Datasets:

---

feedback-to-code

/

mozzarella

like 1

Follow

Feedback-2-Code 7

[ "keyword_pr_to_issue" ]

[ "Definitely not an expert on this, so take any of the following with a grain of salt. \n\nNote that there are two relevant Java toolchain types: the one for the JDK used for compilation (@bazel_tools//tools/jdk:toolchain_type) and the one for the JDK/JRE used as the runtime for executing a java_binary (@bazel_tools//tools/jdk:runtime_toolchain_type).\n\nThe toolchain_type JDK is used for compilation and should thus be selected based on the exec platform. Since Java compilation outputs are platform-independent, the target platform doesn't matter here.\n\nThe runtime_toolchain_type JDK is added to the runfiles of a java_binary and is executed when you do bazel run, so this actually needs to be selected based on the target platform. If you execute a Java tool during the build, there will be a transition to the exec configuration at some point, which will make the exec platform the target platform and thus again select the correct runtime.\n\nBased on my understanding of the points above, java_library should not have a toolchain dependency on @bazel_tools//tools/jdk:runtime_toolchain_type, only on @bazel_tools//tools/jdk:toolchain_type. That would allow you to have a java_library in the deps of an android_* target even though there is no runtime for Android. I checked the Starlark implementation of the java_library rules and it should allow this, but as far as I know the native implementation is enabled by default and I have no idea how it handles toolchains.\n\nIt's worrying that your reproducer seems to indicate that this doesn't work, which looks like it could indeed be a regression in Bazel 6.\n\n@cushon Could you take a look?\n\nRegarding the other toolchain resolution failures you are seeing: Is it possible that some custom rules or genrules you are using aren't correctly transitioning to the exec configuration? Do you also have a reproducer for these non-Android issues?", "Just noticed that the reproducer you linked has an empty main branch, but the content lives on the master branch.", "Thanks @fmeum for the attention. I fixed up the reproducer branch and split it into two issues, one showing the problem with simply a java_library and a bespoke target platform.\r\n\r\nI think I loosely understand the goal, but some issues I still see:\r\n* The reproducer breaks at java_library, not even java_binary needing to build out its runfiles tree\r\n* Even if only java_binary should have the runtime_toolchain_type dependency, I think you should still be able to `build` a java_binary without requiring a matching available toolchain?\r\n* When using `bazel run`, If you're going to be platform-sensitive with runfiles, wouldn't you want runfiles compatible with the host platform, not target? (If you're not using a hermetic jre, you're going to get the locally installed host-platform-compatible jre)\r\n\r\nI also pulled in the starlark rules_java and the error persists (resolving the current_java_runtime alias). You can toggle the load statement in not_android_example/BUILD. I think an issue is that since the rule declares its need for the runtime_toolchain_type, that must always be valid at analysis time.\r\n\r\n> Regarding the other toolchain resolution failures you are seeing: Is it possible that some custom rules or genrules you are using aren't correctly transitioning to the exec configuration? Do you also have a reproducer for these non-Android issues?\r\n\r\nAre you referring to the toolchain_resolution_debug output from above?", "```\r\nERROR: <output_base>/external/bazel_tools/tools/jdk/BUILD:28:19: While resolving toolchains for target @bazel_tools//tools/jdk:current_java_runtime\r\n```\r\n\r\nYou can use cquery to see the patch to the target that's resolving the toolchain:\r\n\r\n```\r\nbazel cquery 'somepath(//not_android_example/..., @bazel_tools//tools/jdk:current_java_runtime)'\r\n...\r\n//not_android_example:java_test (93797ec)\r\n@bazel_tools//tools/jdk:toolchain_java8 (93797ec)\r\n@bazel_tools//tools/jdk:platformclasspath (93797ec)\r\n@bazel_tools//tools/jdk:current_java_runtime (9d8e7aa)\r\n```\r\n\r\nThe `platformclasspath` stuff depends on a JDK in the target configuration so Java compilation can cross-compile to the right JDK version (this corresponds to javac's `--bootclasspath`/`--system` flags). The Java compilation output is architecture-independent, so we only really need a JDK with the same major version as the target platform, not necessarily the same architecture, but the way the provided Java toolchains are set up right now the two are tied together.", "Waiting for bazel 6.4 to try #18262 ", "A fix for this issue has been included in [Bazel 7.0.0 RC2](https://releases.bazel.build/7.0.0/rc2/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "P2", "team-Rules-Java", "next_month" ]

[ ".bazelrc", "BUILD", "MODULE.bazel", "MODULE.bazel.lock", "distdir_deps.bzl", "src/BUILD", "src/MODULE.tools", "tools/jdk/local_java_repository.bzl", "tools/jdk/remote_java_repository.bzl" ]

[ ".bazelrc", "BUILD", "MODULE.bazel", "MODULE.bazel.lock", "distdir_deps.bzl", "src/BUILD", "src/MODULE.tools", "tools/jdk/local_java_repository.bzl", "tools/jdk/remote_java_repository.bzl" ]

[ "src/test/java/com/google/devtools/build/android/dexer/BUILD", "src/test/py/bazel/test_base.py", "src/test/shell/bazel/bazel_java17_test.sh", "src/test/shell/bazel/bazel_java_test_defaults.sh", "src/test/shell/bazel/bazel_with_jdk_test.sh", "src/test/shell/integration/bazel_java_test.sh", "src/test/shell/testenv.sh.tmpl", "src/test/tools/bzlmod/MODULE.bazel.lock" ]

[ "keyword_pr_to_issue" ]

[ "@bazel-io flag", "CC @meteorcloudy @Wyverald ", "I think I have a fix for this, will submit it soon.", "@bazel-io fork 7.0.0", "Congrats on winning the 20k issue/pr number! 🎉 😄", "A fix for this issue has been included in [Bazel 7.0.0 RC5](https://releases.bazel.build/7.0.0/rc5/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "team-Rules-Java", "untriaged" ]

[ "src/main/java/com/google/devtools/build/lib/skyframe/BzlLoadFunction.java" ]

[ "src/main/java/com/google/devtools/build/lib/skyframe/BzlLoadFunction.java" ]

[ "src/test/py/bazel/bzlmod/bazel_module_test.py" ]

[ "keyword_pr_to_issue" ]

[ "@SalmaSamy @meteorcloudy ", "I'm going to preemptively mark this as a release blocker. 3x slowdown is unacceptable", "@bazel-io fork 7.0.0", "@purkhusid Could you share the output of `bazel info` after each of these builds?", "With `lockfile_mode=off` (25s runtime):\r\n```\r\nbazel-bin: /workspaces/bazel_output_base/execroot/_main/bazel-out/k8-fastbuild/bin\r\nbazel-genfiles: /workspaces/bazel_output_base/execroot/_main/bazel-out/k8-fastbuild/bin\r\nbazel-testlogs: /workspaces/bazel_output_base/execroot/_main/bazel-out/k8-fastbuild/testlogs\r\ncharacter-encoding: file.encoding = ISO-8859-1, defaultCharset = ISO-8859-1\r\ncommand_log: /workspaces/bazel_output_base/command.log\r\ncommitted-heap-size: 6954MB\r\nexecution_root: /workspaces/bazel_output_base/execroot/_main\r\ngc-count: 59\r\ngc-time: 1137ms\r\ninstall_base: /workspaces/bazel_user_root/install/e0df88ff6ee693b65a8a08e8e987fb40\r\njava-home: /workspaces/bazel_user_root/install/e0df88ff6ee693b65a8a08e8e987fb40/embedded_tools/jdk\r\njava-runtime: OpenJDK Runtime Environment (build 21+35) by Azul Systems, Inc.\r\njava-vm: OpenJDK 64-Bit Server VM (build 21+35, mixed mode) by Azul Systems, Inc.\r\nlocal_resources: RAM=64290MB, CPU=16.0\r\nmax-heap-size: 16861MB\r\noutput_base: /workspaces/bazel_output_base\r\noutput_path: /workspaces/bazel_output_base/execroot/_main/bazel-out\r\npackage_path: %workspace%\r\nrelease: release 7.0.0rc2\r\nrepository_cache: null\r\nserver_log: /workspaces/bazel_output_base/java.log.codespaces-4c3412.vscode.log.java.20231026-201446.5072\r\nserver_pid: 5072\r\nused-heap-size: 2798MB\r\nworkspace: /workspaces/monorepo\r\n```\r\n\r\nWith `--lockfile_mode=update` (96s runtime):\r\n```\r\nbazel-bin: /workspaces/bazel_output_base/execroot/_main/bazel-out/k8-fastbuild/bin\r\nbazel-genfiles: /workspaces/bazel_output_base/execroot/_main/bazel-out/k8-fastbuild/bin\r\nbazel-testlogs: /workspaces/bazel_output_base/execroot/_main/bazel-out/k8-fastbuild/testlogs\r\ncharacter-encoding: file.encoding = ISO-8859-1, defaultCharset = ISO-8859-1\r\ncommand_log: /workspaces/bazel_output_base/command.log\r\ncommitted-heap-size: 4445MB\r\nexecution_root: /workspaces/bazel_output_base/execroot/_main\r\ngc-count: 143\r\ngc-time: 2837ms\r\ninstall_base: /workspaces/bazel_user_root/install/e0df88ff6ee693b65a8a08e8e987fb40\r\njava-home: /workspaces/bazel_user_root/install/e0df88ff6ee693b65a8a08e8e987fb40/embedded_tools/jdk\r\njava-runtime: OpenJDK Runtime Environment (build 21+35) by Azul Systems, Inc.\r\njava-vm: OpenJDK 64-Bit Server VM (build 21+35, mixed mode) by Azul Systems, Inc.\r\nlocal_resources: RAM=64290MB, CPU=16.0\r\nmax-heap-size: 16861MB\r\noutput_base: /workspaces/bazel_output_base\r\noutput_path: /workspaces/bazel_output_base/execroot/_main/bazel-out\r\npackage_path: %workspace%\r\nrelease: release 7.0.0rc2\r\nrepository_cache: null\r\nserver_log: /workspaces/bazel_output_base/java.log.codespaces-4c3412.vscode.log.java.20231026-201700.7299\r\nserver_pid: 7299\r\nused-heap-size: 3670MB\r\nworkspace: /workspaces/monorepo\r\n```\r\n\r\nWith `--lockfile_mode=error` (96s runtime):\r\n```\r\nbazel-bin: /workspaces/bazel_output_base/execroot/_main/bazel-out/k8-fastbuild/bin\r\nbazel-genfiles: /workspaces/bazel_output_base/execroot/_main/bazel-out/k8-fastbuild/bin\r\nbazel-testlogs: /workspaces/bazel_output_base/execroot/_main/bazel-out/k8-fastbuild/testlogs\r\ncharacter-encoding: file.encoding = ISO-8859-1, defaultCharset = ISO-8859-1\r\ncommand_log: /workspaces/bazel_output_base/command.log\r\ncommitted-heap-size: 3019MB\r\nexecution_root: /workspaces/bazel_output_base/execroot/_main\r\ngc-count: 139\r\ngc-time: 3255ms\r\ninstall_base: /workspaces/bazel_user_root/install/e0df88ff6ee693b65a8a08e8e987fb40\r\njava-home: /workspaces/bazel_user_root/install/e0df88ff6ee693b65a8a08e8e987fb40/embedded_tools/jdk\r\njava-runtime: OpenJDK Runtime Environment (build 21+35) by Azul Systems, Inc.\r\njava-vm: OpenJDK 64-Bit Server VM (build 21+35, mixed mode) by Azul Systems, Inc.\r\nlocal_resources: RAM=64290MB, CPU=16.0\r\nmax-heap-size: 16861MB\r\noutput_base: /workspaces/bazel_output_base\r\noutput_path: /workspaces/bazel_output_base/execroot/_main/bazel-out\r\npackage_path: %workspace%\r\nrelease: release 7.0.0rc2\r\nrepository_cache: null\r\nserver_log: /workspaces/bazel_output_base/java.log.codespaces-4c3412.vscode.log.java.20231026-201940.9941\r\nserver_pid: 9941\r\nused-heap-size: 1371MB\r\nworkspace: /workspaces/monorepo\r\n```\r\n\r\nI did `bazel shutdown` between each invocation.", "I am able to reproduce the issue with a synthetic module extension that generates 2000 repos. This is the call stack responsible for the slowdown:\n\n\n", "@purkhusid Are you able to run Bazel built from source (`bazel build //src:bazel`)? If so, could you build and test https://github.com/bazelbuild/bazel/pull/19970?", "Ran the proposed fix on on our repo:\r\nWith `lockfile_mode=off`: 50s\r\nWith `lockfile_mode=update`: 41s\r\n\r\nNote that it's not the same machine as I used for the earlier runs so the overall slowdown compared to that is expected\r\n\r\nThe proposed fix seems to have worked and then some additional perf gain!\r\n", "A fix for this issue has been included in [Bazel 7.0.0 RC5](https://releases.bazel.build/7.0.0/rc5/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "P1", "team-ExternalDeps", "area-Bzlmod" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/BazelModuleResolutionEvent.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionResolutionEvent.java", "src/main/java/com/google/devtools/build/lib/events/Reportable.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/BazelModuleResolutionEvent.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionResolutionEvent.java", "src/main/java/com/google/devtools/build/lib/events/Reportable.java" ]

[]

[ "keyword_pr_to_issue" ]

[ "I also think that this should be solved in 7.0 since this will make migrations to bzlmod more painful.", "Exposing the repo mapping on `FilesToRunProvider` seems very low risk to me, but only allows you to get the repo mapping for existing executable targets, not arbitrary collections of runfiles. When packaging multiple binary targets this would require some kind of merging logic.\n\n@purkhusid Would this be sufficient for your use case?\n\nEdit: Whatever we decide to do here, the change should be straightforward and cherry-pickable.", "My exact use case i only packaging a single executable with it's runfiles so that would be sufficient for me. It could be a first step if you want to think a bit better about an public API for generating mappings files for arbitrary runfiles.", "A fix for this issue has been included in [Bazel 7.0.0 RC5](https://releases.bazel.build/7.0.0/rc5/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: feature request", "untriaged", "team-Rules-API" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/FilesToRunProvider.java", "src/main/java/com/google/devtools/build/lib/starlarkbuildapi/FilesToRunProviderApi.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/FilesToRunProvider.java", "src/main/java/com/google/devtools/build/lib/starlarkbuildapi/FilesToRunProviderApi.java" ]

[ "src/test/java/com/google/devtools/build/lib/analysis/RunfilesRepoMappingManifestTest.java" ]

[ "keyword_pr_to_issue" ]

[ "Which commands besides `build` (and its logical descendents, like `test` and `cquery`) do you think need support for `--flag_alias`?", "> Which commands besides `build` (and its logical descendents, like `test` and `cquery`) do you think need support for `--flag_alias`?\r\n\r\nFor our use case, none and I'm not sure where it might be useful in other commands. I was just basing this off of how `common` seems to be working for us with other flags. We've been able to replace usage of `build` in bazelrc with `common` except for this case.", "The code that checks this is [in `BlazeOptionHandler`](https://cs.opensource.google/bazel/bazel/+/master:src/main/java/com/google/devtools/build/lib/runtime/BlazeOptionHandler.java;drc=8833fd75d54c1763c221f1ed9d2e3f24ccfcc3d8;l=616), and the tests are in [`starlark_configurations_test.sh`](https://cs.opensource.google/bazel/bazel/+/master:src/test/shell/integration/starlark_configurations_test.sh;drc=4dc0d119a6f4901aa8353e75941c19205542743e;l=643).\r\n\r\nThe test makes it look like this is deliberate, and @gregestren added the check that it isn't supported for `test`, so I'm going to assign to him for further triage.", "I'm also re-classifying this as a Feature Request instead of a Bug, since the behavior seems to be intentional.", "Thats fair, I was considering this a bug based on this: https://github.com/bazelbuild/bazel/issues/3054", "https://github.com/bazelbuild/bazel/commit/329b22a980ad7a0835c53938afe06a111f40ac96 is the original change.\r\n\r\nThe intention was to not allow flag alias definitions for subsets of `build` (like `test`). Motivated specifically by CI systems that start with a `$ bazel canonicalize-flags` to vet flags. `canonicalize-flags` itself is a subset of `build` and any setting has to be reflected there \r\n\r\n`common` goes the other way up the inheritance tree and sounds safe to me. ", "This is basically a lowest common ancestor problem.\r\n\r\nWe can only apply flag alias definitions on nodes in `canonicalize-flags`'s ancestor path.", "A fix for this issue has been included in [Bazel 7.0.0 RC5](https://releases.bazel.build/7.0.0/rc5/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: feature request", "P2", "team-Configurability" ]

[ "src/main/java/com/google/devtools/build/lib/runtime/BlazeOptionHandler.java" ]

[ "src/main/java/com/google/devtools/build/lib/runtime/BlazeOptionHandler.java" ]

[ "src/test/shell/integration/starlark_configurations_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "@lberki I created this issue to get downstream coverage. Assuming CI is green, would you support flipping this for 7.0.0?", "Flipping this flag SGTM. It looks like the incompatible changes pipeline didn't find any incompatibilities." ]

[]

[ "type: process", "incompatible-change", "team-Local-Exec", "migration-ready", "breaking-change-7.0" ]

[ "src/main/java/com/google/devtools/build/lib/sandbox/SandboxOptions.java" ]

[ "src/main/java/com/google/devtools/build/lib/sandbox/SandboxOptions.java" ]

[ "src/test/java/com/google/devtools/build/lib/buildtool/EditDuringBuildTest.java", "src/test/shell/bazel/bazel_sandboxing_networking_test.sh", "src/test/shell/integration/execution_strategies_test.sh", "src/test/shell/integration/sandboxing_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "This reproduces with bazel version set to `7.0.0rc2`", "cc @fmeum ", "You are passing a dict from strings to tuples to `sdk`, but that's not a supported attribute type. Instead, you should pass a dict from strings to lists. But yes, Bazel shouldn't crash in this case.", "I see, so we should also probably adjust the `sdks` [README example](https://github.com/bazelbuild/rules_go/blob/master/go/toolchains.rst#go_download_sdk), which uses tuples", "Actually, I do see the semantic appeal of this now and if BUILD files support this and it's reasonably simple to add support for MODULE.bazel files, then why not. We need to handle the crash in some way anyway. I submitted https://github.com/bazelbuild/bazel/pull/20129.", "A fix for this issue has been included in [Bazel 7.0.0 RC5](https://releases.bazel.build/7.0.0/rc5/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "team-ExternalDeps", "untriaged", "area-Bzlmod" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/AttributeValuesAdapter.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/AttributeValuesAdapter.java" ]

[ "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/AttributeValuesAdapterTest.java" ]

[ "keyword_pr_to_issue" ]

[ "I think of this a bit different than @fmeum (in the closed PR).\r\n\r\nImo for WORKSPACE downloads, the download cache deduplication should only matter “much” across source code revisions(git commits) and not within the same revision. In practice, you would want to group most duplication within a single http_archive with multiple URLs, and there is good value into having a URL invalidation scheme.\r\n\r\nHowever I could see a potential fix to this:\r\n- Create a CAS store for download content, potentially the same CAS as disk_cache.\r\n- Make repository_cache (the download cache, not the new proposed cache) to track pointer record to CAS.\r\n\r\nThis means upon download, the archives should be written to CAS, then Bazel should create a pointer to the CAS in repository_cache dir. We could name this repository_download_cache or http_download_cache for good. This means that different http_archive rules downloading the same blob would create different file pointer records pointing to the same object in CAS.", "@sluongng This additional indirection is a good idea to avoid unnecessary duplication (in both downloads and storage). \r\n\r\nI think that we should figure out these details before we enable the invalidation in a release. I mostly see two problems with the behavior of `--repository_cache_urls_as_default_canonical_id`:\r\n\r\n1. The invalidation behavior can only realistically catch mistakes if both the list of URLs and the hash are specified by the end user. However, the flag applies on the level of `repository_ctx.download` and thus affects all repository rules, with the vast majority not reaping the benefits as they don't satisfy this condition. Restricting this behavior to `http_*` rules would avoid these kinds of silent performance degradations without benefits.\r\n2. The behavior of the flag is trivial to implement in Starlark: `canonical_id = \" \".join(urls)` is really all it does. Repository rules that satisfy the condition in 1. could just implement this logic themselves.\r\n3. The invalidation scheme makes it so that adding a mirror to an existing repository rule invalidates the cache, which is pretty counter-intuitive. @sluongng's suggestion would help here.\r\n\r\nEven if we think that 3. is acceptable in order to catch a common mistake (I can get behind this), we would probably still be better off by implementing this for `http_*` and in Starlark only.\r\n\r\n(I will flag this for 7.0.0 so that we end up making a conscious decision rather than running out of time)", "@bazel-io flag", "@bazel-io fork 7.0.0", "1 and 2 are good suggestions. 3 is rare enough that I think we could just slow roll it and target for 7.1, 7.2.\r\n\r\nIn the long run, i think it would be beneficial if we could provide a bit more information of the repository cache (real and download) via repository_ctx api. It would help us implement custom downloaders(i.e. go_repository fetcher) while still leveraging the cache. ", "Okay, I can implement the Starlark feature in `http_*`.\r\n\r\n@meteorcloudy Could you roll back https://github.com/bazelbuild/bazel/commit/fc3460e92597091997fa4154173e2d4e973ede31 as a first step? I can then replace the flag with an env var and Starlark logic.", "@fmeum Is it sufficient to implement this in starlark I wonder.\r\nThe problem is not just about cache invalidation, the problem is in the Java layer, the downloader will bias toward cache entry with SHA256 declared in the `http_archive`. So even if the starlark invalidate and re-trigger the downloader, the downloader would still grab the same binary out of the cache.\r\n\r\nI think the only concrete way to guard against this issue, where user updates the URLs but forgot to update SHA256, is to have the flag enabled. It does come with the reduction in repository_cache deduplication… however introducing a re-direction layer, such as a pointer system referencing a CAS, should solve the deduplication issue.", "> So even if the starlark invalidate and re-trigger the downloader, the downloader would still grab the same binary out of the cache.\r\n\r\nit would not use the cache if the canonical id is different, even if the digest is the same ", "Ah true. For some reason i thought we are still passing sha256 down to java. I has been a long day 🥲", "@fmeum Thanks for the analyzing! I'll take care of the rollback.", "I submitted a Starlark implementation as https://github.com/bazelbuild/bazel/pull/20047.", "A fix for this issue has been included in [Bazel 7.0.0 RC5](https://releases.bazel.build/7.0.0/rc5/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "P2", "team-ExternalDeps" ]

[ "MODULE.bazel.lock", "scripts/bootstrap/bootstrap.sh", "src/main/java/com/google/devtools/build/lib/bazel/BazelRepositoryModule.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/RepositoryOptions.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/DownloadManager.java", "tools/build_defs/repo/BUILD", "tools/build_defs/repo/http.bzl", "tools/build_defs/repo/jvm.bzl" ]

[ "MODULE.bazel.lock", "scripts/bootstrap/bootstrap.sh", "src/main/java/com/google/devtools/build/lib/bazel/BazelRepositoryModule.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/RepositoryOptions.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/DownloadManager.java", "tools/build_defs/repo/BUILD", "tools/build_defs/repo/cache.bzl", "tools/build_defs/repo/http.bzl", "tools/build_defs/repo/jvm.bzl" ]

[ "src/test/java/com/google/devtools/build/lib/blackbox/bazel/DefaultToolsSetup.java", "src/test/py/bazel/bzlmod/bazel_fetch_test.py", "src/test/py/bazel/test_base.py", "src/test/shell/bazel/bazel_repository_cache_test.sh", "src/test/shell/testenv.sh.tmpl", "src/test/tools/bzlmod/MODULE.bazel.lock" ]

[ "keyword_pr_to_issue" ]

[ "It looks like this might be related to https://github.com/bazelbuild/bazel/issues/17123 and https://github.com/bazelbuild/bazel/issues/17124\r\n\r\nI think this is more urgent than P2, since a mismatching/incorrect value is printed when the integrity is wrong.", "this has been helping us locally:\r\n\r\n**in `~/bin/integrity`** (or wherever you prefer your scripts):\r\n```bash\r\n#!/bin/bash\r\n\r\n#\r\n# SRI hash generator utility from github.com/bazelbuild/bazel/issues/17803\r\n#\r\n\r\nset -eo pipefail\r\n\r\n>&2 echo \"Computing integrity for $1\";\r\necho -n \"sha256-\";\r\ncurl -fsSL -tlsv1.2 \"$1\" | openssl dgst -sha256 -binary | openssl base64 -A\r\n```\r\n\r\n**usage:**\r\n```\r\n$> integrity github.com/sgammon/rules_graalvm/archive/v0.9.0.tar.gz\r\n\r\nComputing integrity for github.com/sgammon/rules_graalvm/archive/v0.9.0.tar.gz\r\nsha256-ljI6wbeludsa4aOIxe0fuDDUYo06tLfwlThVgyHgMRE=\r\n```\r\n\r\npiping the output will yield the integrity hash only, so it works with `pbcopy` on a mac:\r\n```\r\n$> integrity github.com/sgammon/rules_graalvm/archive/v0.9.0.tar.gz | pbcopy\r\nComputing integrity for github.com/sgammon/rules_graalvm/archive/v0.9.0.tar.gz\r\n\r\n# On your clipboard:\r\n# sha256-ljI6wbeludsa4aOIxe0fuDDUYo06tLfwlThVgyHgMRE=\r\n```", ":wave: I ran into this error. I found this really confusing. After consulting the docs and not seeing any guidance, I searched around github, but didn't find any actually using the integrity field. It is not clear at all what is wrong, and it seems like it is broken.\r\n\r\nI asked in slack and was helpfully shown the blessed way to translate the sha sum bazel provides ([thread](https://bazelbuild.slack.com/archives/C014RARENH0/p1699633751197809)).\r\n\r\nImproving the usability would be really helpful! Perhaps a \"legacy\" `sha256` field that allows it to work like `http_archive` would be fantastic.", "Should be fixed in 7.0.0 (except the doc part, which is covered by https://github.com/bazelbuild/bazel/issues/17124).", "A fix for this issue has been included in [Bazel 7.0.0 RC5](https://releases.bazel.build/7.0.0/rc5/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "P2", "team-ExternalDeps", "bad error messaging", "area-Bzlmod" ]

