Datasets:

---

feedback-to-code

/

mozzarella

like 1

Follow

Feedback-2-Code 7

[ "keyword_pr_to_issue" ]

[ "A simpler workaround than setting `linkopts` as a parameter to `cc_binary` is set the `BAZEL_LINKOPTS` environment variable to `-lc++`:\r\n```shell\r\nnoa@Noas-MacBook-Air stage1 % BAZEL_LINKOPTS=-lc++ bazelisk build //main:hello-world\r\nINFO: Analyzed target //main:hello-world (0 packages loaded, 0 targets configured).\r\nINFO: Found 1 target...\r\nTarget //main:hello-world up-to-date:\r\n bazel-bin/main/hello-world\r\nINFO: Elapsed time: 0.078s, Critical Path: 0.00s\r\nINFO: 1 process: 1 internal.\r\nINFO: Build completed successfully, 1 total action\r\n% ./bazel-bin/main/hello-world\r\nHello world\r\nSun Nov 14 20:22:54 2021\r\n```", "The `local_config_cc/BUILD` that gets generated on macOS 12 will use the `unix_cc_toolchain_config.bzl` which will silently drop any `link_libs` being passed to the cc_toolchain_config on systems that has `target_libc == \"macosx\"`.\r\n\r\nAs I don't have access to a macOS 11 Big Sur, it is a bit unclear what has changed, but it seems like a good idea to me to have the `user_link_flags_feature` be added to the features list being declared towards the end of `_impl()` in `unix_cc_toolchain_config.bzl` regardless of whether is_linux (i.e. target_libc != \"macosx\") is true or not. The consequence of user_link_flags_feature being added is that the values in the link_libs parameter list gets added to the command line arguments of the linker, which seems to me like the right thing to do.", "On macOS, Bazel should be using `tools/cpp/osx_cc_configure.bzl` to configure the local C++ toolchain, so the problem is more likely that Bazel is picking the wrong file when building on Monterey.", "Thanks for the fix, we should cherry-pick it for 5.0. \r\nPlease send a PR to branch https://github.com/bazelbuild/bazel/tree/release-5.0.0rc2 after https://github.com/bazelbuild/bazel/pull/14275 is merged. ", "@meteorcloudy Can this get into a 4.2.2 as well? I believe it qualifies as OS support.", "@meteorcloudy I cherry-picked the commit that landed to master and opened a PR to land it in `release-5.0.0rc2` https://github.com/bazelbuild/bazel/pull/14313 Please have a look.", "@brentleyjones I agree that it looks bad that all released bazel versions fails to compile the example cpp hello world in the current stable macOS release that has been out for almost a month.\r\n\r\nOne aspect that makes macOS boxes with just the developer tools so common is that as soon as a developer types for example 'git' in a Terminal window, a popup will suggest installing the developer tools. ", "@nresare can you submit the 4.x PR as well? Not sure if they'll accept it but it should be the exact same as the others", "@nresare Thanks, merged!", "Also cherry-picking this for 4.2.2" ]

[]

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

[ "tools/cpp/unix_cc_toolchain_config.bzl" ]

[ "tools/cpp/unix_cc_toolchain_config.bzl" ]

[]

[ "keyword_pr_to_issue" ]

[ "A simpler workaround than setting `linkopts` as a parameter to `cc_binary` is set the `BAZEL_LINKOPTS` environment variable to `-lc++`:\r\n```shell\r\nnoa@Noas-MacBook-Air stage1 % BAZEL_LINKOPTS=-lc++ bazelisk build //main:hello-world\r\nINFO: Analyzed target //main:hello-world (0 packages loaded, 0 targets configured).\r\nINFO: Found 1 target...\r\nTarget //main:hello-world up-to-date:\r\n bazel-bin/main/hello-world\r\nINFO: Elapsed time: 0.078s, Critical Path: 0.00s\r\nINFO: 1 process: 1 internal.\r\nINFO: Build completed successfully, 1 total action\r\n% ./bazel-bin/main/hello-world\r\nHello world\r\nSun Nov 14 20:22:54 2021\r\n```", "The `local_config_cc/BUILD` that gets generated on macOS 12 will use the `unix_cc_toolchain_config.bzl` which will silently drop any `link_libs` being passed to the cc_toolchain_config on systems that has `target_libc == \"macosx\"`.\r\n\r\nAs I don't have access to a macOS 11 Big Sur, it is a bit unclear what has changed, but it seems like a good idea to me to have the `user_link_flags_feature` be added to the features list being declared towards the end of `_impl()` in `unix_cc_toolchain_config.bzl` regardless of whether is_linux (i.e. target_libc != \"macosx\") is true or not. The consequence of user_link_flags_feature being added is that the values in the link_libs parameter list gets added to the command line arguments of the linker, which seems to me like the right thing to do.", "On macOS, Bazel should be using `tools/cpp/osx_cc_configure.bzl` to configure the local C++ toolchain, so the problem is more likely that Bazel is picking the wrong file when building on Monterey.", "Thanks for the fix, we should cherry-pick it for 5.0. \r\nPlease send a PR to branch https://github.com/bazelbuild/bazel/tree/release-5.0.0rc2 after https://github.com/bazelbuild/bazel/pull/14275 is merged. ", "@meteorcloudy Can this get into a 4.2.2 as well? I believe it qualifies as OS support.", "@meteorcloudy I cherry-picked the commit that landed to master and opened a PR to land it in `release-5.0.0rc2` https://github.com/bazelbuild/bazel/pull/14313 Please have a look.", "@brentleyjones I agree that it looks bad that all released bazel versions fails to compile the example cpp hello world in the current stable macOS release that has been out for almost a month.\r\n\r\nOne aspect that makes macOS boxes with just the developer tools so common is that as soon as a developer types for example 'git' in a Terminal window, a popup will suggest installing the developer tools. ", "@nresare can you submit the 4.x PR as well? Not sure if they'll accept it but it should be the exact same as the others", "@nresare Thanks, merged!", "Also cherry-picking this for 4.2.2" ]

[]

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

[ "tools/cpp/unix_cc_toolchain_config.bzl" ]

[ "tools/cpp/unix_cc_toolchain_config.bzl" ]

[]

[ "keyword_pr_to_issue" ]

[ "This issue also exists in the latest rolling release.", "cc @coeuvre \r\n\r\nAlso seeing this after upgrading to 4.2.1 from 4.0.0. These are the changes that have gone in since the 4.1.0 release:\r\n```\r\n$ git log HEAD ^4.1.0 --oneline src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java\r\nbf5e2b8ec0 Remote: Async upload (Part 8)\r\nb51b31dbe7 Remote: Fix a race when reporting action progresses.\r\n0f812eb5e5 Remote: Display download progress when actions are downloading outputs from remote cache.\r\nf0983dfddb Remote: Report checking cache status before the action is scheduled to run remotely.\r\n```\r\n\r\nIt looks like there was an attempt at fixing the UI bug in `b51b31dbe7`.\r\n\r\nBy adding the below patch you can reproduce this issue:\r\n```\r\nid 'File:[[<execution_root>]bazel-out/k8-opt/.../foo' is already finished, and event Downloading foo, 5.2 KiB / 5.2 KiB is being processed\r\n...\r\n```\r\n\r\nThe problem seems to be like @DavidANeil mentioned -- `actionProgress(ActionProgressEvent event)` is probably called in ThreadA before `actionCompletion(ActionCompletionEvent event)` in ThreadB but the ThreadA goes to sleep and when it revives it calls `getActionState()` which calls `activeActions.computeIfAbsent()` which re-adds the actions that was just removed by ThreadB. The other possibility is events are being fired out-of-order.\r\n\r\nI wonder if the right fix is for `UiEventHandler` to queue and only fire 1 in-progress event per actionId and wait for each event completion before firing the next. Presumably `computeIfAbsent` is being used because of this exact problem -- events can be processed out-of-order so `actionStarted()` cannot be assumed to have already ran & created an action in `activeActions`.\r\n\r\nPatch:\r\n```\r\n$ git diff\r\ndiff --git a/src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java b/src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java\r\nindex eef7ea5c64..99ae5e36d2 100644\r\n--- a/src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java\r\n+++ b/src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java\r\n@@ -361,6 +361,8 @@ final class UiStateTracker {\r\n \r\n private final Map<Artifact, ActionState> activeActions;\r\n \r\n+ private final Map<Artifact, Boolean> finishedActions;\r\n+\r\n // running downloads are identified by the original URL they were trying to access.\r\n private final Deque<String> runningDownloads;\r\n private final Map<String, Long> downloadNanoStartTimes;\r\n@@ -396,6 +398,7 @@ final class UiStateTracker {\r\n \r\n UiStateTracker(Clock clock, int targetWidth) {\r\n this.activeActions = new ConcurrentHashMap<>();\r\n+ this.finishedActions = new ConcurrentHashMap<>();\r\n \r\n this.actionsCompleted = new AtomicInteger();\r\n this.testActions = new ConcurrentHashMap<>();\r\n@@ -582,6 +585,9 @@ final class UiStateTracker {\r\n ActionExecutionMetadata action = event.action();\r\n Artifact actionId = event.action().getPrimaryOutput();\r\n long now = clock.nanoTime();\r\n+ if (finishedActions.containsKey(actionId)) {\r\n+ System.out.println(String.format(\"id '%s' is already finished, and event %s is being processed\", actionId, event.progress()));\r\n+ }\r\n getActionState(action, actionId, now).onProgressEvent(event, now);\r\n }\r\n \r\n@@ -589,6 +595,7 @@ final class UiStateTracker {\r\n Action action = event.getAction();\r\n Artifact actionId = action.getPrimaryOutput();\r\n checkNotNull(activeActions.remove(actionId), \"%s not active after %s\", actionId, event);\r\n+ finishedActions.put(actionId, true);\r\n \r\n // As callers to the experimental state tracker assume we will fully report the new state once\r\n // informed of an action completion, we need to make sure the progress receiver is aware of the\r\n@@ -604,6 +611,7 @@ final class UiStateTracker {\r\n Artifact actionId = action.getPrimaryOutput();\r\n \r\n checkNotNull(activeActions.remove(actionId), \"%s not active after %s\", actionId, event);\r\n+ finishedActions.put(actionId, true);\r\n \r\n if (action.getOwner() != null) {\r\n Label owner = action.getOwner().getLabel();\r\n```", "Also for anyone who OK with compiling bazel from source and applying a patch (patch based off `4.2.1`), this fixes it for me:\r\n```\r\n$ git diff\r\ndiff --git a/src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java b/src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java\r\nindex eef7ea5c64..9e5570e8a7 100644\r\n--- a/src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java\r\n+++ b/src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java\r\n@@ -526,6 +526,11 @@ final class UiStateTracker {\r\n return activeActions.computeIfAbsent(actionId, (key) -> new ActionState(action, nanoTimeNow));\r\n }\r\n \r\n+ @Nullable\r\n+ private ActionState getActionStateIfPresent(Artifact actionId) {\r\n+ return activeActions.get(actionId);\r\n+ }\r\n+\r\n void actionStarted(ActionStartedEvent event) {\r\n Action action = event.getAction();\r\n Artifact actionId = action.getPrimaryOutput();\r\n@@ -582,7 +587,10 @@ final class UiStateTracker {\r\n ActionExecutionMetadata action = event.action();\r\n Artifact actionId = event.action().getPrimaryOutput();\r\n long now = clock.nanoTime();\r\n- getActionState(action, actionId, now).onProgressEvent(event, now);\r\n+ ActionState actionState = getActionStateIfPresent(actionId);\r\n+ if (actionState != null) {\r\n+ actionState.onProgressEvent(event, now);\r\n+ }\r\n }\r\n \r\n void actionCompletion(ActionScanningCompletedEvent event) {\r\n```", "@bjacklyn Would you mind sending your patch as a pull request?" ]

[]

[ "type: bug", "P1", "team-Remote-Exec" ]

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

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

[ "src/test/java/com/google/devtools/build/lib/runtime/UiStateTrackerTest.java" ]

[ "keyword_pr_to_issue" ]

