cushon2017-01-20
| | | | | | | | Fixes bazelbuild/bazel#1405 -- PiperOrigin-RevId: 145015113 MOS_MIGRATED_REVID=145015113
* Introduces --apple_crosstool_transition, which turns controls the activation ofGravatar Cal Peyser2017-01-20
| | | | | | | | the AppleCrosstoolTransition. -- PiperOrigin-RevId: 145006581 MOS_MIGRATED_REVID=145006581
* Add new names for depset traversal ordersGravatar Jon Brandvein2017-01-20
| | | | | | | | | | The old names are deprecated but still supported for now. A flag for enforcing the deprecation will come in a later change. RELNOTES[INC]: The traversal orders for depsets have been renamed. The old names are deprecated and will be removed in the future. New names: "stable" -> "default", "compile" -> "postorder", "link" -> "topological", "naive_link" -> "preorder". -- PiperOrigin-RevId: 145004718 MOS_MIGRATED_REVID=145004718
* Remove check that deps of cc target share its configuration.Gravatar Cal Peyser2017-01-20
| | | | | | -- PiperOrigin-RevId: 144992642 MOS_MIGRATED_REVID=144992642
* Change SpawnAction members back to privateGravatar Googler2017-01-20
| | | | | | | | | | | As pointed out in the new review comments on commit df726eae7b6ef11df3b962d720df36f63d5cd3a3, the SpawnAction members did not need to be changed from private to protected, they just needed to be accessed via accessor methods in the derived LTOBackendAction class. -- PiperOrigin-RevId: 144989490 MOS_MIGRATED_REVID=144989490
* Description redacted.Gravatar Carmi Grushko2017-01-20
| | | | | | -- PiperOrigin-RevId: 144986187 MOS_MIGRATED_REVID=144986187
* Description redacted.Gravatar Googler2017-01-20
| | | | | | -- PiperOrigin-RevId: 144983864 MOS_MIGRATED_REVID=144983864
* Rename and reorder depset testsGravatar Jon Brandvein2017-01-20
| | | | | | -- PiperOrigin-RevId: 144980245 MOS_MIGRATED_REVID=144980245
* Rollback of commit 8436354b93e0b611931dfc965d3e718980143dee.Gravatar Vladimir Moskva2017-01-19
| | | | | | | | | | | | | | *** Reason for rollback *** Didn't fix the problem. *** Original change description *** Tag targets as jdk8 to avoid building them for jdk7 -- PiperOrigin-RevId: 144974286 MOS_MIGRATED_REVID=144974286
* maven_rules.bzl: Fix maven coordinates order mismatch Gravatar David Ostrovsky2017-01-19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Classifier should be provided in the form: "groupId:artifactId:version[:packaging][:classifier]" because that's what maven-dependency-plugin expects and not in the form: "groupId:artifactId[:packaging][:classifier]:version" as documented here: [1]. Also it was missed to reflect the classifier in the output artifact name. TEST PLAN: WORKSPACE: load("//tools:maven_rules.bzl", "maven_jar", "maven_dependency_plugin") maven_dependency_plugin() maven_jar( name = "jetty_server_sources", artifact = "org.eclipse.jetty:jetty-server:9.3.11.v20160721:jar:sources", sha1 = "b23cac190808baed928260b2c9beca3b1ed232b4", ) BUILD: java_library( name = "jetty-server-sources", exports = ["@jetty_server_sources//jar"], ) $ bazel build :jetty-server-sources [...] Target //:jetty-server-sources up-to-date: bazel-bin/libjetty-server-sources.jar [1] https://maven.apache.org/pom.html#Maven_Coordinates Fixes #2049. -- Change-Id: I3297fb3676324cc6b4bb6ff6b2b6e18ce33f633c Reviewed-on: https://cr.bazel.build/7213 PiperOrigin-RevId: 144972944 MOS_MIGRATED_REVID=144972944
* Mark ConcurrentTests as flaky (flakiness is because of []). This is ↵Gravatar Nathan Harmata2017-01-19
| | | | | | | | preferable to disabling the test, but of course the ideal thing to do is actually fix the code. -- PiperOrigin-RevId: 144970101 MOS_MIGRATED_REVID=144970101
* Fix testLinkerParamFileIsExported on WindowsGravatar Yun Peng2017-01-19
| | | | | | | | Fixed: https://github.com/bazelbuild/bazel/issues/2376 -- PiperOrigin-RevId: 144969681 MOS_MIGRATED_REVID=144969681
* Description redacted.Gravatar Vladimir Moskva2017-01-19
| | | | | | -- PiperOrigin-RevId: 144965845 MOS_MIGRATED_REVID=144965845
* Description redacted.Gravatar Damien Martin-Guillerez2017-01-19
| | | | | | -- PiperOrigin-RevId: 144964816 MOS_MIGRATED_REVID=144964816