[ "MODULE.bazel.lock", "tools/build_defs/repo/http.bzl" ]

[ "MODULE.bazel.lock", "tools/build_defs/repo/http.bzl" ]

[ "src/test/shell/bazel/bazel_repository_cache_test.sh", "src/test/shell/bazel/workspace_resolved_test.sh", "src/test/tools/bzlmod/MODULE.bazel.lock" ]

[ "keyword_pr_to_issue" ]

[ "Is this a blocker for flipping the flag? If one of the android rules would fail with it?", "I've marked it as a blocker for Bazel 7 release (not for starting the release, but for finishing it), but we need someone from the rules_android team to investigate because the code is pretty confusing.", "@bazel-io fork 7.0.0", "Error log from `bazel test //src/test/shell/bazel/android:android_local_test_integration_test_with_platforms` (with the test re-enabled):\r\n```\r\n** test_android_local_test *****************************************************\r\n$TEST_TMPDIR defined: output root default is '/usr/local/google/home/jcater/.cache/bazel/_bazel_jcater/a85a0dacd9e20e18c47928db0f1530e0/sandbox/linux-sandbox/1280/execroot/_main/_tmp/c0008185402992cea4ecb54c46211ee6' and max_idle_secs default is '15'.\r\nWARNING: The following rc files are no longer being read, please transfer their contents or import their path into one of the standard rc files:\r\n/etc/bazel.bazelrc\r\nExtracting Bazel installation...\r\nStarting local Bazel server and connecting to it...\r\nINFO: Starting clean (this may take a while). Consider using --async if the clean takes more than several minutes.\r\n-- Test log: -----------------------------------------------------------\r\n$TEST_TMPDIR defined: output root default is '/usr/local/google/home/jcater/.cache/bazel/_bazel_jcater/a85a0dacd9e20e18c47928db0f1530e0/sandbox/linux-sandbox/1280/execroot/_main/_tmp/c0008185402992cea4ecb54c46211ee6' and max_idle_secs default is '15'.\r\nWARNING: The following rc files are no longer being read, please transfer their contents or import their path into one of the standard rc files:\r\n/etc/bazel.bazelrc\r\nComputing main repo mapping: \r\nWARNING: --enable_bzlmod is set, but no MODULE.bazel file was found at the workspace root. Bazel will create an empty MODULE.bazel file. Please consider migrating your external dependencies from WORKSPACE to MODULE.bazel. For more details, please refer to https://github.com/bazelbuild/bazel/issues/18958.\r\nLoading: \r\nLoading: 0 packages loaded\r\nAnalyzing: target //javatests/com/bin:test (1 packages loaded, 0 targets configured)\r\nAnalyzing: target //javatests/com/bin:test (1 packages loaded, 0 targets configured)\r\n[0 / 1] checking cached actions\r\nAnalyzing: target //javatests/com/bin:test (5 packages loaded, 6 targets configured)\r\n[1 / 1] checking cached actions\r\nAnalyzing: target //javatests/com/bin:test (62 packages loaded, 12 targets configured)\r\n[1 / 1] checking cached actions\r\nERROR: /usr/local/google/home/jcater/.cache/bazel/_bazel_jcater/a85a0dacd9e20e18c47928db0f1530e0/sandbox/linux-sandbox/1280/execroot/_main/_tmp/c0008185402992cea4ecb54c46211ee6/workspace.FJHlJe57/javatests/com/bin/BUILD:5:19: While resolving toolchains for target //javatests/com/bin:test (d3e1c6f): No matching toolchains found for types @bazel_tools//tools/jdk:runtime_toolchain_type.\r\nTo debug, rerun with --toolchain_resolution_debug='@bazel_tools//tools/jdk:runtime_toolchain_type'\r\nIf platforms or toolchains are a new concept for you, we'd encourage reading https://bazel.build/concepts/platforms-intro.\r\nERROR: Analysis of target '//javatests/com/bin:test' failed; build aborted\r\nINFO: Elapsed time: 3.703s, Critical Path: 0.04s\r\nINFO: 1 process: 1 internal.\r\nERROR: Build did NOT complete successfully\r\nERROR: No test targets were found, yet testing was requested\r\n------------------------------------------------------------------------\r\ntest_android_local_test FAILED: Tests for //javatests/com/bin:test failed.\r\n/usr/local/google/home/jcater/.cache/bazel/_bazel_jcater/a85a0dacd9e20e18c47928db0f1530e0/sandbox/linux-sandbox/1280/execroot/_main/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/android/android_local_test_integration_test_with_platforms.runfiles/_main/src/test/shell/bazel/android/android_local_test_integration_test_with_platforms:162: in call to test_android_local_test\r\n\r\nTear down:\r\n\r\nINFO[android_local_test_integration_test_with_platforms 2023-11-06 21:16:31 (+0000)] Cleaning up workspace\r\n\r\nFAILED: test_android_local_test\r\n```\r\n\r\nThe problem (I think) is that the test sets `--platforms` to an Android platform (in android_helper.sh), but there's no JDK available for that platform.", "Hullo! We're trying to get 7.0 blockers under control and this one seems to not have been investigated yet. Could we get an initial triage and an estimate on how long this'll take to fix?", "@ahumesky was investigating, I'll ask him to update.", "I have a change out for review that should address this" ]

[]

[ "type: bug", "P1", "team-Android" ]

[ "tools/android/android_sdk_repository_template.bzl" ]

[ "tools/android/android_sdk_repository_template.bzl" ]

[ "src/test/shell/bazel/android/android_helper.sh", "src/test/shell/bazel/android/android_local_test_integration_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "Could you check whether this is fixed with `USE_BAZEL_VERSION=d15ea929e2382b7354cb900b89fc22480b60ff4d`? Could be the same root cause as https://github.com/bazelbuild/bazel/issues/19822.", "No, unfortunately USE_BAZEL_VERSION=d15ea929e2382b7354cb900b89fc22480b60ff4d does not fix it, exactly the same error message is produced.\r\n\r\n```\r\nD:\\udu\\units\\cb_1012_bazel>bazel info release\r\n2023/10/23 16:41:15 Using unreleased version at commit d15ea929e2382b7354cb900b89fc22480b60ff4d\r\nWARNING: Option 'experimental_remote_build_event_upload' is deprecated: Use --remote_build_event_upload instead\r\nWARNING: Option 'experimental_remote_cache_compression' is deprecated: Use --remote_cache_compression instead\r\nINFO: Invocation ID: b90d014e-9200-45cf-ae56-2a2b3812852e\r\ndevelopment version\r\n```", "@bazel-io flag", "@bazel-io fork 7.0.0", "Hitting the exact same issue here, with the only (potential) difference that I disable bzlmod. How can we help debugging this ?\r\n\r\nAlso, turning the cc_shared_library into a cc_library works, with the obvious issue that it does not build the same thing in the end.", "CC @oquenchil ", "Fix is on the way\r\n\r\n@bazel-io fork 7.0.0", "Reopening since the fix has been rolled back.", "Hi all, is this still on track for 7.0? We're trying to close out the open release blockers and prepare for the final RC on Wed.", "Cc @buildbreaker2021 ", "@peakschris I have a potential fix, however it would be nice to have a minimal example to reproduce this to make sure that the fix is correct.", "@peakschris We have the potential fix submitted, we will backport it into 7.0 and have a new rc today. Please help verify the fix when it's available. You can also test with Bazel built from HEAD with Bazelisk by setting `USE_BAZEL_VERSION=last_green`.", "@meteorcloudy \r\n\r\nLet me know if you'd like for me to open a new issue for this, but I'm still hitting errors on 7.1.0.\r\n\r\nI've created a small repro here: https://github.com/Ryang20718/repro-cpp-linker\r\n\r\nif you clone it and run `bazelisk test //rust:rust_example_test` you'll hit a similar error \r\n\r\n```\r\nError in fail: The following libraries were linked statically by different cc_shared_libraries but not exported:\r\ncc_shared_library @proj//:proj_so:\r\n```\r\n", "@Ryang20718 Yes, please file a separate issue.\n\n@meteorcloudy I suspect that some of the issues could be due to the `--enable_bzlmod` flip and its effect on Label stringification. https://github.com/bazelbuild/bazel/pull/21622 would fix that." ]

[]

[ "type: bug", "team-Rules-CPP", "untriaged" ]

[ "src/main/starlark/builtins_bzl/common/cc/cc_shared_library.bzl", "src/main/starlark/builtins_bzl/common/cc/semantics.bzl" ]

[ "src/main/starlark/builtins_bzl/common/cc/cc_shared_library.bzl", "src/main/starlark/builtins_bzl/common/cc/semantics.bzl" ]

[ "src/main/starlark/tests/builtins_bzl/cc/cc_shared_library/test_cc_shared_library/BUILD.builtin_test" ]

[ "keyword_pr_to_issue" ]

[ "The issue seems not to be present in 77eacce3f56ce6b5b466337c123a5cdfd671f2ee .", "Bisection says that the culprit is unfortunately 51bddee96bc357d8131efba606420227c7a8ebc5 , an Abseil version bump. I'm pretty certain: after this change, I got 2/20 restarts, before, 0/50 .", "Interesting,\r\n\r\n> This issue doesn't seem to happen with 7.0.0rc4.\r\n\r\nThe abseil-cpp upgrade is also backported to 7.0.0rc4 in https://github.com/bazelbuild/bazel/pull/20221, if that's the root cause, 7.0.0rc4 should have the same issue?", "well that's alarming. I checked again and the confirmed the result of the bisection: 51bddee contains the bug, 39084edb2718865b6648274240ea94a29e62cdf7, its parent, does not and 7.0.0rc4 doesn't either (I did 50 runs each, 51bddee fails very frequently so I'm pretty sure 39084edb is not a false negative)\r\n\r\nMaybe we started using some Abseil functionality slightly differently between HEAD and 7.0.0 that happens to break with the upgrade?", "@lberki How did you build the binary at 51bddee96bc357d8131efba606420227c7a8ebc5?\r\n\r\nI actually couldn't reproduce this issue with Bazelisk with `USE_BAZEL_VERSION=51bddee96bc357d8131efba606420227c7a8ebc5`", "Wat. `git checkout 51bddee; bazel build //src:bazel`, then use the resulting Bazel binary. I put the two binaries I have at `/google/data/ro/users/lb/lberki/bazel.restart/*` (commithashes in their names).\r\n\r\nThe issue readily reproduces on my workstation (asciinema recording on request :) )\r\n\r\nAlso, automatic bisection readily zeroed in on the right commit, which is a good indication that the issue is reproducible at will.", "The difference is that Bazelisk uses `-c opt` builds. Without `-c opt`, the issue reproduces for me at 51bddee, with it, it doesn't. More alarmingly, the issue also reproduces when I build Bazel from `release-7.0.0pre4` (8c4b1c908e5b520126a2480b0855c0c1e58d1ae0) without `-c opt`.", "The good news is that this is reproducible also with `bazel version`. The bad news is that it becomes much less frequent with `GRPC_TRACE=all GRPC_VERBOSITY=INFO` and not reproducible with `GRPC_TRACE=all GRPC_VERBOSITY=DEBUG`.\r\n\r\nI teased these lines out from the gRPC logs:\r\n\r\n```\r\nI1121 14:04:48.250881833 4115421 retry_filter.cc:896] chand=0x564e161284f0 calld=0x564e16111070 attempt=0x564e160f0cc0: adding batch (start replayable pending batch on call attempt): SEND_INITIAL_METADATA{:path: /command_server.CommandServer/Ping, grpc-timeout: 30S} SEND_MESSAGE:flags=0x00000000:len=33 SEND_TRAILING_METADATA{} RECV_INITIAL_METADATA RECV_MESSAGE RECV_TRAILING_METADATA\r\nI1121 14:04:48.250898079 4115421 retry_filter.cc:1113] chand=0x564e161284f0 calld=0x564e16111070 attempt=0x564e160f0cc0: starting 1 retriable batches on lb_call=0x564e16111900\r\nI1121 14:04:48.250907137 4115421 call_combiner.h:172] CallCombinerClosureList executing closure while already holding call_combiner 0x564e1610ceb0: closure=0x564e16111840 error=OK reason=start replayable pending batch on call attempt\r\nI1121 14:04:48.250942134 4115421 handshaker.cc:93] handshake_manager 0x564e16119eb0: error=UNKNOWN:Invalid argument {file:\"external/grpc~1.48.1.bcr.1/src/core/lib/iomgr/tcp_client_posix.cc\", file_line:352, created_time:\"2023-11-21T14:04:48.250314978+00:00\", errno:22, os_error:\"Invalid argument\", syscall:\"connect\", target_address:\"ipv6:%5B::1%5D:35397\"} shutdown=0 index=1, args={endpoint=(nil), args=(nil) {size=0: {}}, read_buffer=(nil) (length=0), exit_early=0}\r\nI1121 14:04:48.250972032 4115421 handshaker.cc:126] handshake_manager 0x564e16119eb0: handshaking complete -- scheduling on_handshake_done with error=UNKNOWN:Invalid argument {file:\"external/grpc~1.48.1.bcr.1/src/core/lib/iomgr/tcp_client_posix.cc\", file_line:352, created_time:\"2023-11-21T14:04:48.250314978+00:00\", errno:22, os_error:\"Invalid argument\", syscall:\"connect\", target_address:\"ipv6:%5B::1%5D:35397\"}\r\n```\r\n\r\nNext step is `strace`, I guess, to see what exact system call fails and how it's different from the case when it doesn't fail?\r\n", "@bazel-io fork 7.0.0", "A quick look with strace on the client says:\r\n\r\n```\r\nconnect(8, {sa_family=AF_INET6, sin6_port=htons(42657), sin6_flowinfo=htonl(0), inet_pton(AF_INET6, \"::1\", &sin6_addr), sin6_scope_id=0}, 28) = -1 EINPROGRESS (Operation now in progress)\r\nuname({sysname=\"Linux\", nodename=\"<REDACTED>\", ...}) = 0\r\nmmap(NULL, 65536, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f48234c3000\r\nprctl(PR_SET_VMA, PR_SET_VMA_ANON_NAME, 0x7f48234c3000, 65536, \"absl\") = -1 EINVAL (Invalid argument)\r\nepoll_ctl(6, EPOLL_CTL_ADD, 8, {events=EPOLLIN|EPOLLOUT|EPOLLET, data={u32=734601009, u64=93996093873969}}) = 0\r\n```\r\n\r\nAnd I have to stop debugging for today; it looks like the reported `EINVAL` comes from a `prctl()` call, but what that system call is, I know not. There are a number of other similar `prctl()` calls, so it could very well be a red herring.", "After pondering this a bit, I propose that we treat this as a release blocker until we find out what's going on. Given that this bug manifests itself without `-c opt`, but not with it, I find it difficult to come up with a theory that makes it benign.\r\n\r\n\r\n", "wdyt @meteorcloudy @meisterT ?", "I can try to check if newer abseil-cpp version fixes it.", "`20230125.1` does not have the issue but I don't know how to verify whether it still includes the fix for https://github.com/bazelbuild/bazel/issues/20076", "@meisterT Thanks, I can confirm `20230125.1` also fixes https://github.com/bazelbuild/bazel/issues/20076. I can send a change to update abseil to this version.", "This is a gRPC bug: https://github.com/grpc/grpc/pull/35064", "@benjaminp impressive sleuthing!\r\n\r\nMy guess is that you're right and this is in fact a gRPC bug and the reason why the Abseil update tickled it is that https://github.com/abseil/abseil-cpp/commit/3a46229c3cdd945ccc8bae1458148115ff8f88e7 added a `prctl()` call in some low-level allocation code that gets called somehow in between (dunno, e.g. by `grpc_sockaddr_to_uri()`). That commit was pushed in 2023 April, which is conveniently within the 2022 Jun 23 - 2023 Aug 2 time span covered by the Abseil update (it's a long time, though, so it's not a smoking gun). It also explains why using Abseil 20230125.1 fixes this issue.\r\n\r\nThe only part that's missing is explaining why this bug does not reproduce with `-c opt`. I haven't found a suspicious `#ifdef` but that doesn't mean that there isn't any and given that it's low level memory allocation code, it's not surprising that it gets called at different times depending on how much the compiler optimizes things.\r\n\r\n", "A fix for this issue has been included in [Bazel 7.0.0 RC5](https://releases.bazel.build/7.0.0/rc5/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "P1", "untriaged", "team-Core" ]

[ "MODULE.bazel", "MODULE.bazel.lock" ]

[ "MODULE.bazel", "MODULE.bazel.lock" ]

[]

[ "keyword_pr_to_issue" ]

[ "`/DBAZEL_CURRENT_REPOSITORY=\"\"` actually isn't the escaped form, the value of this preprocessor variable is really meant to be an empty string *literal*, not the empty string.\r\n\r\nThis class in Bazel is responsible for the escaping and mentions double quotes as special characters that `gcc` and `clang` require to be escaped: https://github.com/bazelbuild/bazel/blob/818c5c8693c43fe490c9f6b2c05149eb8f45cf52/src/main/java/com/google/devtools/build/lib/util/GccParamFileEscaper.java#L32\r\n\r\nSince the newline character itself can appear in arguments, there has to be some kind of encoding scheme in parameter files that doesn't apply to command-line arguments. What kind of scheme does your compiler support, e.g., how would it represent this literal argument in the param file?\r\n\r\n```\r\n/DMY_VALUE_WITH_NEWLINE=\"fooo\r\nbar\"\r\n```", "Thanks for the details.\r\n\r\nIt appears my compiler is doing something with the first and last \" in the param files, so to achieve an empty string literal, it would want \r\n`-DBAZEL_CURRENT_REPOSITORY=\"\\\"\\\"\"` \r\nin the param file.\r\n\r\nI can file a bug with the compiler vendor, but this will take months to get fixed that way. Is there any workaround to disable the definiton BAZEL_CURRENT_REPOSITORY in a custom toolchain? I do see that this would also be an issue with the redacted_dates feature for \\_\\_DATE\\_\\_, \\_\\_TIMESTAMP\\_\\_, and \\_\\_TIME__ defines as well, but our \"workaround\" is to disable that feature (Its not needed anyways for us).", "While one could add a feature that disables the addition of `BAZEL_CURRENT_REPOSITORY`, that could end up breaking rulesets and tools from third parties that rely on runfiles libraries and thus this variable. That doesn't sound like a great feature to have since the problem is clearly unrelated to this particular variable.\r\n\r\nAlternatives I see:\r\n1. Use a custom Bazel binary until this is fixed in the compiler. This isn't difficult, you just have to modify the line https://github.com/bazelbuild/bazel/blob/a3c677dfea2de636a719d50345a5a97af96fae60/src/main/starlark/builtins_bzl/common/cc/cc_helper.bzl#L971 and then build `bazel build -c opt //src:bazel`. The C++ rules don't exist outside the Bazel binary yet, otherwise you could just load a custom version of `rules_cc`.\r\n2. In your C/C++ toolchain, wrap the compiler into a binary or shell script that fixes up the param file before invoking the compiler with it (see https://github.com/bazelbuild/apple_support/blob/master/crosstool/wrapped_clang.cc for an example that used to be part of Bazel itself).\r\n3. If this quoting style is common enough and not just a bug, maybe support for it could be added to the C++ rules.\r\n\r\nWould any of that work for you?\r\n\r\nCC @buildbreaker2021 who might have more ideas", "The only input I might add is that out of the options @fmeum has listed my favorite would be the second one.\r\nYou will have to write shell script once and forget about the bug until it is fixed.", "Sharing my opinion on that topic since it poped up on our end as well:\r\n\r\n- Adding a define to a compilation might not affect the final binary - but it could. From saftey perspective it is a problem. \r\n- In some industry sectors where you have to track each single define, compiler option/flag to pass an assesment, this will become a problem and therefore we decided to not upgrade to Bazel 6.0.0 +. \r\n- We (internal) still need to make a decision if we want to go on long term with a custom bazel binary, which would be not a one time effort in case this behaviour will be not changed.\r\n\r\nTherefore I was happy to find this thread. Coudln't we add a feature that explicitly need to be enabled to overcome that this define will be set ? ", "In my original proposal for runfiles handling, I proposed to add the define only to targets that directly depend on the runfiles library in `@bazel_tools//tools/cpp/runfiles`. I could see that being a better solution than introducing a feature.\r\n\r\n@buildbreaker2021 What do you think? It would be a slightly incompatible change.", "@bazel-io flag", "> While one could add a feature that disables the addition of `BAZEL_CURRENT_REPOSITORY`, that could end up breaking rulesets and tools from third parties that rely on runfiles libraries and thus this variable. That doesn't sound like a great feature to have since the problem is clearly unrelated to this particular variable.\r\n\r\nYes, I agree with all this. However, since my team is developing an embedded application, we do not rely on runfiles, so disabling this _should_ not effect our implementation, but I do understand the thought here.\r\n\r\n> \r\n> Alternatives I see:\r\n> \r\n> 1. Use a custom Bazel binary until this is fixed in the compiler. This isn't difficult, you just have to modify the line https://github.com/bazelbuild/bazel/blob/a3c677dfea2de636a719d50345a5a97af96fae60/src/main/starlark/builtins_bzl/common/cc/cc_helper.bzl#L971\r\n> and then build `bazel build -c opt //src:bazel`. The C++ rules don't exist outside the Bazel binary yet, otherwise you could just load a custom version of `rules_cc`.\r\n\r\nThis is probably the \"quick\" solution that I will implement if I can't get the command line length down below where a param file is required.\r\n\r\n> 2. In your C/C++ toolchain, wrap the compiler into a binary or shell script that fixes up the param file before invoking the compiler with it (see https://github.com/bazelbuild/apple_support/blob/master/crosstool/wrapped_clang.cc for an example that used to be part of Bazel itself).\r\n\r\nAnd this will be the longer term solution we will target if needed.\r\n\r\n> 3. If this quoting style is common enough and not just a bug, maybe support for it could be added to the C++ rules.\r\n\r\nEvaluating our other compilers, it seems that this quoting is unique to one, so I don't think option 3 is the right path for Bazel to take.\r\n\r\n\r\n> In my original proposal for runfiles handling, I proposed to add the define only to targets that directly depend on the runfiles library in `@bazel_tools//tools/cpp/runfiles`. I could see that being a better solution than introducing a feature.\r\n> \r\n> @buildbreaker2021 What do you think? It would be a slightly incompatible change.\r\n\r\nHowever, I think this is a more correct solution, which would also fix my issue.\r\n", "@bazel-io fork 7.0.0", "@lwelch @McAmun If possible, could you check whether https://github.com/bazelbuild/bazel/pull/20388 fixes the issue for your builds? It should also be cherry-pickable to Bazel 6 (but not 5).", "@fmeum, I cherry-picked #20388 on top of 6.4.0 and it meets my needs.\r\n\r\nI'm not sure how easy/hard releasing patch is, but is there any chance of getting this as a patch on top of any 6.x version?\r\n\r\n", "> @fmeum, I cherry-picked this on top of 6.4.0 and it meets my needs.\r\n> \r\n> I'm not sure how easy/hard releasing patch is, but is there any chance of getting this as a patch on top of any 6.x version?\r\n\r\ncc: @Wyverald", "@iancha1992 As this is a (mildly) backwards incompatible change, we could either cherry-pick it into 7.0.0 or may have to wait for Bazel 8. \r\n\r\nIt could be argued that we never promised that `BAZEL_CURRENT_REPOSITORY` is available everywhere as we only document it in the C++ runfiles library docs, which *may* make it okay to cherry-pick into 7.1.0.", "> It could be argued that we never promised that BAZEL_CURRENT_REPOSITORY is available everywhere as we only document it in the C++ runfiles library docs, which may make it okay to cherry-pick into 7.1.0.\r\n\r\nYes, I think https://github.com/bazelbuild/bazel/pull/20388 is more of a fix for an existing issue instead of a breaking change of an existing feature. And I guess the actual risk of breaking user should be small and easy to fix anyway? Let's cherry-pick it into 7.1.0.", "Sounds good!", "A fix for this issue has been included in [Bazel 7.0.0 RC6](https://releases.bazel.build/7.0.0/rc6/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!", "> A fix for this issue has been included in [Bazel 7.0.0 RC6](https://releases.bazel.build/7.0.0/rc6/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!\r\n\r\nI tested [Bazel 7.0.0 RC6] and it also resolved this issue. Thanks." ]

[]

[ "type: bug", "team-Rules-CPP", "untriaged" ]

[ "src/main/starlark/builtins_bzl/common/cc/cc_binary.bzl", "src/main/starlark/builtins_bzl/common/cc/cc_helper.bzl", "src/main/starlark/builtins_bzl/common/cc/cc_library.bzl", "tools/cpp/runfiles/runfiles_src.h" ]

