aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Make android tools stop accepting --annotationJar.Gravatar ajmichael2018-03-22
| | | | | | | Bazel stopped passing it in https://github.com/bazelbuild/bazel/commit/1a6ca6f47aef36d56b5cb2f9da114af75dde583d. RELNOTES: None PiperOrigin-RevId: 190064697
* Forward deps artifacts from FeatureFlagSetterRule.Gravatar mstaib2018-03-22
| | | | | | | This is helpful for testing the behavior of rules underneath the flag setter without going around Bazel internals. RELNOTES: None. PiperOrigin-RevId: 190064449
* Add a new SandboxfsSandboxedSpawn to spawn actions using sandboxfs.Gravatar jmmv2018-03-22
| | | | | RELNOTES: None. PiperOrigin-RevId: 190062172
* Remove current contents of the default global bazelrc.Gravatar John Cater2018-03-22
| | | | | | | | | | | | | The file is left as an empty file, so make upgrades easier. Site administrators can add their own content as needed, still. Fixes #4850. Closes #4857. TESTED: Created a Debian VM and tested installing the new version as an upgrade and as the first Bazel, observed that the global bazelrc had the expected content. RELNOTES: Remove the default content of the global bazelrc. Change-Id: I5f285f0c59d18179005ffe5bfc35931299635bbb PiperOrigin-RevId: 190060838
* Local Disk Cache: Only write to disk if target hash doesn't already exist.Gravatar robinnabel2018-03-22
| | | | | | | | | | Under Windows, the default permissions used when writing to the local disk cache prevent the files from being overwritten. This CR adds a check: if the target file already exists, return early. This is a performance improvement and will fix the error described above as existing files will no longer need to be overwritten. Similar features have been implemented on the remote gRPC cache implementations, see bazel issue #4789. Closes #4886. PiperOrigin-RevId: 190060233
* Assume toolchain supportsDynamicLinker from presence of ↵Gravatar hlopko2018-03-22
| | | | | | | | | | | | | | | "dynamic_linking_mode" feature Toolchains no longer have to provide linking_mode_flags { mode: DYNAMIC } to state that they support dynamic linking mode. It is enough to provide a feature. This is part of ongoing work to get rid of linking_mode_flags from the CROSSTOOL. This is rollforward of https://github.com/bazelbuild/bazel/commit/3dab964407a2f1e12feb439d8507079bc06cb769. RELNOTES: None. PiperOrigin-RevId: 190059334
* Adding olaola to OWNERS on all remote libraries and tests.Gravatar olaola2018-03-22
| | | | | | TESTED=presubmit RELNOTES: None PiperOrigin-RevId: 190056498
* Readd SkylarkCcToolchainConfigureTest after fixGravatar hlopko2018-03-22
| | | | | | | This is "rollforward" of unknown commit. RELNOTES: None. PiperOrigin-RevId: 190055662
* Remove the --incompatible_disallow_uncalled_set_constructor flagGravatar vladmos2018-03-22
| | | | | | | | RELNOTES[INC]: The --incompatible_disallow_uncalled_set_constructor flag is no longer available, the `set` constructor` is completely removed from Skylark. Use `depset` instead. PiperOrigin-RevId: 190053331
* Deprecate BuildConfiguration.Fragment#topLevelConfigurationHook().Gravatar lberki2018-03-22
| | | | | RELNOTES: None. PiperOrigin-RevId: 190052728
* Remove ResourceFilterFactory#getOutputDirectorySuffix().Gravatar lberki2018-03-22
| | | | | | | | | I have a bit of a PTSD wrt. output directory naming (although stritcly speaking, the trauma is not over yet...) so I'm very happy to remove functionality that reminds me of having that problem. RELNOTES: None. PiperOrigin-RevId: 190052059
* remote/http: properly support read timeoutGravatar Jakob Buchgraber2018-03-22
| | | | | | | | | | Also, remove unused SO_TIMEOUT. Fixes #4890 cc @benjaminp Closes #4895. PiperOrigin-RevId: 190051030
* Remove overloaded function in QueryEnvironmentFactoryGravatar dbabkin2018-03-22
| | | | | RELNOTES: none PiperOrigin-RevId: 190046225
* Improve error message when a persistent worker sends an unparseable response.Gravatar philwo2018-03-22
| | | | | | Fixes #4897. PiperOrigin-RevId: 190045930