[ "I've found the need for this as well, when trying to target x86_64 macOS RBE from arm64 macOS. Repository rules incorrectly provide arm64 binaries which fail to run on RBE.", "Also see [rules_rust](https://github.com/bazelbuild/rules_rust/blob/14468f97b9a06ab6b33e94bc87c1d24b2433e5f8/crate_universe/private/util.bzl#L18).", "This seems like a reasonable request. It shouldn't be too hard to add a field to `repository_os`, so if anyone wants to send a PR, it would be greatly appreciated.", "I started working on this in #14738, but then noticed that this is more subtle than expected: The natural implementation would be to just return `System.getProperty(\"os.arch\").toLowerCase(Locale.ROOT)`, but that is just the architecture of the JDK Bazel is running on, which is not necessarily the same as the architecture of the operating system Bazel is running on. For example, which value should this have in the case of an x86_64 JDK executing Bazel using Rosetta on an aarch64/arm64 M1 Mac?\r\n\r\n@uhthomas @brentleyjones Could you clarify what exactly you would want to know the architecture of? You also referred to RBE, but I think that repository rules could only ever provide the _host_ CPU, not the _exec_ CPU (which isn't even known during the loading phase since it depends on the result of toolchain resolution). Is that sufficient for your use cases?", "For a M1 mac running through Rosetta x86_64 is right IMO, since the emulation should affect everything. ", "The intent of this issue is to solve the multi cpu/os repository problem where rule maintainers are forced to write odd hacks to detect the exec (host?) CPU. For example the Go rules declare a signal repository, `go_sdk`, which will automatically detect the correct SDK to use. The `go_download_sdk()` rule uses [uname](https://github.com/bazelbuild/rules_go/blob/fd11dd2768669dc2cc1f3a11f2b0b81d84e81c32/go/private/sdk.bzl#L253) to do this detection.\r\n\r\nIt's been suggested that multiple repositories for the corresponding cpu and os should be declared ([ref](https://github.com/bazelbuild/bazel/issues/11655#issuecomment-734212603)), but this can be really messy and weird.\r\n\r\n`select` does not work with repository rules and there is no clear guidance on how to actually do this, which is odd considering this is extremely common.\r\n\r\nThe M1 situation is a bit weird and I don't really know what the right answer is. I think it should depend on whether the architecture of the Bazel binary, as there is no other way to differentiate between native and rosetta at current.\r\n\r\nDoes what I've written make sense, and is it clear what the goal of this issue is?", ">Could you clarify what exactly you would want to know the architecture of? You also referred to RBE, but I think that repository rules could only ever provide the host CPU, not the exec CPU (which isn't even known during the loading phase since it depends on the result of toolchain resolution). Is that sufficient for your use cases?\r\n\r\nI would like the know the architecture of `--host_platform`. If I don't run Bazel through Rosetta, all repository rules will see arm64 on an M1 Mac, even if I set `--host_platform` to a platform with `x86_64` cpu.", "The architecture of the Bazel binary, at least assuming it uses the shipped JDK, the default `cpu` constraint on the auto-detected host platform and the value of the `os.arch` Java system property are all the same as far as I know, so @uhthomas and @keith requests would be handled by my PR.\r\n\r\n@brentleyjones I understand your use case, but this seems difficult to do: Strictly speaking, the platform specified by `--host_platform` is an opaque collection of constraint values, which may theoretically not even contain one for the constraint setting `@platforms//cpu`. And even if we come up with logic for this (admittedly very special) case, I fear getting a constraint value of a platform identified by a label from the context of a repository rule is pretty difficult.", "Awesome @fmeum, your PR looks great. Thanks for taking a look at this! :)\r\n\r\nIs there some reference which lists the possible values of `os.arch` and `os.name`?\r\n\r\nFor my own sanity also, are repository rules executed on the host platform, or the execution platform? More specifically, will the values of `os.arch` and `os.name` be of the host platform, or execution platform?", "Additionally, how does this all fit into platforms? If at all. I think it might be strange to use the Platforms API in most cases, but the repository_os struct in others. What are your thoughts? ", "> Is there some reference which lists the possible values of `os.arch` and `os.name`?\r\n\r\nThe two sources I usually refer to are the OpenJDK sources and Bazel's [CPU](https://github.com/bazelbuild/bazel/blob/master/src/main/java/com/google/devtools/build/lib/util/CPU.java) and [OS](https://github.com/bazelbuild/bazel/blob/master/src/main/java/com/google/devtools/build/lib/util/OS.java) classes.\r\n\r\n> For my own sanity also, are repository rules executed on the host platform, or the execution platform? More specifically, will the values of `os.arch` and `os.name` be of the host platform, or execution platform?\r\n\r\n> Additionally, how does this all fit into platforms? If at all. I think it might be strange to use the Platforms API in most cases, but the repository_os struct in others. What are your thoughts?\r\n\r\nMy understanding is that repository rules are always executed locally (i.e., on the host), even with RBE, and don't really have a concept of platforms. Rather, platforms could be used to select the correct repository among the instances of a repository rule for the different platforms. Bazel would then only fetch the repository rules that are relevant for the current configuration. In this context, `repository_os` is more of a shortcut geared towards local execution, whereas sophisticated RBE setups would probably favor the approach described in https://github.com/bazelbuild/bazel/issues/11655#issuecomment-734212603.\r\n", " I think https://github.com/bazelbuild/bazel/pull/14738 solves exactly the request of this issue. The `repository_ctx.os.name` and `repository_ctx.os.arch` return the values for the host platform (which is the one Bazel runs on), this should be enough as in most cases repository rules run locally. \r\n\r\nBut we do have a way to make `repository.execute` [run commands on RBE](https://docs.bazel.build/versions/main/skylark/lib/globals.html#repository_rule(implementation,%20attrs,%20local,%20environ,%20configure,%20remotable,%20doc)), but this is a very much experimental feature and currently unmaintained.\r\n\r\nI also agree if Bazel runs in a simulation it should just get the simulated arch. We do provide native arm64 binaries for Linux and macOS, arm64 support for Windows has also been [added](https://github.com/bazelbuild/bazel/issues/14339) recently. ", "@bazel-io flag", "(I miss-read what that command was for, this isn't a blocker, sorry.)", "OK, I removed the label. But I guess we could still cherry-pick this to 5.1? @Wyverald ", "Yeah, that was what I wanted to do. Per the process, it seems we just open the PR instead of flagging it.", "@bazel-io fork 5.1", "The fork command will create an associate issue to track the cherry-pick, feel free to send a PR!", "Amazing, thank you for the work on this. I look forward to using it! I'll make contributions for the major rules to also take advantage of this new attribute when available.\n\nFor clarification; this only provides info for the host OS, correct? What happens in a remote execution context? Are those actions still executed locally? I suppose it depends on the declared toolchains.\n\nIs there any official guidance on how best to write repository rules for compatibility with remote execution? It seems to me like it may still be the case that rule authors must declare every cpu/os combination as repositories and reference them via toolchains.\n\nMaybe I'm missing something 🙂\n\n" ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkOS.java", "tools/cpp/lib_cc_configure.bzl", "tools/jdk/local_java_repository.bzl", "tools/osx/xcode_configure.bzl" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkOS.java", "tools/cpp/lib_cc_configure.bzl", "tools/jdk/local_java_repository.bzl", "tools/osx/xcode_configure.bzl" ]

[]

[ "keyword_pr_to_issue" ]

[ "Is `--experimental_remote_cache_compression` being used as well?", "Actually the reason for this error seems to be `--experimental_remote_cache_compression`. From my early tests locally that flag reduced build times by 20-25%, but I see `BulkTransferException` in every single CI build. I've removed `--experimental_remote_cache_async` and only enabled `--experimental_remote_cache_compression` and the problem persists.", "This is where those errors are thrown:\r\nhttps://github.com/bazelbuild/bazel/blob/72d525575a7a6c22abee5e7791f1072bb334340e/src/main/java/com/google/devtools/build/lib/remote/ByteStreamUploader.java#L466-L476", "Highlighting the relevant parts of the error:\r\n\r\n```\r\n\tSuppressed: java.io.IOException: write incomplete: committed_size 721 for 374 total\r\n\tSuppressed: java.io.IOException: write incomplete: committed_size 144 for 125 total\r\n\tSuppressed: java.io.IOException: write incomplete: committed_size 2 for 14 total\r\n```\r\n\r\nI am not super familiar with the byte stream uploader (attempting to grok it now), but is it possible that the `!chunker.hasNext()` condition does not necessarily imply that the uploader was the one who completed the upload, as the comment states? (Could it be that `chunker.hasNext()` returns false in some cases when there was only 1 chunk to flush, and the server returned that the chunk already exists, returning the blob's uncompressed size as the committed size?)\r\n\r\nAsking because those error messages look as though the server is returning uncompressed blob sizes as the committed size (e.g. a 721 byte chunk compressing to 374 bytes looks about right, and a 2 byte chunk compressing to 14 bytes looks accurate as well since the compression overhead probably requires more space than the bytes themselves).", "CC @coeuvre , @benjaminp , @buchgr , @Wyverald ", "CC @bazelbuild/remote-execution ", "I started a draft PR which has a failing test illustrating the issue, and I am fairly certain that the fake ByteStream impl in the test adheres to protocol, but please correct me if I'm wrong: https://github.com/bazelbuild/bazel/pull/14655\r\n\r\nI think there is an ambiguity with the current way that the `ByteStream.Write` protocol is specified. The exact spec is here: https://github.com/bazelbuild/remote-apis/blob/636121a32fa7b9114311374e4786597d8e7a69f3/build/bazel/remote/execution/v2/remote_execution.proto#L256-L264\r\n\r\n```\r\n// When attempting an upload, if another client has already completed the upload\r\n// (which may occur in the middle of a single upload if another client uploads\r\n// the same blob concurrently), the request will terminate immediately with\r\n// a response whose `committed_size` is the full size of the uploaded file\r\n// (regardless of how much data was transmitted by the client). If the client\r\n// completes the upload but the\r\n// [Digest][build.bazel.remote.execution.v2.Digest] does not match, an\r\n// `INVALID_ARGUMENT` error will be returned. In either case, the client should\r\n// not attempt to retry the upload.\r\n```\r\n\r\nIn summary, one of two things can happen in response to clients sending their final chunk:\r\n\r\n1. The server can respond with `committed_size = client_uploaded_size` which means that the client successfully performed the upload, or\r\n2. the server can respond with `committed_size = \"full size of the uploaded file\"` \r\n\r\nBut, what does \"full size of the uploaded file\" mean?\r\n* If it means \"compressed size of the file, as uploaded by the other client,\" then the \"full size\" could be pretty arbitrary since there is more than one valid way to compress a file, and it's useless to return this to the client.\r\n* But if it means \"uncompressed size of the file\" (i.e. `Digest.size_bytes`), then the client needs to check that, once they have received a `WriteResponse` and they have already uploaded all local chunks, then _either_ the `WriteResponse.committed_size` equals `Digest.size_bytes`, _or_ it equals the total number of bytes written by the local client (`chunker.getOffset()`) -- because the protocol gives us no way of knowing whether the successful upload of the last chunk was due to the local client's chunk being accepted, or if it was due to some other client completing their upload.\r\n\r\nGiven that the former interpretation would be useless to clients, I imagine the latter interpretation is what was originally intended. If so, I think it implies that the current implementation of `ByteStreamUploader` has a bug, because it does not check whether the `committed_size` could equal `Digest.size_bytes`. The current Bazel implementation seems to have used the first interpretation, based on [this comment](https://github.com/bazelbuild/bazel/blob/90abed5a31ae75761252c7e082572babafeb0ae6/src/main/java/com/google/devtools/build/lib/remote/ByteStreamUploader.java#L466-L468):\r\n\r\n```\r\n // Only check for matching committed size if we have completed the upload.\r\n // If another client did, they might have used a different compression\r\n // level/algorithm, so we cannot know the expected committed offset\r\n```\r\n\r\nCC @mostynb who may be able to help clarify the spec", "You're right that `WriteResponse.committed_size` is kind of useless for compressed blobs, but I don't think clients need to check that value- the server should return an error status code if and only if the write failed, and the client can check that instead of the `committed_size` field.", "@mostynb Thanks for your input!\r\n\r\nI would be happy to send a PR so that Bazel stops checking the `committed_size` in the `WriteResponse`, so long as the folks familiar with the ByteStreamUploader don't see any problems with removing the check. I am not 100% sure whether this would a breaking change, but it seems like it wouldn't be. If anything, it might break some expectations on the remote cache side, so it might help to have input from some remote cache impl maintainers as well.\r\n\r\nSeparately, I am trying to figure out whether the `compressed_size` check on the Bazel side is possible for a remote cache implementation to work around in the meantime. Specifically, I'm thinking about the following edge case and not sure how to handle it:\r\n\r\n```\r\n1. Bazel sends chunk #1, FinishWrite=false\r\n2. Bazel sends chunk #2, FinishWrite=true\r\n3. Server receives chunk #1, parses ResourceName, sees that the digest already exists in the cache\r\n Server immediately returns `WriteResponse{committed_size = digest_size}`\r\n4. Bazel receives WriteResponse from the server. Since it has already sent all chunks, it thinks it has completed\r\n the upload, and then asserts `committed_size == chunker.offset`. This fails.\r\n```\r\n\r\nFirstly, is this scenario is even possible? (not sure if client-streaming gRPC works the same way in Java as it does in Go, but IIUC the client does not need to wait for the server to receive and respond to each message before proceeding to send the next message)\r\n\r\nSecondly, is the only workaround on the cache side just to wait for Bazel to upload the entire stream, regardless of whether the digest already exists in cache or not, and then return the total length of the compressed stream so that it satisfies Bazel's committed_size expectation?", "@mostynb FYI, I think this issue might affect `bazel-remote` as well:\r\n\r\nhttps://github.com/buchgr/bazel-remote/blob/a3c6189b64cff12065750073b11a116c8c1ed00a/server/grpc_bytestream.go#L415-L421\r\n\r\nIt looks like on the first WriteRequest, if the digest already exists in cache, a `WriteResponse` with `committed_size=uncompressed_size` is effectively returned.", "Maybe the best we can do is update the REAPI spec to advise clients to ignore `committed_size` for compressed writes and to rely on the error status instead in that case?\r\n\r\nI'm not sure how the early-exit mechanism is useful in practice actually. As you mentioned the client calls `Send` until it thinks it has sent all the data, and only then calls `CloseAndRecv` to get the `WriteResponse` (at least in the go bindings). At this point the client has sent all the data even if the server decided to return early. So instead of returning early the server could have discarded all the received data and just counted how much compressed data was sent and returned that number. So maybe we should instead update the REAPI spec to advise servers to do that for compressed-blobs writes instead of returning early?", "> instead of returning early the server could have discarded all the received data and just counted how much compressed data was sent and returned that number. \r\n\r\nThis is exactly the workaround that I'm currently implementing :+1: \r\n\r\n> advise servers to do that for compressed-blobs writes instead of returning early\r\n\r\nI think it's good advice for servers that want to be compatible with Bazel's current behavior, so it's probably worth mentioning this approach in the REAPI spec as a notable client quirk. I'm not sure whether it'd be the best advice though if/when Bazel removes the `committed_size` check, since I don't have a great intuition (nor any data) as to whether the early-exit mechanism provides significant savings in practice.", "Let's open an REAPI issue to get more input- can you do that or would you like me to?", "I'll go ahead and open one", "@bazel-io fork 5.1", "Still seeing this, removing `--experimental_remote_cache_compression` works around it", "@keith: your remote cache might require an update- is it something opensource? If so which one and which version?", "Using Google's 🙃 ", "cc @bergsieker", "This will be fixed in Bazel 5.3 😍." ]

[]

[ "untriaged", "team-Remote-Exec" ]

[ "CONTRIBUTORS", "src/main/java/com/google/devtools/build/lib/remote/ByteStreamUploader.java" ]

[ "CONTRIBUTORS", "src/main/java/com/google/devtools/build/lib/remote/ByteStreamUploader.java" ]

[]

[ "keyword_pr_to_issue" ]