[ "src/main/starlark/builtins_bzl/common/cc/cc_binary.bzl", "src/main/starlark/builtins_bzl/common/cc/cc_helper.bzl", "src/main/starlark/builtins_bzl/common/cc/cc_library.bzl", "tools/cpp/runfiles/runfiles_src.h" ]

[ "src/test/shell/bazel/cc_integration_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "cc @meteorcloudy ", "One possible solution I see is to change the order of registration to the following in order of decreasing precedence:\r\n1. Toolchains registered in the root module's MODULE.bazel file\r\n2. Toolchains registered in the WORKSPACE file\r\n3. Toolchains registered in dependencies' MODULE.bazel files in BFS order\r\n\r\nIndependent of that particular issue, the auto-detecting Python toolchain should have constraints to match on the auto-detected host platform, otherwise it probably breaks remote execution of Python tools.", "> RegisteredToolchainsFunction creates the list of registered toolchains by concatenating two lists: those from bzlmod, then those from the WORKSPACE file.\r\n\r\nYes, this is WAI. But we probably should document the priorities of toolchain registration more clearly.\r\n\r\n> Toolchains registered in the root module's MODULE.bazel file\r\n> Toolchains registered in the WORKSPACE file\r\n> Toolchains registered in dependencies' MODULE.bazel files in BFS order\r\n\r\nThis does look like a better order since it gives more control to users. But still could be surprising and should be documented.\r\n\r\n> Independent of that particular issue, the auto-detecting Python toolchain should have constraints to match on the auto-detected host platform, otherwise it probably breaks remote execution of Python tools.\r\n\r\n/cc @rickeylev ", "Yep, I find the order proposed by @fmeum more intuitive than what's at HEAD.", "There is a potential problem with this new order though: We can't distinguish top-level toolchain registrations from those coming from WORKSPACE macros or even the suffix. This could result in unexpected results of a different kind.", "Ideally toolchains registered in the workplace suffix would have the lowest priority (that's what effectively happened before bzlmod was a thing).\r\n\r\nWe've discussed (and I can't currently find the issue) adding explicit priority levels to `register_toolchains`: that could be used with some defaults to set the order of toolchains as described, but I'd want to take some time and flesh out the problem.\r\n\r\nFixing the python toolchains should be very easy, if @rickeylev wants help I can provide pointers.\r\n\r\nWhat's the priority on fixing bzlmod/workspace toolchain ordering?", "If you told me the behavior at HEAD is by design, I'd grudgingly accept that and ask how we could help people like the Kleaf folks with their migration. They do have a point that this (and other behavior similarly surprising, if it exists) doesn't make migration to bzlmod easy. I have a number of ideas which could make this more intuitive, but given that I you and @Wyverald follow toolchains and bzlmod much closer than I do, I'd rather not voice them before being asked.\r\n\r\nPractically speaking, I don't think there is a rush, at least from this project, given that they also need bzlmod vendoring to be able to turn bzlmod on. But let's not fall into the trap where both issues get deferred because they expect the other one to progress slowly :)\r\n", "Went ahead and wrote up #20356 for the python part: @lberki, while tests are running, can you see if this addresses the original issue?", "It doesn't fix the repro in the description of this bug, so I guess it does not?", "That's unfortunate. Can you share the repro with me?", "It's in the description of https://github.com/bazelbuild/bazel/issues/20354 , complete with a shell snippet to produce a repository from scratch and example Bazel invocations.", "Whoops, sorry", "Is it feasible for Kleaf to move/copy their toolchain registrations from WORKSPACE to MODULE.bazel? That would be the immediate fix.", "Using `--extra_toolchains` in a bazelrc would also be a workaround.", "> rules_python is by default and immutably (except if bzlmod is disabled) a bzlmod module and it registers @bazel_tools//tools/python:_autodetecting_py_runtime_pair as a Python toolchain.\r\n\r\nWell, yes and no. There's three different mechanisms that can cause it to get registered when you don't expect it:\r\n\r\n1. As part of Bazel itself. I don't know where this puts it in the toolchain priorities, but it happens very early, when Bazel is initializing the rule types it knows about. See https://github.com/bazelbuild/bazel/blob/38e8da1b571de21e51d92cce750875f0a9f280ca/src/main/java/com/google/devtools/build/lib/bazel/rules/BazelRuleClassProvider.java#L479 and https://github.com/bazelbuild/bazel/blob/38e8da1b571de21e51d92cce750875f0a9f280ca/src/main/java/com/google/devtools/build/lib/bazel/rules/python/python.WORKSPACE#L4\r\n2. As part of older rules_python \"releases\" (0.4 to 0.12). I put releases in quotes because, from what I can tell, those releases are of rules_python versions that didn't actually have bzlmod support. I'm guessing they were manually created to help bootstrap things. After that, the autodetecting toolchain is no longer registered. Starting in 0.23.0, a hermetic toolchain is automatically registered. Also note that Bazel itself depends on rules_python 0.4.0 right now (I have a pending change to upgrade that to 0.22.0, which doesn't register the autodetecting toolchain). See https://github.com/search?q=repo%3Abazelbuild%2Fbazel-central-registry+path%3Amodules%2Frules_python+autodetecting_toolchain&type=code\r\n4. The rules_bazel_integration_test module registers the autodetecting toolchain. So if that's in the module graph, that'll happen. See https://github.com/search?q=repo%3Abazelbuild%2Fbazel-central-registry+-path%3Amodules%2Frules_python+autodetecting_toolchain&type=code\r\n\r\n> https://github.com/bazelbuild/rules_python/issues/1313 \r\n\r\nThat actually sounds exactly like this issue: the user has an empty module, a workspace registering a toolchain, but then when they enable bzlmod, their toolchain is ignored.\r\n\r\nAs an aside, don't use the stuff in `@bazel_tools//tools/python`. Use `@rules_python` instead.", "> There is a potential problem with this new order though: We can't distinguish top-level toolchain registrations from those coming from WORKSPACE macros or even the suffix. This could result in unexpected results of a different kind.\r\n\r\nThis is a fixable problem, actually. We could annotate toolchains registered in WORKSPACE suffixes with a \"low priority\" bit (messy, but technically possible), then order the toolchains per 1) root MODULE.bazel 2) WORKSPACE sans suffix 3) non-root Bzlmod stuff 4) WORKSPACE suffix. This would address the Kleaf use case.", "I believe that this is catching more users by surprise: https://bazelbuild.slack.com/archives/CA306CEV6/p1701394561232659\r\n\r\nIt's pretty hard to come up with why this is failing if users are not aware of `MODULE.bazel` at all, as in this case.", "@bazel-io fork 7.0.0", "Since Bzlmod is enabled by default in Bazel 7 and that could trigger this issue for many users, we should probably address this in 7.0.0.", "A fix for this issue has been included in [Bazel 7.0.0 RC6](https://releases.bazel.build/7.0.0/rc6/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "P2", "team-ExternalDeps", "area-Bzlmod" ]

[ "site/en/external/migration.md", "src/main/java/com/google/devtools/build/lib/packages/Package.java", "src/main/java/com/google/devtools/build/lib/packages/WorkspaceFactory.java", "src/main/java/com/google/devtools/build/lib/packages/WorkspaceFactoryHelper.java", "src/main/java/com/google/devtools/build/lib/packages/WorkspaceGlobals.java", "src/main/java/com/google/devtools/build/lib/skyframe/WorkspaceFileFunction.java", "src/main/java/com/google/devtools/build/lib/skyframe/toolchains/BUILD", "src/main/java/com/google/devtools/build/lib/skyframe/toolchains/RegisteredExecutionPlatformsFunction.java", "src/main/java/com/google/devtools/build/lib/skyframe/toolchains/RegisteredToolchainsFunction.java" ]

[ "site/en/external/migration.md", "src/main/java/com/google/devtools/build/lib/packages/Package.java", "src/main/java/com/google/devtools/build/lib/packages/WorkspaceFactory.java", "src/main/java/com/google/devtools/build/lib/packages/WorkspaceFactoryHelper.java", "src/main/java/com/google/devtools/build/lib/packages/WorkspaceGlobals.java", "src/main/java/com/google/devtools/build/lib/skyframe/WorkspaceFileFunction.java", "src/main/java/com/google/devtools/build/lib/skyframe/toolchains/BUILD", "src/main/java/com/google/devtools/build/lib/skyframe/toolchains/RegisteredExecutionPlatformsFunction.java", "src/main/java/com/google/devtools/build/lib/skyframe/toolchains/RegisteredToolchainsFunction.java" ]

[ "src/test/java/com/google/devtools/build/lib/analysis/mock/BazelAnalysisMock.java", "src/test/java/com/google/devtools/build/lib/packages/util/LoadingMock.java", "src/test/java/com/google/devtools/build/lib/repository/ExternalPackageHelperTest.java", "src/test/java/com/google/devtools/build/lib/rules/platform/ToolchainTestCase.java", "src/test/java/com/google/devtools/build/lib/skyframe/toolchains/BUILD", "src/test/java/com/google/devtools/build/lib/skyframe/toolchains/RegisteredExecutionPlatformsFunctionTest.java", "src/test/java/com/google/devtools/build/lib/skyframe/toolchains/RegisteredToolchainsFunctionTest.java" ]

[ "keyword_pr_to_issue" ]

[ "Interesting observation thanks. I'll dig in some to find the root cause and report back...", "Some breakpoint debugging shows me this is what fails:\r\n\r\nhttps://github.com/bazelbuild/bazel/blob/09c621e4cf5b968f4c6cdf905ab142d5961f9ddc/src/main/java/com/google/devtools/build/lib/rules/config/ConfigSetting.java#L477-L480\r\n\r\n`optionDetails` has the registered value for `:enable_feature_a` (the original flag). But it's being called with `specifiedLabel=: enable_feature_a_aliased`. So `optionDetails.getOptionValue` returns null, which means it returns the flag's default value (`False`).\r\n\r\nhttps://github.com/bazelbuild/bazel/commit/fd2c682a6a6bb0759f92476e533bffd2883b9c27 has some precedent. I think extending some of its setup logic to user-defined flags can fix this easily enough.", "I prototyped the change, basically change to:\r\n\r\n```\r\nObject configurationValue =\r\n optionDetails.getOptionValue(actualLabel) != null\r\n ? optionDetails.getOptionValue(actualLabel)\r\n : provider.getDefaultValue(); \r\n```\r\n\r\nIt works, with one blocking caveat. There's already code to support `label_flag` in the opposite way:\r\nhttps://github.com/bazelbuild/bazel/commit/83f8d1846938f19976c2f69f39941975303372b8\r\n\r\n`label_flag` is implemented internally as an alias. And that code expects to be able to select() against its name vs. what it points to. Inside `ConfigSetting` there's no way to distinguish this from any other alias.\r\n\r\nSo we'd need either some way to distinguish them, and implement different semantics accordingly, or change the semantics so they're both interpreted the same way.", "Re: `label_flag` implementation, see https://github.com/bazelbuild/bazel/blob/c9f9eedc313fc043755dd30ae7e01478eed7c417/src/main/java/com/google/devtools/build/lib/rules/LabelBuildSettings.java#L92.", "Hi @gregestren , I'm new here and I'm looking through this ticket. Would it be fine if I pick this ticket up? ", "Hi @nikhilpothuru - you're welcome to. Do you feel like you can navigate effectively the conceptual challenge above of\r\n\r\n> So we'd need either some way to distinguish them, and implement different semantics accordingly, or change the semantics so they're both interpreted the same way.\r\n\r\n\r\n?\r\n\r\nThe smallest user footprint way would probably be to tell whether an alias is a `label_flag` vs. a `bool_flag`, then implement each one's logic accordingly. ", "Hi @gregestren, I take a look at this over the week and let you know if this is something I can handle. As I look through this ticket, I'll add any questions I have on this thread.", "Hi @gregestren, Is there any documentation for setting up the development environment? (i.e. setting up the debugger)", "Check out https://bazel.build/basics/getting_started.html.", "Thank you for contributing to the Bazel repository! This issue has been marked as stale since it has not had any activity in the last 1+ years. It will be closed in the next 14 days unless any other activity occurs or one of the following labels is added: \"not stale\", \"awaiting-bazeler\". Please reach out to the triage team (`@bazelbuild/triage`) if you think this issue is still relevant or you are interested in getting the issue resolved.", "@bazelbuild/triage Not stale", "Could you please prioritize this? Kleaf hits this issue too, so we just put in a wonky workaround here:\r\n\r\nhttps://cs.android.com/android/kernel/superproject/+/common-android-mainline:build/kernel/kleaf/common_kernels.bzl;l=929;drc=8e7727c052aef0b328cd6dfa94121f7ef7d9f77f\r\n\r\nIt is not urgent, but definitely annoying.", "@lberki noted this as a possible task for people looking for tasks. Still happily accepting contributors. This bug has some nice guidance in its history.", "We also encountered this on ChromeOS. \r\n\r\nWe are generating alias so developers don't have to type the full package paths. For example we have this alias set up:\r\n\r\n```\r\n% BOARD=amd64-generic bazel query @portage//target/sys-apps/attr:2.5.1 --output=build\r\n(22:43:22) INFO: Current date is 2023-11-21\r\n# /usr/local/google/home/aaronyu/.cache/bazel/_bazel_aaronyu/a5468eaf077d9e25965baa9c750fa428/external/_main~portage~portage/target/sys-apps/attr/BUILD.bazel:7:6\r\nalias(\r\n name = \"2.5.1\",\r\n visibility = [\"//bazel:internal\"],\r\n actual = select({\"//bazel/portage:stage1\": \"@_main~portage~portage//internal/packages/stage1/target/board/portage-stable/sys-apps/attr:2.5.1\", \"//bazel/portage:stage2\": \"@_main~portage~portage//internal/packages/stage2/target/board/portage-stable/sys-apps/attr:2.5.1\"}),\r\n)\r\n```\r\n\r\n\r\nWe hit this problem when we want to generate some target-specific build flags, along with the alias where `@portage//target/sys-apps/attr:incremental` points to the `bool_flag` `\"@_main~portage~portage//internal/packages/stage2/target/board/portage-stable/sys-apps/attr:incremental\"`\r\n\r\n\r\n`--@portage//target/sys-apps/attr:incremental` doesn't work in the command line.", "I have an internal change out to fix this. Hilariously, @jin just started working on this 2.5 years old bug the exact moment I did :(", "Any chance we get this in Bazel 7?", "@bazel-io flag", "@afq984 Bazel 7 has already been released, would Bazel 7.1 work?", "Apologies, I confused this issue with #20582. We were aliasing on the flag, not config_setting.\r\nI think we don't need this one.", "@bazel-io fork 7.1.0", "@iancha1992 Candidate for 7.0.1?", "@brentleyjones I don't think so, unless @meteorcloudy pushes hard for it. 7.0.1 is to fix serious regressions that slipped through into 7.0.0 and this doesn't look like one.", "A fix for this issue has been included in [Bazel 7.1.0 RC1](https://releases.bazel.build/7.1.0/rc1/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "P3", "team-Configurability", "not stale" ]

[ "src/main/java/com/google/devtools/build/lib/rules/config/ConfigSetting.java" ]

[ "src/main/java/com/google/devtools/build/lib/rules/config/ConfigSetting.java" ]

[ "src/test/java/com/google/devtools/build/lib/rules/LabelBuildSettingTest.java", "src/test/java/com/google/devtools/build/lib/rules/config/ConfigSettingTest.java" ]

[ "keyword_pr_to_issue" ]

[ "Thanks for the highly detailed report! I will look into this tomorrow.\n\nCc @lberki ", "Woah, that was a very detailed report, thanks for investing the time and energy into it :) It helped me to zero in on the problem:\r\n\r\n\r\nI can reproduce this. It's due to this snippet:\r\n\r\nhttps://github.com/bazelbuild/bazel/blob/c6a5ebb9af3d3e114a055c566f9aca777263a945/src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java#L506\r\n\r\nIt doesn't recognize `VirtualActionInput`s as something in the execroot, therefore, it doesn't transform its root to `/tmp/bazel-execroot`. Let me see if I can fix this with a well-aimed tiny change (I probably won't have the time to put together a test case, though)\r\n", "The good news is that this tiny patch:\r\n\r\n```\r\ndiff --git a/src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java b/src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java\r\nindex fb72d591e0..845d69be7b 100644\r\n--- a/src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java\r\n+++ b/src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java\r\n@@ -503,6 +503,8 @@ public final class SandboxHelpers {\r\n\r\n if (actionInput instanceof EmptyActionInput) {\r\n inputPath = null;\r\n+ } else if (actionInput instanceof VirtualActionInput) {\r\n+ inputPath = RootedPath.toRootedPath(withinSandboxExecRoot, actionInput.getExecPath());\r\n } else if (actionInput instanceof Artifact) {\r\n Artifact inputArtifact = (Artifact) actionInput;\r\n if (inputArtifact.isSourceArtifact() && sandboxSourceRoots != null) {\r\n```\r\n\r\nfixes the issue with the params files. The bad news is that it's _still_ not working, but the deploy jar creation (which is the action that fails for me) now fails with a different error. I probably won't have time to debug this any further today.\r\n\r\n", "Reproduction:\r\n\r\n```\r\ntouch WORKSPACE\r\nmkdir -p java/a\r\n\r\ncat > java/a/BUILD <<'EOF'\r\njava_library(\r\n name = \"a\",\r\n srcs = [\"A.java\"],\r\n)\r\n\r\njava_binary(\r\n name = \"b\",\r\n srcs = [\"B.java\"],\r\n deps = [\":a\"],\r\n)\r\nEOF\r\n\r\ncat > java/a/A.java <<'EOF'\r\npackage a;\r\n\r\npublic class A {\r\n public static int getValue() {\r\n return 3;\r\n }\r\n}\r\nEOF\r\n\r\ncat > java/a/B.java <<'EOF'\r\nimport a.A;\r\n\r\npublic class B {\r\n public static void main(String[] args) {\r\n System.err.println(A.getValue());\r\n }\r\n}\r\nEOF\r\n\r\nbazel --output_base=/tmp/ob build //java/a:b_deploy.jar\r\n\r\n```", "I continued debugging @lberki's reproducer and it turns out that the next issue is most likely https://github.com/bazelbuild/bazel/issues/20518, given that the failure points to this file:\r\nhttps://github.com/bazelbuild/bazel/blob/48374dfd677248be8c1318b5c6fbe885d6c690a7/tools/build_defs/build_info/bazel_java_build_info.bzl#L70", "@fmeum I think you're right -- I didn't get even far enough to find out that the missing file is a symlink but it indeed is.", "BTW: I'm running it either this issue or a very similar one showing up when `external/bazel_tools/tools/test/generate-xml.sh` tries to run:\r\n\r\n```\r\n$ bazel --output_base=$(mktemp -d -t tmp-bazel-base-XXXXXXXXXX) test //:some_test --sandbox_debug\r\n[...]\r\n1702872574.903264941: src/main/tools/linux-sandbox-pid1.cc:302: bind mount: /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/162/_hermetic_tmp -> /tmp\r\n1702872574.903285799: src/main/tools/linux-sandbox-pid1.cc:311: writable: /tmp/bazel-execroot/wd\r\n[...]\r\n1702872574.973004420: src/main/tools/linux-sandbox-pid1.cc:302: bind mount: /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/_hermetic_tmp -> /tmp\r\n1702872574.973021869: src/main/tools/linux-sandbox-pid1.cc:311: writable: /tmp/bazel-execroot/wd\r\n[...]\r\nERROR: /home/bcs/wd/BUILD:9:12: Testing //:some_test failed: (Exit 1): linux-sandbox failed: error executing TestRunner command \r\n (cd /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/execroot/wd && \\\r\n[...]\r\n /home/bcs/.cache/bazel/_bazel_bcs/install/89a68939cbf63eb54205fdf943a58b15/linux-sandbox \\\r\n -W /tmp/bazel-working-directory/wd \\\r\n -t 15 \\\r\n -w /tmp/bazel-execroot/wd \\\r\n -w /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/execroot/wd/_tmp/270135f98d0932ba0f76c9b90a5c8ea3 \\\r\n -w /tmp \\\r\n -w /dev/shm \\\r\n -M /tmp/tmp-bazel-base-ms8mQWfOXq/execroot \\\r\n -m /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/_hermetic_tmp/bazel-execroot \\\r\n -M /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/execroot \\\r\n -m /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/_hermetic_tmp/bazel-working-directory \\\r\n -M /tmp/tmp-bazel-base-ms8mQWfOXq/external/bazel_tools \\\r\n -m /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/_hermetic_tmp/bazel-source-roots/0 \\\r\n -M /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/_hermetic_tmp \\\r\n -m /tmp \\\r\n -S /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/stats.out \\\r\n -D /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/debug.out \\\r\n -- \\\r\n external/bazel_tools/tools/test/generate-xml.sh \\\r\n bazel-out/k8-fastbuild/testlogs/some_test/test.log \\\r\n bazel-out/k8-fastbuild/testlogs/some_test/test.xml 0 1)\r\n[...]\r\n$ ll /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/execroot/wd/external/bazel_tools/tools/test/generate-xml.sh\r\n```\r\n\r\nEvery path starting with `/tmp` that isn't under `/tmp/tmp-bazel-base-ms8mQWfOXq/` seems to be wrong. I'm seeing at least these three:\r\n- `/tmp`\r\n- `/tmp/bazel-working-directory/`\r\n- `/tmp/bazel-execroot/`\r\n\r\n", "@bazel-io flag", "@bazel-io fork 7.1.0", "@iancha1992 Forget to say that we should really have this target 7.0.1 - maybe not the full fix, but something that makes the default configuration work with it.", "https://github.com/bazelbuild/bazel/pull/20603 is a draft fix for this (it contains the aforementioned two-line change and the more involved fix for #20518)", "A fix for this issue has been included in [Bazel 7.1.0 RC1](https://releases.bazel.build/7.1.0/rc1/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!\r\n" ]

[]

[ "type: bug", "team-Local-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/actions/FileArtifactValue.java", "src/main/java/com/google/devtools/build/lib/sandbox/BUILD", "src/main/java/com/google/devtools/build/lib/sandbox/DarwinSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/DockerSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/ProcessWrapperSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java", "src/main/java/com/google/devtools/build/lib/sandbox/WindowsSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/skyframe/ActionOutputMetadataStore.java", "src/main/java/com/google/devtools/build/lib/worker/WorkerSpawnRunner.java" ]

[ "src/main/java/com/google/devtools/build/lib/actions/FileArtifactValue.java", "src/main/java/com/google/devtools/build/lib/sandbox/BUILD", "src/main/java/com/google/devtools/build/lib/sandbox/DarwinSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/DockerSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/ProcessWrapperSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java", "src/main/java/com/google/devtools/build/lib/sandbox/WindowsSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/skyframe/ActionOutputMetadataStore.java", "src/main/java/com/google/devtools/build/lib/worker/WorkerSpawnRunner.java" ]

[ "src/test/java/com/google/devtools/build/lib/sandbox/BUILD", "src/test/java/com/google/devtools/build/lib/sandbox/SandboxHelpersTest.java", "src/test/java/com/google/devtools/build/lib/skyframe/ActionOutputMetadataStoreTest.java", "src/test/shell/bazel/bazel_sandboxing_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "CC @lberki \r\n\r\nA simple workaround would be to just disable the hermetic `/tmp` in this case, similar to what we do for tmpfs paths. What do you think of this?\r\n\r\n```java\r\ndiff --git a/src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java b/src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java\r\nindex 3f6e49c72c..1e9b58ae00 100644\r\n@@ -206,17 +207,21 @@ final class LinuxSandboxedSpawnRunner extends AbstractSandboxSpawnRunner {\r\n return false;\r\n }\r\n \r\n- Optional<PathFragment> tmpfsPathUnderTmp =\r\n- getSandboxOptions().sandboxTmpfsPath.stream()\r\n+ Optional<PathFragment> mountUnderTmp =\r\n+ Stream.concat(\r\n+ getSandboxOptions().sandboxTmpfsPath.stream(),\r\n+ getSandboxOptions().sandboxAdditionalMounts.stream()\r\n+ .map(Map.Entry::getKey)\r\n+ .map(PathFragment::create))\r\n .filter(path -> path.startsWith(SLASH_TMP))\r\n .findFirst();\r\n- if (tmpfsPathUnderTmp.isPresent()) {\r\n+ if (mountUnderTmp.isPresent()) {\r\n if (warnedAboutNonHermeticTmp.compareAndSet(false, true)) {\r\n reporter.handle(\r\n Event.warn(\r\n String.format(\r\n- \"Falling back to non-hermetic '/tmp' in sandbox due to '%s' being a tmpfs path\",\r\n- tmpfsPathUnderTmp.get())));\r\n+ \"Falling back to non-hermetic '/tmp' in sandbox due to '%s' being a tmpfs path or mount source.\",\r\n+ mountUnderTmp.get())));\r\n }\r\n \r\n return false;\r\n```", "Flipping source and target in my example also results in a failure, not sure why I couldn't reproduce this earlier. I have edited the issue description accordingly.", "@bazel-io flag ", "@bazel-io fork 7.0.1", "I have a fix for this one. It's a one-line fix, modulo the test.", "@lberki Feel free to reuse the integration tests from https://github.com/bazelbuild/bazel/pull/20583.", "Sorry, too late, I already wrote my own :) (very similar to yours, though)", "@lberki Thanks for pushing the fix, but my mount targets under `/tmp` are still failing. I trimmed my PR at https://github.com/bazelbuild/bazel/pull/20583 down to just the new tests.", "Mount *targets*? I seem to remember discussing this with you at some point in time, and IIRC the consensus was that mount targets under `/tmp` aren't a good idea: the whole point of `--incompatible_sandbox_hermetic_tmp` is that `/tmp` is pristine and mount points under it seem to compromise that.", "I would say that's still entirely true, but with the flag flip, the user is no longer the one who specified both `--incompatible_sandbox_hermetic_tmp` *and* the mount pair. I think that it would be entirely reasonable to just automatically disable the hermetic `/tmp` in that case, but not doing so results in a confusing error.", "WDYT about emitting a non-confusing error?\r\n\r\nI agree that the current situation isn't fantastic and technically speaking, allowing mount targets under `/tmp` is not impossible, but I'd rather not allow that, at least on the first attempt.\r\n\r\nI also have an ardent desire to delete the old code path so reverting back to it under certain conditions isn't my favorite thing, either.", "@bazel-io fork 7.1.0", "A fix for this issue has been included in [Bazel 7.0.1 RC2](https://releases.bazel.build/7.0.1/rc2/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "team-Local-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java" ]