* Remove the flag --incompatible_show_all_print_messagesGravatar vladmos2018-03-22
| | | | | | | | RELNOTES[INC]: The flag --incompatible_show_all_print_messages is removed. Messages generated by `print` statements from any package will be displayed as DEBUG messages. PiperOrigin-RevId: 190043451
* add Bazel's Buildkite configurationGravatar Jakob Buchgraber2018-03-22
| | | | | | Closes #4900. PiperOrigin-RevId: 190043375
* Windows: Expose setup_vc_env_vars function from windows_cc_configure.bzlGravatar Yun Peng2018-03-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | `setup_vc_env_vars` runs vcvarsall.bat to setup environment varibles for Visuall C++ build tools. It returns the values of PATH, LIB, INCLUDE and WINDOWSSDKDIR. Expose this function so that users can use load("@bazel_tools/tools/cpp:windows_cc_configure.bzl", "setup_vc_env_vars") to access this function. Fixed https://github.com/bazelbuild/bazel/issues/4542 RELNOTES[NEW]: Now you can access three functions in windows_cc_configure.bzl by: load("@bazel_tools/tools/cpp:windows_cc_configure.bzl", "<function_name>") 1. find_vc_path(repository_ctx): Detect the VC installation path on your machine, respect BAZEL_VC first, return None if VC not found. 2. setup_vc_env_vars(repository_ctx, vc_path): Get PATH, LIB, INCLUDE, WINDOWSSDKDIR env vars set by VCVARSALL.bat, return None if VCVARSALL.bat not found. 3. find_msvc_tool(repository_ctx, vc_path, tool): Find a certain VC build tool (eg. cl.exe, lib.exe, link.exe), return None if tool not found. Change-Id: Iec5770165aaa4766470bb3b4d818a2cf90678586 Closes #4892. Change-Id: Iec5770165aaa4766470bb3b4d818a2cf90678586 PiperOrigin-RevId: 190039950
* Windows doc: Update docs about building Java and Python on WindowsGravatar Yun Peng2018-03-22
| | | | | | | | | | Now we are using the native launcher to run binary for Java and Python, so the outputs of these targets have changed. Closes #4898. Change-Id: Id2ea33d7c02b5608215b5cd4173ed917df0cb217 PiperOrigin-RevId: 190037369
* Remove an unused flag.Gravatar lberki2018-03-22
| | | | | | | I don't know what it *was* used for, but it sure is never writen at the moment, nor is it useful on the command line. RELNOTES: None. PiperOrigin-RevId: 190031826
* Store RuleClassType in RuleClass. Shouldn't use much more memory, and allows ↵Gravatar janakr2018-03-21
| | | | | | serialization to preserve the type when reconstituting the RuleClass. PiperOrigin-RevId: 190015323
* Include manifest size when presizing ijar outputGravatar cushon2018-03-21
| | | | PiperOrigin-RevId: 190010996
* Clean up unnecessary "additional data" from memoizing deserialization. Since ↵Gravatar janakr2018-03-21
| | | | | | memoization is now a simple on-off switch, change semantics to have at most one memoizing frame: starting memoization is now an idempotent operation. PiperOrigin-RevId: 189993914
* Use PATH and LD_LIBRARY_PATH from the client's environment if possibleGravatar ulfjack2018-03-21
| | | | | | | | This is technically an incompatible change, but I think it's unlikely to affect a lot of users. Note that this change leaves out Windows, where we set the PATH to the server env PATH plus the MSYS root determined from the shell path. This is a cleanup, and it also makes unknown commit slightly safer. PiperOrigin-RevId: 189981959
* Clarify migration instructions for providersGravatar brandjon2018-03-21
| | | | | | | | | This makes legacy providers a subsection so it can be referenced from linter error messages. It also adds some prose explaining the exact correspondence between legacy provider field keys and modern providers. RELNOTES: None PiperOrigin-RevId: 189980562
* Introduce string.elem() method.Gravatar laurentlb2018-03-21
| | | | | | | RELNOTES[NEW]: Strings have a new .elems() method, that provides an iterator on the characters of the string. PiperOrigin-RevId: 189980269
* Update the client's skeleton logging framework to use it for --client_debug.Gravatar ccalvarin2018-03-21