[ "This looks like a dupe of #8736." ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/analysis/BUILD", "src/main/java/com/google/devtools/build/lib/analysis/BaseRuleClasses.java", "src/main/java/com/google/devtools/build/lib/rules/core/CoreRules.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/BUILD", "src/main/java/com/google/devtools/build/lib/analysis/BaseRuleClasses.java", "src/main/java/com/google/devtools/build/lib/rules/core/CoreRules.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/test/CoverageConfiguration.java", "src/main/java/com/google/devtools/build/lib/starlarkbuildapi/test/BUILD", "src/main/java/com/google/devtools/build/lib/starlarkbuildapi/test/CoverageConfigurationApi.java", "src/test/shell/bazel/bazel_coverage_starlark_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "@gregestren , FYI.\r\n\r\nWhen I next have some free time I'll work on fixing this before the refactor. This is stopping us from upgrading.", "I should clarify that this didn't happen in 4.2.2:\r\n```console\r\n$ bazel --nohome_rc coverage //:all --javabase=@bazel_tools//tools/jdk:remote_jdk11 --combined_report=lcov\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/home/pschrader/.bazelrc\r\nINFO: Using default value for --instrumentation_filter: \"^//\".\r\nINFO: Override the above default with --instrumentation_filter\r\nINFO: Build options --java_runtime_version and --javabase have changed, discarding analysis cache.\r\nERROR: Cannot generate coverage report - no coverage information was collected\r\nINFO: Analyzed 2 targets (4 packages loaded, 405 targets configured).\r\nINFO: Found 1 target and 1 test target...\r\nINFO: Elapsed time: 3.302s, Critical Path: 0.05s\r\nINFO: 2 processes: 1 internal, 1 linux-sandbox.\r\nINFO: Build completed successfully, 2 total actions\r\n//:example_test SKIPPED\r\n\r\nINFO: Build completed successfully, 2 total actions\r\n```", "This bug affects my organization, we might have to downgrade to 4.2.\r\n\r\nI just checked and the issue is in 5.0.0 as well, it's not really a \"5.1\" bug as it says in the title. But I haven't tracked further than that to find a specific commit.", "If you compile your own version, I think you can use this as a quick hack to work around the issue:\r\n```patch\r\ndiff --git a/src/main/java/com/google/devtools/build/lib/analysis/BuildView.java b/src/main/java/com/google/devtools/build/lib/analysis/BuildView.java\r\nindex 581a655a65..67afcb3ab0 100644\r\n--- a/src/main/java/com/google/devtools/build/lib/analysis/BuildView.java\r\n+++ b/src/main/java/com/google/devtools/build/lib/analysis/BuildView.java\r\n@@ -531,8 +531,9 @@ public class BuildView {\r\n\r\n Set<ConfiguredTarget> allTargetsToTest = null;\r\n if (testsToRun != null) {\r\n- // Determine the subset of configured targets that are meant to be run as tests.\r\n+ // Determine the subset of compatible configured targets that are meant to be run as tests.\r\n allTargetsToTest = filterTestsByTargets(configuredTargets, testsToRun);\r\n+ allTargetsToTest.removeAll(targetsToSkip);\r\n }\r\n\r\n ImmutableSet.Builder<Artifact> artifactsToBuild = ImmutableSet.builder();\r\n```\r\nThe problem is that this approach doesn't pass the skipped targets information to the final test reporter (`TerminalTestResultNotifier`). That means it'll skip the incompatible tests, but not report on them. A proper fix would pass the information through farther so the test reporter reports on both executed tests and skipped tests.", "Thanks for the update! I don't think it's likely I can get go-ahead on deploying an inofficial Bazel version internally but it's good to hear that this is being worked on and that the problem is understood.\r\n\r\nDo you think a fix will be added to a 5.1.2 release in the near future (say, a couple of weeks)?", "I'll start investigating to identify the cause, but I can't judge on the time to fix just yet.", "@c-mita , WDYT about #15419?", "I would like to understand the regression better before judging a fix.\r\n\r\nBisecting indicates that 5b216b27435aeb9eb9c3bd3c552d6498e1050cc7 triggered the regression (the flag flip for\r\n`--experimental_forward_instrumentation_files_info_by_default`).\r\n\r\nUnfortunately the flag was removed in 5.0, so there's no way of disabling this behaviour in that version.\r\n\r\nThis wouldn't be the first time the change activated by that flag has broken something, but at least this gives an indication of where something might be going wrong.\r\n\r\nFWIW, the provided reproduction only starts passing properly after ed01f42a8ff533967c8850821fd8058d38377ff1 (bisecting between 4.0.0 and 4.2.0).", "So the cause of the regression is straight-forward.\r\n\r\nWhen 21179b2a82154b870ad8c5a7ad22b4662067cb29 added the check that forwards an `InstrumentedFilesInfo`, it changed the behaviour of the case of a test that normally provides its own but didn't because it was \"incompatible\".\r\n\r\nThis trips up the coverage code down the line.\r\n\r\nThe fix could either be what you've done in https://github.com/bazelbuild/bazel/pull/15419 or to change the check in `RuleConfiguredTargetBuilder::build` to also check for the existence of an `IncompatibleTargetProvider` and not create the forwarding `InstrumentedFilesInfo` in that case.\r\n\r\nI think I prefer the former as it strikes me as a little more \"robust\".\r\n" ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/bazel/coverage/BUILD", "src/main/java/com/google/devtools/build/lib/bazel/coverage/CoverageReportActionBuilder.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/coverage/BUILD", "src/main/java/com/google/devtools/build/lib/bazel/coverage/CoverageReportActionBuilder.java" ]

[ "src/test/shell/bazel/BUILD", "src/test/shell/bazel/bazel_coverage_compatibility_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "I believe it was waiting for xcode-locator, that can be super long if the machine has many Xcodes. \n\nTry this hack: https://www.smileykeith.com/2021/03/08/locking-xcode-in-bazel/", "Thanks, I will try that. Do you think that this could also be related to Issue 2?", "It could be, since that file is only created if Xcode is successfully detected.\r\n\r\nhttps://github.com/bazelbuild/bazel/blob/b4b0c321910bc968736ef48e8140528ea7d323cd/tools/cpp/cc_configure.bzl#L73", "@fmeum did @thii's suggestion fix the issue?", "I just checked again and found no new failures in the two weeks since I started using a checked-in `xcode_version_config`. I will close this now as the workaround seems to have worked and open a new issue if it should show up again.", "@sventiffe Reopening since the exact same issue appeared again today over at [rules_jni](https://github.com/fmeum/rules_jni).\r\n\r\nI started pinning with `--xcode_version_config` in https://github.com/fmeum/rules_jni/commit/b5021809cd5ea4393291d0d592b2f195c6f4e1f5. Today, [the runs](https://github.com/fmeum/rules_jni/runs/4823744403) on macos-latest with Bazel 4.0.0 and Bazel 5.0.0rc3 failed. I attached the [CI logs](https://github.com/bazelbuild/bazel/files/7874710/logs.zip), the failing step is [here](https://github.com/fmeum/rules_jni/blob/ff4e93e428869246f4779becd56b027bb13b3003/.github/workflows/run-tests-externally.yml#L75) with the message (in the case of Bazel 5.0.0rc3 with bzlmod):\r\n\r\n```\r\n2022/01/15 00:35:14 Downloading https://releases.bazel.build/4.0.0/release/bazel-4.0.0-darwin-x86_64...\r\nExtracting Bazel installation...\r\nStarting local Bazel server and connecting to it...\r\nINFO: Invocation ID: 0e7ba86d-1a19-4b3a-a770-f0c3d58d166f\r\nLoading: \r\nLoading: 0 packages loaded\r\nLoading: 11 packages loaded\r\n currently loading: analysis/cc_jni_library ... (2 packages)\r\nLoading: 11 packages loaded\r\n currently loading: analysis/cc_jni_library ... (2 packages)\r\nAnalyzing: 21 targets (13 packages loaded, 0 targets configured)\r\nAnalyzing: 21 targets (15 packages loaded, 1 target configured)\r\nAnalyzing: 21 targets (18 packages loaded, 8 targets configured)\r\nAnalyzing: 21 targets (18 packages loaded, 8 targets configured)\r\nAnalyzing: 21 targets (19 packages loaded, 8 targets configured)\r\nAnalyzing: 21 targets (21 packages loaded, 9 targets configured)\r\nAnalyzing: 21 targets (21 packages loaded, 9 targets configured)\r\nAnalyzing: 21 targets (21 packages loaded, 9 targets configured)\r\nAnalyzing: 21 targets (21 packages loaded, 9 targets configured)\r\nAnalyzing: 21 targets (21 packages loaded, 9 targets configured)\r\nAnalyzing: 21 targets (38 packages loaded, 206 targets configured)\r\nAnalyzing: 21 targets (38 packages loaded, 206 targets configured)\r\nAnalyzing: 21 targets (38 packages loaded, 206 targets configured)\r\nERROR: /Users/runner/work/rules_jni/rules_jni/tests/native_loader/src/test/java/com/example/math/BUILD.bazel:3:17: Target '//native_loader/src/test/java/com/example/math:math_remove_this_part_' depends on toolchain '@local_config_cc//:cc-compiler-darwin', which cannot be found: error loading package '@local_config_cc//': cannot load '@local_config_cc_toolchains//:osx_archs.bzl': no such file'\r\nERROR: Analysis of target '//native_loader/src/test/java/com/example/math:math' failed; build aborted: Analysis failed\r\nINFO: Elapsed time: 216.940s\r\nINFO: 0 processes.\r\nFAILED: Build did NOT complete successfully (39 packages loaded, 206 targets configured)\r\nERROR: Couldn't start the build. Unable to run tests\r\nFAILED: Build did NOT complete successfully (39 packages loaded, 206 targets configured)\r\nError: Process completed with exit code 1.\r\n```\r\n\r\n@thii Are there any additional diagnostics that I could enable that would help diagnose the underlying issue?", "@keith I just encountered another failure and attached the [the logs](https://github.com/bazelbuild/bazel/files/8044694/rules_jni.log). They include the content of the BUILD file, which turns out not to be empty.\r\n", "That file actually looks pretty good to me, since yours didn't fail with osx_archs import, it's because that file doesn't contain `darwin` https://github.com/bazelbuild/bazel/blob/master/tools/osx/crosstool/osx_archs.bzl\r\n\r\nI'm not sure where that's coming from but it should be darwin_x86_64 probably instead", "> That file actually looks pretty good to me, since yours didn't fail with osx_archs import, it's because that file doesn't contain `darwin` https://github.com/bazelbuild/bazel/blob/master/tools/osx/crosstool/osx_archs.bzl\r\n> \r\n> I'm not sure where that's coming from but it should be darwin_x86_64 probably instead\r\n\r\nNice catch. I think I found the root cause and submitted a fix as https://github.com/bazelbuild/bazel/pull/14796.", "Nice, I'll keep an eye on that. My assumption for why this is similar to the flaky CI case is because the toolchain only falls back to this codepath on macOS when Xcode cannot be found, which could happen in the case of a timeout running the discovery logic for that. This would likely workaround that issue if folks needed full Xcode in their cases https://www.smileykeith.com/2021/03/08/locking-xcode-in-bazel/" ]

[]

[ "type: bug", "more data needed", "platform: apple" ]

[ "tools/osx/crosstool/BUILD.tpl" ]

[ "tools/osx/crosstool/BUILD.tpl" ]

[]

[ "keyword_pr_to_issue" ]

[ "Hello @xhebox, Could you please provide complete steps to reproduce the above issue at our end. Thanks! ", "> Hello @xhebox, Could you please provide complete steps to reproduce the above issue at our end. Thanks!\r\n\r\nEdited, it really needs nothing more than normal building instructions. You could try this on alpine linux with docker. I am on my own custom linux distro with musl libc.", "I will send a patch removing the strdupa, since there is exactly one usage in whole of bazel.", "@bazel-io flag", "@bazel-io fork 5.3.0" ]

[]

[ "P3", "type: support / not a bug (process)", "team-OSS", "help wanted" ]

[ "src/main/tools/linux-sandbox-pid1.cc" ]

[ "src/main/tools/linux-sandbox-pid1.cc" ]

[]

[ "keyword_pr_to_issue" ]

[ "people are reporting this is causing the issue\r\n\r\nhttps://bazelbuild.slack.com/archives/CA31HN1T3/p1655638147979559?thread_ts=1655544708.176529&cid=CA31HN1T3\r\n\r\nhttps://github.com/bazelbuild/bazel/commit/4d900ceea12919ad62012830a95e51f9ec1a48bb", "@coeuvre ", "What's error message looks like? Does it also the case for HEAD?", "I am not sure, it seems the cache is not uploading artifacts in certain scenarios, there are no diagnostic messages. I can't tell if people reporting the issues are only using http caching though, but the person in the slack thread and this report are both using http cache.", "I also confirm this is a problem. We didn't notice at the beginning of the upgrade as we were using the previously populated remote cache but after a couple of days we were having mostly cache misses and less and less cache hits. The problem happens both when using the buildbuddy grpc cache or a simple GCS Bucket. Reverting to 5.1.1 makes everything working again.", "I am not able to reproduce this locally. Do you know which actions are not uploading artifacts? ", "@coeuvre it actually happens in rules from `rules_nodejs` we are using like `TsProject` https://github.com/bazelbuild/rules_nodejs/blob/stable/packages/typescript/index.bzl#L109", "We reverted back to 5.1.1 due to this, however we do not use `rules_nodejs` (mainly `java_library` and `genrule`).\r\n\r\nWe noticed because the bazel-remote logs were **all** GET 404.", "> \r\n\r\nSame error.", "@nickbreen Are you using directory inputs/outputs? `rules_nodejs` does and 5.2 includes https://github.com/bazelbuild/bazel/commit/26f878325e915e0905626a0e4c8bbacffd72f875, but if you are facing the same issues without relying on TreeArtifacts, then that commit is less likely to be the culprit (which would make it rather likely it's actually https://github.com/bazelbuild/bazel/commit/4d900ceea12919ad62012830a95e51f9ec1a48bb).", "We have various rules using TreeArtifacts (`ctx.actions.declare_directory`) and rules that produce a directory output using `ctx.outputs` and genrules doing the same.", "We have also noticed this similar bug where caching seems to have broken on 5.2.0. When I was tracing the request logs something seemed odd in terms of the action cache `GET` missing and content addressable `PUT`. I could be missing a detail here, maybe this will help narrow down to what other folks might be seeing.\r\n\r\n\r\nBefore doing a build there is a miss on the action cache `GET 404 127.0.0.1 /ac/877dcafd4dc610e6cc734aa39f91aaf9a8e9a7f730e5bd1021c557c5c74bfda5` \r\nhttps://gist.github.com/adam-singer/0c73500e47d1a2410924dc5be4da0a3a#file-gistfile1-txt-L1 \r\n\r\nBazel/rules go along their business and build/calculate results, then I would assume that the next time `877dcafd4dc610e6cc734aa39f91aaf9a8e9a7f730e5bd1021c557c5c74bfda5` is seen in the request logs it would be a `PUT` to `/ac/877dcafd4dc610e6cc734aa39f91aaf9a8e9a7f730e5bd1021c557c5c74bfda5`, but what I've seen in our request logs is the remote uploader is making the request to `PUT 200 127.0.0.1 /cas/877dcafd4dc610e6cc734aa39f91aaf9a8e9a7f730e5bd1021c557c5c74bfda5` https://gist.github.com/adam-singer/0c73500e47d1a2410924dc5be4da0a3a#file-gistfile1-txt-L62 the content addressable cache.\r\n\r\nCould someone confirm or verify seeing similar behavior and if that is expected? \r\n\r\n**Edit:**\r\n\r\nhttps://github.com/bazelbuild/bazel/commit/4d900ceea12919ad62012830a95e51f9ec1a48bb#diff-5556e28152f91eac5403663774979225b86e64133da164366b56f91d1daa641cR536 The `Step` is set `action.getRemoteActionExecutionContext().setStep(Step.CHECK_ACTION_CACHE);` but when comes time to upload via `DiskAndRemoteCacheClient.uploadActionResult` https://github.com/bazelbuild/bazel/commit/4d900ceea12919ad62012830a95e51f9ec1a48bb#diff-f826b1e1a71d4a2d9f9524894df3069dcb54964998277721021093ff7d55d194R62 the check is for `Step.UPLOAD_OUTPUTS`, that seems to be the reason uploading to action cache isn't happening.\r\n\r\n<img width=\"1469\" alt=\"Screen Shot 2022-07-05 at 10 55 07 PM\" src=\"https://user-images.githubusercontent.com/654526/177479745-75124c77-c2ef-4dd5-9153-dc31b35ce19e.png\">\r\n\r\nThis patch fixed uploading to action cache https://gist.github.com/adam-singer/6500f268885f1c2c604fd4f9f1fec439, would prefer if owners verify/confirm bug." ]