[ "src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java" ]

[ "src/test/shell/bazel/bazel_sandboxing_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "Thanks for the highly detailed report! I will look into this tomorrow.\n\nCc @lberki ", "Woah, that was a very detailed report, thanks for investing the time and energy into it :) It helped me to zero in on the problem:\r\n\r\n\r\nI can reproduce this. It's due to this snippet:\r\n\r\nhttps://github.com/bazelbuild/bazel/blob/c6a5ebb9af3d3e114a055c566f9aca777263a945/src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java#L506\r\n\r\nIt doesn't recognize `VirtualActionInput`s as something in the execroot, therefore, it doesn't transform its root to `/tmp/bazel-execroot`. Let me see if I can fix this with a well-aimed tiny change (I probably won't have the time to put together a test case, though)\r\n", "The good news is that this tiny patch:\r\n\r\n```\r\ndiff --git a/src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java b/src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java\r\nindex fb72d591e0..845d69be7b 100644\r\n--- a/src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java\r\n+++ b/src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java\r\n@@ -503,6 +503,8 @@ public final class SandboxHelpers {\r\n\r\n if (actionInput instanceof EmptyActionInput) {\r\n inputPath = null;\r\n+ } else if (actionInput instanceof VirtualActionInput) {\r\n+ inputPath = RootedPath.toRootedPath(withinSandboxExecRoot, actionInput.getExecPath());\r\n } else if (actionInput instanceof Artifact) {\r\n Artifact inputArtifact = (Artifact) actionInput;\r\n if (inputArtifact.isSourceArtifact() && sandboxSourceRoots != null) {\r\n```\r\n\r\nfixes the issue with the params files. The bad news is that it's _still_ not working, but the deploy jar creation (which is the action that fails for me) now fails with a different error. I probably won't have time to debug this any further today.\r\n\r\n", "Reproduction:\r\n\r\n```\r\ntouch WORKSPACE\r\nmkdir -p java/a\r\n\r\ncat > java/a/BUILD <<'EOF'\r\njava_library(\r\n name = \"a\",\r\n srcs = [\"A.java\"],\r\n)\r\n\r\njava_binary(\r\n name = \"b\",\r\n srcs = [\"B.java\"],\r\n deps = [\":a\"],\r\n)\r\nEOF\r\n\r\ncat > java/a/A.java <<'EOF'\r\npackage a;\r\n\r\npublic class A {\r\n public static int getValue() {\r\n return 3;\r\n }\r\n}\r\nEOF\r\n\r\ncat > java/a/B.java <<'EOF'\r\nimport a.A;\r\n\r\npublic class B {\r\n public static void main(String[] args) {\r\n System.err.println(A.getValue());\r\n }\r\n}\r\nEOF\r\n\r\nbazel --output_base=/tmp/ob build //java/a:b_deploy.jar\r\n\r\n```", "I continued debugging @lberki's reproducer and it turns out that the next issue is most likely https://github.com/bazelbuild/bazel/issues/20518, given that the failure points to this file:\r\nhttps://github.com/bazelbuild/bazel/blob/48374dfd677248be8c1318b5c6fbe885d6c690a7/tools/build_defs/build_info/bazel_java_build_info.bzl#L70", "@fmeum I think you're right -- I didn't get even far enough to find out that the missing file is a symlink but it indeed is.", "BTW: I'm running it either this issue or a very similar one showing up when `external/bazel_tools/tools/test/generate-xml.sh` tries to run:\r\n\r\n```\r\n$ bazel --output_base=$(mktemp -d -t tmp-bazel-base-XXXXXXXXXX) test //:some_test --sandbox_debug\r\n[...]\r\n1702872574.903264941: src/main/tools/linux-sandbox-pid1.cc:302: bind mount: /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/162/_hermetic_tmp -> /tmp\r\n1702872574.903285799: src/main/tools/linux-sandbox-pid1.cc:311: writable: /tmp/bazel-execroot/wd\r\n[...]\r\n1702872574.973004420: src/main/tools/linux-sandbox-pid1.cc:302: bind mount: /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/_hermetic_tmp -> /tmp\r\n1702872574.973021869: src/main/tools/linux-sandbox-pid1.cc:311: writable: /tmp/bazel-execroot/wd\r\n[...]\r\nERROR: /home/bcs/wd/BUILD:9:12: Testing //:some_test failed: (Exit 1): linux-sandbox failed: error executing TestRunner command \r\n (cd /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/execroot/wd && \\\r\n[...]\r\n /home/bcs/.cache/bazel/_bazel_bcs/install/89a68939cbf63eb54205fdf943a58b15/linux-sandbox \\\r\n -W /tmp/bazel-working-directory/wd \\\r\n -t 15 \\\r\n -w /tmp/bazel-execroot/wd \\\r\n -w /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/execroot/wd/_tmp/270135f98d0932ba0f76c9b90a5c8ea3 \\\r\n -w /tmp \\\r\n -w /dev/shm \\\r\n -M /tmp/tmp-bazel-base-ms8mQWfOXq/execroot \\\r\n -m /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/_hermetic_tmp/bazel-execroot \\\r\n -M /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/execroot \\\r\n -m /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/_hermetic_tmp/bazel-working-directory \\\r\n -M /tmp/tmp-bazel-base-ms8mQWfOXq/external/bazel_tools \\\r\n -m /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/_hermetic_tmp/bazel-source-roots/0 \\\r\n -M /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/_hermetic_tmp \\\r\n -m /tmp \\\r\n -S /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/stats.out \\\r\n -D /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/debug.out \\\r\n -- \\\r\n external/bazel_tools/tools/test/generate-xml.sh \\\r\n bazel-out/k8-fastbuild/testlogs/some_test/test.log \\\r\n bazel-out/k8-fastbuild/testlogs/some_test/test.xml 0 1)\r\n[...]\r\n$ ll /tmp/tmp-bazel-base-ms8mQWfOXq/sandbox/linux-sandbox/163/execroot/wd/external/bazel_tools/tools/test/generate-xml.sh\r\n```\r\n\r\nEvery path starting with `/tmp` that isn't under `/tmp/tmp-bazel-base-ms8mQWfOXq/` seems to be wrong. I'm seeing at least these three:\r\n- `/tmp`\r\n- `/tmp/bazel-working-directory/`\r\n- `/tmp/bazel-execroot/`\r\n\r\n", "@bazel-io flag", "@bazel-io fork 7.1.0", "@iancha1992 Forget to say that we should really have this target 7.0.1 - maybe not the full fix, but something that makes the default configuration work with it.", "https://github.com/bazelbuild/bazel/pull/20603 is a draft fix for this (it contains the aforementioned two-line change and the more involved fix for #20518)", "A fix for this issue has been included in [Bazel 7.1.0 RC1](https://releases.bazel.build/7.1.0/rc1/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!\r\n" ]

[]

[ "type: bug", "team-Local-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/actions/FileArtifactValue.java", "src/main/java/com/google/devtools/build/lib/sandbox/BUILD", "src/main/java/com/google/devtools/build/lib/sandbox/DarwinSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/DockerSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/ProcessWrapperSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java", "src/main/java/com/google/devtools/build/lib/sandbox/WindowsSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/skyframe/ActionOutputMetadataStore.java", "src/main/java/com/google/devtools/build/lib/worker/WorkerSpawnRunner.java" ]

[ "src/main/java/com/google/devtools/build/lib/actions/FileArtifactValue.java", "src/main/java/com/google/devtools/build/lib/sandbox/BUILD", "src/main/java/com/google/devtools/build/lib/sandbox/DarwinSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/DockerSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/ProcessWrapperSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/sandbox/SandboxHelpers.java", "src/main/java/com/google/devtools/build/lib/sandbox/WindowsSandboxedSpawnRunner.java", "src/main/java/com/google/devtools/build/lib/skyframe/ActionOutputMetadataStore.java", "src/main/java/com/google/devtools/build/lib/worker/WorkerSpawnRunner.java" ]

[ "src/test/java/com/google/devtools/build/lib/sandbox/BUILD", "src/test/java/com/google/devtools/build/lib/sandbox/SandboxHelpersTest.java", "src/test/java/com/google/devtools/build/lib/skyframe/ActionOutputMetadataStoreTest.java", "src/test/shell/bazel/bazel_sandboxing_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "CC @lberki \r\n\r\nA simple workaround would be to just disable the hermetic `/tmp` in this case, similar to what we do for tmpfs paths. What do you think of this?\r\n\r\n```java\r\ndiff --git a/src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java b/src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java\r\nindex 3f6e49c72c..1e9b58ae00 100644\r\n@@ -206,17 +207,21 @@ final class LinuxSandboxedSpawnRunner extends AbstractSandboxSpawnRunner {\r\n return false;\r\n }\r\n \r\n- Optional<PathFragment> tmpfsPathUnderTmp =\r\n- getSandboxOptions().sandboxTmpfsPath.stream()\r\n+ Optional<PathFragment> mountUnderTmp =\r\n+ Stream.concat(\r\n+ getSandboxOptions().sandboxTmpfsPath.stream(),\r\n+ getSandboxOptions().sandboxAdditionalMounts.stream()\r\n+ .map(Map.Entry::getKey)\r\n+ .map(PathFragment::create))\r\n .filter(path -> path.startsWith(SLASH_TMP))\r\n .findFirst();\r\n- if (tmpfsPathUnderTmp.isPresent()) {\r\n+ if (mountUnderTmp.isPresent()) {\r\n if (warnedAboutNonHermeticTmp.compareAndSet(false, true)) {\r\n reporter.handle(\r\n Event.warn(\r\n String.format(\r\n- \"Falling back to non-hermetic '/tmp' in sandbox due to '%s' being a tmpfs path\",\r\n- tmpfsPathUnderTmp.get())));\r\n+ \"Falling back to non-hermetic '/tmp' in sandbox due to '%s' being a tmpfs path or mount source.\",\r\n+ mountUnderTmp.get())));\r\n }\r\n \r\n return false;\r\n```", "Flipping source and target in my example also results in a failure, not sure why I couldn't reproduce this earlier. I have edited the issue description accordingly.", "@bazel-io flag ", "@bazel-io fork 7.0.1", "I have a fix for this one. It's a one-line fix, modulo the test.", "@lberki Feel free to reuse the integration tests from https://github.com/bazelbuild/bazel/pull/20583.", "Sorry, too late, I already wrote my own :) (very similar to yours, though)", "@lberki Thanks for pushing the fix, but my mount targets under `/tmp` are still failing. I trimmed my PR at https://github.com/bazelbuild/bazel/pull/20583 down to just the new tests.", "Mount *targets*? I seem to remember discussing this with you at some point in time, and IIRC the consensus was that mount targets under `/tmp` aren't a good idea: the whole point of `--incompatible_sandbox_hermetic_tmp` is that `/tmp` is pristine and mount points under it seem to compromise that.", "I would say that's still entirely true, but with the flag flip, the user is no longer the one who specified both `--incompatible_sandbox_hermetic_tmp` *and* the mount pair. I think that it would be entirely reasonable to just automatically disable the hermetic `/tmp` in that case, but not doing so results in a confusing error.", "WDYT about emitting a non-confusing error?\r\n\r\nI agree that the current situation isn't fantastic and technically speaking, allowing mount targets under `/tmp` is not impossible, but I'd rather not allow that, at least on the first attempt.\r\n\r\nI also have an ardent desire to delete the old code path so reverting back to it under certain conditions isn't my favorite thing, either.", "@bazel-io fork 7.1.0", "A fix for this issue has been included in [Bazel 7.0.1 RC2](https://releases.bazel.build/7.0.1/rc2/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "team-Local-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java" ]

[ "src/main/java/com/google/devtools/build/lib/sandbox/LinuxSandboxedSpawnRunner.java" ]

[ "src/test/shell/bazel/bazel_sandboxing_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "@bazel-io fork 7.0.1", "@Wyverald @meteorcloudy This, combined with https://github.com/bazelbuild/bazel/commit/f4be0b2e7b5539502661033bcf6f47722aaa7e47#diff-ef4da5cc2da76cded220009a0dcc95373700d704629a1a2e9ebe84a9be076854, actually breaks `jvm_maven_import_external` and `http_jar` when `rules_java` is updated, only workaround is to manually refetch the `maven` repo.", "If you don't plan to solve this for 7.0.1, we could alternatively fix the more immediate issue by getting rid of the `rules_java` loads in all Bazel-provided repo rules, instead relying on the native symbols.", "Feel free to downgrade this to 7.1.0 if you think that https://github.com/bazelbuild/bazel/pull/20810 is reasonable in the meantime.", "@bazel-io fork 7.1.0", "A fix for this issue has been included in [Bazel 7.1.0 RC1](https://releases.bazel.build/7.1.0/rc1/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "P2", "team-ExternalDeps" ]

[ "MODULE.bazel.lock", "tools/build_defs/repo/http.bzl", "tools/build_defs/repo/jvm.bzl" ]

[ "MODULE.bazel.lock", "tools/build_defs/repo/http.bzl", "tools/build_defs/repo/jvm.bzl" ]

[ "src/test/tools/bzlmod/MODULE.bazel.lock" ]

[ "keyword_pr_to_issue" ]

[ "This is how Kleaf works around the issue right now (by defining multiple repositories):\r\n\r\nhttps://cs.android.com/android/kernel/superproject/+/common-android-mainline:build/kernel/kleaf/download_repo.bzl;l=356;drc=93375eb99ad216911bca8366cd8b70dd3d00def8\r\n\r\nThis feature request blocks DDKv2 development ([Internal bug 298416462](https://issuetracker.google.com/issues/298416462))", "@meteorcloudy @Wyverald WDYT?\r\n\r\nI could imagine implementing something like `ctx.download_multiple()`; it would be a non-trivial but small addition to our API surface, although the official Bazel Way at the moment is, as you say, to define multiple repositories and achieve parallelism that way.\r\n\r\n@jacky8hyf How big are your downloads and how many of them are there and what pain do you experience from this? As long as you don't have an enormous number, repository-level parallelization should work, even if it is, as you say, a bit clunky.\r\n\r\nWhat I have pretty strong feelings about is that I'd much rather not get into the business of implementing something like Futures (`tokens = [ctx.download(url) for url in urls]; [token.wait() for token in tokens]`) in Starlark.\r\n\r\nFrom that internal bug, do I understand correctly that this is not a hard blocker, \"only\" a significant usability hurdle?\r\n\r\n", "I agree this would be a nice feature. But the team is quite overloaded currently, community contribution is welcome!", "I am working on rules to download kernel prebuilts from ci.android.com, and here are some numbers that may provide some context:\r\n\r\nFor a typical GKI build, I need to download **24** files to support external driver developments. The files range from empty (modules_blocklist), a few bytes (e.g. toolchain_version) to hundreds of megabytes (vmlinux, 294M)\r\n\r\nFor a certain build number that involves driver development kit (DDK):\r\n\r\n```\r\n42983565 gki_prebuilts_boot_img_tar_gz/file/boot-img.tar.gz\r\n35205632 gki_prebuilts_Image/file/Image\r\n14210442 gki_prebuilts_Image_gz/file/Image.gz\r\n16654085 gki_prebuilts_Image_lz4/file/Image.lz4\r\n146731 gki_prebuilts_kernel_aarch64_config_outdir_tar_gz/file/kernel_aarch64_config_outdir.tar.gz\r\n191 gki_prebuilts_kernel_aarch64_ddk_headers_archive_tar_gz/metadata.bzl\r\n9607113 gki_prebuilts_kernel_aarch64_ddk_headers_archive_tar_gz/file/kernel_aarch64_ddk_headers_archive.tar.gz\r\n13099 gki_prebuilts_kernel_aarch64_env_sh/file/kernel_aarch64_env.sh\r\n181195 gki_prebuilts_kernel_aarch64_internal_outs_tar_gz/file/kernel_aarch64_internal_outs.tar.gz\r\n8875380 gki_prebuilts_kernel_aarch64_module_env_tar_gz/file/kernel_aarch64_module_env.tar.gz\r\n1673 gki_prebuilts_kernel_aarch64_modules/file/kernel_aarch64_modules\r\n1661306 gki_prebuilts_kernel_uapi_headers_tar_gz/file/kernel-uapi-headers.tar.gz\r\n20328 gki_prebuilts_modules_builtin/file/modules.builtin\r\n133459 gki_prebuilts_modules_builtin_modinfo/file/modules.builtin.modinfo\r\n3922357 gki_prebuilts_modules_prepare_outdir_tar_gz/file/modules_prepare_outdir.tar.gz\r\n7640754 gki_prebuilts_modules_staging_dir_tar_gz/file/modules_staging_dir.tar.gz\r\n0 gki_prebuilts_system_dlkm_modules_blocklist/file/system_dlkm.modules.blocklist\r\n2135 gki_prebuilts_system_dlkm_modules_load/file/system_dlkm.modules.load\r\n7684967 gki_prebuilts_system_dlkm_staging_archive_tar_gz/file/system_dlkm_staging_archive.tar.gz\r\n4710929 gki_prebuilts_System_map/file/System.map\r\n9 gki_prebuilts_toolchain_version/file/toolchain_version\r\n35398730 gki_prebuilts_unstripped_modules_tar_gz/file/unstripped_modules.tar.gz\r\n307406096 gki_prebuilts_vmlinux/file/vmlinux\r\n779305 gki_prebuilts_vmlinux_symvers/file/vmlinux.symvers\r\n```\r\n\r\n> From that internal bug, do I understand correctly that this is not a hard blocker, \"only\" a significant usability hurdle?\r\n\r\nThis is correct.\r\n", "I think I'm hitting into a use case for this as well, though in my case splitting the downloads across multiple repos doesn't help, as passing a list of labels to a subsequent repo rule seems to still force downloads to happen serially. Details copied from the [Slack thread][slack].\r\n\r\nTo add some detail, I'm attempting to write a module extension that downloads several artifacts, extracts them, and then manipulates the output (in a way that's not amenable to e.g. a `genrule`).\r\n\r\nInitially I went for the obvious implementation (`rctx.download_and_extract()` in a loop). Which worked, but had the disadvantage that prompted this FR, i.e. the downloads being serial.\r\n\r\nI tried to work around this by changing it to more of a hub-and-spoke model: the `module_extension` generates a number of `http_archive`s, and then another repo rule takes a label list with the targets from the `http_archive`s.\r\n\r\nHowever, this also seems to result in downloads happening serially -- I'm guessing this is because the label attr prefetch mechanism mentioned in the [docs][prefetch].\r\n\r\nI've got a minimal-ish repro [here][repro] if it helps clarify the use-case at all. Please let me know if there's anything I've missed that would make the workaround as described ineffective.\r\n\r\n[slack]: https://bazelbuild.slack.com/archives/CA31HN1T3/p1700269436249229\r\n[repro]: https://gist.github.com/colatkinson/d3c0fbe0860d759b7ae88747a43eb67c\r\n[prefetch]: https://bazel.build/extending/repo#restarting_the_implementation_function\r\n", "Draft pull request: https://github.com/bazelbuild/bazel/pull/20309", "@bazel-io fork 7.1.0", "Just wanted to say thanks for developing this feature, it makes it much easier to implement `bzlmod` extensions that can do a few lightweight queries to the internet before calling repository rules without forcing users to wait a long time.", "We're glad we could help :)", "May I ask if this is enabled on 7.1.0 onwards? We are on 7.0.2 right now so we are looking for the version number to update to.", "Yes, this is in 7.1.0.", "A fix for this issue has been included in [Bazel 7.1.0 RC1](https://releases.bazel.build/7.1.0/rc1/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: feature request", "P2", "team-ExternalDeps", "team-Core", "help wanted" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/BUILD", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/DownloadManager.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/HttpDownloader.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkBaseExternalContext.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkRepositoryFunction.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/BUILD", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/DownloadManager.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/HttpDownloader.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkBaseExternalContext.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkRepositoryFunction.java" ]

[ "src/test/shell/bazel/external_integration_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "@bazel-io fork 7.0.1", "@fmeum Are you looking into a fix? This seems to be the only blocker for 7.0.1", "@meteorcloudy Yes, the speculative fix is https://github.com/bazelbuild/bazel/pull/20833. Unfortunately the reporter is unable to test this and it's also very challenging to cover in integration tests.", "A fix for this issue has been included in [Bazel 7.0.1 RC2](https://releases.bazel.build/7.0.1/rc2/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!\r\n", "Verified that 7.0.1 fixes this issue.", "@bazel-io fork 7.1.0" ]

[]

[ "type: bug", "team-Rules-CPP", "untriaged" ]

[ "MODULE.bazel.lock", "tools/cpp/unix_cc_configure.bzl" ]

[ "MODULE.bazel.lock", "tools/cpp/unix_cc_configure.bzl" ]

[ "src/test/tools/bzlmod/MODULE.bazel.lock" ]

[ "keyword_pr_to_issue" ]

[ "It doesn't out of the box, but it can be made to with a bit of boiler-plate:\r\n\r\n```starlark\r\nload(\r\n \"@bazel_tools//tools/build_defs/repo:utils.bzl\",\r\n \"read_netrc\",\r\n \"use_netrc\",\r\n)\r\nload(\"@bazel_skylib//lib:versions.bzl\", \"versions\")\r\n\r\ndef _use_netrc(netrc, urls, patterns):\r\n if versions.is_at_most(\"3.0.0\", versions.get()):\r\n return use_netrc(netrc, urls)\r\n return use_netrc(netrc, urls, patterns)\r\n\r\ndef get_auth(ctx, urls):\r\n \"\"\"Given the list of URLs obtain the correct auth dict.\r\n Args:\r\n ctx: The repository context.\r\n urls: A list of URLs.\r\n Returns:\r\n A dictionary of workspace attributes to update.\r\n \"\"\"\r\n\r\n # Mostly copied from https://github.com/bazelbuild/bazel/blob/3.7.2/tools/build_defs/repo/http.bzl\r\n if ctx.attr.netrc:\r\n netrc = read_netrc(ctx, ctx.attr.netrc)\r\n return _use_netrc(netrc, urls, ctx.attr.auth_patterns)\r\n\r\n if not ctx.os.name.startswith(\"windows\"):\r\n if \"HOME\" in ctx.os.environ:\r\n netrcfile = \"%s/.netrc\" % (ctx.os.environ[\"HOME\"],)\r\n if ctx.execute([\"test\", \"-f\", netrcfile]).return_code == 0:\r\n netrc = read_netrc(ctx, netrcfile)\r\n return _use_netrc(netrc, urls, ctx.attr.auth_patterns)\r\n elif \"USERPROFILE\" in ctx.os.environ and ctx.os.name.startswith(\"windows\"):\r\n netrcfile = \"%s/.netrc\" % (ctx.os.environ[\"USERPROFILE\"])\r\n if ctx.path(netrcfile).exists:\r\n netrc = read_netrc(ctx, netrcfile)\r\n return _use_netrc(netrc, urls, ctx.attr.auth_patterns)\r\n netrcfile = \"%s/_netrc\" % (ctx.os.environ[\"USERPROFILE\"])\r\n if ctx.path(netrcfile).exists:\r\n netrc = read_netrc(ctx, netrcfile)\r\n return _use_netrc(netrc, urls, ctx.attr.auth_patterns)\r\n\r\n return {}\r\n \r\ndef _my_rule_implementation(ctx):\r\n auth = get_auth(ctx, all_urls)\r\n\r\n download_info = ctx.download_and_extract(\r\n ctx.attr.urls,\r\n ctx.attr.sha256,\r\n auth = auth,\r\n )\r\n```\r\n\r\nThat said it isn't ideal for every rule implementation to have to do this.", "In recent versions, it's even simpler:\r\n\r\n```bzl\r\nload(\"@bazel_tools//tools/build_defs/repo:utils.bzl\", \"read_netrc\", \"read_user_netrc\", \"use_netrc\")\r\n\r\ndef _get_auth(ctx, urls):\r\n \"\"\"Given the list of URLs obtain the correct auth dict.\"\"\"\r\n if ctx.attr.netrc:\r\n netrc = read_netrc(ctx, ctx.attr.netrc)\r\n else:\r\n netrc = read_user_netrc(ctx)\r\n return use_netrc(netrc, urls, ctx.attr.auth_patterns)\r\n\r\ndef _my_rule_implementation(ctx):\r\n auth = _get_auth(ctx, all_urls)\r\n\r\n download_info = ctx.download_and_extract(\r\n ctx.attr.urls,\r\n ctx.attr.sha256,\r\n auth = auth,\r\n )\r\n\r\nmy_rule = rule(\r\n attrs = {\r\n \"auth_patterns\": attr.string_dict(),\r\n \"netrc\": attr.string(),\r\n \"urls\": attr.string_list(),\r\n },\r\n implementation = _my_rule_implementation,\r\n)\r\n```", "> In recent versions, it's even simpler:\r\n> \r\n> ```python\r\n> load(\"@bazel_tools//tools/build_defs/repo:utils.bzl\", \"read_netrc\", \"read_user_netrc\", \"use_netrc\")\r\n> \r\n> def _get_auth(ctx, urls):\r\n> \"\"\"Given the list of URLs obtain the correct auth dict.\"\"\"\r\n> if ctx.attr.netrc:\r\n> netrc = read_netrc(ctx, ctx.attr.netrc)\r\n> else:\r\n> netrc = read_user_netrc(ctx)\r\n> return use_netrc(netrc, urls, ctx.attr.auth_patterns)\r\n> \r\n> def _my_rule_implementation(ctx):\r\n> auth = _get_auth(ctx, all_urls)\r\n> \r\n> download_info = ctx.download_and_extract(\r\n> ctx.attr.urls,\r\n> ctx.attr.sha256,\r\n> auth = auth,\r\n> )\r\n> \r\n> my_rule = rule(\r\n> attrs = {\r\n> \"auth_patterns\": attr.string_dict(),\r\n> \"netrc\": attr.string(),\r\n> \"urls\": attr.string_list(),\r\n> },\r\n> implementation = _my_rule_implementation,\r\n> )\r\n> ```\r\n\r\nI would have never found this in a billion years. Any link to documentation or your process on how you figured this out?", "Everything in `@bazel_tools//`, including the built-in `http_archive`, lives in https://github.com/bazelbuild/bazel/tree/6.1.2/tools/. You can just read the code there. Documentation would be nice, though.", "The always-comprehensive, always-accurate documentation: the Source.\r\n\r\n:/", "Note for anyone blocked on this: using the `NETRC` environment variable or the `--credential_helper` flag should work around the issue, until we implement a proper fix. See integration test at https://github.com/bazelbuild/bazel/blob/62799ed724391ef3cd792fae0d81910c06daf245/src/test/py/bazel/bzlmod/bzlmod_credentials_test.py#L135 for an example." ]