* Bazel client: abstract away mtime handlingGravatar Laszlo Csomor2017-01-19
| | | | | | | | | | | | | | | | | | | | | Create an IFileMtime class and platform-specific implementations to deal with mtime handling. Since epochs and time granularity vary from platform to platform, and we only care about setting a file's/directory's mtime to the current time or to a future time plus querying whether something is in the future, we can easily create an interface for these operations and that's exactly what IFileMtime is. Implement PosixFileMtime and WindowsFileMtime. See https://github.com/bazelbuild/bazel/issues/2107 -- PiperOrigin-RevId: 144956966 MOS_MIGRATED_REVID=144956966
* Remove JavaBuilder from the test dependencies Gravatar Damien Martin-Guillerez2017-01-19
| | | | | | | | | | | This dependency is actually inside the embedded tools of Bazel and no longer needed and cause issue with JDK 7 builds. -- Change-Id: I4810de68f0cc9c8d05a1aedbfb0594a69dc48d20 Reviewed-on: https://cr.bazel.build/8391 PiperOrigin-RevId: 144956010 MOS_MIGRATED_REVID=144956010
* Inject the digest of the marker file in the RepositoryDirectoryValue Gravatar Damien Martin-Guillerez2017-01-19
| | | | | | | | | | | This digest make sure that we produce a different RepositoryDirectoryValue for different marker data, thus invalidating the node. -- Change-Id: Ibc6286c76de7a8d07ece2bb8125bfb75095f7784 Reviewed-on: https://cr.bazel.build/8136 PiperOrigin-RevId: 144948233 MOS_MIGRATED_REVID=144948233
* Windows JNI: refactoringsGravatar Laszlo Csomor2017-01-19
| | | | | | | | | | | | | | | | | | | In this change: - rename //src/main/native:windows_jni_utils to //src/main/native:windows_jni_lib and make it visible to //src/main/cpp:__subpackages__ because we will use some methods there from this library - move AutoHandle into windows_util.h, we'll use it from blaze_util_windows.cc / file_windows.cc later See https://github.com/bazelbuild/bazel/issues/2107 -- PiperOrigin-RevId: 144946842 MOS_MIGRATED_REVID=144946842
* Refactor BlazeJavacMain to emit structured diagnosticsGravatar Liam Miller-Cushon2017-01-19
| | | | | | | | | instead of producing text output containing formatted diagnostics. This is necessary to support robust, locale-independent diagnostic post-processing. -- PiperOrigin-RevId: 144907133 MOS_MIGRATED_REVID=144907133
* Refactor a traversal over @SkylarkModule ancestorsGravatar Jon Brandvein2017-01-19
| | | | | | -- PiperOrigin-RevId: 144890965 MOS_MIGRATED_REVID=144890965
* Make frame bindings use LinkedHashMap for determinismGravatar Jon Brandvein2017-01-19
| | | | | | | | | This matters when the same rule (or other exportable) is bound to multiple variables, since the identifier of the first variable will become its name. -- PiperOrigin-RevId: 144881310 MOS_MIGRATED_REVID=144881310
* Set correct ENV variables for macOS builds.Gravatar Dmitry Shevchenko2017-01-19
| | | | | | | | * This allows xcrunwrapper to set the correct SDKROOT when building darwin targets. -- PiperOrigin-RevId: 144874291 MOS_MIGRATED_REVID=144874291
* Introduce --experimental_objc_crosstool, which replacesGravatar Cal Peyser2017-01-19
| | | | | | | | | | --experimental_objc_library and --experimental_objc_use_crosstool_for_binary. This flag will allow testing of the complete (compilation and linking) rollout of the objc crosstool. -- PiperOrigin-RevId: 144864301 MOS_MIGRATED_REVID=144864301
* Make the Path to FileSystem cache non-static for persistent workersGravatar Liam Miller-Cushon2017-01-19
| | | | | | -- PiperOrigin-RevId: 144858495 MOS_MIGRATED_REVID=144858495
* Fix javabuilder top-level exception handling with workers enabledGravatar Liam Miller-Cushon2017-01-19
| | | | | | -- PiperOrigin-RevId: 144857011 MOS_MIGRATED_REVID=144857011
* Remove end position hackGravatar Liam Miller-Cushon2017-01-19
| | | | | | | | javac enabled end positions by default when invoke using the API. -- PiperOrigin-RevId: 144850857 MOS_MIGRATED_REVID=144850857
* --cpu no longer reads its value from the CROSSTOOL.Gravatar Greg Estren2017-01-18