[]

[ "type: bug", "untriaged", "team-Remote-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/remote/RemoteExecutionService.java" ]

[ "src/main/java/com/google/devtools/build/lib/remote/RemoteExecutionService.java" ]

[ "src/test/shell/bazel/remote/remote_execution_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "@meteorcloudy ", "/cc @Wyverald ", "Yeah there's a bunch of quality-of-life changes we should make around the Starlark-running part of bzlmod before official launch.", "This goes even further: If a module extension specifies incorrect attributes for a repository rule, an error is printed but the build doesn't fail." ]

[]

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

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

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

[]

[ "keyword_pr_to_issue" ]

[ "@bazel-io flag", "@bazel-io fork 5.3.0", "Since this feature is implemented and cherry picked to 5.3.0, should we close this one?", "The proposal hasn't been fully implemented yet (e.g. we're missing support for repository fetching).", "Marking this as a release blocker since we want the repository fetching support to make it into 6.0.0 if possible.", "@tjgq What's the ETA for implementing this? Ideally, we don't want to delay the 6.0 release further more, and since this should be a backwards compatible feature, we can cherry pick it in 6.x releases?", "I was planning to get this submitted in time for 6.0.0, but other things got in the way. Let's do it for a 6.x release instead.", "This has been fully implemented in 7.x and backported to 6.x." ]

[]

[ "type: feature request", "P2", "team-Remote-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/BUILD", "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelper.java", "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/GetCredentialsResponse.java" ]

[ "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/BUILD", "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelper.java", "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelperEnvironment.java", "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/GetCredentialsResponse.java" ]

[ "src/test/java/com/google/devtools/build/lib/authandtls/credentialhelper/BUILD", "src/test/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelperTest.java", "src/test/java/com/google/devtools/build/lib/authandtls/credentialhelper/test_credential_helper.py" ]

[ "keyword_pr_to_issue" ]

[ "@bazel-io flag", "@bazel-io fork 5.3.0", "Since this feature is implemented and cherry picked to 5.3.0, should we close this one?", "The proposal hasn't been fully implemented yet (e.g. we're missing support for repository fetching).", "Marking this as a release blocker since we want the repository fetching support to make it into 6.0.0 if possible.", "@tjgq What's the ETA for implementing this? Ideally, we don't want to delay the 6.0 release further more, and since this should be a backwards compatible feature, we can cherry pick it in 6.x releases?", "I was planning to get this submitted in time for 6.0.0, but other things got in the way. Let's do it for a 6.x release instead.", "This has been fully implemented in 7.x and backported to 6.x." ]

[]

[ "type: feature request", "P2", "team-Remote-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/authandtls/AuthAndTLSOptions.java", "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelper.java", "src/main/java/com/google/devtools/build/lib/remote/BUILD", "src/main/java/com/google/devtools/build/lib/remote/RemoteModule.java" ]

[ "src/main/java/com/google/devtools/build/lib/authandtls/AuthAndTLSOptions.java", "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelper.java", "src/main/java/com/google/devtools/build/lib/remote/BUILD", "src/main/java/com/google/devtools/build/lib/remote/RemoteModule.java" ]

[ "src/test/java/com/google/devtools/build/lib/authandtls/BUILD", "src/test/java/com/google/devtools/build/lib/authandtls/UnresolvedScopedCredentialHelperConverterTest.java", "src/test/java/com/google/devtools/build/lib/remote/BUILD", "src/test/java/com/google/devtools/build/lib/remote/RemoteModuleTest.java" ]

[ "keyword_pr_to_issue" ]

[ "Interesting. Are you comfortable if it still uses the same syntax `build_setting_default=\"\"`? And that semantically means an empty list?\r\n\r\nI get your point, and agree this is worth addressing. I'd want to be cautious to make sure whatever solution doesn't create some unknown other consistency in some other API piece with different expectations.", "@gregestren What would you think of allowing both `list of <type>` and `<type>` as default values, with the latter continuing to be converted to a singleton list for backwards compatibility? Since there is already a `string_list` type, having the `BuildSetting` return a type that dynamically depends on the value of its `allowMultiple` member may actually not be that much work.\r\n\r\nEdit: I now see that the problem is not that `BuildSetting` can't return a different type, it's that `build_setting_default` has to go through the usual `attr`, which does not allow for this kind of fallback behavior.", "I ultimately came to the conclusion that the source of issues is that repeatable string flags should be represented as `string_list` flags with a different CLI argument passing style rather than `string` flags that conditionally provide a list value. I created #14911 to create a replacement for the current `allow_multiple` flag that requires significantly less special casing internally and should also be more transparent to users.", "Thanks. I'll take a look. Also cc @aranguyen FYI who's been looking recently at flag parsing logic in general.", "@fmeum thanks for sending out the proposal. I don't know how to make comments on the proposal you linked so I'm dropping them here instead.\r\n1) Have you considered what the overriding/accumulating behavior would look like if users do the following `bazel build --flag=[\"a\", \"b\"] --flag=[\"c\", \"d\"]`? Would the final evaluation be `--flag=[\"c\", \"d\"]` or would it be `--flag=[\"a\", \"b\", \"c\", \"d\"]`? \r\n2) If the answer to the previous question is `--flag=[\"c\", \"d\"]`. Would it break for the case of inheritance (test inheriting from build for example)? Some users might want to pass in additional flags for testing. \r\n", "@aranguyen My proposal doesn't allow for this syntax, so it (luckily) doesn't have to specify this interaction. My proposal only affects how repeated flags are represented in Starlark, not how they are specified on the command-line: You still either specify a `string_list` as `--flag=first,second` (where later occurrences of `--flag` override all previous occurrences) or as `--flag=first --first=second`, depending on the flag style. The starlark value would be `[\"first\", \"second\"]` in both cases." ]

[]

[ "P2", "type: support / not a bug (process)", "team-Configurability" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/starlark/StarlarkConfig.java", "src/main/java/com/google/devtools/build/lib/packages/BuildSetting.java", "src/main/java/com/google/devtools/build/lib/runtime/StarlarkOptionsParser.java", "src/main/java/com/google/devtools/build/lib/starlarkbuildapi/StarlarkConfigApi.java", "src/main/java/com/google/devtools/build/skydoc/fakebuildapi/FakeConfigApi.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/starlark/StarlarkConfig.java", "src/main/java/com/google/devtools/build/lib/packages/BuildSetting.java", "src/main/java/com/google/devtools/build/lib/runtime/StarlarkOptionsParser.java", "src/main/java/com/google/devtools/build/lib/starlarkbuildapi/StarlarkConfigApi.java", "src/main/java/com/google/devtools/build/skydoc/fakebuildapi/FakeConfigApi.java" ]

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

[ "keyword_pr_to_issue" ]

[ "cc @tjgq " ]

[]

[ "type: feature request", "untriaged", "team-Remote-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/authandtls/BUILD", "src/main/java/com/google/devtools/build/lib/authandtls/GoogleAuthUtils.java", "src/main/java/com/google/devtools/build/lib/buildeventservice/BUILD", "src/main/java/com/google/devtools/build/lib/buildeventservice/BazelBuildEventServiceModule.java", "src/main/java/com/google/devtools/build/lib/buildeventservice/BuildEventServiceModule.java", "src/main/java/com/google/devtools/build/lib/remote/RemoteModule.java" ]

[ "src/main/java/com/google/devtools/build/lib/authandtls/BUILD", "src/main/java/com/google/devtools/build/lib/authandtls/GoogleAuthUtils.java", "src/main/java/com/google/devtools/build/lib/buildeventservice/BUILD", "src/main/java/com/google/devtools/build/lib/buildeventservice/BazelBuildEventServiceModule.java", "src/main/java/com/google/devtools/build/lib/buildeventservice/BuildEventServiceModule.java", "src/main/java/com/google/devtools/build/lib/remote/RemoteModule.java" ]

[ "src/test/java/com/google/devtools/build/lib/authandtls/BUILD", "src/test/java/com/google/devtools/build/lib/authandtls/GoogleAuthUtilsTest.java", "src/test/java/com/google/devtools/build/lib/remote/BUILD", "src/test/java/com/google/devtools/build/lib/remote/GrpcCacheClientTest.java", "src/test/java/com/google/devtools/build/lib/remote/RemoteModuleTest.java", "src/test/java/com/google/devtools/build/lib/remote/RemoteServerCapabilitiesTest.java", "src/test/java/com/google/devtools/build/lib/remote/RemoteSpawnRunnerWithGrpcRemoteExecutorTest.java", "src/test/shell/bazel/remote/remote_execution_http_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "@bazel-io flag", "@bazel-io fork 5.3.0", "Since this feature is implemented and cherry picked to 5.3.0, should we close this one?", "The proposal hasn't been fully implemented yet (e.g. we're missing support for repository fetching).", "Marking this as a release blocker since we want the repository fetching support to make it into 6.0.0 if possible.", "@tjgq What's the ETA for implementing this? Ideally, we don't want to delay the 6.0 release further more, and since this should be a backwards compatible feature, we can cherry pick it in 6.x releases?", "I was planning to get this submitted in time for 6.0.0, but other things got in the way. Let's do it for a 6.x release instead.", "This has been fully implemented in 7.x and backported to 6.x." ]

[]

[ "type: feature request", "P2", "team-Remote-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/runtime/CommandEnvironment.java", "src/main/java/com/google/devtools/build/lib/runtime/CommandLinePathFactory.java" ]

[ "src/main/java/com/google/devtools/build/lib/runtime/CommandEnvironment.java", "src/main/java/com/google/devtools/build/lib/runtime/CommandLinePathFactory.java" ]

[]

[ "keyword_pr_to_issue" ]

[ "@gregestren , FYI.\r\n\r\nI'll work on this once the robotics season is over at the end of this month and after I finish with #14096.", "Makes total sense, and I agree. Thanks for the report and commitment, and good luck with robotics!" ]

[]

[ "P2", "team-Configurability" ]

[ "src/main/java/com/google/devtools/build/lib/skyframe/AspectFunction.java", "src/main/java/com/google/devtools/build/lib/skyframe/BUILD" ]

[ "src/main/java/com/google/devtools/build/lib/skyframe/AspectFunction.java", "src/main/java/com/google/devtools/build/lib/skyframe/BUILD" ]

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

[ "keyword_pr_to_issue" ]

[ "Does https://github.com/bazelbuild/bazel/pull/14600 fix this issue?", "Aw, I was hopeful. I fixed some merge conflicts (looks like \"../\".replace() -> Strings.replace(\"../\", ...) and I still get the same failure.\r\n\r\n```\r\n==================== Test output for @aos//aos:condition_test:\r\n/var/lib/worker/work/3/exec/bazel-out/k8-opt/bin/external/aos/aos/condition_test.runfiles/repo/../aos/aos/condition_test: error while loading shared libraries: libexternal_Saos_Saos_Slibcondition.so: cannot open shared object file: No such file or directory\r\n```", "Could you check whether https://github.com/bazelbuild/bazel/pull/16008 fixes the issue? It includes an integration test that I distilled from your reproducer.", "It does! Great work, thanks for the prompt response and prompt resolution. I really appreciate it.\r\n\r\nFYI, there's the same merge conflict with Strings.replace. Not a hard thing to fix though.", "Assigning to @oquenchil since the linked fix is about cc rules.", "Aw, @fmeum , looks like this fixed almost all of the tests except one which expects $(location) and RPATH to agree on the path to the shared libraries. https://github.com/bazelbuild/bazel/issues/16108 is the issue since it smells different enough to be a new bug.", "@bazel-io fork 5.3.1", "The fix introduced a regression (see https://github.com/bazelbuild/bazel/pull/16008#issuecomment-1224267553), which we will probably need a patch release for.", "Seeing a very similar problem when upgrading from 5.2.0 to 5.3.0:\r\n```\r\nerror while loading shared libraries: libexternal_Sopenvkl_Slibispc_Uutil_Uispc.so: cannot open shared object file: No such file or directory\r\n```\r\nBut trying out 5.3.1rc2, it looks like that particular problem is resolved for us.", "Hello @AustinSchuh, Are you still seeing this issue with Release 5.3.1 ? ", "5.3.1 works for me. I'm hitting https://github.com/bazelbuild/bazel/issues/16108 but that feels separate.", "Thanks for the update @AustinSchuh!" ]

[]

[ "type: bug", "P2", "team-Rules-CPP" ]

[ "src/main/java/com/google/devtools/build/lib/rules/cpp/CppLinkActionBuilder.java", "src/main/java/com/google/devtools/build/lib/rules/cpp/LibrariesToLinkCollector.java" ]

[ "src/main/java/com/google/devtools/build/lib/rules/cpp/CppLinkActionBuilder.java", "src/main/java/com/google/devtools/build/lib/rules/cpp/LibrariesToLinkCollector.java" ]