[]

[ "type: bug", "P2", "team-ExternalDeps" ]

[ "MODULE.bazel.lock", "tools/build_defs/repo/http.bzl", "tools/build_defs/repo/jvm.bzl", "tools/build_defs/repo/utils.bzl" ]

[ "MODULE.bazel.lock", "tools/build_defs/repo/http.bzl", "tools/build_defs/repo/jvm.bzl", "tools/build_defs/repo/utils.bzl" ]

[ "src/test/tools/bzlmod/MODULE.bazel.lock" ]

[ "keyword_pr_to_issue" ]

[ "@bazel-io fork 7.0.1", "@Wyverald @meteorcloudy This, combined with https://github.com/bazelbuild/bazel/commit/f4be0b2e7b5539502661033bcf6f47722aaa7e47#diff-ef4da5cc2da76cded220009a0dcc95373700d704629a1a2e9ebe84a9be076854, actually breaks `jvm_maven_import_external` and `http_jar` when `rules_java` is updated, only workaround is to manually refetch the `maven` repo.", "If you don't plan to solve this for 7.0.1, we could alternatively fix the more immediate issue by getting rid of the `rules_java` loads in all Bazel-provided repo rules, instead relying on the native symbols.", "Feel free to downgrade this to 7.1.0 if you think that https://github.com/bazelbuild/bazel/pull/20810 is reasonable in the meantime.", "@bazel-io fork 7.1.0", "A fix for this issue has been included in [Bazel 7.1.0 RC1](https://releases.bazel.build/7.1.0/rc1/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: bug", "P2", "team-ExternalDeps" ]

[ "MODULE.bazel.lock", "tools/build_defs/repo/http.bzl", "tools/build_defs/repo/jvm.bzl" ]

[ "MODULE.bazel.lock", "tools/build_defs/repo/http.bzl", "tools/build_defs/repo/jvm.bzl" ]

[ "src/test/tools/bzlmod/MODULE.bazel.lock" ]

[ "keyword_pr_to_issue" ]

[ "cc @Wyverald ", "Note, we'll need to add a workaround in rules_oci for now, however it uses a requires using another tool during the repository rule, making it require a fake toolchain resolution and also the download it makes is not cached anywhere.", "@meteorcloudy Do you see any potential issues with allowing `rctx.download[_and_extract]` to set arbitrary headers? \r\n\r\n@thesayyn Would you be interested in contributing this functionality if you get the okay from maintainers? I could help with pre-reviews and references if needed.", "> Do you see any potential issues with allowing rctx.download[_and_extract] to set arbitrary headers?\r\n\r\nNot really. @tjgq integrated credential helper for repository download, do you have any concern here?", "We already [merge](https://cs.opensource.google/bazel/bazel/+/master:src/test/shell/bazel/starlark_repository_test.sh?q=test_auth_from_credential_helper_overrides_starlark) authentication headers from the credential helper and the argument to `rctx.download*` so adding additional headers from the Starlark side should work.", "> @thesayyn Would you be interested in contributing this functionality if you get the okay from maintainers? I could help with pre-reviews and references if needed.\r\n\r\nYes, I'd be willing to do that. ", "@fmeum I'd appreciate your points on https://github.com/bazelbuild/bazel/pull/19501", "I know this may not be directly related to bazel, but this is a message for others who are directed here by the rules_oci error log.\r\n\r\nAn checksum failed was throwed, and the error log looks like below:\r\n```\r\nWARNING: Could not fetch the manifest. Either there was an authentication issue or trying to pull an image with OCI image media types.\r\nFalling back to using `curl`. See https://github.com/bazelbuild/bazel/issues/17829 for the context.\r\nAnalyzing: target ... (89 packages loaded, 6572 targets configured)\r\nINFO: Repository rules_oci~1.4.0~oci~distroless_static_linux_amd64 instantiated at:\r\ncallstack not available\r\nRepository rule oci_pull defined at:\r\n/home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0/oci/private/pull.bzl:427:27: in <toplevel>\r\nERROR: An error occurred during the fetch of repository 'rules_oci~1.4.0~oci~distroless_static_linux_amd64':\r\nTraceback (most recent call last):\r\nFile \"/home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0/oci/private/pull.bzl\", line 378, column 57, in _oci_pull_impl\r\nmf, mf_len, mf_digest = downloader.download_manifest(rctx.attr.identifier, \"manifest.json\")\r\nFile \"/home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0/oci/private/pull.bzl\", line 319, column 74, in lambda\r\ndownload_manifest = lambda identifier, output: _download_manifest(rctx, state, identifier, output),\r\nFile \"/home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0/oci/private/pull.bzl\", line 281, column 18, in _download_manifest\r\n_download(\r\nFile \"/home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0/oci/private/pull.bzl\", line 247, column 23, in _download\r\nreturn download_fn(\r\nFile \"/home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0/oci/private/download.bzl\", line 129, column 29, in _download\r\ncache_it = rctx.download(\r\nError in download: com.google.devtools.build.lib.bazel.repository.downloader.UnrecoverableHttpException: Checksum was e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 but wanted 6706c73aae2afaa8201d63cc3dda48753c09bcd6c300762251065c0f7e602b25\r\nERROR: <builtin>: fetching oci_pull rule //:rules_oci~1.4.0~oci~distroless_static_linux_amd64: Traceback (most recent call last):\r\nFile \"/home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0/oci/private/pull.bzl\", line 378, column 57, in _oci_pull_impl\r\nmf, mf_len, mf_digest = downloader.download_manifest(rctx.attr.identifier, \"manifest.json\")\r\nFile \"/home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0/oci/private/pull.bzl\", line 319, column 74, in lambda\r\ndownload_manifest = lambda identifier, output: _download_manifest(rctx, state, identifier, output),\r\nFile \"/home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0/oci/private/pull.bzl\", line 281, column 18, in _download_manifest\r\n_download(\r\nFile \"/home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0/oci/private/pull.bzl\", line 247, column 23, in _download\r\nreturn download_fn(\r\nFile \"/home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0/oci/private/download.bzl\", line 129, column 29, in _download\r\ncache_it = rctx.download(\r\nError in download: com.google.devtools.build.lib.bazel.repository.downloader.UnrecoverableHttpException: Checksum was e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 but wanted 6706c73aae2afaa8201d63cc3dda48753c09bcd6c300762251065c0f7e602b25\r\nERROR: /home/admin/.cache/bazel/_bazel_admin/23e125c776c0411833d0c469fe9d82fd/external/rules_oci~1.4.0~oci~distroless_static/BUILD.bazel:1:6: @rules_oci~1.4.0~oci~distroless_static//:distroless_static depends on @rules_oci~1.4.0~oci~distroless_static_linux_amd64//:distroless_static_linux_amd64 in repository @rules_oci~1.4.0~oci~distroless_static_linux_amd64 which failed to fetch. no such package '@rules_oci~1.4.0~oci~distroless_static_linux_amd64//': com.google.devtools.build.lib.bazel.repository.downloader.UnrecoverableHttpException: Checksum was e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 but wanted 6706c73aae2afaa8201d63cc3dda48753c09bcd6c300762251065c0f7e602b25\r\n```\r\n\r\nI believe the reason that causes this problem is [here](https://github.com/bazel-contrib/rules_oci/blob/v1.4.0/oci/private/download.bzl#L130). To cache the file downloaded from `curl`, here using `ctx.download` downloads file from local which is impossible to fetch from remote downloader machine.\r\n\r\nTo workaround this, flag `--experimental_remote_downloader_local_fallback=true` could be helpful.\r\n\r\n\r\nreleate: \r\n- https://github.com/bazelbuild/bazel/issues/17771\r\n- https://github.com/bazel-contrib/rules_oci/issues/275", "Is this something that will require changes to `FetchBlob` RPC? That is, https://github.com/bazelbuild/remote-apis/blob/7e33c12ee96192f12b7fbb089e6fe1af6a29ef14/build/bazel/remote/asset/v1/remote_asset.proto#L154 doesn't include a mechanism for the client to specify headers.", "I don't there's a mechanism for it for remote downloaders. ", "I agree - it sounds like it'll require an addition to the protocol.", "A fix for this issue has been included in [Bazel 7.1.0 RC1](https://releases.bazel.build/7.1.0/rc1/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!", "I've opened https://github.com/bazelbuild/bazel/pull/21490 which is necessary so that rulesets which adopt this feature don't break users of the remote downloader service. If it looks good, I'd appreciate it if we could pull into either 7.1.0 or 7.1.1.", "> I've opened #21490 which is necessary so that rulesets which adopt this feature don't break users of the remote downloader service. If it looks good, I'd appreciate it if we could pull into either 7.1.0 or 7.1.1.\r\n\r\ncc: @Wyverald @meteorcloudy", "> We already [merge](https://cs.opensource.google/bazel/bazel/+/master:src/test/shell/bazel/starlark_repository_test.sh?q=test_auth_from_credential_helper_overrides_starlark) authentication headers from the credential helper and the argument to `rctx.download*` so adding additional headers from the Starlark side should work.\r\n\r\n@tjgq Just want to flag that `credential_helper` do not override the headers, it just adds another header line. The headers that Bazel sends when `credential_helper` returns `Accept` are something like this:\r\n```\r\n Accept: text/html, image/gif, image/jpeg, */*\\r\\n\r\n User-Agent: Bazel/release 7.1.0\\r\\n\r\n Authorization: Bearer XXXXXX\\r\\n\r\n Accept-Encoding: gzip\\r\\n\r\n Accept: application/octet-stream\\r\\n\r\n```\r\n\r\nThe tests in [`starlark_repository_test.sh`](https://github.com/bazelbuild/bazel/blob/master/src/test/shell/bazel/starlark_repository_test.sh) pass only since [`testing_server.py`](https://github.com/bazelbuild/bazel/blob/master/src/test/shell/bazel/testing_server.py) does not correctly merge headers despite [RFC9110](https://www.rfc-editor.org/rfc/rfc9110.html#name-field-order).\r\n\r\nIf you\r\n\r\n```bash\r\npython src/test/shell/bazel/testing_server.py --dump_headers headers.txt always /dev/null\r\n```\r\n\r\nand then \r\n\r\n```bash\r\ncurl -L http://localhost:43726/ -H \"Accept: this\" -H \"Accept: that\" \r\ncat header.txt\r\n{\"Accept\": \"that\"}\r\n```\r\n\r\nOnly the last header is captured. " ]

[]

[ "type: feature request", "P3", "team-ExternalDeps" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/DelegatingDownloader.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/DownloadManager.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/Downloader.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/HttpConnectorMultiplexer.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/HttpDownloader.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkBaseExternalContext.java", "src/main/java/com/google/devtools/build/lib/remote/downloader/GrpcRemoteDownloader.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/DelegatingDownloader.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/DownloadManager.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/Downloader.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/HttpConnectorMultiplexer.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/HttpDownloader.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkBaseExternalContext.java", "src/main/java/com/google/devtools/build/lib/remote/downloader/GrpcRemoteDownloader.java" ]

[ "src/test/java/com/google/devtools/build/lib/bazel/repository/downloader/HttpDownloaderTest.java", "src/test/java/com/google/devtools/build/lib/remote/downloader/GrpcRemoteDownloaderTest.java", "src/test/shell/bazel/remote_helpers.sh", "src/test/shell/bazel/starlark_repository_test.sh", "src/test/shell/bazel/testing_server.py" ]

[ "keyword_pr_to_issue" ]

[ "@bazel-io fork 7.0.1", "@fmeum Are you looking into a fix? This seems to be the only blocker for 7.0.1", "@meteorcloudy Yes, the speculative fix is https://github.com/bazelbuild/bazel/pull/20833. Unfortunately the reporter is unable to test this and it's also very challenging to cover in integration tests.", "A fix for this issue has been included in [Bazel 7.0.1 RC2](https://releases.bazel.build/7.0.1/rc2/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!\r\n", "Verified that 7.0.1 fixes this issue.", "@bazel-io fork 7.1.0" ]

[]

[ "type: bug", "team-Rules-CPP", "untriaged" ]

[ "MODULE.bazel.lock", "tools/cpp/unix_cc_configure.bzl" ]

[ "MODULE.bazel.lock", "tools/cpp/unix_cc_configure.bzl" ]

[ "src/test/tools/bzlmod/MODULE.bazel.lock" ]

[ "keyword_pr_to_issue" ]

[ "I started working on this in the past in https://github.com/bazelbuild/bazel/pull/18990, but noticed a problems with my implementation: All repositories generated by a module extension share the same repository mapping containing all these repos. If stored naively, this results in memory usage that is quadratic in the number of extension repos, which can easily be in the 1,000s. Thus, I think it's necessary to \"intern\" the repo mappings by having one array containing the deduplicated mappings and a map that references them by ID, e.g.:\r\n```json\r\n{\r\n \"mappings\": [{\"foo\": \"foo~1.2.3, ...}, {\"bar\": \"bar~4.5.6\", ...}, ...],\r\n \"repoToMappingIndex\": {\"foo~1.2.3\": 0, ...}\r\n}\r\n```\r\n\r\nHappy to support anyone who wants to pick this up!", "CC @mai93 as this format will also be needed to make IntelliJ's \"Go to Definition\" features work for labels with Bzlmod.", "Looks like this one is done! \\o/", "A fix for this issue has been included in [Bazel 7.1.0 RC1](https://releases.bazel.build/7.1.0/rc1/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[]

[ "type: feature request", "P1", "team-ExternalDeps" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/modcommand/ModOptions.java", "src/main/java/com/google/devtools/build/lib/bazel/commands/BUILD", "src/main/java/com/google/devtools/build/lib/bazel/commands/ModCommand.java", "src/main/java/com/google/devtools/build/lib/bazel/commands/mod.txt" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/modcommand/ModOptions.java", "src/main/java/com/google/devtools/build/lib/bazel/commands/BUILD", "src/main/java/com/google/devtools/build/lib/bazel/commands/ModCommand.java", "src/main/java/com/google/devtools/build/lib/bazel/commands/mod.txt" ]

[ "src/test/py/bazel/bzlmod/mod_command_test.py" ]

[ "keyword_pr_to_issue" ]

[ "I simply didn't consider this case when I submitted https://github.com/bazelbuild/bazel/commit/ef3f05852b3bf542903f091184da96b96354a13d. Happy to (pre-)review a PR.", "A fix for this issue has been included in [Bazel 7.1.0 RC1](https://releases.bazel.build/7.1.0/rc1/index.html). Please test out the release candidate and report any issues as soon as possible. Thanks!" ]

[ "This should not be changed." ]

[ "type: bug", "P3", "team-Rules-CPP" ]

[ "MODULE.bazel.lock", "tools/cpp/unix_cc_configure.bzl" ]

[ "MODULE.bazel.lock", "tools/cpp/unix_cc_configure.bzl" ]

[ "src/test/tools/bzlmod/MODULE.bazel.lock" ]

[ "timestamp(timedelta=66336.0, similarity=0.8547909199217864)" ]

[ "Thanks, should all be fixed.\n" ]

[]

[]

[ "README.md" ]

[ "README.md" ]

[]

[ "timestamp(timedelta=407.0, similarity=0.9395480498605061)" ]

[ "Nice catch. Can you submit your patch with a github pull request?\n", "OK, I sent a pull request. I haven't done that before, so let me know if anything is fishy.\n", "https://github.com/google/bazel/commit/e93638335955ff8d7e948f2d385319e458ee0ca4\n" ]

[]

[]

[ "compile.sh" ]

[ "compile.sh" ]

[]

[ "timestamp(timedelta=87.0, similarity=0.8616929189821002)" ]

[ "As mentioned on https://groups.google.com/d/msg/bazel-discuss/25X8NKHv4H8/pS2so6AjJAAJ, downloading the javadoc jar would also be nice.\n", "+1\n", "+1\n", "+1\n", "+1\n", "+1\n", "Why is this P4 priority issue? Quoting Gerrit Code Review priority descriptions: [1], that apparently common for all Google projects:\n\n```\nPriority-4\n\nPonies and icebox.\nUnfortunate: it's a legitimate issue, but the team never plans to fix this.\n```\n\nAnd even better question: what is the appropriate workaround for this missing feature? What we are surely not going to do is to spam `WORKSPACE` file with two rule invocations for each and every artifact:\n\n```\nmaven_jar(\n name = 'javax_validation',\n artifact = 'javax.validation:validation-api:1.0.0.GA',\n sha1 = 'b6bd7f9d78f6fdaa3c37dae18a4bd298915f328e',\n)\n\nhttp_jar(\n name = \"javax_validation_src\",\n url = \"http://repo1.maven.org/maven2/javax/validation/validation-api/1.0.0.GA/validation-api-1.0.0.GA-sources.jar\",\n)\n```\n\nInstead, Skylark `maven_jar2` rule that is doing something similar to this `maven_jar` Bucklet rule would be nice: [2].\n- [1] https://www.gerritcodereview.com/issues.md#Priority_4\n- [2] https://github.com/davido/bucklets/blob/master/maven_jar.bucklet#L60-L73\n", "P4 is almost that yes. For us it mostly means \"It makes sense, we have no time to work on it but we would welcome any contribution to make it happens\"\n\nWhich is weird because it is P4 and assigned. Anyway, defering to \n@kchodorow to reasssess its priority and ping @jin for the reimplementation of maven_jar in skylark.\n", "Probably much easier to add once #1410 is in.\n", "7a7c41d adds this capability! I'm closing this for now but definitely reopen if there are any lingering qs (:" ]

[]

[ "type: feature request" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/DecompressorDescriptor.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/FileDecompressor.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/JarDecompressor.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/MavenDownloader.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/MavenJarFunction.java", "src/main/java/com/google/devtools/build/lib/bazel/rules/workspace/MavenJarRule.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/DecompressorDescriptor.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/FileDecompressor.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/JarDecompressor.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/MavenDownloader.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/MavenJarFunction.java", "src/main/java/com/google/devtools/build/lib/bazel/rules/workspace/MavenJarRule.java" ]

[ "src/test/java/com/google/devtools/build/lib/rules/repository/JarDecompressorTest.java" ]

[ "timestamp(timedelta=158.0, similarity=0.918334548218563)" ]

[ "I'm sorry to dig up this issue from the history but I'm trying to introduce [jvm_flags](https://bazel.build/versions/master/docs/be/java.html#java_binary.jvm_flags) to rules_scala and I could use a bit of understanding about the motivation of this issue.\r\nThis change added the ability to configure the `scalac` options **and** JVM flags of `scalac`. I'm not sure why one would need configuring the JVM flags of `scalac`. I've looked hard but can't find a corollary option in the Java Rules.\r\nAFAIU `jvm_flags` in the [java rules](https://bazel.build/versions/master/docs/be/java.html#java_binary) only apply to the binaries and are flags for the executable. `java_library` doesn't support that and only supports `javacopts`. \r\nI can see 3 options:\r\n1. This behavior exists in the JavaRules under a different name which I did not find.\r\n2. It doesn't exist in JavaRules and it was an honest mistake.\r\n3. It doesn't exist in JavaRules but has a valid use-case (which I'd like then to understand to document it).\r\n\r\nAny inputs would be appreciated.\r\ncc @johnynek @damienmg @ulfjack for visibility and maybe additional knowledge", "I think it is irrelevant in java because the standard compiler is not\nwritten in Java, is it? Unlike scala where the we run the scala compiler on\na JVM, and sometimes we may need to tune memory issues (for instance we\nhave had issues with thrift targets with tons of files which are rarely\ncompiled).\n\nI think this should be rarely needed, but that and just plain carelessness\nresulted in the state we have today I think.\n\nTo do it again, I think we would want scalac_jvm_arg or something to\nsignify options to scalac. We may want to set some default ones with a\nscala_toolchain rule if we can make one.\nOn Sat, Apr 29, 2017 at 19:35 Ittai Zeidman <[email protected]>\nwrote:\n\n> I'm sorry to dig up this issue from the history but I'm trying to\n> introduce jvm_flags\n> <https://bazel.build/versions/master/docs/be/java.html#java_binary.jvm_flags>\n> to rules_scala and I could use a bit of understanding about the motivation\n> of this issue.\n> This change added the ability to configure the scalac options *and* JVM\n> flags of scalac. I'm not sure why one would need configuring the JVM\n> flags of scalac. I've looked hard but can't find a corollary option in\n> the Java Rules.\n> AFAIU jvm_flags in the java rules\n> <https://bazel.build/versions/master/docs/be/java.html#java_binary> only\n> apply to the binaries and are flags for the executable. java_library\n> doesn't support that and only supports javacopts.\n> I can see 3 options:\n>\n> 1. This behavior exists in the JavaRules under a different name which\n> I did not find.\n> 2. It doesn't exist in JavaRules and it was an honest mistake.\n> 3. It doesn't exist in JavaRules but has a valid use-case (which I'd\n> like then to understand to document it).\n>\n> Any inputs would be appreciated.\n> cc @johnynek <https://github.com/johnynek> @damienmg\n> <https://github.com/damienmg> @ulfjack <https://github.com/ulfjack> for\n> visibility and maybe additional knowledge\n>\n> —\n> You are receiving this because you were mentioned.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/bazelbuild/bazel/issues/457#issuecomment-298212635>,\n> or mute the thread\n> <https://github.com/notifications/unsubscribe-auth/AAEJdiiVHPW-egWJMzREXPxChxhH7nUrks5r1B2OgaJpZM4F-Hje>\n> .\n>\n", "ok fair enough. I see your point. In that case I'll continue like we discussed with scalac_jvm_flags to keep with the current convention of jvm_flags but to prefix it with scalac to signal it's flags for it.", "@johnynek why do we pass then the jvm_flags to javac? We use `-J`/`--jvm_flag` to pass it to javac itself. So we currently overload the same attribute with 3 responsibilities:\r\n1. Scalac jvm flags\r\n2. Javac jvm flags\r\n3. Target exectubale jvm flags\r\nClearly 3 doesn't belong with 1 and 2. Should 1 and 2 be the same attribute? I think they should but a second opinion would be great here." ]

[]

[]

[ "tools/build_defs/scala/scala.bzl" ]

[ "tools/build_defs/scala/scala.bzl" ]

[]

[ "timestamp(timedelta=2377.0, similarity=0.8508197549457626)" ]

[ "you could create a scala_test that expands into java_test that imports the jar generated by scala_binary?\n", "Added this issue as b/25141720.\n" ]