| | | | | | | | | | | | When the code below was first written, the default value for @Option cpu was null. This value was then overwritten with whatever target_cpu got loaded from the CROSSTOOL, which became the basis for all code queries of the value "cpu". This behavior was deprecated some months ago after the default @Option cpu was changed to auto-read from the host machine. So the CROSSTOOL "override" no longer ever happens. Also made the relevant test more hermetic. -- PiperOrigin-RevId: 144850329 MOS_MIGRATED_REVID=144850329
* Disallow proto_library deps for android_* rules.Gravatar Greg Estren2017-01-18
| | | | | | -- PiperOrigin-RevId: 144849686 MOS_MIGRATED_REVID=144849686
* Tag targets as jdk8 to avoid building them for jdk7Gravatar Vladimir Moskva2017-01-18
| | | | | | -- PiperOrigin-RevId: 144849308 MOS_MIGRATED_REVID=144849308
* Support mapping of Paths to URIsGravatar Klaus Aehlig2017-01-18
| | | | | | | | | | | Bazel-created files (like log files of test runs) are internally reported as Paths. However, this is not always the most useful representation of the location of that artifact for a consumer of build events. Therefore, support a mapping of paths to more useful URIs. -- PiperOrigin-RevId: 144843525 MOS_MIGRATED_REVID=144843525
* Pass AutoFDO profiles to LTO backend.Gravatar Googler2017-01-18
| | | | | | -- PiperOrigin-RevId: 144841772 MOS_MIGRATED_REVID=144841772
* --Gravatar Carmi Grushko2017-01-18
| | | | | PiperOrigin-RevId: 144839489 MOS_MIGRATED_REVID=144839489
* If --strict_proto_deps=default and a proto_library didn't specify ↵Gravatar Carmi Grushko2017-01-18
| | | | | | | | strict_proto_deps, then strictness is enforced. -- PiperOrigin-RevId: 144836042 MOS_MIGRATED_REVID=144836042
* Tests: empty windows_util_test on other platformsGravatar Laszlo Csomor2017-01-18
| | | | | | | | | | | | //src/test/native:windows_util_test now has a dummy source file on platforms other than Windows. This allows continuous integration to run all tests in this package on any platform. -- PiperOrigin-RevId: 144834522 MOS_MIGRATED_REVID=144834522
* Change the ActionInputPrefetcher API to bulk.Gravatar Ulf Adams2017-01-18
| | | | | | | | | | This interface is part of the module API, and this makes more sense than the previous per-file API, IMO, and is more future proof, even if there's no immediate benefit to the change. -- PiperOrigin-RevId: 144827566 MOS_MIGRATED_REVID=144827566
* Windows, JNI: add tests for windows_utilGravatar Laszlo Csomor2017-01-18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add tests for the AsExecutablePathForCreateProcess method, since its logic is pretty complex. Unfortunately testing it also requires complex logic, as we need to test what exactly happens when the input path is shorter than MAX_PATH or when it's longer than it. To test that reliably, we need a base path that we know will not get shortened. Creating that base path under the temp directory is a nightmare, we need to: (1) retrieve the temp dir, shorten it so we know that it won't be shortened further (2) keep creating subdirectories that have a short name so they also won't get shortened, but keep the entire path below MAX_PATH while leaving enough space for a file name in the end (3) append a file name such that the path is just below MAX_PATH, or is exactly that long, or is longer than it. Because of steps (1) and (2) we can be sure that no other component in the path will get shortened, so we can test exactly what's going on with the shortener logic and its error handling. But oh boy is it complicated. Side note, we need to use the Widechar WinAPI functions to create/delete the directories and files, because the POSIX API on Windows appears to be backed by the ASCII API functions, so attempting to `mkdir` with a path longer than CreateDirectoryA's limit is going to fail. But on the positive side, adding tests caught two bugs in the method, so we have that going for us which is nice. See https://github.com/bazelbuild/bazel/issues/2107 See https://github.com/bazelbuild/bazel/issues/2181 -- PiperOrigin-RevId: 144823029 MOS_MIGRATED_REVID=144823029
* Jarjar the third-party dependencies of the Jacoco test runner, except on ↵Gravatar Lukacs Berki2017-01-18