[ "src/test/shell/bazel/remote/remote_execution_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "@bazel-io flag", "Sorry about that. This corner of Bazel is poorly tested; I guess something like this was bound to happen.\r\n\r\n> When enabling verbose failures, the stack trace pointed to a failed precondition because the host portion of the URI was null\r\n\r\nCould you please share the full stack trace and the exact values of `--remote_executor` and `--remote_cache` you're using?", "Here's the command line and stack trace. The line numbers might be off a bit -- I recompiled bazel from the head and added a couple printf statements.\r\n```\r\n$ bazel build --remote_cache=unix:///home/coder/.cache/fax/3a162c244f546dff591bc0d335e9c174/jail.sock --remote_executor=unix:///home/coder/.cache/fax/3a162c244f546dff591bc0d335e9c174/jail.sock :all --verbose_failures\r\nExtracting Bazel installation...\r\nStarting local Bazel server and connecting to it...\r\nINFO: Invocation ID: 55dd5c7b-43ea-4a6f-a6cf-0c05ce656dbe\r\nERROR: java.io.IOException: io.grpc.StatusRuntimeException: UNAUTHENTICATED: Failed computing credential metadata\r\n at com.google.devtools.build.lib.remote.RemoteServerCapabilities.get(RemoteServerCapabilities.java:90)\r\n at com.google.devtools.build.lib.remote.RemoteModule.verifyServerCapabilities(RemoteModule.java:196)\r\n at com.google.devtools.build.lib.remote.RemoteModule.beforeCommand(RemoteModule.java:492)\r\n at com.google.devtools.build.lib.runtime.BlazeCommandDispatcher.execExclusively(BlazeCommandDispatcher.java:378)\r\n at com.google.devtools.build.lib.runtime.BlazeCommandDispatcher.exec(BlazeCommandDispatcher.java:232)\r\n at com.google.devtools.build.lib.server.GrpcServerImpl.executeCommand(GrpcServerImpl.java:550)\r\n at com.google.devtools.build.lib.server.GrpcServerImpl.lambda$run$1(GrpcServerImpl.java:614)\r\n at io.grpc.Context$1.run(Context.java:566)\r\n at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(Unknown Source)\r\n at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)\r\n at java.base/java.lang.Thread.run(Unknown Source)\r\nCaused by: io.grpc.StatusRuntimeException: UNAUTHENTICATED: Failed computing credential metadata\r\n at io.grpc.stub.ClientCalls.toStatusRuntimeException(ClientCalls.java:262)\r\n at io.grpc.stub.ClientCalls.getUnchecked(ClientCalls.java:243)\r\n at io.grpc.stub.ClientCalls.blockingUnaryCall(ClientCalls.java:156)\r\n at build.bazel.remote.execution.v2.CapabilitiesGrpc$CapabilitiesBlockingStub.getCapabilities(CapabilitiesGrpc.java:215)\r\n at com.google.devtools.build.lib.remote.RemoteServerCapabilities.lambda$get$0(RemoteServerCapabilities.java:85)\r\n at com.google.devtools.build.lib.remote.ReferenceCountedChannel.lambda$withChannelBlocking$2(ReferenceCountedChannel.java:85)\r\n at com.google.devtools.build.lib.remote.ReferenceCountedChannel.lambda$withChannel$4(ReferenceCountedChannel.java:108)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleUsing.subscribeActual(SingleUsing.java:59)\r\n at io.reactivex.rxjava3.core.Single.subscribe(Single.java:4855)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleFlatMap$SingleFlatMapCallback.onSuccess(SingleFlatMap.java:85)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleFlatMap$SingleFlatMapCallback$FlatMapSingleObserver.onSuccess(SingleFlatMap.java:112)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleMap$MapSingleObserver.onSuccess(SingleMap.java:65)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleDoOnDispose$DoOnDisposeObserver.onSuccess(SingleDoOnDispose.java:84)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleDoOnError$DoOnError.onSuccess(SingleDoOnError.java:52)\r\n at io.reactivex.rxjava3.internal.operators.observable.ObservableSingleSingle$SingleElementObserver.onComplete(ObservableSingleSingle.java:110)\r\n at io.reactivex.rxjava3.internal.observers.DeferredScalarDisposable.complete(DeferredScalarDisposable.java:85)\r\n at io.reactivex.rxjava3.subjects.AsyncSubject.subscribeActual(AsyncSubject.java:233)\r\n at io.reactivex.rxjava3.core.Observable.subscribe(Observable.java:13176)\r\n at io.reactivex.rxjava3.internal.operators.observable.ObservableSingleSingle.subscribeActual(ObservableSingleSingle.java:36)\r\n at io.reactivex.rxjava3.core.Single.subscribe(Single.java:4855)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleDoOnError.subscribeActual(SingleDoOnError.java:35)\r\n at io.reactivex.rxjava3.core.Single.subscribe(Single.java:4855)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleDoOnDispose.subscribeActual(SingleDoOnDispose.java:38)\r\n at io.reactivex.rxjava3.core.Single.subscribe(Single.java:4855)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleMap.subscribeActual(SingleMap.java:35)\r\n at io.reactivex.rxjava3.core.Single.subscribe(Single.java:4855)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleFlatMap$SingleFlatMapCallback.onSuccess(SingleFlatMap.java:85)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleCreate$Emitter.onSuccess(SingleCreate.java:68)\r\n at com.google.devtools.build.lib.remote.grpc.TokenBucket$1.onNext(TokenBucket.java:79)\r\n at io.reactivex.rxjava3.internal.util.NotificationLite.accept(NotificationLite.java:247)\r\n at io.reactivex.rxjava3.subjects.BehaviorSubject$BehaviorDisposable.test(BehaviorSubject.java:507)\r\n at io.reactivex.rxjava3.subjects.BehaviorSubject$BehaviorDisposable.emitFirst(BehaviorSubject.java:468)\r\n at io.reactivex.rxjava3.subjects.BehaviorSubject.subscribeActual(BehaviorSubject.java:224)\r\n at io.reactivex.rxjava3.core.Observable.subscribe(Observable.java:13176)\r\n at com.google.devtools.build.lib.remote.grpc.TokenBucket.lambda$acquireToken$0(TokenBucket.java:64)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleCreate.subscribeActual(SingleCreate.java:40)\r\n at io.reactivex.rxjava3.core.Single.subscribe(Single.java:4855)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleFlatMap.subscribeActual(SingleFlatMap.java:37)\r\n at io.reactivex.rxjava3.core.Single.subscribe(Single.java:4855)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleDefer.subscribeActual(SingleDefer.java:43)\r\n at io.reactivex.rxjava3.core.Single.subscribe(Single.java:4855)\r\n at io.reactivex.rxjava3.internal.operators.single.SingleFlatMap.subscribeActual(SingleFlatMap.java:37)\r\n at io.reactivex.rxjava3.core.Single.subscribe(Single.java:4855)\r\n at io.reactivex.rxjava3.core.Single.blockingGet(Single.java:3644)\r\n at com.google.devtools.build.lib.remote.ReferenceCountedChannel.withChannelBlocking(ReferenceCountedChannel.java:85)\r\n at com.google.devtools.build.lib.remote.RemoteServerCapabilities.lambda$get$1(RemoteServerCapabilities.java:84)\r\n at com.google.devtools.build.lib.remote.Retrier.execute(Retrier.java:244)\r\n at com.google.devtools.build.lib.remote.RemoteRetrier.execute(RemoteRetrier.java:125)\r\n at com.google.devtools.build.lib.remote.RemoteRetrier.execute(RemoteRetrier.java:114)\r\n at com.google.devtools.build.lib.remote.RemoteServerCapabilities.get(RemoteServerCapabilities.java:82)\r\n ... 10 more\r\nCaused by: java.lang.NullPointerException\r\n at com.google.common.base.Preconditions.checkNotNull(Preconditions.java:889)\r\n at com.google.devtools.build.lib.authandtls.credentialhelper.CredentialHelperProvider.findCredentialHelper(CredentialHelperProvider.java:74)\r\n at com.google.devtools.build.lib.authandtls.credentialhelper.CredentialHelperCredentials$CredentialHelperCacheLoader.load(CredentialHelperCredentials.java:132)\r\n at com.google.devtools.build.lib.authandtls.credentialhelper.CredentialHelperCredentials$CredentialHelperCacheLoader.load(CredentialHelperCredentials.java:114)\r\n at com.github.benmanes.caffeine.cache.LocalLoadingCache.lambda$newMappingFunction$2(LocalLoadingCache.java:141)\r\n at com.github.benmanes.caffeine.cache.BoundedLocalCache.lambda$doComputeIfAbsent$14(BoundedLocalCache.java:2413)\r\n at java.base/java.util.concurrent.ConcurrentHashMap.compute(Unknown Source)\r\n at com.github.benmanes.caffeine.cache.BoundedLocalCache.doComputeIfAbsent(BoundedLocalCache.java:2411)\r\n at com.github.benmanes.caffeine.cache.BoundedLocalCache.computeIfAbsent(BoundedLocalCache.java:2394)\r\n at com.github.benmanes.caffeine.cache.LocalCache.computeIfAbsent(LocalCache.java:108)\r\n at com.github.benmanes.caffeine.cache.LocalLoadingCache.get(LocalLoadingCache.java:54)\r\n at com.google.devtools.build.lib.authandtls.credentialhelper.CredentialHelperCredentials.getRequestMetadataFromCredentialHelper(CredentialHelperCredentials.java:90)\r\n at com.google.devtools.build.lib.authandtls.credentialhelper.CredentialHelperCredentials.getRequestMetadata(CredentialHelperCredentials.java:74)\r\n at com.google.auth.Credentials.blockingGetToCallback(Credentials.java:112)\r\n at com.google.auth.Credentials$1.run(Credentials.java:98)\r\n at io.grpc.stub.ClientCalls$ThreadlessExecutor.waitAndDrain(ClientCalls.java:740)\r\n at io.grpc.stub.ClientCalls.blockingUnaryCall(ClientCalls.java:149)\r\n ... 57 more\r\n\r\nERROR: Failed to query remote execution capabilities: UNAUTHENTICATED: Failed computing credential metadata\r\n```", "Thanks, this is helpful. It appears that `new URI(\"unix:///foo/bar).getHost() == null` and `new URI(\"unix:/foo/bar\").getHost() == null`, but `new URI(\"unix://foo/bar\").getHost() == \"foo\"`. I wonder if using one or two slashes after `unix:` instead of three is a valid workaround. Hopefully the gRPC library isn't too picky...\r\n\r\nThat being said, I totally agree that we shouldn't crash here! I think the right fix is to replace the precondition check in `findCredentialHelper` with an early exit if the host part is empty/null. That way we don't have to make any assumptions about URI schemes. I'll put a fix together on Monday.", "Regardless of the number of slashes, it used to work with versions of bazel older than 5.3. I've used it on 5.2.0, 5.1.0, 5.0.0, and 4.2.2 to name a few. But I'll look into seeing how it behaves when we have only 1 slash and let you know what I find.", "Yep, regardless of the number of slashes, it fails at the same precondition check. (I tried 1, 2, and 3.) Also, for unix domain sockets, there is no host -- the entire remainder of the string is the pathname of the socket file, so getHost() probably should return null when the scheme is \"unix:\".", "@tjgq do you think this warrants a patch release?", "@Wyverald It's a regression that affects existing users and there doesn't seem to be a workaround for it, so probably yes? (I don't know what's our patch release policy.)\r\n\r\n@srago I'd like to go with #16184 as the fix. Could you please confirm that it works for you?\r\n\r\ncc @Yannic - until #16185 is fixed, we should manually verify that a UDS proxy works when making changes to credential helpers.", "@bazel-io fork 5.3.1", "Confirmed #16184 fixes the problem with Unix domain sockets. Thanks for the quick response time!" ]

[]

[ "type: bug", "untriaged", "team-Remote-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelperProvider.java" ]

[ "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelperProvider.java" ]

[ "src/test/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelperProviderTest.java" ]

[ "keyword_pr_to_issue" ]

[]

[]

[ "type: feature request", "P2", "team-Remote-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelperProvider.java" ]

[ "src/main/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelperProvider.java" ]

[ "src/test/java/com/google/devtools/build/lib/authandtls/credentialhelper/CredentialHelperProviderTest.java" ]

[ "keyword_pr_to_issue" ]

[ "@beasleyr-vmw If you would like to see this fix in 5.3.1, you could cherry-pick it onto the `release-5.3.1` branch and submit a PR.", "Exactly what @fmeum said. Please send a PR." ]

[]

[ "type: bug", "team-Configurability", "untriaged" ]

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

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

[]

[ "keyword_pr_to_issue" ]

[ "@Wyverald Could you assign me and add the appropriate labels?", "~~@comius Do you expect `java_binary` to be starlarkified in time for Bazel 6? That would simplify the changes to the Java rules.~~ Most likely no longer necessary, as the changes will apply to `buildjar` itself.", "@Wyverald Oops, I missed the use of `Closes ...` in the PR description - we should use `Work towards ...` instead. Could you reopen? For some I lack the permission to do so.", "It's the \"To fix #NNN\" line... GitHub trying to be clever here", "The most likely future seems to be that we'll standardize on Python imports _not_ containing the repository name, in which case bzlmod is a no-op.\r\n\r\nIt's not an optimal solution because it requires an entry in `$PYTHONPATH` for each repository in the transitive closure of the Python binary, but it's the status quo and I'd much prefer thinking up a solution to that (e.g. binary-specific symlink trees that do not look like runfiles trees, dunno) than to standardize on having the repository names in imports because the concept of \"repository\" is specific to Bazel.\r\n\r\nPut differently: the only way for bzlmod to cause trouble for Python is if we decide to require repository names in imports, which seems to be unlikely and it looks like the decision on that will take a while and I prefer not blocking either Bazel 6.0 or the removal of bzlmod from the experimental domain to avoid an unlikely failure mode. And should that happen, we can always paper over it with some sort of import hook.\r\n\r\n/cc @rickeylev ", "@oquenchil I'm thinking about whether we should use a macro to pass `BAZEL_CURRENT_REPOSITORY` to the `Rlocation` function automatically (e.g. do `#define Rlocation(path) RlocationReal(path, BAZEL_CURRENT_REPOSITORY)`). Do you think that would be too intrusive/magical?", "Would users have to write that line or is it part of a file that is automatically generated?\r\n\r\nI don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.", "> Would users have to write that line or is it part of a file that is automatically generated?\r\n> \r\n> I don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.\r\n\r\n@oquenchil I would want to make that magic part of the runfiles library header so that users don't have to write it themselves.\r\n\r\nWhat I am worried about is having an `Rlocation` macro mess with people's existing codebase since there doesn't seem to be a way to put it in a namespace or restrict it to member functions only.", "Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?", "> Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?\r\n\r\nWe can call it something else, we just won't be able to have existing C++ codebases use the new repo mapping-aware lookup automatically if we do so. But of course having to replace all `runfiles::create()` calls with something else won't be too bad.", "If the macro name is Rlocation and that may start clashing with people's code in C++ files then I think it would be prudent to rename it if we decide to go with the macro.\r\n\r\nI guess the call to make here is how much of a breaking change this is compared to the benefit of passing BAZEL_CURRENT_REPOSITORY automatically for users. I would say that the benefit is not too great but as far as I understood the use of Bazel runfiles in C++ code wasn't too wide-spread either?\r\n\r\n> But of course having to replace all runfiles::create() calls with something else won't be too bad.\r\n\r\nHow do you know? Is it based on search results from GH? If a release note would be enough for it not to be too disruptive then let's rename it.", "I have experimented a bit more with APIs while working on the corresponding change for Java (https://github.com/bazelbuild/bazel/pull/16549).\r\n\r\nWhat I now like more than the macro approach is to have users (optionally) pass in a canonical repository name when creating the `Runfiles` instance and default to that repo in `Rlocation` calls, while providing ways to override the default with an overload of `Rlocation` that accepts a repository name as a second parameter. In this way, users only have to pass in `BAZEL_CURRENT_REPOSITORY` only once in the common case, which should render a macro unnecessary.", "Update: All changes relevant for Bazel 6.0.0 are now in review.", "This is mostly done now, with only the bugfix https://github.com/bazelbuild/bazel/pull/16755 and a number of cherry-picks to 6.0.0 missing.", "@Wyverald You can remove this from the \"6.0.0 nice to have\" milestone as the only remaining issue is not tied to a Bazel release (Stardoc).", "Downgrading because the most important points have been addressed; the only outstanding issue is Stardoc, which is undergoing a rewrite currently by @tetromino." ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/analysis/BUILD", "src/main/java/com/google/devtools/build/lib/analysis/RunfilesLibraryInfo.java", "src/main/java/com/google/devtools/build/lib/analysis/starlark/StarlarkModules.java", "src/main/java/com/google/devtools/build/lib/starlarkbuildapi/RunfilesLibraryInfoApi.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/BUILD", "src/main/java/com/google/devtools/build/lib/analysis/starlark/StarlarkModules.java" ]