[ "are these needed? It would be nice to default to discovering all the suites in the jar and running them. That seems like the preferred way to go. If you don't want to run a suite, it should be in its own target, and just not test that target.\n\nOr at least, that's my view.\n", "I think it is better to download rather than add the jar to the git repo. This can be done by adding to the WORKSPACE.\n", "same comment about it being better to not check in the jar.\n", "Sounds good. Added a download instead.\n", "Added a download here as well :)\n", "I agree that it would be nice. As discussed on the email thread, the big challenge is that, in Scala, the test suite name cannot easily be inferred from the path (because packages are independent of the file sytem). So I see two options:\n- Add a custom test runner (written in Scala/Java) which wraps the usual scalatest runner and does this discovery at runtime.\n- Specify the suites by hand.\n\nI went for the second option for now because the first one seemed controversial on the thread. Also, supporting the second one seems like a good idea regardless of whether we end up adding the first one in addition.\n\nWhat do you think?\n", "I looked at the dependencies here:\n\nhttp://central.maven.org/maven2/org/scalatest/scalatest_2.11/2.2.4/scalatest_2.11-2.2.4.pom\n\nIt looks like everything other than scala is optional or test. This does mean using generator/scalacheck style won't work here, but I think that's okay for a first draft.\n", "I have this:\nhttps://bazel-review.googlesource.com/#/c/2410/1\n\nwhich pulls scala 2.11 and gets the scalac and scalalib there. It would be nice to use the jar provided there once both this and that are merged.\n", "I didn't find the email thread. Can you link here for completeness?\n", "Certainly, here it is: https://groups.google.com/forum/#!msg/bazel-dev/5Kbbwr11XOA/znTY-H1DBgAJ\n\nIt's also in the top-level description of the change but maybe it's more visible here inline.\n", "Oh, nice! Specifying the scala version to use in inside the workspace rather than a symlink sounds great.\n\n+1 regarding the plan once both are merged.\n", "Yup, I figured if we need some of the optionals in there, we can still add them manually when the need comes up.\n", "@johnynek @dinowernli Added a comment to the thread, which I think basically implements what @johnynek said above. Would love to hear what you think (preferably in the thread?).\n", "Thanks Orr, responded on the main thread. Would it make sense to get this PR in as a \"basic\" way of specifying tests and possibly add more runners in the future?\n", "+1 to something basic that works is better than the current state. We can definitely iterate.\n", "these should no longer be here right?\n", "After thinking more, leave it for now.\n", "revert\n", "revert, our style guide requires 2 empty lines between global def.\n", "oops mixed up with python styleguide, you can leave it as it is.\n", "revert\n", "Ack :)\n", "Don't depends on java being on the PATH, instead add dependency to //tools/defaults:jdk and to @bazel_tools//tools/jdk:java and use the latter as your java binary here.\n" ]

[ "type: feature request", "P3" ]

[ "WORKSPACE", "tools/build_defs/scala/README.md", "tools/build_defs/scala/scala.bzl" ]

[ "WORKSPACE", "tools/build_defs/scala/README.md", "tools/build_defs/scala/scala.bzl" ]

[ "tools/build_defs/scala/test/BUILD", "tools/build_defs/scala/test/HelloLib.scala", "tools/build_defs/scala/test/HelloLibTest.scala" ]

[ "timestamp(timedelta=1.0, similarity=0.8929275300915026)" ]

[ "To sort of reiterate a comment I made on bazel-discuss (possibly redundant, as I'm not sure what \"add versioning\" means), it would be helpful if:\r\n\r\n- The docs prominently display the Bazel version to which they apply.\r\n- Docs for multiple versions, including `master` are readily available (e.g. see Python docs, Qt docs, ...)\r\n- The default version displayed is *the latest stable release*.\r\n- The site sets a cookie so that the last viewed version is displayed when the user comes back.\r\n\r\nAs a stop-gap, it would be nice if the main documentation was for the latest release, with the \"bleeding edge\" documentation published somewhere less obvious or less \"official\"." ]

[ "What are these hashes?\r\nHow do we update them?", "why do we need `exit` at the end?\r\n\r\nIf it's because of the trap, can you instead put the `mv` here?\r\nOr put the `sed` only when the condition is true?", "Add comment that it must be in sync with the config file?", "good idea. fixed.", "added instructions on what they are and how to get them." ]

[ "type: documentation (cleanup)", "P2", "team-Bazel" ]

[ "WORKSPACE", "site/BUILD", "site/_config.yml", "site/_layouts/documentation.html", "site/_sass/sidebar.scss", "site/jekyll-tree.sh" ]

[ "WORKSPACE", "scripts/docs/doc_versions.bzl", "scripts/docs/generate_versioned_docs.sh", "site/BUILD", "site/_config.yml", "site/_layouts/documentation.html", "site/_sass/sidebar.scss", "site/jekyll-tree.sh" ]

[]

[ "timestamp(timedelta=35552.0, similarity=0.8496915557676555)" ]

[ "Quick update on the situation: After cold starting Bazel multiple times (due to rebooting the OS) it seems the problem has disappeared entirely. However, let me leave the bug open for now.\n", "Hit ctrl-\\ while bazel is \"stuck\" and you'll get a thread dump.\nPosting that here may help us diagnose the problem. I'll update the\nuser manual to mention this feature.\n\nOn Jan 3, 2016 4:52 AM, \"Michael Nett\" [email protected] wrote:\n\n> Quick update on the situation: After cold starting Bazel multiple times (due to rebooting the OS) it seems the problem has disappeared entirely. However, let me leave the bug open for now.\n> \n> —\n> Reply to this email directly or view it on GitHub.\n", "Thanks Janak, that's really helpful. Since the problem has gone away, shall we mark this as not reproducible for the time being?\n", "@janakdr Could you please close this bug once the documentation has been updated?\n@mnett Sounds good - please reopen if you reproduce it again in the future!\n", "Hi,\n\nthe problem cropped up again and this time I got a the thread dump after getting stuck mid-analysis\n\n``` bash\nINFO: Loading complete. Analyzing...\n^\\\nSending SIGQUIT to JVM process 3802 (see /home/mnett/.cache/bazel/_bazel_mnett/630833098f940540d7dc43bb06a2df6a/server/jvm.out).\n```\n\nsee https://gist.github.com/mnett/e71b2a9e5a36382194b0\n\n@janakdr can you reopen the issue?\n", "Looks like you had two stack traces there. First one looks like it was\nduring loading. The only work being done is:\n\n\"skyframe-evaluator 120\" #221 prio=5 os_prio=0 tid=0x00007f0b1802a000\nnid=0xfbd runnable [0x00007f0ad1ad8000]\njava.lang.Thread.State: RUNNABLE\nat java.net.SocketInputStream.socketRead0(Native Method)\nat java.net.SocketInputStream.socketRead(SocketInputStream.java:116)\nat java.net.SocketInputStream.read(SocketInputStream.java:170)\nat java.net.SocketInputStream.read(SocketInputStream.java:141)\nat sun.security.ssl.InputRecord.readFully(InputRecord.java:465)\nat sun.security.ssl.InputRecord.readV3Record(InputRecord.java:593)\nat sun.security.ssl.InputRecord.read(InputRecord.java:532)\nat sun.security.ssl.SSLSocketImpl.readRecord(SSLSocketImpl.java:973)\n- locked <0x00000000e67fbf60> (a java.lang.Object)\n at sun.security.ssl.SSLSocketImpl.readDataRecord(SSLSocketImpl.java:930)\n at sun.security.ssl.AppInputStream.read(AppInputStream.java:105)\n- locked <0x00000000e68aa508> (a sun.security.ssl.AppInputStream)\n at java.io.BufferedInputStream.fill(BufferedInputStream.java:246)\n at java.io.BufferedInputStream.read1(BufferedInputStream.java:286)\n at java.io.BufferedInputStream.read(BufferedInputStream.java:345)\n- locked <0x00000000e68aa520> (a java.io.BufferedInputStream)\n at sun.net.www.http.ChunkedInputStream.fastRead(ChunkedInputStream.java:244)\n at sun.net.www.http.ChunkedInputStream.read(ChunkedInputStream.java:689)\n- locked <0x00000000e68aa548> (a sun.net.www.http.ChunkedInputStream)\n at java.io.FilterInputStream.read(FilterInputStream.java:133)\n at sun.net.www.protocol.http.HttpURLConnection$HttpInputStream.read(HttpURLConnection.java:3336)\n at org.eclipse.jgit.util.io.UnionInputStream.read(UnionInputStream.java:145)\n at org.eclipse.jgit.transport.SideBandInputStream.read(SideBandInputStream.java:143)\n at org.eclipse.jgit.transport.PackParser.fill(PackParser.java:1131)\n at org.eclipse.jgit.transport.PackParser.access$000(PackParser.java:97)\n at org.eclipse.jgit.transport.PackParser$InflaterStream.read(PackParser.java:1664)\n at java.io.InputStream.read(InputStream.java:101)\n at org.eclipse.jgit.transport.PackParser.whole(PackParser.java:983)\n at org.eclipse.jgit.transport.PackParser.indexOneObject(PackParser.java:916)\n at org.eclipse.jgit.transport.PackParser.parse(PackParser.java:487)\n at org.eclipse.jgit.internal.storage.file.ObjectDirectoryPackParser.parse(ObjectDirectoryPackParser.java:194)\n at org.eclipse.jgit.transport.PackParser.parse(PackParser.java:448)\n at org.eclipse.jgit.transport.BasePackFetchConnection.receivePack(BasePackFetchConnection.java:762)\n at org.eclipse.jgit.transport.BasePackFetchConnection.doFetch(BasePackFetchConnection.java:363)\n at org.eclipse.jgit.transport.TransportHttp$SmartHttpFetchConnection.doFetch(TransportHttp.java:779)\n at org.eclipse.jgit.transport.BasePackFetchConnection.fetch(BasePackFetchConnection.java:301)\n at org.eclipse.jgit.transport.BasePackFetchConnection.fetch(BasePackFetchConnection.java:291)\n at org.eclipse.jgit.transport.FetchProcess.fetchObjects(FetchProcess.java:245)\n at org.eclipse.jgit.transport.FetchProcess.executeImp(FetchProcess.java:161)\n at org.eclipse.jgit.transport.FetchProcess.execute(FetchProcess.java:122)\n at org.eclipse.jgit.transport.Transport.fetch(Transport.java:1138)\n at org.eclipse.jgit.api.FetchCommand.call(FetchCommand.java:130)\n at org.eclipse.jgit.api.CloneCommand.fetch(CloneCommand.java:193)\n at org.eclipse.jgit.api.CloneCommand.call(CloneCommand.java:133)\n at com.google.devtools.build.lib.bazel.repository.GitCloneFunction.compute(GitCloneFunction.java:126)\n at com.google.devtools.build.skyframe.ParallelEvaluator$Evaluate.run(ParallelEvaluator.java:938)\n\nSo, waiting for an http download, I guess. The second dump is during\nanalysis, and all the threads appear to be doing useful work. Was the\nlong delay before Bazel printed the \"Analyzing\" message, or after?\n" ]

[ "Why this change? The include path is relative to the root of the workspace.\n", "Have you tested this? The `:hello-world` target is in the `//examples/cpp` package. There is no `:hello-world` target in the root BUILD file.\n\nIf you want to build `:hello-world` from inside the `examples/cpp` directory, then the correct command would be `bazel build :hello-world`, and Bazel would look for a target called `:hello-world` in the BUILD file in the current directory. Labels beginning with `//` are absolute paths relative to the workspace root.\n" ]

[]

[ "examples/cpp/README.md", "examples/cpp/hello-lib.cc", "examples/cpp/hello-world.cc" ]

[ "examples/cpp/README.md", "examples/cpp/hello-lib.cc", "examples/cpp/hello-world.cc" ]

[]

[ "timestamp(timedelta=109140.0, similarity=0.8482036932530586)" ]

[ "Thanks for the report! We recently moved all of these rules to their own repos (e.g., https://github.com/bazelbuild/rule_appengine), and I think we're still updating how their documentation is generated.\n" ]

[]

[ "type: documentation (cleanup)" ]

[ "src/main/java/com/google/devtools/build/docgen/templates/be/be-nav.vm" ]

[ "src/main/java/com/google/devtools/build/docgen/templates/be/be-nav.vm" ]

[]

[ "timestamp(timedelta=112365.0, similarity=0.8625023770678207)" ]

[ "@aehlig this is done, right?", "It was addressed in https://github.com/bazelbuild/bazel/pull/4368 in the skylark rules only (not the built-in)." ]

[ "I know you didn't start it, but I would have preferred email adresses in the rfc2606-reserved domains for fictional characters. (This also applies to other parts of that change.)\r\n ", "I shall enter a new PR to fix those (since it will be a big diff but mostly noise). Thanks!" ]

[ "type: feature request", "P4" ]

[ "tools/build_defs/repo/git.bzl" ]

[ "tools/build_defs/repo/git.bzl" ]

[ "src/test/shell/bazel/skylark_git_repository_test.sh", "src/test/shell/bazel/testdata/pluto.git_log", "src/test/shell/bazel/testdata/refetch.git_log" ]

[ "timestamp(timedelta=444.0, similarity=0.8789615018699354)" ]

[ "The default value of `use_testrunner` is 1 and default behavior is to use `BazelTestRunner`, if a user needs to use this property, he would set `use_testrunner = 0` to use `GoogleTestRunner`.\n\nMaybe we should make it more clear. ;)\n", "Sorry, what I said above was totally wrong. The GoogleTestRunner doesn't exist in Bazel code. Setting `use_testrunner = 0` will just run the test java binary without any test runner. We should definitely update the document!\n", "PR?\nOn יום ג׳, 7 ביוני 2016 at 17:01 Yun Peng [email protected] wrote:\n\n> Sorry, what I said above was totally wrong. The GoogleTestRunner doesn't\n> exist in Bazel code. Setting use_testrunner = 0 will just run the test\n> java binary without any test runner. We should definitely update the\n> document!\n> \n> —\n> You are receiving this because you authored the thread.\n> Reply to this email directly, view it on GitHub\n> https://github.com/bazelbuild/bazel/issues/1345#issuecomment-224289479,\n> or mute the thread\n> https://github.com/notifications/unsubscribe/ABUIF5huEY8z07V-SIbEoCjaBjXtbjjqks5qJXm2gaJpZM4Iv4Fh\n> .\n", "Yeah, that would be nice. :)\n" ]

[]

[ "type: documentation (cleanup)", "P3" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/rules/java/BazelJavaRuleClasses.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/rules/java/BazelJavaRuleClasses.java" ]

[]

[ "timestamp(timedelta=2132.0, similarity=0.8414718166154534)" ]

[ "Isn't this fixed already?", "Andreas's fix is still pending: https://github.com/bazelbuild/bazel/pull/1524", "Is merged.", "Thank you!" ]

[]

[ "type: feature request" ]

[ "src/main/java/com/google/devtools/build/lib/syntax/BinaryOperatorExpression.java", "src/main/java/com/google/devtools/build/lib/syntax/SkylarkList.java" ]

[ "src/main/java/com/google/devtools/build/lib/syntax/BinaryOperatorExpression.java", "src/main/java/com/google/devtools/build/lib/syntax/SkylarkList.java" ]

[ "src/test/java/com/google/devtools/build/lib/syntax/EvaluationTest.java" ]

[ "timestamp(timedelta=0.0, similarity=0.9419610559888507)" ]

[]

[ "replace to www.bazel.io?\n" ]

[ "type: documentation (cleanup)", "P2" ]

[ ".gitignore", "site/_includes/head.html" ]

[ ".gitignore", "site/_includes/head.html" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.8777530327961319)" ]

[]

[]

[]

[ "src/main/java/com/google/devtools/build/lib/skyframe/PackageFunction.java" ]

[ "src/main/java/com/google/devtools/build/lib/skyframe/PackageFunction.java" ]

[ "src/test/shell/integration/BUILD", "src/test/shell/integration/prelude_test.sh" ]

[ "timestamp(timedelta=0.0, similarity=0.8691097587201603)" ]

[ "@petemounce @laszlocsomor \n", "Woah woah woah, regex find and replace in powershell? What have I let myself in for. Are different drives really necessary? :p\n\nYeah, sure. I'll get that done.\n", "Help us Obi-Wan, you're the only PS expert! :)\n", "Done.\n" ]

[ "I think you don't need to match $2 if you're outputting it unaltered.\n", "That's true. I like this pattern more since it's more explicit, but I don't feel strongly.\n", "ok\n" ]

[ "P3", "platform: windows" ]

[ "scripts/packages/chocolatey/tools/chocolateyinstall.ps1" ]

[ "scripts/packages/chocolatey/tools/chocolateyinstall.ps1" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.852035181509092)" ]

[ "this (in combination with #1752) sounds valuable since it lowers the entry costs of an organization. People can check out and best practices are laid out in the shared bazelrc ", "+1" ]

[ "Please add an `OS.getCurrent()` check around this `if`, and only replace \"~\" on Linux/macOS.", "here too", "(`OS` in the same Java package)", "AFAIK, `user.home` should be resolved just fine on Windows and recent JDK versions: [1]. On Windows, on Java 8, we have:\r\n\r\n```java\r\npublic class Test {\r\n public static void main(String[] args) {\r\n String path = \"~/foo/\";\r\n path = path.replace(\"~\", System.getProperty(\"user.home\"));\r\n System.out.println(path);\r\n }\r\n}\r\n```\r\n\r\nRunning it produces:\r\n\r\n```shell\r\nC:\\>java Test\r\nC:\\Users\\davido/foo/\r\n```\r\n\r\nFTR: I backported this solution from that Buck feature: [Add support for user home in cache directory name](https://github.com/facebook/buck/commit/a1ba00118666c39f87354e3c0f2226e6f87c6970) that I added to Buck 4 years ago. And of course, Buck is platform independent.\r\n\r\n[1] https://bugs.java.com/view_bug.do?bug_id=4787931\r\n[2] https://stackoverflow.com/questions/2134338/java-user-home-is-being-set-to-userprofile-and-not-being-resolved\r\n", "Right, but `~abc` is a valid directory name on Windows. So I'm reluctant to expand \"~\" to user.home on Windows too.", "We need a working and platform independent solution to configure local action path per default and commit it in (D)VCS. I am open to any suggestions, but we just can't solve the problem at hand for Linux/macOS, and ignore Windows users.\r\n\r\nShould we compare for \"~/\" instead on all platforms, a lá:\r\n\r\n```java\r\n if (path.startsWith(\"~/\")) {\r\n [...]\r\n```\r\n?", "Yes please, that'd be better.", "Done." ]

[ "type: feature request", "P3", "category: misc > misc" ]

[ "src/main/java/com/google/devtools/build/lib/util/OptionsUtils.java" ]

[ "src/main/java/com/google/devtools/build/lib/util/OptionsUtils.java" ]

[ "src/test/java/com/google/devtools/build/lib/util/OptionsUtilsTest.java" ]

[ "timestamp(timedelta=166854.0, similarity=0.8431258311203514)" ]

[ "Hi Shahms,\r\n\r\nThank you for the report, I'll try to improve the clang support in cc_configure.bzl when I'll have spare cycles. If in the meantime you feel like it, feel free to tackle it yourself, cc_configure.bzl is not too complicated. If not, no worries, I will get to it eventually :)\r\nCheers", "Piggybacking on this, since it's the same issue, just on different OS: builds on macOS fail if GCC is installed in the system, because Bazel prefers it over Clang (even though `/usr/bin/cc` still points to `clang`), which leads to a case when Clang-only flags are passed to GCC, because `if darwin` check in `cc_configure.bzl` should really be a `if clang` check.", "Hi Piotr,\r\n\r\nbazel doesn't look at /usr/bin/cc, but what you could do is to set CC env var to clang and blaze will take it. Altough it is probably reasonable to autodetect gcc on macOS, cc_configure.bzl can only get you so far. At some point you will have to provide your own CROSSTOOL file with all the specific requirements of your environment. Luckily, we are making CROSSTOOL more expressive and nicer to work with almost every day now :)", "Marcel, but my problem is not with (lack of) auto-detection of GCC on macOS, it's with the fact that Bazel prefers GCC over Clang (i.e. behavior changes without any changes to Bazel usage, depending on whether or not GCC is available in the `$PATH`), which results in failed builds, because the checks in `cc_configre.bzl` are platform-specific and not compiler-specific.", "I see, that is indeed incorrect. I'll take a look once I get rid of some higher priority issues in my queue. Thanks Piotr!", "Created a minimal patch adding `-pass-exit-codes` through the `_add_option_if_supported` guard.\r\nhttps://bazel-review.googlesource.com/#/c/8070/", "Thanks Markus so much! Importing.", "This looks fixed to me, please shout if it's not true." ]

[]

[ "type: feature request", "P3", "team-Rules-CPP" ]

[ "tools/cpp/windows_cc_configure.bzl" ]

[ "tools/cpp/windows_cc_configure.bzl" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.8828373293694733)" ]

[]

[]

[ "type: documentation (cleanup)" ]

[ "site/_layouts/contribute.html", "site/index.html", "site/users.md" ]

[ "site/_layouts/contribute.html", "site/index.html" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.912943864111191)" ]

[]

[ "s/install/installed/\r\n\r\nAlso, `bazel version` instead of `$ bazel` (don't include the $ for inline snippets).", "s/install/installed/\r\n\r\nAlso, `bazel version` instead of `$ bazel` (don't include the $ for inline snippets).", "This seems like useful info to have.", "fixed", "fixed\r\n", "fixed" ]

[ "type: documentation (cleanup)", "P2" ]

[ "site/versions/master/docs/install.md" ]

[ "site/versions/master/docs/install.md" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.9664825861331298)" ]

[ "@philwo can you take a look at this?", "I would like to take this issue up.. ", "@philwo, @iirina any progress on this yet?", "@aljabrr I don't think anyone of us is working on it. Please feel free to send a patch via Gerrit. :)\r\n\r\nThe help string for that option is here:\r\n\r\nhttps://github.com/bazelbuild/bazel/blob/94fbad5b230ae0c02a88100f1978d40d17d57b71/src/main/java/com/google/devtools/build/lib/bazel/rules/BazelRulesModule.java#L51\r\n\r\nIt is used to automatically build the documentation on the website, so just changing that string to a better explanation should be enough." ]

[ "Not only for tests, I would change this to:\r\n\r\n'standalone' means run all of them locally without any kind of sandboxing", "This is also wrong by now, it should read:\r\n\r\n'sandboxed' means to run them in a sandboxed environment with limited privileges (details depend on platform support)", "Thanks for the review. Updated.", "Thanks for the review. Updated" ]

[ "type: feature request", "type: documentation (cleanup)", "P4", "category: sandboxing", "good first issue" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/rules/BazelStrategyModule.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/rules/BazelStrategyModule.java" ]

[]

[ "timestamp(timedelta=1.0, similarity=0.9331296354628537)" ]

[ "You mean that having a ~/.bazelrc is no longer mandatory so the installer\nshouldn't try to create one?\nOn Sat, 6 May 2017 at 10:29 Philipp Wollermann <[email protected]>\nwrote:\n\n> The Bazel installer currently moves ~/.bazelrc to ~/.bazelrc.bak and\n> creates an empty file instead, which is rather annoying. We should change\n> the installer to not touch the file as that mechanism is no longer needed.\n>\n> —\n> You are receiving this because you are subscribed to this thread.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/bazelbuild/bazel/issues/2958>, or mute the thread\n> <https://github.com/notifications/unsubscribe-auth/ABUIF0cdPZ5QOWcjcBL7paMGWpth2imdks5r3CFmgaJpZM4NSqOC>\n> .\n>\n", "@ittaiz I don't think a ~/.bazelrc was ever mandatory, Bazel should work just fine without any configuration files like /etc/bazelrc or ~/.bazelrc. :)\r\n\r\nThe installer currently always wants to create an empty ~/.bazelrc file. If the user has already created a ~/.bazelrc file (and possibly put some config flags there), it will be moved to ~/.bazelrc.bak, which is annoying, because your next step is probably to just move it back to ~/.bazelrc.\r\n\r\nInstead the installer should just not touch the file at all - if it doesn't exist already, don't create one. If one exists, leave it as is.", "On May 6, 2017 4:37 PM, \"Philipp Wollermann\" <[email protected]>\nwrote:\n\n@ittaiz <https://github.com/ittaiz> I don't think a ~/.bazelrc was ever\nmandatory, Bazel should work just fine without any configuration files like\n/etc/bazelrc or ~/.bazelrc. :)\n\nThe installer currently always wants to create an empty ~/.bazelrc file. If\nthe user has already created one (and possibly put some config flags\nthere), it will be moved to ~/.bazelrc.bak, which is annoying, because you\nbasically have to immediately move it back.\n\nInstead it should just not touch the file at all - if it doesn't exist\nalready, don't create one. If one exists, leave it as is.\n\n+1\n\n\n—\nYou are receiving this because you are subscribed to this thread.\nReply to this email directly, view it on GitHub\n<https://github.com/bazelbuild/bazel/issues/2958#issuecomment-299664397>,\nor mute the thread\n<https://github.com/notifications/unsubscribe-auth/AAAIwlomGv_P6P-fhkfrcd_9sfkesGVYks5r3NoKgaJpZM4NSqOC>\n.\n", "Ok. I just don't understand what you meant by \"that mechanism is no longer\nneeded\" but I guess it's not that important.\nOn Sun, 7 May 2017 at 0:04 doug tangren <[email protected]> wrote:\n\n> On May 6, 2017 4:37 PM, \"Philipp Wollermann\" <[email protected]>\n> wrote:\n>\n> @ittaiz <https://github.com/ittaiz> I don't think a ~/.bazelrc was ever\n> mandatory, Bazel should work just fine without any configuration files like\n> /etc/bazelrc or ~/.bazelrc. :)\n>\n> The installer currently always wants to create an empty ~/.bazelrc file. If\n> the user has already created one (and possibly put some config flags\n> there), it will be moved to ~/.bazelrc.bak, which is annoying, because you\n> basically have to immediately move it back.\n>\n> Instead it should just not touch the file at all - if it doesn't exist\n> already, don't create one. If one exists, leave it as is.\n>\n> +1\n>\n>\n> —\n> You are receiving this because you are subscribed to this thread.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/bazelbuild/bazel/issues/2958#issuecomment-299664397>,\n> or mute the thread\n> <\n> https://github.com/notifications/unsubscribe-auth/AAAIwlomGv_P6P-fhkfrcd_9sfkesGVYks5r3NoKgaJpZM4NSqOC\n> >\n> .\n>\n> —\n> You are receiving this because you were mentioned.\n>\n>\n> Reply to this email directly, view it on GitHub\n> <https://github.com/bazelbuild/bazel/issues/2958#issuecomment-299665809>,\n> or mute the thread\n> <https://github.com/notifications/unsubscribe-auth/ABUIF5bMaAvWmwGSuY78O4Vytc1IxsiZks5r3OBGgaJpZM4NSqOC>\n> .\n>\n", "@ittaiz No problem and sorry for the misunderstanding. With \"that mechanism\", I just meant \"the part of the installer that deals with the ~/.bazelrc\", but re-reading my issue description I can see now how that might be confusing. I removed the part of the sentence." ]