| | | | | | | | | We are still unable to turn this on to write to files, but there are currently 2 logging systems in use in the client: the inactive one, and the print-to-stderr option triggered by --client_debug. Combine these, so we can use the same logging format for both. Also combine it with the VerboseLogging functionality - it was not documented anywhere and seems redundant. RELNOTES: None. PiperOrigin-RevId: 189979051
* A tiny binary to print all rules in a package, after BUILD macro expansion.Gravatar carmi2018-03-21
| | | | | RELNOTES: None PiperOrigin-RevId: 189973158
* Stop objc_proto_library from returning the generated sources.Gravatar kaipi2018-03-21
| | | | PiperOrigin-RevId: 189971511
* Automatic code cleanup.Gravatar Googler2018-03-21
| | | | PiperOrigin-RevId: 189970123
* Increase the timeout of SkylarkTests, which should make it no longer flaky.Gravatar cparsons2018-03-21
| | | | | RELNOTES: None. PiperOrigin-RevId: 189949312
* Precompute Attribute#hashCode to remove gc churn from the Object[] ↵Gravatar nharmata2018-03-21
| | | | | | | | | allocation done by Attribute#hashCode. Attribute#hashCode is called a lot by the visitation done by e.g. Rule#getTransition. RELNOTES: None PiperOrigin-RevId: 189947421
* Memoize non-rule configured targets. This is probably a good idea in ↵Gravatar janakr2018-03-21
| | | | | | general, but especially necessary because OutputFileConfiguredTarget (which is not a RuleConfiguredTarget) contains a reference to its generating configured target, so all the optimizations needed for RuleConfiguredTarget apply to OutputFileConfiguredTarget. PiperOrigin-RevId: 189942424
* Opensource some basic client tests.Gravatar ccalvarin2018-03-21
| | | | | RELNOTES: None. PiperOrigin-RevId: 189938838
* Propagating and printing server logs on demand.Gravatar olaola2018-03-21
| | | | | | | WANT_LGTM=all TESTED=RBE, unit tests RELNOTES: None PiperOrigin-RevId: 189938345
* Remove obsolete phrase about SDKs and aaptGravatar Googler2018-03-21
| | | | | | | All SDKs now have aapt2 in them. RELNOTES: None. PiperOrigin-RevId: 189935587
* Change error messaging of @SkylarkCallable invocations to match ↵Gravatar cparsons2018-03-21
| | | | | | | | | @SkylarkSignature more closely. Also clarify the method representation in these error messages is for the attempted method *call*, not the actual method signature. RELNOTES: None. PiperOrigin-RevId: 189935148
* Remove no-op flag 'incompatible_load_argument_is_label'.Gravatar laurentlb2018-03-21
| | | | | | | No relnotes, because deletion was already announced. RELNOTES: None. PiperOrigin-RevId: 189934041
* Automated rollback of commit 3dab964407a2f1e12feb439d8507079bc06cb769.Gravatar hlopko2018-03-21
| | | | | | | | | | | | | | | | | | | *** Reason for rollback *** Break bazel ci. *** Original change description *** Assume toolchain supportsDynamicLinker from presence of "dynamic_linking_mode" feature Toolchains no longer have to provide linking_mode_flags { mode: DYNAMIC } to state that they support dynamic linking mode. It is enough to provide a feature. This is part of ongoing work to get rid of linking_mode_flags from the CROSSTOOL. RELNOTES: None. PiperOrigin-RevId: 189928991
* Remove support for dynamically configured resource filteringGravatar asteinb2018-03-21
| | | | | | | | | | | | Dynamically configured resource filtering was never turned on, as dynamic configuration turned out to be too slow. While we're at it, do a bit of related cleanup (remove the unused flag that controlled resource filtering in analysis, and switch from a now-two-valued enum to a boolean to specify whether resources should be filtered in analysis). RELNOTES: none PiperOrigin-RevId: 189923588
* Manual rollback of ↵Gravatar hlopko2018-03-21
| | | | | | | https://github.com/bazelbuild/bazel/commit/5cee7f4afbadf8f463b33cf47dcd2fde096c6a4b#diff-76ac2009811ac59b4f76dcfb6d17be2b RELNOTES: None. PiperOrigin-RevId: 189917189
* Automated rollback of commit 67549a752c2316234b19d7b50ea84a8ae80d3669.Gravatar hlopko2018-03-21