| | | | | | | | | | | | | | Windows. This way, we don't need a new release to make coverage collection happen. This is the fifth try on commit 823091f7516abf7d854021edc765daf1467f1647 . This time, it's disabled on Windows because two bugs (#2306 and #2342) collude to make it impossible to run Java binaries during the build on Windows and jarjar is a Java binary. Tested by building //src:bazel on a Windows machine with --output_base=<something with a $ sign in it>. I also verified that the genrule gets run by adding an "exit 1" at its beginning. Fifth time is the charm! Hopefully we don't get to seven. -- PiperOrigin-RevId: 144818587 MOS_MIGRATED_REVID=144818587
* Bazel client, Windows: use CreateFileWGravatar Laszlo Csomor2017-01-18
| | | | | | | | | | | | | | | | | | | | | | Use CreateFileW in blaze_util_windows.cc when opening the "jvm.out" file. This allows supporting long paths. Also use AsWindowsPathWithUncPrefix instead of just AsWindowsPath plus manually adding the UNC prefix. Also fix a compilation error in file_windows_test.cc, I'm surprised the CI system didn't catch this, maybe we aren't running this test there. See https://github.com/bazelbuild/bazel/issues/2107 See https://github.com/bazelbuild/bazel/issues/2181 -- PiperOrigin-RevId: 144813245 MOS_MIGRATED_REVID=144813245
* With interleaving now enabled, clean up our tests.Gravatar Ulf Adams2017-01-18
| | | | | | | | | | | | | | | | | | | | | | | | | Remove many calls from tests to SkyframeLabelVisitor.sync, which is generally no longer necessary. In particular, BuildViewTestCase.ensureTarget(s)Visited was calling it, so removing that, as well as simulateLoadingPhase. I moved some tests for TransitiveTargetFunction to the corresponding test class, where they belong. I also made sure that we have test coverage in the BuildViewTest for some cases that were previously only tested with TTF. I dropped some tests which no longer seem applicable / for which we have coverage elsewhere. Also, as a drive-by change, I fixed some warnings in BuildViewTestCase. Note that TTF is only used for genquery and in an experimental code path, but generally superseded by ConfiguredTargetFunction, which has taken over the responsibility for loading in all other cases (and interleaves loading and analysis). Some of the tests are not externally visible yet. -- PiperOrigin-RevId: 144813237 MOS_MIGRATED_REVID=144813237
* Bazel client, Windows: implement envvar handlingGravatar Laszlo Csomor2017-01-18
| | | | | | | | | | | Implement GetEnv, SetEnv, UnsetEnv, and use it in some locations, plus add some tests. See https://github.com/bazelbuild/bazel/issues/2107 -- PiperOrigin-RevId: 144808435 MOS_MIGRATED_REVID=144808435
* Change TargetContext.findDirectPrerequisite from config1 == config2 to ↵Gravatar Greg Estren2017-01-18
| | | | | | | | | | config1.equals(c2). This fixes an obscure bug between dynamic configurations, host configuration caching, and Skyframe skyKey interning that makes Bazel crash under certain sequences of builds. See changes for details. -- PiperOrigin-RevId: 144766296 MOS_MIGRATED_REVID=144766296
* Remove CC_LIBRARY values which are covered via LIBRARY when subtracting ↵Gravatar Chris Parsons2017-01-18
| | | | | | | | ObjcProvider subtrees -- PiperOrigin-RevId: 144747299 MOS_MIGRATED_REVID=144747299
* Rename symbolsTxt to symbols reflecting it's actual usage.Gravatar Googler2017-01-18
| | | | | | -- PiperOrigin-RevId: 144741831 MOS_MIGRATED_REVID=144741831
* Description redacted.Gravatar Adam Michael2017-01-18
| | | | | | -- PiperOrigin-RevId: 144739068 MOS_MIGRATED_REVID=144739068
* Add an attribute to java_toolchain to forcibly disable header compilationGravatar Liam Miller-Cushon2017-01-18
| | | | | | | | | | The attribute overrides the value of --java_header_compilation so we can enable it by default without breaking platforms that don't support it (currently only JDK 7 Bazel). -- PiperOrigin-RevId: 144737908 MOS_MIGRATED_REVID=144737908
* Disable --experimental_java_header_compilation_direct_classpathGravatar Liam Miller-Cushon2017-01-18
| | | | | | -- PiperOrigin-RevId: 144733615 MOS_MIGRATED_REVID=144733615