[]

[ "P2" ]

[ "scripts/packages/template_bin.sh" ]

[ "scripts/packages/template_bin.sh" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.8717960914532032)" ]

[ "any updates on this?", "No updates yet -- we'll be chatting about this over the next few business days to prioritize.", "Is there a hint so that we could fix it and perhaps submit a PR ?", "We're not likely to get to this ourselves; we deemed this low priority internally.\r\nIf you'd like to go ahead and create a PR, you'll need to trace through ObjcLibrary.java and understand where the module name gets set; you can then hook through an attribute value to optionally change the default. \r\n\r\n(I don't know off the top of my head what decides the default module name in terms of the end command line.. You might start there? You can build an example with `-s` to see the command lines invoked in your build)", "Makes total sense that's it's low priority.\r\n\r\nThat's exactly what I needed, we'll see how bad we want it and if so we'll submit a PR. Thanks!" ]

[]

[ "under investigation", "category: rules > ObjC / iOS / J2ObjC" ]

[ "src/main/java/com/google/devtools/build/lib/rules/cpp/CppModuleMapAction.java", "src/main/java/com/google/devtools/build/lib/rules/objc/CompilationSupport.java", "src/main/java/com/google/devtools/build/lib/rules/objc/IntermediateArtifacts.java", "src/main/java/com/google/devtools/build/lib/rules/objc/ObjcRuleClasses.java" ]

[ "src/main/java/com/google/devtools/build/lib/rules/cpp/CppModuleMapAction.java", "src/main/java/com/google/devtools/build/lib/rules/objc/CompilationSupport.java", "src/main/java/com/google/devtools/build/lib/rules/objc/IntermediateArtifacts.java", "src/main/java/com/google/devtools/build/lib/rules/objc/ObjcRuleClasses.java" ]

[ "src/test/java/com/google/devtools/build/lib/rules/objc/ObjcLibraryTest.java" ]

[ "timestamp(timedelta=0.0, similarity=0.894002445849351)" ]

[ "I'm actually not sure why it doesn't pick this up automatically, but it seems to only be looking for crosstools in `@local_config_cc`, which doesn't contain a ppc crosstool (?). You have to give the `--crosstool_top=//tools/cpp:default-toolchain` argument to make it look at //tools/cpp, which _does_ contain a ppc toolchain.\r\n\r\nHowever, I'm not actually sure how to give extra arguments to the ./compile.sh script, so assigning to @damienmg for clarification on:\r\n\r\n1. How to pass `--crosstool_top` into `./compile.sh`.\r\n2. Why `@local_config_cc` doesn't contain the same toolchain as `@bazel_tools`.\r\n3. Why they aren't both used to find toolchains.", "I solved it making two small modifications:\r\n\r\n1- added \r\n default_toolchain {\r\n \t cpu: \"ppc64\"\r\n \t toolchain_identifier: \"local_linux\"\r\n\t}\r\n\r\n to tools/cpp/CROSSTOOL\r\n\r\n2- in the file \"tools/cpp/cc_configure.bzl\" changed \r\n\r\n \"if result.stdout.strip() in [\"power\", \"ppc64le\", \"ppc\"]:\"\r\n\r\n into\r\n\r\n \"if result.stdout.strip() in [\"power\", \"ppc64le\", \"ppc\", \"ppc64\"]:\"\r\n\r\nI am not sure if both the modifications are actually necessary, but that's what I did and I successfully built bazel", "@davideleoni90: would you be willing to send a PR with those modification? They looks reasonable to me.", "(re-opened the bug because the code is not fixed)" ]

[ "nit: add missing space after the comma." ]

[ "type: bug", "P2", "platform: other" ]

[ "tools/cpp/CROSSTOOL", "tools/cpp/cc_configure.bzl" ]

[ "tools/cpp/CROSSTOOL", "tools/cpp/cc_configure.bzl" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.9105872551600119)" ]

[ "Thanks for the bug report! Unfortunately we have no capacity to fix it now, so I'm marking it P4 to indicate that we need external help to fix it. Would you be interested in fixing it?" ]

[ "Why not use Locale.US here. That seems to more clearly document the purpose. Also, might want to add a comment on why we provide a locale here.", "Maybe the test should reset the locale before it returns?" ]

[ "type: feature request", "P4", "category: misc > misc", "good first issue" ]

[ "src/main/java/com/google/devtools/build/lib/profiler/chart/HtmlChartVisitor.java" ]

[ "src/main/java/com/google/devtools/build/lib/profiler/chart/HtmlChartVisitor.java" ]

[ "src/test/java/com/google/devtools/build/lib/profiler/ProfilerChartTest.java" ]

[ "timestamp(timedelta=0.0, similarity=0.8405917286274984)" ]

[ "You're right, that's a holdover from the internal docs, and should be removed. Would you like to send a patch to remove it?" ]

[]

[ "type: documentation (cleanup)", "P2" ]

[ "site/docs/bazel-user-manual.html" ]

[ "site/docs/bazel-user-manual.html" ]

[]

[ "timestamp(timedelta=2790.0, similarity=0.8870933720382193)" ]

[ "SHA1 is a bit outdated now, as is MD5. We should probably add support for - at least - SHA256.", "A good option might be `blake2`, which is secure and fast in software.", "SHA256 isn't exactly slow either and recent CPUs might even already have native instructions: https://en.wikipedia.org/wiki/Intel_SHA_extensions" ]

[ "You probably meant SHA256 here", "Fixed!" ]

[ "type: feature request", "P2" ]

[ "src/main/java/com/google/devtools/build/lib/remote/README.md", "src/main/java/com/google/devtools/build/lib/vfs/FileSystem.java", "src/tools/remote_worker/BUILD", "src/tools/remote_worker/README.md" ]

[ "src/main/java/com/google/devtools/build/lib/remote/README.md", "src/main/java/com/google/devtools/build/lib/vfs/FileSystem.java", "src/tools/remote_worker/BUILD", "src/tools/remote_worker/README.md" ]

[ "src/test/shell/bazel/remote_execution_sandboxing_test.sh", "src/test/shell/bazel/remote_execution_test.sh" ]

[ "timestamp(timedelta=0.0, similarity=0.9268525109128359)" ]

[ "```\r\n$ bazel analyze-profile profile\r\njava.lang.NullPointerException\r\n\tat com.google.devtools.build.lib.runtime.commands.ProfileCommand.exec(ProfileCommand.java:251)\r\n\tat com.google.devtools.build.lib.runtime.BlazeCommandDispatcher.execExclusively(BlazeCommandDispatcher.java:552)\r\n\tat com.google.devtools.build.lib.runtime.BlazeCommandDispatcher.exec(BlazeCommandDispatcher.java:331)\r\n\tat com.google.devtools.build.lib.runtime.BlazeRuntime.batchMain(BlazeRuntime.java:760)\r\n\tat com.google.devtools.build.lib.runtime.BlazeRuntime.main(BlazeRuntime.java:552)\r\n\tat com.google.devtools.build.lib.bazel.BazelMain.main(BazelMain.java:57)\r\njava.lang.NullPointerException\r\n\tat com.google.devtools.build.lib.runtime.commands.ProfileCommand.exec(ProfileCommand.java:251)\r\n\tat com.google.devtools.build.lib.runtime.BlazeCommandDispatcher.execExclusively(BlazeCommandDispatcher.java:552)\r\n\tat com.google.devtools.build.lib.runtime.BlazeCommandDispatcher.exec(BlazeCommandDispatcher.java:331)\r\n\tat com.google.devtools.build.lib.runtime.BlazeRuntime.batchMain(BlazeRuntime.java:760)\r\n\tat com.google.devtools.build.lib.runtime.BlazeRuntime.main(BlazeRuntime.java:552)\r\n\tat com.google.devtools.build.lib.bazel.BazelMain.main(BazelMain.java:57)\r\n```" ]

[ "There's no need to call new String() here. Just pass in \"<workspace>\"." ]

[ "type: bug", "P1", "category: misc > misc" ]

[ "src/main/java/com/google/devtools/build/lib/runtime/commands/ProfileCommand.java" ]

[ "src/main/java/com/google/devtools/build/lib/runtime/commands/ProfileCommand.java" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.8864842583364211)" ]

[ "+1!\nOn Wed, 28 Jun 2017 at 18:21 Kristina <[email protected]> wrote:\n\n> Right now we print something like:\n>\n> ** Please add the following dependencies:\n> //:C to //:A\n>\n> Now that buildozer is open sourced, it would be nice to give people a\n> one-liner to run.\n>\n> —\n> You are receiving this because you are subscribed to this thread.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/bazelbuild/bazel/issues/3285>, or mute the thread\n> <https://github.com/notifications/unsubscribe-auth/ABUIFxn0FL_wRhVUmDkFCrhFcJOiYPOdks5sIm9-gaJpZM4OIH7W>\n> .\n>\n", "For future reference:\r\nThis seems to be the test: https://github.com/bazelbuild/bazel/blob/66437a0173488ca4e77aab8c41d5454871c8c6a2/src/test/shell/bazel/local_repository_test_jdk8.sh#L98\r\nThis seems to be the prod code:\r\nhttps://github.com/bazelbuild/bazel/blob/66437a0173488ca4e77aab8c41d5454871c8c6a2/src/java_tools/buildjar/java/com/google/devtools/build/buildjar/javac/plugins/dependency/DependencyModule.java#L337" ]

[]

[ "type: feature request", "P3" ]

[ "src/java_tools/buildjar/java/com/google/devtools/build/buildjar/javac/plugins/dependency/DependencyModule.java" ]

[ "src/java_tools/buildjar/java/com/google/devtools/build/buildjar/javac/plugins/dependency/DependencyModule.java" ]

[]

[ "timestamp(timedelta=169720.0, similarity=0.8471509188515058)" ]

[ "The install docs are now versioned, and I also don't see hardcoded Bazel version numbers in the installation instructions anymore.\r\n\r\nhttps://docs.bazel.build/versions/0.20.0/install-windows.html\r\nhttps://docs.bazel.build/versions/0.20.0/install-ubuntu.html" ]

[]

[ "type: feature request", "type: documentation (cleanup)" ]

[ "scripts/docs/doc_versions.bzl", "site/_config.yml" ]

[ "scripts/docs/doc_versions.bzl", "site/_config.yml" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.8480277078258937)" ]

[ "I remember @ianthehat asked for this." ]

[ "What's the use case for defaults? Are they required for a constraint_setting or are they optional?\r\n\r\nFor example, if have `constraint_setting(name = \"macos_version\", default_constraint_value = \":osx10_11_6\")` and we build for platform `\"osx10_11_6\"`, does that risk pulling in toolchains that don't even support OSX because they automatically take the default?\r\n\r\nMore generally, is there risk in a toolchain \"indirectly\" advertising support for a constraint the toolchain doesn't even know exists?", "Defaults are optional.\r\n\r\nDuring toolchain resolution, only the constraint settings that are present in either the platform or the toolchain are considered. Therefore, if the platform does not declare a constraint value for \"macos_version\", and the toolchain doesn't declare a constraint value for \"macos_version\", the setting isn't considered. If either of them does use that setting, the other would be considered to have the default value, which the constraint author should have chosen as a safe value.\r\n\r\nThis feature is intended less for thing like \"macos_version\", and more for tunable constraints such as \"compilation_mode\" or \"enable_race_detection\", to be used as toolchain modes during resolution.", "Does \"enable_race_detection\" really need a default, or can its absence in a toolchain definition alone imply lack of support? Are there any precise motivating examples?\r\n\r\nSorry, I'm not trying to be difficult about this. There's a very real argument for defaults under circumstances where the alternative is having to, say, add a new mode that just one toolchain variant uses, then being forced to explicitly tag all other variants with the \"old\" mode setting. That can get unwieldy.\r\n\r\nBut I've also learned to be more cautious when using defaults. They provide behavior at a distance that can be hard to isolate and in a worst case toolchains, etc. could get selected on properties the toolchain designers didn't even consider.\r\n\r\nWe've seen this tug and pull with existing compatible_with / restricted_to deployments.", "In addition from the uses rules_go wants, there are some uses that RBE wants as well. In both cases, the initial advice given was \"use the absence of the value as the default\", and that works okay, but in both cases users end up tagging several platform and toolchain definitions with comments like \"constraint race_detection disabled by default\", where it would be easier to just add the constraint value.\r\n\r\nAlso, having an explicit constraint value for the default makes it easier to switch the default later." ]

[ "type: feature request", "P2" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/platform/ConstraintCollection.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/platform/ConstraintCollection.java" ]

[ "src/test/java/com/google/devtools/build/lib/rules/platform/ConstraintCollectionApiTest.java", "src/test/java/com/google/devtools/build/lib/rules/platform/PlatformInfoApiTest.java", "src/test/shell/bazel/toolchain_test.sh" ]

[ "timestamp(timedelta=1.0, similarity=0.9352260033689938)" ]

[]

[]

[ "type: documentation (cleanup)", "P2", "team-Bazel" ]

[ "site/_sass/style.scss", "src/main/java/com/google/devtools/common/options/OptionsUsage.java" ]

[ "site/_sass/style.scss", "src/main/java/com/google/devtools/common/options/OptionsUsage.java" ]

[ "src/test/java/com/google/devtools/common/options/OptionsUsageTest.java" ]

[ "timestamp(timedelta=0.0, similarity=0.9092550907745788)" ]

[]

[]

[ "type: feature request", "P3" ]

[ "src/main/java/com/google/devtools/build/lib/rules/cpp/CcCommon.java" ]

[ "src/main/java/com/google/devtools/build/lib/rules/cpp/CcCommon.java" ]

[ "src/test/java/com/google/devtools/build/lib/rules/cpp/CcCommonTest.java" ]

[ "timestamp(timedelta=1.0, similarity=0.8960968417948798)" ]

[ "PR: https://github.com/bazelbuild/bazel/pull/7207" ]

[]

[ "type: feature request", "P2" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/debug/WorkspaceRuleEvent.java", "src/main/java/com/google/devtools/build/lib/bazel/debug/workspace_log.proto", "src/main/java/com/google/devtools/build/lib/bazel/repository/skylark/SkylarkRepositoryContext.java", "src/main/java/com/google/devtools/build/lib/skylarkbuildapi/repository/SkylarkRepositoryContextApi.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/debug/WorkspaceRuleEvent.java", "src/main/java/com/google/devtools/build/lib/bazel/debug/workspace_log.proto", "src/main/java/com/google/devtools/build/lib/bazel/repository/skylark/SkylarkRepositoryContext.java", "src/main/java/com/google/devtools/build/lib/skylarkbuildapi/repository/SkylarkRepositoryContextApi.java" ]

[ "src/test/shell/bazel/bazel_workspaces_test.sh" ]

[ "timestamp(timedelta=66037.0, similarity=1.0)" ]

[ "`bazel info release` suffices for this, recommend it be closed.", "@werkt Sure, but it doesn't solve the problem that it's non-intuitive to people who are so used to the GNU convention of -v and --version, which in my experience is by far the most common way of allowing people to query version information in other command-line tools.\r\n\r\nSo I vote to keep this open until https://github.com/bazelbuild/bazel/pull/3671 is closed.", "I added a --gnu_format flag to version in 938a53d3d54ede8132109df3c492c47afc63f5a6.", "However, changing the default is tricky and will require more work.", "`--version` works as of 371a2e343199ee799404fd2bb8319d990a2ad572, and without starting a server! :tada:" ]

[]

[ "type: feature request", "P4", "good first issue", "team-OSS" ]

[ "src/main/cpp/blaze_util.cc", "src/main/cpp/option_processor.cc", "src/main/java/com/google/devtools/build/lib/runtime/BlazeCommandDispatcher.java" ]

[ "src/main/cpp/blaze_util.cc", "src/main/cpp/option_processor.cc", "src/main/java/com/google/devtools/build/lib/runtime/BlazeCommandDispatcher.java" ]

[ "src/test/cpp/option_processor_test.cc" ]

[ "timestamp(timedelta=0.0, similarity=0.9011354485736779)" ]

[ "Yes, I am sure it can, will looking into this." ]

[ "We don't need --cpu=x64_windows_msvc anymore, please remove this." ]

[ "type: bug", "P3", "platform: windows" ]

[ "tools/cpp/CROSSTOOL.tpl", "tools/cpp/unix_cc_configure.bzl", "tools/cpp/windows_cc_configure.bzl" ]

[ "tools/cpp/CROSSTOOL.tpl", "tools/cpp/unix_cc_configure.bzl", "tools/cpp/windows_cc_configure.bzl" ]

[ "src/test/py/bazel/bazel_windows_test.py" ]

[ "timestamp(timedelta=0.0, similarity=0.8991682871191983)" ]

[ "Sounds like a good idea! @philwo , @aehlig : WDYT? ", "@jmillikin-stripe Considering that java8-jdk-headless doesn't exist, should we then change the dependency list from:\r\n\r\n> google-jdk | java8-jdk | java8-sdk | oracle-java8-installer\r\n\r\ninto:\r\n\r\n> google-jdk | java8-sdk-headless | oracle-java8-installer\r\n\r\n?\r\n\r\nThe java8-sdk-headless package seems to exist and do the job for Debian and Ubuntu:\r\nhttps://packages.debian.org/sid/java8-sdk-headless\r\nhttps://packages.ubuntu.com/zesty/java8-sdk-headless\r\n\r\nNot sure why we even had the java8-jdk one in there.", "`java8-jdk` exists in Debian: https://packages.debian.org/sid/java8-jdk\r\n\r\nI'm slightly nervous about changing/removing virtual packages, because they can be mapped to different things on different systems. I think a safe, minimally-intrusive change would be to change the dep to:\r\n\r\n`google-jdk | java8-sdk-headless | java8-jdk | java8-sdk | oracle-java8-installer`\r\n\r\nThat should prioritize the headless version if available, without breaking any existing dependency resolution on Dpkg-based non-Debian systems.", "Gentle ping.\r\n\r\nI could send a PR for this if you want, but someone in the Bazel side would still need to handle updating the Debian packages and such.", "Note for future readers: the Debian package deps are duplicated, and need to be updated in both `debian/control` and `debian/BUILD`. See https://github.com/bazelbuild/bazel/pull/4686." ]

[]

[ "type: feature request", "P3", "category: misc > misc" ]

[ "scripts/packages/debian/control" ]

[ "scripts/packages/debian/control" ]

[]

[ "timestamp(timedelta=1.0, similarity=0.8473022587967347)" ]

[ "This sounds related to (though definitely not a duplicate of) #3438", "Has there been any progress on this, or prioritization? It's worse than the issue description when you consider windows compatibility, as \"cat\" isn't an external executable on that platform (it being a powershell alias to Get-Content, which is also not an external executable file. So for windows, one has to implement platform-specific toolchains to supply a call to powershell with that command, all to read text file contents with a repository_ctx. ", "PR: https://github.com/bazelbuild/bazel/pull/7309", "What is the motivation for reading the file into Starlark rule?", "Getting metadata in that you can use to generate the repository more effectively. In my case, for example, getting teh appropriate maven metadata (pom files) to ensure that we generate dependencies properly in the generated bazel targets. Right now I'm using \"cat\" which doesn't work on windows. I don't want to have to set up a per-platform toolchain just to read a file. ", "I've been reading files in repository rules for two reasons:\r\n\r\n1. Sometimes the choice of which artifacts to download is driven by a file from the main workspace. For example, Python or Node.JS projects save their download URLs to \"lock files\". The only way to access the content of a file from the main workspace is via the symlink hack described in the first post.\r\n\r\n2. Some external dependencies require running a setup tool, which outputs files with non-deterministic paths. I want to inspect the log file generated by those tools to discover the paths, so they can be inserted into the generated `BUILD` file.", "+1, would be nice to have an OS-independent `cat` like feature in a repository rule. ", "This came up in a discussion with @dkelmer earlier on again. Currently, to do cross-platform `cat` in a repository rule, we'll need to do this:\r\n\r\n```python\r\ndef _cat_file(repository_ctx, filepath):\r\n # For Windows, use cat from msys.\r\n cat = \"C:\\\\msys64\\\\usr\\\\bin\\\\cat\" if (_is_windows(repository_ctx)) else repository_ctx.which(\"cat\")\r\n exec_result = repository_ctx.execute([cat, repository_ctx.path(filepath)])\r\n if (exec_result.return_code != 0):\r\n fail(\"Error while trying to read %s: %s\" % (filepath, exec_result.stderr))\r\n return exec_result.stdout\r\n```\r\n\r\nIt'll be nice to shorten this to `content = repository_ctx.read(file)`.", "I mean, it's that or add toolchains. :) \r\n" ]

[ "We want 8859-1 as the default? ", "The default is set to UTF-8 in `SkylarkRepositoryContextApi`. I'm using `StandardCharsets.ISO_8859_1` to mean raw bytes, for when the user calls `ctx.read(\"somefile\", encoding=\"\")`.\r\n\r\nMy assumption is that this function will almost always be used to ingest text files in ASCII / UTF-8, with `encoding = \"\"` available as an escape hatch for odd cases.", "Oh, I see! Ok. Gotcha. ", "Starlark strings are raw bytes (https://github.com/bazelbuild/starlark/blob/master/spec.md#strings). Therefore this method should not have any encoding passed in.", "`repository_ctx.file()` assumes strings are UTF-8:\r\n```java\r\npublic void createFile(Object path, String content, Boolean executable, Location location) {\r\n ...\r\n try (OutputStream stream = p.getPath().getOutputStream()) {\r\n stream.write(content.getBytes(StandardCharsets.UTF_8));\r\n }\r\n```\r\n\r\nIf the encoding does not default to UTF-8, then `ctx.file(\"foo\", s); ctx.read(\"foo\") == s` will be false.\r\n\r\nIf we do not allow the user to specify no encoding, then it won't be possible to properly read a binary file.\r\n\r\nIf you're certain that `ctx.read()` should not have an encoding parameter, should `ctx.file()` also be changed to treat its input as ISO-8859-1 (i.e. plain bytes)?", "> repository_ctx.file() assumes strings are UTF-8:\r\n\r\nThat's a bug in `ctx.file()`\r\n\r\n> If we do not allow the user to specify no encoding, then it won't be possible to properly read a binary file.\r\n\r\nWe should read in ISO-8859-1 - then we get raw bytes.\r\n\r\n> If you're certain that ctx.read() should not have an encoding parameter, should ctx.file() also be changed to treat its input as ISO-8859-1 (i.e. plain bytes)?\r\n\r\nYes (this is an incompatible change, so the way to do it is to add `legacy_utf8` boolean to `ctx.file`, defaulting to `true` and go through incompatible change process to make the default `false`.", "Or rename `ctx.file()` to `ctx.write()` to be symmetrical with `read()` and make these paired functions behave in the new way, and deprecate `file()`\r\n\r\nStill requires the incompatible change process to deal with the deprecation, but it's probably more instructive and better named.", "OK, changes made in new commit:\r\n* Remove `encoding=`, and reads be decoded as ISO-8859-1.\r\n* Add `ctx.write_file()`, which is like `ctx.file()` but encodes with ISO-8859-1.\r\n * I also had to use a new `workspace_log.proto` message type, because the old `FileEvent` has its contents in `string` instead of `bytes`.\r\n* Renamed new function to `ctx.read_file()` to match.", "Let's keep the `ctx.file()` and add `legacy_utf8` as I suggested. Why? Migration will be no-op for most (I expect 100% of) people, where with function rename people will have to modify their code.\r\n\r\nLet's keep the name `ctx.read()` the function you are adding.\r\n\r\nI am not happy about the new event in BEP. Add bytes to `FileEvent` instead.", "@dslomov Above changes applied.", "Removing a field is a breaking change to the protocol.\r\nInstead, let's have *two* fields for now: `string content` and `bytes bytes`, deprecating the first one. As part of an incompatible change to make legecy_utf8 default to false, we can set content to empty string and then remove that field.", "Add a test with non-Latin-1 UTF-8 characters. Add a test for legacy_utf8=False", "Done and done.", "After further thought, I've reverted this message to only containing `string content`. The protobuf field has always been set from the `String content` directly, without encoding, so it doesn't make sense for it to change as part of this PR.\r\n\r\nIt might be reasonable to add a new field `bytes encoded_content` that differs depending on `legacy_utf8` mode, but I think that is best done separately.", "This really should be a bool or enum instead of a string. If you use a bool, the call sites should specify the param name in a comment, as in\r\n\r\n```\r\ncheckInOutputDirectory(/*isRead=*/ true, ...)\r\n```", "What are we testing? Should this be named `test_read_file_nonascii`?", "These call sites can use some param name comments.", "I think a `bool` would be inappropriate here, because there's no reason it wouldn't use more specific error messages for other checks in the future. I only copy-pasted `\"write\"` into existing callers to avoid significant diffs in unrelated tests.\r\n\r\nI'm not familiar enough with Java to know when an enum, set of constants, or inline literals would be more idiomatic. Feel free to append commits to this PR for any stylistic changes (I care only about user-visible behavior).", "The pre-existing test `test_file()` verifies the behavior of `repository_ctx.file()`. This test verifies the behavior of `repository_ctx.file()` when the file content contains codepoints outside the ASCII range.", "Sure. That sounds like a reasonable cleanup that could be applied to this test file in a separate PR, or internal CL.", "I wasn't suggesting changing the error message for the existing case, I just wanted to avoid passing strings if it's only ever allowed to take on two possible values (e.g. to avoid subtle misspellings leading to errors). That said I see that this is only used for the error message, not to control the actual operation being done, so it's fine to leave it a string.", "Oh, fair enough. That was my lack of domain knowledge in this area showing through.", "Granted they should've been there in the pre-existing code. I mention it because you're creating an additional parameter which makes the readability slightly worse, but I won't block this PR on it." ]