[ "src/test/java/com/google/devtools/build/lib/starlark/BUILD", "src/test/java/com/google/devtools/build/lib/starlark/StarlarkIntegrationTest.java" ]

[ "keyword_pr_to_issue" ]

[ "@Wyverald Could you assign me and add the appropriate labels?", "~~@comius Do you expect `java_binary` to be starlarkified in time for Bazel 6? That would simplify the changes to the Java rules.~~ Most likely no longer necessary, as the changes will apply to `buildjar` itself.", "@Wyverald Oops, I missed the use of `Closes ...` in the PR description - we should use `Work towards ...` instead. Could you reopen? For some I lack the permission to do so.", "It's the \"To fix #NNN\" line... GitHub trying to be clever here", "The most likely future seems to be that we'll standardize on Python imports _not_ containing the repository name, in which case bzlmod is a no-op.\r\n\r\nIt's not an optimal solution because it requires an entry in `$PYTHONPATH` for each repository in the transitive closure of the Python binary, but it's the status quo and I'd much prefer thinking up a solution to that (e.g. binary-specific symlink trees that do not look like runfiles trees, dunno) than to standardize on having the repository names in imports because the concept of \"repository\" is specific to Bazel.\r\n\r\nPut differently: the only way for bzlmod to cause trouble for Python is if we decide to require repository names in imports, which seems to be unlikely and it looks like the decision on that will take a while and I prefer not blocking either Bazel 6.0 or the removal of bzlmod from the experimental domain to avoid an unlikely failure mode. And should that happen, we can always paper over it with some sort of import hook.\r\n\r\n/cc @rickeylev ", "@oquenchil I'm thinking about whether we should use a macro to pass `BAZEL_CURRENT_REPOSITORY` to the `Rlocation` function automatically (e.g. do `#define Rlocation(path) RlocationReal(path, BAZEL_CURRENT_REPOSITORY)`). Do you think that would be too intrusive/magical?", "Would users have to write that line or is it part of a file that is automatically generated?\r\n\r\nI don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.", "> Would users have to write that line or is it part of a file that is automatically generated?\r\n> \r\n> I don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.\r\n\r\n@oquenchil I would want to make that magic part of the runfiles library header so that users don't have to write it themselves.\r\n\r\nWhat I am worried about is having an `Rlocation` macro mess with people's existing codebase since there doesn't seem to be a way to put it in a namespace or restrict it to member functions only.", "Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?", "> Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?\r\n\r\nWe can call it something else, we just won't be able to have existing C++ codebases use the new repo mapping-aware lookup automatically if we do so. But of course having to replace all `runfiles::create()` calls with something else won't be too bad.", "If the macro name is Rlocation and that may start clashing with people's code in C++ files then I think it would be prudent to rename it if we decide to go with the macro.\r\n\r\nI guess the call to make here is how much of a breaking change this is compared to the benefit of passing BAZEL_CURRENT_REPOSITORY automatically for users. I would say that the benefit is not too great but as far as I understood the use of Bazel runfiles in C++ code wasn't too wide-spread either?\r\n\r\n> But of course having to replace all runfiles::create() calls with something else won't be too bad.\r\n\r\nHow do you know? Is it based on search results from GH? If a release note would be enough for it not to be too disruptive then let's rename it.", "I have experimented a bit more with APIs while working on the corresponding change for Java (https://github.com/bazelbuild/bazel/pull/16549).\r\n\r\nWhat I now like more than the macro approach is to have users (optionally) pass in a canonical repository name when creating the `Runfiles` instance and default to that repo in `Rlocation` calls, while providing ways to override the default with an overload of `Rlocation` that accepts a repository name as a second parameter. In this way, users only have to pass in `BAZEL_CURRENT_REPOSITORY` only once in the common case, which should render a macro unnecessary.", "Update: All changes relevant for Bazel 6.0.0 are now in review.", "This is mostly done now, with only the bugfix https://github.com/bazelbuild/bazel/pull/16755 and a number of cherry-picks to 6.0.0 missing.", "@Wyverald You can remove this from the \"6.0.0 nice to have\" milestone as the only remaining issue is not tied to a Bazel release (Stardoc).", "Downgrading because the most important points have been addressed; the only outstanding issue is Stardoc, which is undergoing a rewrite currently by @tetromino." ]

[]

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

[ "src/main/java/com/google/devtools/build/docgen/templates/be/make-variables.vm", "src/main/java/com/google/devtools/build/lib/analysis/LocationExpander.java", "src/main/java/com/google/devtools/build/lib/analysis/LocationTemplateContext.java" ]

[ "src/main/java/com/google/devtools/build/docgen/templates/be/make-variables.vm", "src/main/java/com/google/devtools/build/lib/analysis/LocationExpander.java", "src/main/java/com/google/devtools/build/lib/analysis/LocationTemplateContext.java" ]

[ "src/test/java/com/google/devtools/build/lib/analysis/LocationExpanderIntegrationTest.java", "src/test/java/com/google/devtools/build/lib/analysis/LocationExpanderTest.java", "src/test/java/com/google/devtools/build/lib/analysis/LocationFunctionTest.java", "src/test/py/bazel/runfiles_test.py" ]

[ "keyword_pr_to_issue" ]

[ "As a workaround I wrote the following repository rule:\r\n```python\r\ndef _define_repository_name_impl(repository_ctx):\r\n repository_ctx.file(\"BUILD.bazel\",\r\n \"\"\"package(default_visibility = [\"//visibility:public\"])\r\n \"\"\")\r\n repository_ctx.file(\"defs.bzl\",\r\n \"\"\"main_repository_name=\"{}\"\r\n \"\"\".format(repository_ctx.attr.repo_name))\r\n\r\n_define_repository_name = repository_rule(\r\n implementation = _define_repository_name_impl,\r\n attrs={\"repo_name\": attr.string(mandatory=True)}\r\n )\r\n\r\ndef define_repository_name(repo_name):\r\n _define_repository_name(name = \"main_repository_name\",repo_name = repo_name)\r\n\r\n```\r\n\r\nwhich allows me to load the main repository name by calling:\r\n```python\r\nload(\"@main_repository_name//:defs.bzl\",\"main_repository_name\")\r\n```\r\n\r\nThe `WORKSPACE` file looks like this:\r\n```python\r\nrepository_name = \"xxx\"\r\nworkspace(name = repository_name)\r\n\r\nload(\"@bazel_tooling//rules:define_repository_name.bzl\",\"define_repository_name\")\r\ndefine_repository_name(repository_name)\r\n```\r\n\r\nI don't like it... given that `workspace()` call should be the first line of the WORKSPACE file. But it works (`bazel 2.1.0`)", "hey @laurentlb ... May I ask for your opinion on this matter? :-)", "@or-shachar I have the same issue, did you ever find a resolution to this?", "@or-shachar @jheaff1 Could you describe in more detail what you want to use the name declared in the `WORKSPACE` file for in a macro? Knowing more about the actual problem you are trying to solve would help me answer these questions.", "You most likely want just \"@\" instead. With the recent fixes on repo mapping, \"@workspace_name//:foo\" is equivalent to \"@//:foo\".\r\n\r\n(Unless you're doing exec path or runfiles shenanigans, in which case refer to @fmeum's comment.)", "@fmeum I am trying to write a sh_binary target for the rules_foreign_cc repo that wraps around an executable binary (built by rules_foreign_cc) with shared library dependencies. This would allow such a binary to be used as a build tool, eg the protobuf compiler or various generators part of the Qt framework.\r\n\r\nI would like to use the `rlocation` bash function provided here:\r\n\r\nhttps://github.com/bazelbuild/bazel/blob/a76cf2f4c3aa514dffa1ddc038c196103ae4583c/tools/bash/runfiles/runfiles.bash#L68\r\n\r\nThe `rlocation` bash function requires the workspace name, which I would like to obtain programatically in a macro.\r\n\r\nI’ve tried to use native.repository_name() but that returns `@` rather than the name of the repository.\r\n\r\n", "@jheaff1 Would an analogue of `$(location //foo)` that returns the correct path for `rlocation` help you realize this? I could see that being more user-friendly than yet another \"some type of name of a repo\" thing.", "@fmeum Absolutely that would be very useful. Ive found that using `$(rootpath` rather than `$(location` provides exactly what `$(rlocation` requires apart from the workspace name.", "@lberki I have been thinking about proposing a `$(location ...)` variant that returns the correct path for Rlocation for a while. The way runfiles work with Bzlmod would almost make this a requirement for location expansion to remain useful in `args`. What do you think?", "@fmeum without taking sides on the `$(location) question, it's a feature orthogonal to the one proposed in this bug.\r\n\r\nre: this bug in particular, what is the argument *against* introducing `repository_name()`? Naively, one would think that it's a very logical extension to `package_name()`. @Wyverald maybe you can think of a counter-argument?", "> this bug in particular, what is the argument against introducing repository_name()? Naively, one would think that it's a very logical extension to package_name(). @Wyverald maybe you can think of a counter-argument?\r\n\r\nWe already have `native.repository_name` and it does what it says on the tin (https://bazel.build/rules/lib/native#repository_name). @fmeum's & @jheaff1's question is more about the runfiles/execpath use case, where we sometimes need the *workspace name*, which is only accessible right now in a rule ctx object (https://bazel.build/rules/lib/ctx#workspace_name; also to note, *not* the same as [Label.workspace_name](https://bazel.build/rules/lib/Label#workspace_name)...)", "@Wyverald I'm embarrassed! I specifically looked for the identifier called `repository_name` and didn't find it in the source code. My bad.", "The issue with making the (global) workspace name available in Starlark code is that then the evaluation of BUILD files in external repositories would be a function of the content of the main repository, which isn't too intuitive (it's problematic enough for `ctx.workspace_name()`, but we can't help that because that's how runfiles work at the moment.", "@lberki That plus the fact that having yet another `workspace_name` would be confusing made me think that being able to get Rlocation paths out of location expansion may indeed be the best solution to this issue (see also the comment by @jheaff1). Since `native.repository_name` returns the right identifier for everything *except* runfiles discovery, why not add a way for users to get the full runfiles path right away without having them stitch it together themselves?", "The reason why I was confused is that for some reason, I thought `$(rlocation)` is already a thing in Bazel so I went on a wild goose chase trying to figure out what *else* you want :)\r\n\r\nI think both adding `$(rlocation)` and the same method to File is reasonable. @Wyverald , it looks like you also agree? ", "@lberki `$(rlocation)` seems reasonably easy to do, but `File.rlocation_path` is difficult: Without `--experimental_sibling_repository_layout`, that function needs to know the workspace name. I'm not sure how the API for that would look like, should we just let the user pass in `ctx`?", "Not necessarily; there is `BazelStarlarkContext`, which is available to the Java implementations of Starlark functions without an extra argument and it, weirdly, already contains the repository name of the tools repository. So I could imagine plumbing the workspace name in there that way, too. Rule implementations need to know the name of the main workspace anyway, so it would probably not require too much extra plumbing.", "@lberki I will give that a try. We will have to make a concession though: Using `StarlarkThread` is incompatible with `structField = True`, so this will have to be a function instead.", "Today I Learned. It seems to be an artificial choice and not an inherent limitation of how Starlark works, though, and I don't see an immediate reason for this limitation. @c-parsons , do you remember why you prohibited setting `useStarlarkThread` and `structField` to true at the same time in 73e1016 ?" ]

[]

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

[ "src/main/java/com/google/devtools/build/docgen/templates/be/make-variables.vm", "src/main/java/com/google/devtools/build/lib/analysis/LocationExpander.java", "src/main/java/com/google/devtools/build/lib/analysis/LocationTemplateContext.java" ]

[ "src/main/java/com/google/devtools/build/docgen/templates/be/make-variables.vm", "src/main/java/com/google/devtools/build/lib/analysis/LocationExpander.java", "src/main/java/com/google/devtools/build/lib/analysis/LocationTemplateContext.java" ]

[ "src/test/java/com/google/devtools/build/lib/analysis/LocationExpanderIntegrationTest.java", "src/test/java/com/google/devtools/build/lib/analysis/LocationExpanderTest.java", "src/test/java/com/google/devtools/build/lib/analysis/LocationFunctionTest.java", "src/test/py/bazel/runfiles_test.py" ]

[ "keyword_pr_to_issue" ]