| | | | | | | | | | | | | | | | | | | | | | | | *** Reason for rollback *** Breaks on windows: https://www.google.com/url?sa=D&q=https%3A%2F%2Fbuildkite.com%2Fbazel%2Fgoogle-bazel-presubmit%2Fbuilds%2F624%234a68440b-948b-437b-a633-4f0595721bab *** Original change description *** Automated rollback of commit 3c5a1098af0c5ae80d4e3b1fc52dd1fef6027d43. *** Reason for rollback *** Breaks bazel ci: https://github.com/bazelbuild/bazel/issues/4894#event-1533040075 *** Original change description *** Add crosstool_lib.bzl and crosstool_utils.bzl These will be used to rewrite current crosstool autoconfiguration into action_configs and features. RELNOTES: None. PiperOrigin-RevId: 189906675
* Add behavior to NestedSetCodec to prevent it from running during testing.Gravatar cpeyser2018-03-21
| | | | PiperOrigin-RevId: 189906038
* @AutoCodec SupportData.Gravatar janakr2018-03-21
| | | | PiperOrigin-RevId: 189905795
* Automated rollback of commit 3c5a1098af0c5ae80d4e3b1fc52dd1fef6027d43.Gravatar hlopko2018-03-21
| | | | | | | | | | | | | | | | *** Reason for rollback *** Breaks bazel ci: https://github.com/bazelbuild/bazel/issues/4894#event-1533040075 *** Original change description *** Add crosstool_lib.bzl and crosstool_utils.bzl These will be used to rewrite current crosstool autoconfiguration into action_configs and features. RELNOTES: None. PiperOrigin-RevId: 189901308
* NDK cc_toolchains: include bundled runtime libraries in cc_toolchain.all_filesGravatar Jingwen Chen2018-03-21
| | | | | | | | | | | | Currently, generated NDK `cc_toolchain.all_files` filegroups do not glob the STL's static and dynamic runtime libraries, which includes important files like `libandroid_support.a` and `libc++.{so, a}`. This introduces sandboxed link time errors as @steeve discovered at https://github.com/bazelbuild/bazel/issues/3923#issuecomment-372750002 This PR adds the two filegroups to the respective toolchain's `all-files` filegroup to make them part of the toolchain's output in the sandbox. This generates a BUILD.bazel file like this: https://gist.github.com/jin/7d9fcf924fa33ab8fbc4c43c97676ab3 Fixes https://github.com/bazelbuild/bazel/issues/4864 Closes #4866. PiperOrigin-RevId: 189900428
* Add a test, verifying that paths within the same repo can be hard-codedGravatar Klaus Aehlig2018-03-21
| | | | | | | | | | | | | | To simplify the transition for existing repositories, we keep the property that paths to true source files in your own repository may be hard-coded. Add a test verifying that we keep this property. Also add a test checking that this property is true, if the same repository is built as a remote repository. But disable this test, as this property does not hold yet; we plan to make it true with a change of the execroot layout. Change-Id: I0fb961f8ddb8feb98a6309533fc82f2acd9afd0d PiperOrigin-RevId: 189898773
* blaze.cc: Remove unused had_to_wait variableGravatar David Ostrovsky2018-03-21
| | | | | | Closes #4851. PiperOrigin-RevId: 189897065
* Split PackageProvider to TargetProvider and CachingPackageLocatorGravatar dbabkin2018-03-21
| | | | | | | No need to pass instance of PackageProvider there as two more general interfaces required separately. RELNOTES:none PiperOrigin-RevId: 189895042
* Fix output_file build variable to be correct for fake compile actions.Gravatar hlopko2018-03-21
| | | | | | | | | | | | When this change is released, we can refactor compiler_output_flags feature to use output_file build variable in place of output_object_file, output_preprocess_file, and output_assembly_file for generating -o flag. We still need output_preprocess_file and output_assembly_file to add -S and -E, but we can get rid of output_object_file, and what's better we can stop caring about extension when constructing build variables for compile actions. RELNOTES: None. PiperOrigin-RevId: 189892917
* Assume toolchain supportsDynamicLinker from presence of ↵Gravatar hlopko2018-03-21
| | | | | | | | | | | | | "dynamic_linking_mode" feature Toolchains no longer have to provide linking_mode_flags { mode: DYNAMIC } to state that they support dynamic linking mode. It is enough to provide a feature. This is part of ongoing work to get rid of linking_mode_flags from the CROSSTOOL. RELNOTES: None. PiperOrigin-RevId: 189890583