[ "type: feature request", "P3", "team-ExternalDeps" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/debug/WorkspaceRuleEvent.java", "src/main/java/com/google/devtools/build/lib/bazel/debug/workspace_log.proto", "src/main/java/com/google/devtools/build/lib/bazel/repository/skylark/SkylarkRepositoryContext.java", "src/main/java/com/google/devtools/build/lib/skylarkbuildapi/repository/SkylarkRepositoryContextApi.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/debug/WorkspaceRuleEvent.java", "src/main/java/com/google/devtools/build/lib/bazel/debug/workspace_log.proto", "src/main/java/com/google/devtools/build/lib/bazel/repository/skylark/SkylarkRepositoryContext.java", "src/main/java/com/google/devtools/build/lib/skylarkbuildapi/repository/SkylarkRepositoryContextApi.java" ]

[ "src/test/java/com/google/devtools/build/lib/bazel/repository/skylark/SkylarkRepositoryContextTest.java", "src/test/shell/bazel/bazel_workspaces_test.sh" ]

[ "timestamp(timedelta=0.0, similarity=0.8750544054369024)" ]

[]

[ "add \"python_version\": \"3\"", "add \"python_version\": \"2\"", "I recommend removing the non-linux configuration so you get a less noisy result. E.g. freebsd is a breakage at HEAD.", "Added", "Added", "Done", "I'm not sure what indentation it wants here." ]

[ "type: feature request", "P3", "category: misc > testing" ]

[ "scripts/ci/bazel-tests.json" ]

[ "scripts/ci/bazel-tests.json" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.8606838243996491)" ]

[ "Hi John, internal reviewer asked for tests, could you add some to src/test/shell/bazel/apple/bazel_objc_test.sh please?", "Test added, to `cpp_darwin_integration_test.sh` -- this isn't Objective-C, it's just regular C++.\r\n\r\nI'm some trouble running the test (with changes or at HEAD) because Bazel gets into an infinite loop and consumes all my disk. I'll keep trying." ]

[ "Ah I missed this one, can you pls use ifMac there?", "Can you pls move the fail onto the next line.", "`ifMac` is the opposite of what we want. These lines should only be added on non-MacOS platforms.", "This matches the existing style. Note that I think there might be something wrong with the current test code -- https://github.com/bazelbuild/bazel/issues/4085 -- would you like me to update both as part of this change?", "Then use ifLinux pls, windows crosstool is not patched by this code.", "I believe this should also be applied to FreeBSD.", "Does it? Line 80 in this file is the desired formatting. Let me investigate #4085 first, we run this test on the CI and I don't think we have this issue there. I'll take a look.", "Hmm you know what, just fix it pls, the fail shouldn't be there twice even if it worked. I didn't know assert_build also fails.", "That is covered by 'linux' cppPlatform :) https://github.com/bazelbuild/bazel/blob/275ae45b1228bdd0f912c4fbd634b29ba4180383/src/main/java/com/google/devtools/build/lib/rules/cpp/CppToolchainInfo.java#L353 I know, this is not very robust :) ", "OK, changed to use `ifLinux`. I also added a comment that FreeBSD counts as `LINUX`.", "Done." ]

[ "type: bug", "P2" ]

[ "src/main/java/com/google/devtools/build/lib/rules/cpp/CppActionConfigs.java", "tools/osx/crosstool/CROSSTOOL.tpl" ]

[ "src/main/java/com/google/devtools/build/lib/rules/cpp/CppActionConfigs.java", "tools/osx/crosstool/CROSSTOOL.tpl" ]

[ "src/test/java/com/google/devtools/build/lib/packages/util/MOCK_OSX_CROSSTOOL", "src/test/shell/bazel/cpp_darwin_integration_test.sh" ]

[ "timestamp(timedelta=80465.0, similarity=0.8798468627057965)" ]

[ "Is this still an issue that needs to be corrected?\r\n", "If this uses the “zipper” target underneath then yes. There’s also a\nseparate issue about zipper itself.\n\nOn Sat, 11 Aug 2018 at 5:13 Josh Fischer <[email protected]> wrote:\n\n> Is this still an issue that needs to be corrected?\n>\n> —\n> You are receiving this because you are subscribed to this thread.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/bazelbuild/bazel/issues/4063#issuecomment-412244199>,\n> or mute the thread\n> <https://github.com/notifications/unsubscribe-auth/ABUIF-vVfPFNJfVw0Cx6OG52Fi-BFDaCks5uPj3HgaJpZM4QYdqk>\n> .\n>\n", "I run into the same problem. I did check the `pkg_tar` rule and I see the it uses [`build_tar.py`](https://github.com/bazelbuild/bazel/blob/master/tools/build_defs/pkg/build_tar.py) script underneath. As @sghiaus mentioned, the script uses `=` as delimiter which conflicts with the file name.", "it's interesting to see that the rule/script uses `=` as delimiter for [files](https://github.com/bazelbuild/bazel/blame/master/tools/build_defs/pkg/pkg.bzl#L47), [modes](https://github.com/bazelbuild/bazel/blame/master/tools/build_defs/pkg/pkg.bzl#L57), [owners](https://github.com/bazelbuild/bazel/blame/master/tools/build_defs/pkg/pkg.bzl#L59), and [owner names](https://github.com/bazelbuild/bazel/blame/master/tools/build_defs/pkg/pkg.bzl#L62), but uses colon `:` for [symlinks](https://github.com/bazelbuild/bazel/blame/master/tools/build_defs/pkg/pkg.bzl#L76). \r\n" ]

[]

[ "type: bug", "P2", "good first issue" ]

[ "tools/build_defs/pkg/BUILD", "tools/build_defs/pkg/build_tar.py", "tools/build_defs/pkg/build_test.sh", "tools/build_defs/pkg/pkg.bzl" ]

[ "tools/build_defs/pkg/BUILD", "tools/build_defs/pkg/build_tar.py", "tools/build_defs/pkg/build_test.sh", "tools/build_defs/pkg/pkg.bzl" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.8616565838617605)" ]

[ "I think that #3855 should resolve this issue, but I'll test it to confirm.", "Hey thanks for bringing this to our attention, ccing some people that can help\r\n@jin @dkelmer @ahumesky ", "Thanks @aj-michael - I was able to reproduce @angersson's errors and can confirm that #3855 resolves the first one.\r\n\r\nHowever, the second error persists. It can probably be resolved by updating modules in `tools/android` and `tools/build_defs` for compatibility with python3-style strings. Should be relatively simple, so I can probably make those changes soon.", "As promised, I've created a branch containing [python 3 compatibility updates](https://github.com/bazelbuild/bazel/compare/master...akira-baruah:python3-tools) to bazel tools.\r\n\r\n@angersson @aj-michael @jin @dkelmer @ahumesky you can test the fix by running:\r\n```\r\ngit clone https://github.com/akira-baruah/bazel.git /tmp/bazel-akira\r\ncd /tmp/bazel-akira\r\ngit checkout python3-tools # branch with bugfixes\r\nbazel build //src:bazel\r\n\r\ncd <tensorflow repo>\r\n/tmp/bazel-akira/bazel-bin/src/bazel build //tensorflow/contrib/lite/java/demo/app/src/main:TfLiteCameraDemo --cxxopt='--std=c++11' --force_python=py3 --python_path=$(which python3) --config=android_arm --logging=0\r\n```\r\nIf it looks good to you, I'll make a PR. Please note that this branch depends on PR #3855, so we would probably want to have that merged first.\r\n\r\nThanks!", "I was able to replicate the fixes, but I'll need to fetch NDK 15 to fully test building the app in a clean environment due to #4068.", "I verified build success with nvidia-docker, `gcr.io/tensorflow/tensorflow:latest-devel-gpu`, and this... awful script, run inside the container:\r\n\r\n```shell\r\n#!/usr/bin/env bash\r\n\r\nset -euxo pipefail\r\n\r\nTMPDIR=$(mktemp -d)\r\ntrap '' PIPE\r\n\r\ngit clone https://github.com/akira-baruah/bazel.git ${TMPDIR}/bazel-akira\r\ncd ${TMPDIR}/bazel-akira\r\ngit checkout python3-tools # branch with bugfixes\r\nbazel build //src:bazel\r\n\r\ncd ${TMPDIR}\r\n\r\ncurl -o ${TMPDIR}/sdk-tools.zip https://dl.google.com/android/repository/sdk-tools-linux-3859397.zip\r\nunzip ${TMPDIR}/sdk-tools.zip -d ${TMPDIR}/sdk\r\necho \"y\" | ${TMPDIR}/sdk/tools/bin/sdkmanager \"build-tools;26.0.1\" \"platforms;android-26\" \"extras;android;m2repository\"\r\n\r\npip install virtualenv\r\nvirtualenv --python=python3 --system-site-packages ${TMPDIR}/py3\r\nPS1=\"\" # Avoids `set -u`-related error with bin/activate\r\nsource ${TMPDIR}/py3/bin/activate\r\n\r\ngit clone http://github.com/tensorflow/tensorflow ${TMPDIR}/tf\r\ncd ${TMPDIR}/tf\r\n\r\ncurl -o ${TMPDIR}/ndk.zip https://dl.google.com/android/repository/android-ndk-r14b-linux-x86_64.zip\r\nunzip ${TMPDIR}/ndk.zip -d ${TMPDIR}/ndk/\r\n\r\ncat >> ${TMPDIR}/tf/WORKSPACE <<HEREDOC\r\nandroid_sdk_repository(\r\n name = \"androidsdk\",\r\n api_level = 26,\r\n build_tools_version = \"26.0.1\",\r\n path = \"${TMPDIR}/sdk\",\r\n)\r\nandroid_ndk_repository(\r\n name=\"androidndk\",\r\n path=\"${TMPDIR}/ndk/android-ndk-r14b\",\r\n api_level=14)\r\nHEREDOC\r\n\r\nset +o pipefail\r\nyes '' | ${TMPDIR}/tf/configure\r\nset -o pipefail\r\ncd ${TMPDIR}/tf\r\n${TMPDIR}/bazel-akira/bazel-bin/src/bazel build //tensorflow/contrib/lite/java/demo/app/src/main:TfLiteCameraDemo --cxxopt='--std=c++11' --config=android_arm\r\n```\r\n\r\nSweet.", "@angersson now that gflags supports python 3 imports (see commits fb15f0f4e4885243d354a14997d09463594385bf, d926bc40260549b997a6a5a1e82d9e7999dbb65e, and 4d0e6265911199b1376d0f52e249625180a0500d), I've created PR #4265 to resolve this issue." ]

[]

[ "type: feature request" ]

[ "tools/android/aar_embedded_jars_extractor.py", "tools/android/aar_native_libs_zip_creator.py", "tools/android/merge_manifests.py", "tools/android/merge_manifests_test.py", "tools/android/resource_extractor.py", "tools/android/stubify_manifest.py", "tools/build_defs/pkg/archive.py", "tools/objc/protobuf_compiler.py" ]

[ "tools/android/aar_embedded_jars_extractor.py", "tools/android/aar_native_libs_zip_creator.py", "tools/android/merge_manifests.py", "tools/android/merge_manifests_test.py", "tools/android/resource_extractor.py", "tools/android/stubify_manifest.py", "tools/build_defs/pkg/archive.py", "tools/objc/protobuf_compiler.py" ]

[]

[ "timestamp(timedelta=31452.0, similarity=0.9220350633737152)" ]

[ "I think that #3855 should resolve this issue, but I'll test it to confirm.", "Hey thanks for bringing this to our attention, ccing some people that can help\r\n@jin @dkelmer @ahumesky ", "Thanks @aj-michael - I was able to reproduce @angersson's errors and can confirm that #3855 resolves the first one.\r\n\r\nHowever, the second error persists. It can probably be resolved by updating modules in `tools/android` and `tools/build_defs` for compatibility with python3-style strings. Should be relatively simple, so I can probably make those changes soon.", "As promised, I've created a branch containing [python 3 compatibility updates](https://github.com/bazelbuild/bazel/compare/master...akira-baruah:python3-tools) to bazel tools.\r\n\r\n@angersson @aj-michael @jin @dkelmer @ahumesky you can test the fix by running:\r\n```\r\ngit clone https://github.com/akira-baruah/bazel.git /tmp/bazel-akira\r\ncd /tmp/bazel-akira\r\ngit checkout python3-tools # branch with bugfixes\r\nbazel build //src:bazel\r\n\r\ncd <tensorflow repo>\r\n/tmp/bazel-akira/bazel-bin/src/bazel build //tensorflow/contrib/lite/java/demo/app/src/main:TfLiteCameraDemo --cxxopt='--std=c++11' --force_python=py3 --python_path=$(which python3) --config=android_arm --logging=0\r\n```\r\nIf it looks good to you, I'll make a PR. Please note that this branch depends on PR #3855, so we would probably want to have that merged first.\r\n\r\nThanks!", "I was able to replicate the fixes, but I'll need to fetch NDK 15 to fully test building the app in a clean environment due to #4068.", "I verified build success with nvidia-docker, `gcr.io/tensorflow/tensorflow:latest-devel-gpu`, and this... awful script, run inside the container:\r\n\r\n```shell\r\n#!/usr/bin/env bash\r\n\r\nset -euxo pipefail\r\n\r\nTMPDIR=$(mktemp -d)\r\ntrap '' PIPE\r\n\r\ngit clone https://github.com/akira-baruah/bazel.git ${TMPDIR}/bazel-akira\r\ncd ${TMPDIR}/bazel-akira\r\ngit checkout python3-tools # branch with bugfixes\r\nbazel build //src:bazel\r\n\r\ncd ${TMPDIR}\r\n\r\ncurl -o ${TMPDIR}/sdk-tools.zip https://dl.google.com/android/repository/sdk-tools-linux-3859397.zip\r\nunzip ${TMPDIR}/sdk-tools.zip -d ${TMPDIR}/sdk\r\necho \"y\" | ${TMPDIR}/sdk/tools/bin/sdkmanager \"build-tools;26.0.1\" \"platforms;android-26\" \"extras;android;m2repository\"\r\n\r\npip install virtualenv\r\nvirtualenv --python=python3 --system-site-packages ${TMPDIR}/py3\r\nPS1=\"\" # Avoids `set -u`-related error with bin/activate\r\nsource ${TMPDIR}/py3/bin/activate\r\n\r\ngit clone http://github.com/tensorflow/tensorflow ${TMPDIR}/tf\r\ncd ${TMPDIR}/tf\r\n\r\ncurl -o ${TMPDIR}/ndk.zip https://dl.google.com/android/repository/android-ndk-r14b-linux-x86_64.zip\r\nunzip ${TMPDIR}/ndk.zip -d ${TMPDIR}/ndk/\r\n\r\ncat >> ${TMPDIR}/tf/WORKSPACE <<HEREDOC\r\nandroid_sdk_repository(\r\n name = \"androidsdk\",\r\n api_level = 26,\r\n build_tools_version = \"26.0.1\",\r\n path = \"${TMPDIR}/sdk\",\r\n)\r\nandroid_ndk_repository(\r\n name=\"androidndk\",\r\n path=\"${TMPDIR}/ndk/android-ndk-r14b\",\r\n api_level=14)\r\nHEREDOC\r\n\r\nset +o pipefail\r\nyes '' | ${TMPDIR}/tf/configure\r\nset -o pipefail\r\ncd ${TMPDIR}/tf\r\n${TMPDIR}/bazel-akira/bazel-bin/src/bazel build //tensorflow/contrib/lite/java/demo/app/src/main:TfLiteCameraDemo --cxxopt='--std=c++11' --config=android_arm\r\n```\r\n\r\nSweet.", "@angersson now that gflags supports python 3 imports (see commits fb15f0f4e4885243d354a14997d09463594385bf, d926bc40260549b997a6a5a1e82d9e7999dbb65e, and 4d0e6265911199b1376d0f52e249625180a0500d), I've created PR #4265 to resolve this issue." ]

[]

[ "type: feature request" ]

[ "tools/android/aar_embedded_jars_extractor.py", "tools/android/aar_embedded_jars_extractor_test.py", "tools/android/aar_native_libs_zip_creator_test.py", "tools/android/aar_resources_extractor.py", "tools/android/aar_resources_extractor_test.py", "tools/android/build_incremental_dexmanifest.py", "tools/android/incremental_install.py", "tools/android/incremental_install_test.py", "tools/android/merge_manifests.py", "tools/android/resource_extractor_test.py" ]

[ "tools/android/aar_embedded_jars_extractor.py", "tools/android/aar_embedded_jars_extractor_test.py", "tools/android/aar_native_libs_zip_creator_test.py", "tools/android/aar_resources_extractor.py", "tools/android/aar_resources_extractor_test.py", "tools/android/build_incremental_dexmanifest.py", "tools/android/incremental_install.py", "tools/android/incremental_install_test.py", "tools/android/merge_manifests.py", "tools/android/resource_extractor_test.py" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.865193421544384)" ]

[]

[]

[]

[ "site/docs/skylark/aspects.md" ]

[ "site/docs/skylark/aspects.md" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.9192119149424052)" ]

[]

[]

[]

[ "site/docs/bazel-and-cpp.md" ]

[ "site/docs/bazel-and-cpp.md" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.9283720135274562)" ]

[ "Sounds like a nice improvement.\r\n\r\n@akira-baruah do you want to send your fix over as a pull request? :)", "@philwo Thanks - I made PR #4236, but haven't tested it on Windows yet." ]

[]

[ "type: bug", "P3", "category: misc > misc" ]

[ "src/main/java/com/google/devtools/build/lib/runtime/commands/CleanCommand.java" ]

[ "src/main/java/com/google/devtools/build/lib/runtime/commands/CleanCommand.java" ]

[ "src/test/java/com/google/devtools/build/lib/BUILD", "src/test/java/com/google/devtools/build/lib/runtime/commands/CleanCommandTest.java" ]

[ "timestamp(timedelta=0.0, similarity=1.0000000000000002)" ]

[]

[]

[ "type: feature request", "P2" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/skylark/SkylarkRepositoryContext.java", "src/main/java/com/google/devtools/build/lib/skylarkbuildapi/repository/SkylarkRepositoryContextApi.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/skylark/SkylarkRepositoryContext.java", "src/main/java/com/google/devtools/build/lib/skylarkbuildapi/repository/SkylarkRepositoryContextApi.java" ]

[ "src/test/java/com/google/devtools/build/lib/blackbox/tests/workspace/RepoWithRuleWritingTextGenerator.java", "src/test/java/com/google/devtools/build/lib/blackbox/tests/workspace/WorkspaceBlackBoxTest.java" ]

[ "timestamp(timedelta=1.0, similarity=0.8857976795019705)" ]

[ "cc @razvanm", "@brandjon , not sure if anyone owns this area of Bazel; could you maybe take a look?", "I can look in my spare cycles but I won't be able to get to this soon.", "Forwarding to dslomov as requested.", "BazelWorkspaceStatusAction.equals is incorrect - it doesn't check for equality of the clientEnv.", "@ulfjack : thanks for the analysis!\r\n@dslomov : can you fix it in Q1?", "I sent a fix (based on @ulfjack 's suggestion) in #4705" ]

[ "add a non-stable key as well, assert that changing it will not trigger a rebuild", "Please remove the `.sh` ending here and in other `mktemp` calls.\r\nOn macOS `mktemp` requires that the XXXs be at the end of the template, otherwise the command creates a file called literally `wsc-XXXXXXXX.sh`.", "Good point - thank you! Fixed." ]

[ "type: bug", "P3", "category: misc > misc" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/BazelWorkspaceStatusModule.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/BazelWorkspaceStatusModule.java" ]

[ "src/test/shell/bazel/bazel_workspace_status_test.sh" ]

[ "timestamp(timedelta=0.0, similarity=0.8619646931969689)" ]

[ "I know nothing about zsh.\r\n\r\n@schroederc : I believe you added the zsh completion script to Bazel in commit https://github.com/bazelbuild/bazel/commit/6b2766dd0, so I'm hoping you can help.\r\n@damienmg : you approved Cody's change, you may know about zsh.", "also confused by this sentence", "The `_bazel` file just needs to be in one of the directories listed in your `fpath`. The installation example shows how add a directory to your `fpath` and move the `_bazel` file into it, but if you already have a directory in your `fpath` that you prefer, you can move `_bazel` there instead.\r\n\r\nIt doesn't matter what the name of the directory is at all (`$HOME/.zsh/completion/` is just commonly used). The `_bazel` filename only matters because that is what the completion function is named (i.e. https://github.com/bazelbuild/bazel/blob/64d9a4d6dcd720a3b7a60ff550a17a7707dd41d0/scripts/zsh_completion/_bazel#L247). Otherwise, it's only a convention to name the completion file/function after the command it completes with a leading underscore. When `compinit` is run, zsh will look through all files in all of the `fpath` directories and look for `#compdef` lines at the start of each to register completion functions.", "Oh, I see. So the documentation was written from the perspective of that file and was copy/pasted/generated into documentation here: https://docs.bazel.build/versions/master/install.html#zsh\r\n\r\nIt looks like there may be a need to reword it.", "The document disappeared. Is zsh auto completion not supported anymore?", "Looking at the [history of install.md](https://github.com/bazelbuild/bazel/commits/master/site/docs/install.md), the instructions were there at [3e0c545404be6bce908d035044a5ff0ee41cc472](https://github.com/bazelbuild/bazel/blob/3e0c545404be6bce908d035044a5ff0ee41cc472/site/docs/install.md), but not anymore at [70e84f8926bd2cc38fea9e74af292d19038607b8](https://github.com/bazelbuild/bazel/blob/70e84f8926bd2cc38fea9e74af292d19038607b8/site/docs/install.md).\r\n\r\nLet me see if they are still up-to-date and put them back.", "FYI: alternatively, you guys may try this https://github.com/jackwish/bazel which is to make the process easier.", "on macOS the zsh completions are installed automatically if you install via homebrew" ]

[]

[ "type: documentation (cleanup)", "P3", "team-Bazel" ]

[ "site/_layouts/documentation.html", "site/docs/install-os-x.md", "site/docs/install-ubuntu.md" ]

[ "site/_layouts/documentation.html", "site/docs/completion.md", "site/docs/install-os-x.md", "site/docs/install-ubuntu.md" ]

[]

[ "timestamp(timedelta=0.0, similarity=0.9039208917346069)" ]

[ "We ran into the same problem in all our own windows builder machines. Removing the following two directories fixed the problem: C:\\Users\\builder\\\\_bazel_LOCAL\\ SERVICE/install/dc4fe1b7fd90df4d3976b7aaa4b41899 and cygwin /tmp/_bazel_LOCAL\\ SERVICE/install/dc4fe1b7fd90df4d3976b7aaa4b41899. But it is not addressing the reason for the failure.\r\n\r\nPS, the same failure came back. I had to delete the two directories again. The hash in the paths stayed the same. ", "@meteorcloudy : This is Windows only, right? Then it's likely a bug in this class: https://github.com/bazelbuild/bazel/blob/master/src/main/cpp/util/file_windows.cc#L156 . Yesterday was Jan 2, did you see this yesterday too, or only on Jan 1, or with Bazel installations from last year?", "I saw it on Jan 2 and also today Jan 3, and yes, with every bazel installations from last year.", "Yeah the code looks really suspicious. Let me take a closer look.", "@mhlopko To fix the Windows CI build, we'll have to delete the installation directories.", "@laszlocsomor Thanks!", "One workaround is to delete the install directory (dirname of \"A-server.jar\" in the error message).\r\nAnother workaround is to run `touch -m -t 202801010101 $(find path/to/install/dir -type f)` in MSYS shell.", "Another workaround: rename the install directory.\r\nThis is the easiest approach. It's useful if deleting fails due to long file names. (Delete those directories from MSYS or using Total Commander.)", "I hit this bug in my local Windows 10 machine as well on Jan 3. Some Bazel files have the \"01/01/202**8**\" timestamp.\r\n\r\nReinstalling Bazel fixes the issue, but new Bazel files still have the timestamp of 01/01/2028. See screenshots.\r\n\r\n", "> Reinstalling Bazel fixes the issue, but new Bazel files still have the timestamp of 01/01/2028. See screenshots.\r\n\r\nThat is working as intended.\r\n\r\nWhen you install Bazel, Bazel unpacks embedded binaries and sets their timestamp to Jan 1, 10+$current_year. It's 2018 now, so the timestamps are 2028/01/01.", "Can this be closed? Please re-open if still an issue.", "No, this bug is not yet fixed.", "Just ran into this issue here." ]

[]

[ "type: bug", "P1", "platform: windows", "category: misc > misc", "breakage" ]

[ "src/main/cpp/util/file_windows.cc" ]

[ "src/main/cpp/util/file_windows.cc" ]

[]