[ "@Wyverald Could you assign me and add the appropriate labels?", "~~@comius Do you expect `java_binary` to be starlarkified in time for Bazel 6? That would simplify the changes to the Java rules.~~ Most likely no longer necessary, as the changes will apply to `buildjar` itself.", "@Wyverald Oops, I missed the use of `Closes ...` in the PR description - we should use `Work towards ...` instead. Could you reopen? For some I lack the permission to do so.", "It's the \"To fix #NNN\" line... GitHub trying to be clever here", "The most likely future seems to be that we'll standardize on Python imports _not_ containing the repository name, in which case bzlmod is a no-op.\r\n\r\nIt's not an optimal solution because it requires an entry in `$PYTHONPATH` for each repository in the transitive closure of the Python binary, but it's the status quo and I'd much prefer thinking up a solution to that (e.g. binary-specific symlink trees that do not look like runfiles trees, dunno) than to standardize on having the repository names in imports because the concept of \"repository\" is specific to Bazel.\r\n\r\nPut differently: the only way for bzlmod to cause trouble for Python is if we decide to require repository names in imports, which seems to be unlikely and it looks like the decision on that will take a while and I prefer not blocking either Bazel 6.0 or the removal of bzlmod from the experimental domain to avoid an unlikely failure mode. And should that happen, we can always paper over it with some sort of import hook.\r\n\r\n/cc @rickeylev ", "@oquenchil I'm thinking about whether we should use a macro to pass `BAZEL_CURRENT_REPOSITORY` to the `Rlocation` function automatically (e.g. do `#define Rlocation(path) RlocationReal(path, BAZEL_CURRENT_REPOSITORY)`). Do you think that would be too intrusive/magical?", "Would users have to write that line or is it part of a file that is automatically generated?\r\n\r\nI don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.", "> Would users have to write that line or is it part of a file that is automatically generated?\r\n> \r\n> I don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.\r\n\r\n@oquenchil I would want to make that magic part of the runfiles library header so that users don't have to write it themselves.\r\n\r\nWhat I am worried about is having an `Rlocation` macro mess with people's existing codebase since there doesn't seem to be a way to put it in a namespace or restrict it to member functions only.", "Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?", "> Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?\r\n\r\nWe can call it something else, we just won't be able to have existing C++ codebases use the new repo mapping-aware lookup automatically if we do so. But of course having to replace all `runfiles::create()` calls with something else won't be too bad.", "If the macro name is Rlocation and that may start clashing with people's code in C++ files then I think it would be prudent to rename it if we decide to go with the macro.\r\n\r\nI guess the call to make here is how much of a breaking change this is compared to the benefit of passing BAZEL_CURRENT_REPOSITORY automatically for users. I would say that the benefit is not too great but as far as I understood the use of Bazel runfiles in C++ code wasn't too wide-spread either?\r\n\r\n> But of course having to replace all runfiles::create() calls with something else won't be too bad.\r\n\r\nHow do you know? Is it based on search results from GH? If a release note would be enough for it not to be too disruptive then let's rename it.", "I have experimented a bit more with APIs while working on the corresponding change for Java (https://github.com/bazelbuild/bazel/pull/16549).\r\n\r\nWhat I now like more than the macro approach is to have users (optionally) pass in a canonical repository name when creating the `Runfiles` instance and default to that repo in `Rlocation` calls, while providing ways to override the default with an overload of `Rlocation` that accepts a repository name as a second parameter. In this way, users only have to pass in `BAZEL_CURRENT_REPOSITORY` only once in the common case, which should render a macro unnecessary.", "Update: All changes relevant for Bazel 6.0.0 are now in review.", "This is mostly done now, with only the bugfix https://github.com/bazelbuild/bazel/pull/16755 and a number of cherry-picks to 6.0.0 missing.", "@Wyverald You can remove this from the \"6.0.0 nice to have\" milestone as the only remaining issue is not tied to a Bazel release (Stardoc).", "Downgrading because the most important points have been addressed; the only outstanding issue is Stardoc, which is undergoing a rewrite currently by @tetromino." ]

[]

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

[ "src/main/java/com/google/devtools/build/docgen/templates/be/make-variables.vm", "src/main/java/com/google/devtools/build/lib/analysis/LocationExpander.java", "src/main/java/com/google/devtools/build/lib/analysis/LocationTemplateContext.java" ]

[ "src/main/java/com/google/devtools/build/docgen/templates/be/make-variables.vm", "src/main/java/com/google/devtools/build/lib/analysis/LocationExpander.java", "src/main/java/com/google/devtools/build/lib/analysis/LocationTemplateContext.java" ]

[ "src/test/java/com/google/devtools/build/lib/analysis/LocationExpanderIntegrationTest.java", "src/test/java/com/google/devtools/build/lib/analysis/LocationExpanderTest.java", "src/test/java/com/google/devtools/build/lib/analysis/LocationFunctionTest.java", "src/test/py/bazel/runfiles_test.py" ]

[ "keyword_pr_to_issue" ]

[ "As a workaround I wrote the following repository rule:\r\n```python\r\ndef _define_repository_name_impl(repository_ctx):\r\n repository_ctx.file(\"BUILD.bazel\",\r\n \"\"\"package(default_visibility = [\"//visibility:public\"])\r\n \"\"\")\r\n repository_ctx.file(\"defs.bzl\",\r\n \"\"\"main_repository_name=\"{}\"\r\n \"\"\".format(repository_ctx.attr.repo_name))\r\n\r\n_define_repository_name = repository_rule(\r\n implementation = _define_repository_name_impl,\r\n attrs={\"repo_name\": attr.string(mandatory=True)}\r\n )\r\n\r\ndef define_repository_name(repo_name):\r\n _define_repository_name(name = \"main_repository_name\",repo_name = repo_name)\r\n\r\n```\r\n\r\nwhich allows me to load the main repository name by calling:\r\n```python\r\nload(\"@main_repository_name//:defs.bzl\",\"main_repository_name\")\r\n```\r\n\r\nThe `WORKSPACE` file looks like this:\r\n```python\r\nrepository_name = \"xxx\"\r\nworkspace(name = repository_name)\r\n\r\nload(\"@bazel_tooling//rules:define_repository_name.bzl\",\"define_repository_name\")\r\ndefine_repository_name(repository_name)\r\n```\r\n\r\nI don't like it... given that `workspace()` call should be the first line of the WORKSPACE file. But it works (`bazel 2.1.0`)", "hey @laurentlb ... May I ask for your opinion on this matter? :-)", "@or-shachar I have the same issue, did you ever find a resolution to this?", "@or-shachar @jheaff1 Could you describe in more detail what you want to use the name declared in the `WORKSPACE` file for in a macro? Knowing more about the actual problem you are trying to solve would help me answer these questions.", "You most likely want just \"@\" instead. With the recent fixes on repo mapping, \"@workspace_name//:foo\" is equivalent to \"@//:foo\".\r\n\r\n(Unless you're doing exec path or runfiles shenanigans, in which case refer to @fmeum's comment.)", "@fmeum I am trying to write a sh_binary target for the rules_foreign_cc repo that wraps around an executable binary (built by rules_foreign_cc) with shared library dependencies. This would allow such a binary to be used as a build tool, eg the protobuf compiler or various generators part of the Qt framework.\r\n\r\nI would like to use the `rlocation` bash function provided here:\r\n\r\nhttps://github.com/bazelbuild/bazel/blob/a76cf2f4c3aa514dffa1ddc038c196103ae4583c/tools/bash/runfiles/runfiles.bash#L68\r\n\r\nThe `rlocation` bash function requires the workspace name, which I would like to obtain programatically in a macro.\r\n\r\nI’ve tried to use native.repository_name() but that returns `@` rather than the name of the repository.\r\n\r\n", "@jheaff1 Would an analogue of `$(location //foo)` that returns the correct path for `rlocation` help you realize this? I could see that being more user-friendly than yet another \"some type of name of a repo\" thing.", "@fmeum Absolutely that would be very useful. Ive found that using `$(rootpath` rather than `$(location` provides exactly what `$(rlocation` requires apart from the workspace name.", "@lberki I have been thinking about proposing a `$(location ...)` variant that returns the correct path for Rlocation for a while. The way runfiles work with Bzlmod would almost make this a requirement for location expansion to remain useful in `args`. What do you think?", "@fmeum without taking sides on the `$(location) question, it's a feature orthogonal to the one proposed in this bug.\r\n\r\nre: this bug in particular, what is the argument *against* introducing `repository_name()`? Naively, one would think that it's a very logical extension to `package_name()`. @Wyverald maybe you can think of a counter-argument?", "> this bug in particular, what is the argument against introducing repository_name()? Naively, one would think that it's a very logical extension to package_name(). @Wyverald maybe you can think of a counter-argument?\r\n\r\nWe already have `native.repository_name` and it does what it says on the tin (https://bazel.build/rules/lib/native#repository_name). @fmeum's & @jheaff1's question is more about the runfiles/execpath use case, where we sometimes need the *workspace name*, which is only accessible right now in a rule ctx object (https://bazel.build/rules/lib/ctx#workspace_name; also to note, *not* the same as [Label.workspace_name](https://bazel.build/rules/lib/Label#workspace_name)...)", "@Wyverald I'm embarrassed! I specifically looked for the identifier called `repository_name` and didn't find it in the source code. My bad.", "The issue with making the (global) workspace name available in Starlark code is that then the evaluation of BUILD files in external repositories would be a function of the content of the main repository, which isn't too intuitive (it's problematic enough for `ctx.workspace_name()`, but we can't help that because that's how runfiles work at the moment.", "@lberki That plus the fact that having yet another `workspace_name` would be confusing made me think that being able to get Rlocation paths out of location expansion may indeed be the best solution to this issue (see also the comment by @jheaff1). Since `native.repository_name` returns the right identifier for everything *except* runfiles discovery, why not add a way for users to get the full runfiles path right away without having them stitch it together themselves?", "The reason why I was confused is that for some reason, I thought `$(rlocation)` is already a thing in Bazel so I went on a wild goose chase trying to figure out what *else* you want :)\r\n\r\nI think both adding `$(rlocation)` and the same method to File is reasonable. @Wyverald , it looks like you also agree? ", "@lberki `$(rlocation)` seems reasonably easy to do, but `File.rlocation_path` is difficult: Without `--experimental_sibling_repository_layout`, that function needs to know the workspace name. I'm not sure how the API for that would look like, should we just let the user pass in `ctx`?", "Not necessarily; there is `BazelStarlarkContext`, which is available to the Java implementations of Starlark functions without an extra argument and it, weirdly, already contains the repository name of the tools repository. So I could imagine plumbing the workspace name in there that way, too. Rule implementations need to know the name of the main workspace anyway, so it would probably not require too much extra plumbing.", "@lberki I will give that a try. We will have to make a concession though: Using `StarlarkThread` is incompatible with `structField = True`, so this will have to be a function instead.", "Today I Learned. It seems to be an artificial choice and not an inherent limitation of how Starlark works, though, and I don't see an immediate reason for this limitation. @c-parsons , do you remember why you prohibited setting `useStarlarkThread` and `structField` to true at the same time in 73e1016 ?" ]

[]

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

[ "src/main/java/com/google/devtools/build/docgen/templates/be/make-variables.vm", "src/main/java/com/google/devtools/build/lib/analysis/LocationExpander.java", "src/main/java/com/google/devtools/build/lib/analysis/LocationTemplateContext.java" ]

[ "src/main/java/com/google/devtools/build/docgen/templates/be/make-variables.vm", "src/main/java/com/google/devtools/build/lib/analysis/LocationExpander.java", "src/main/java/com/google/devtools/build/lib/analysis/LocationTemplateContext.java" ]

[ "src/test/java/com/google/devtools/build/lib/analysis/LocationExpanderIntegrationTest.java", "src/test/java/com/google/devtools/build/lib/analysis/LocationExpanderTest.java", "src/test/java/com/google/devtools/build/lib/analysis/LocationFunctionTest.java", "src/test/py/bazel/runfiles_test.py" ]

[ "keyword_pr_to_issue" ]

[ "@Wyverald Could you assign me and add the appropriate labels?", "~~@comius Do you expect `java_binary` to be starlarkified in time for Bazel 6? That would simplify the changes to the Java rules.~~ Most likely no longer necessary, as the changes will apply to `buildjar` itself.", "@Wyverald Oops, I missed the use of `Closes ...` in the PR description - we should use `Work towards ...` instead. Could you reopen? For some I lack the permission to do so.", "It's the \"To fix #NNN\" line... GitHub trying to be clever here", "The most likely future seems to be that we'll standardize on Python imports _not_ containing the repository name, in which case bzlmod is a no-op.\r\n\r\nIt's not an optimal solution because it requires an entry in `$PYTHONPATH` for each repository in the transitive closure of the Python binary, but it's the status quo and I'd much prefer thinking up a solution to that (e.g. binary-specific symlink trees that do not look like runfiles trees, dunno) than to standardize on having the repository names in imports because the concept of \"repository\" is specific to Bazel.\r\n\r\nPut differently: the only way for bzlmod to cause trouble for Python is if we decide to require repository names in imports, which seems to be unlikely and it looks like the decision on that will take a while and I prefer not blocking either Bazel 6.0 or the removal of bzlmod from the experimental domain to avoid an unlikely failure mode. And should that happen, we can always paper over it with some sort of import hook.\r\n\r\n/cc @rickeylev ", "@oquenchil I'm thinking about whether we should use a macro to pass `BAZEL_CURRENT_REPOSITORY` to the `Rlocation` function automatically (e.g. do `#define Rlocation(path) RlocationReal(path, BAZEL_CURRENT_REPOSITORY)`). Do you think that would be too intrusive/magical?", "Would users have to write that line or is it part of a file that is automatically generated?\r\n\r\nI don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.", "> Would users have to write that line or is it part of a file that is automatically generated?\r\n> \r\n> I don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.\r\n\r\n@oquenchil I would want to make that magic part of the runfiles library header so that users don't have to write it themselves.\r\n\r\nWhat I am worried about is having an `Rlocation` macro mess with people's existing codebase since there doesn't seem to be a way to put it in a namespace or restrict it to member functions only.", "Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?", "> Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?\r\n\r\nWe can call it something else, we just won't be able to have existing C++ codebases use the new repo mapping-aware lookup automatically if we do so. But of course having to replace all `runfiles::create()` calls with something else won't be too bad.", "If the macro name is Rlocation and that may start clashing with people's code in C++ files then I think it would be prudent to rename it if we decide to go with the macro.\r\n\r\nI guess the call to make here is how much of a breaking change this is compared to the benefit of passing BAZEL_CURRENT_REPOSITORY automatically for users. I would say that the benefit is not too great but as far as I understood the use of Bazel runfiles in C++ code wasn't too wide-spread either?\r\n\r\n> But of course having to replace all runfiles::create() calls with something else won't be too bad.\r\n\r\nHow do you know? Is it based on search results from GH? If a release note would be enough for it not to be too disruptive then let's rename it.", "I have experimented a bit more with APIs while working on the corresponding change for Java (https://github.com/bazelbuild/bazel/pull/16549).\r\n\r\nWhat I now like more than the macro approach is to have users (optionally) pass in a canonical repository name when creating the `Runfiles` instance and default to that repo in `Rlocation` calls, while providing ways to override the default with an overload of `Rlocation` that accepts a repository name as a second parameter. In this way, users only have to pass in `BAZEL_CURRENT_REPOSITORY` only once in the common case, which should render a macro unnecessary.", "Update: All changes relevant for Bazel 6.0.0 are now in review.", "This is mostly done now, with only the bugfix https://github.com/bazelbuild/bazel/pull/16755 and a number of cherry-picks to 6.0.0 missing.", "@Wyverald You can remove this from the \"6.0.0 nice to have\" milestone as the only remaining issue is not tied to a Bazel release (Stardoc).", "Downgrading because the most important points have been addressed; the only outstanding issue is Stardoc, which is undergoing a rewrite currently by @tetromino." ]

[]

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

[ "tools/cpp/runfiles/runfiles_src.cc", "tools/cpp/runfiles/runfiles_src.h", "tools/cpp/runfiles/runfiles_test.cc" ]

[ "tools/cpp/runfiles/runfiles_src.cc", "tools/cpp/runfiles/runfiles_src.h", "tools/cpp/runfiles/runfiles_test.cc" ]

[]

[ "keyword_pr_to_issue" ]

[ "I think that the root cause is a more fundamental inconsistency between how Starlark and native rules collect and declare data dependencies:\r\n\r\nAs per the [Starlark rules guidelines](https://docs.bazel.build/versions/main/skylark/rules.html#runfiles-features-to-avoid), Starlark rules should handle targets in `data`-like attributes by merging in both the files produced by these targets as well as their default runfiles. As a result, Starlark rules often return some files in their `DefaultInfo`'s `files` attribute that are not also contained in its `runfiles`.\r\n\r\nHowever, native rules only collect transitive runfiles and not the output files from rule targets (see [addTargetIncludingFileTargets](https://github.com/bazelbuild/bazel/blob/4c56447fe18e31098ac14a80e2a98253de9c190d/src/main/java/com/google/devtools/build/lib/analysis/Runfiles.java#L931), which only adds output files for file targets). \r\n\r\n@linzhp Could you check whether https://github.com/bazelbuild/bazel/pull/15052 fixes the issue? It adapts the native logic to collect runfiles in the same way as is recommended for Starlark rules, but of course that may very well regress performance for native rules in general. I do not know whether changing this behavior that is so fundamental to Bazel's native rules is realistically possible. But if it doesn't fail too many tests, maybe there is a way to restrict it to Starlark rule targets or otherwise limit its impact on native rule targets.\r\n", "Sorry for the delay. I pull from #15052, and built `//src:bazel-dev`. But when I used the built binary to build `//docs:update` of rules_docker, I got:\r\n\r\n```\r\nDEBUG: /private/var/tmp/_bazel_zplin/d6323f9e3146540840037c805ccd73ff/external/bazel_toolchains/rules/rbe_repo/version_check.bzl:59:14: \r\nCurrent running Bazel is not a release version and one was not defined explicitly in rbe_autoconfig target. Falling back to '4.0.0'\r\nERROR: /private/var/tmp/_bazel_zplin/d6323f9e3146540840037c805ccd73ff/external/bazel_tools/src/tools/launcher/BUILD:9:14: While resolving toolchains for target @bazel_tools//src/tools/launcher:launcher: No matching toolchains found for types @bazel_tools//tools/cpp:toolchain_type. Maybe --incompatible_use_cc_configure_from_rules_cc has been flipped and there is no default C++ toolchain added in the WORKSPACE file? See https://github.com/bazelbuild/bazel/issues/10134 for details and migration instructions.\r\nERROR: Analysis of target '//docs:update' failed; build aborted: \r\nINFO: Elapsed time: 6.538s\r\nINFO: 0 processes.\r\nFAILED: Build did NOT complete successfully (50 packages loaded, 309 targets configured)\r\n```", "@linzhp The build succeeds for me following the instructions from your initial post. Are you on Linux and does `bazel build //...` succeed for you on the repo on the commit prior to mine? The change may cause building new transitive dependencies for `//docs:update`, but if those weren't broken before, they shouldn't be now. The result of rerunning the command with `--toolchain_resolution_debug='.*'` would also be useful to figure out what broke.", "Switched to macOS for Intel and tried your fix again. It worked. Thanks!\r\n\r\nI was on macOS for Apple Silicon before. The commit before yours fails too, but Bazel 5.0.0 works. I saw:\r\n\r\n```\r\nINFO: ToolchainResolution: Type @bazel_tools//tools/cpp:toolchain_type: target platform @local_config_platform//:host: Rejected toolchain @local_config_cc//:cc-compiler-darwin_arm64; mismatching values: arm64\r\n```\r\n", "That looks more like an unrelated regression. Could you create a separate issue and perhaps rerun the test with 5.1.0rc2? 5.1.0 will be released soon.", "Is this bug a dup (though more clearly restating the issue) of https://github.com/bazelbuild/bazel/issues/12348 ?", "> Is this bug a dup (though more clearly restating the issue) of https://github.com/bazelbuild/bazel/issues/12348 ?\n\nPretty sure it is.", "Dupping against #12348.", "Sorry, I missed that #15052 is being actively developed, so I'll keep this issue alive as the author's preferred tracking bug.", "As an FYI, there may be some dependencies on this behavior between the Python rules and some other Google-internal rules. i.e. enabling the flag introduced by this will likely show failures in the Google global regression tests (TGP). See the large comment block in blaze/python/common.bzl#collect_runfiles for more details. The gist, though, is some Google-internal rules have different default outputs vs data/default runfiles, and have come to rely on this so that otherwise unused outputs (some of which can be very large) don't get included in the runfiles.", "@rickeylev Thanks for mentioning this. I added a workaround to the migration guide at https://github.com/bazelbuild/bazel/issues/16654. Could that even work for Google?" ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/analysis/Runfiles.java", "src/main/java/com/google/devtools/build/lib/analysis/config/BuildConfigurationValue.java", "src/main/java/com/google/devtools/build/lib/analysis/config/CoreOptions.java", "src/main/java/com/google/devtools/build/lib/rules/android/AndroidInstrumentationTestBase.java", "src/main/java/com/google/devtools/build/lib/rules/android/AndroidLocalTestBase.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaBinary.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaCommon.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaImport.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/Runfiles.java", "src/main/java/com/google/devtools/build/lib/analysis/config/BuildConfigurationValue.java", "src/main/java/com/google/devtools/build/lib/analysis/config/CoreOptions.java", "src/main/java/com/google/devtools/build/lib/rules/android/AndroidInstrumentationTestBase.java", "src/main/java/com/google/devtools/build/lib/rules/android/AndroidLocalTestBase.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaBinary.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaCommon.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaImport.java" ]

[ "src/main/java/com/google/devtools/build/lib/rules/test/TestSuite.java", "src/test/java/com/google/devtools/build/lib/starlark/StarlarkIntegrationTest.java" ]

[ "keyword_pr_to_issue" ]

[ "@Wyverald ", "My guess (it really is only a guess) is that we have to explicitly add the repo mapping manifest to the input mapping in [`SpawnInputExpander`](https://cs.opensource.google/bazel/bazel/+/bbc221f60bc8c9177470529d85c3e47a5d9aaf21:src/main/java/com/google/devtools/build/lib/exec/SpawnInputExpander.java;l=119;drc=0a23d46976c3fc999d44fbd1e37732ec2442d485) for it to show up in sandboxed and remote execution.", "@bazel-io flag", "Fix cherry picked in https://github.com/bazelbuild/bazel/pull/16733, removing the `potential release blocker` label." ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/analysis/Runfiles.java", "src/main/java/com/google/devtools/build/lib/analysis/RunfilesSupport.java", "src/main/java/com/google/devtools/build/lib/analysis/SingleRunfilesSupplier.java", "src/main/java/com/google/devtools/build/lib/analysis/SourceManifestAction.java", "src/main/java/com/google/devtools/build/lib/analysis/actions/SymlinkTreeAction.java", "src/main/java/com/google/devtools/build/lib/exec/SymlinkTreeStrategy.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaBinary.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/Runfiles.java", "src/main/java/com/google/devtools/build/lib/analysis/RunfilesSupport.java", "src/main/java/com/google/devtools/build/lib/analysis/SingleRunfilesSupplier.java", "src/main/java/com/google/devtools/build/lib/analysis/SourceManifestAction.java", "src/main/java/com/google/devtools/build/lib/analysis/actions/SymlinkTreeAction.java", "src/main/java/com/google/devtools/build/lib/exec/SymlinkTreeStrategy.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaBinary.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/test/TestActionBuilder.java", "src/test/java/com/google/devtools/build/lib/analysis/SingleRunfilesSupplierTest.java", "src/test/java/com/google/devtools/build/lib/analysis/actions/SpawnActionTest.java", "src/test/java/com/google/devtools/build/lib/analysis/actions/SymlinkTreeActionTest.java", "src/test/java/com/google/devtools/build/lib/analysis/util/AnalysisTestUtil.java", "src/test/java/com/google/devtools/build/lib/exec/SymlinkTreeStrategyTest.java", "src/test/java/com/google/devtools/build/lib/rules/cpp/LtoBackendActionTest.java", "src/test/java/com/google/devtools/build/lib/starlark/StarlarkRuleImplementationFunctionsTest.java", "src/test/py/bazel/bzlmod/bazel_module_test.py" ]

[ "keyword_pr_to_issue" ]

[ "@Wyverald Could you assign me and add the appropriate labels?", "~~@comius Do you expect `java_binary` to be starlarkified in time for Bazel 6? That would simplify the changes to the Java rules.~~ Most likely no longer necessary, as the changes will apply to `buildjar` itself.", "@Wyverald Oops, I missed the use of `Closes ...` in the PR description - we should use `Work towards ...` instead. Could you reopen? For some I lack the permission to do so.", "It's the \"To fix #NNN\" line... GitHub trying to be clever here", "The most likely future seems to be that we'll standardize on Python imports _not_ containing the repository name, in which case bzlmod is a no-op.\r\n\r\nIt's not an optimal solution because it requires an entry in `$PYTHONPATH` for each repository in the transitive closure of the Python binary, but it's the status quo and I'd much prefer thinking up a solution to that (e.g. binary-specific symlink trees that do not look like runfiles trees, dunno) than to standardize on having the repository names in imports because the concept of \"repository\" is specific to Bazel.\r\n\r\nPut differently: the only way for bzlmod to cause trouble for Python is if we decide to require repository names in imports, which seems to be unlikely and it looks like the decision on that will take a while and I prefer not blocking either Bazel 6.0 or the removal of bzlmod from the experimental domain to avoid an unlikely failure mode. And should that happen, we can always paper over it with some sort of import hook.\r\n\r\n/cc @rickeylev ", "@oquenchil I'm thinking about whether we should use a macro to pass `BAZEL_CURRENT_REPOSITORY` to the `Rlocation` function automatically (e.g. do `#define Rlocation(path) RlocationReal(path, BAZEL_CURRENT_REPOSITORY)`). Do you think that would be too intrusive/magical?", "Would users have to write that line or is it part of a file that is automatically generated?\r\n\r\nI don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.", "> Would users have to write that line or is it part of a file that is automatically generated?\r\n> \r\n> I don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.\r\n\r\n@oquenchil I would want to make that magic part of the runfiles library header so that users don't have to write it themselves.\r\n\r\nWhat I am worried about is having an `Rlocation` macro mess with people's existing codebase since there doesn't seem to be a way to put it in a namespace or restrict it to member functions only.", "Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?", "> Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?\r\n\r\nWe can call it something else, we just won't be able to have existing C++ codebases use the new repo mapping-aware lookup automatically if we do so. But of course having to replace all `runfiles::create()` calls with something else won't be too bad.", "If the macro name is Rlocation and that may start clashing with people's code in C++ files then I think it would be prudent to rename it if we decide to go with the macro.\r\n\r\nI guess the call to make here is how much of a breaking change this is compared to the benefit of passing BAZEL_CURRENT_REPOSITORY automatically for users. I would say that the benefit is not too great but as far as I understood the use of Bazel runfiles in C++ code wasn't too wide-spread either?\r\n\r\n> But of course having to replace all runfiles::create() calls with something else won't be too bad.\r\n\r\nHow do you know? Is it based on search results from GH? If a release note would be enough for it not to be too disruptive then let's rename it.", "I have experimented a bit more with APIs while working on the corresponding change for Java (https://github.com/bazelbuild/bazel/pull/16549).\r\n\r\nWhat I now like more than the macro approach is to have users (optionally) pass in a canonical repository name when creating the `Runfiles` instance and default to that repo in `Rlocation` calls, while providing ways to override the default with an overload of `Rlocation` that accepts a repository name as a second parameter. In this way, users only have to pass in `BAZEL_CURRENT_REPOSITORY` only once in the common case, which should render a macro unnecessary.", "Update: All changes relevant for Bazel 6.0.0 are now in review.", "This is mostly done now, with only the bugfix https://github.com/bazelbuild/bazel/pull/16755 and a number of cherry-picks to 6.0.0 missing.", "@Wyverald You can remove this from the \"6.0.0 nice to have\" milestone as the only remaining issue is not tied to a Bazel release (Stardoc).", "Downgrading because the most important points have been addressed; the only outstanding issue is Stardoc, which is undergoing a rewrite currently by @tetromino." ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/analysis/Runfiles.java", "src/main/java/com/google/devtools/build/lib/analysis/RunfilesSupport.java", "src/main/java/com/google/devtools/build/lib/analysis/SingleRunfilesSupplier.java", "src/main/java/com/google/devtools/build/lib/analysis/SourceManifestAction.java", "src/main/java/com/google/devtools/build/lib/analysis/actions/SymlinkTreeAction.java", "src/main/java/com/google/devtools/build/lib/exec/SymlinkTreeStrategy.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaBinary.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/Runfiles.java", "src/main/java/com/google/devtools/build/lib/analysis/RunfilesSupport.java", "src/main/java/com/google/devtools/build/lib/analysis/SingleRunfilesSupplier.java", "src/main/java/com/google/devtools/build/lib/analysis/SourceManifestAction.java", "src/main/java/com/google/devtools/build/lib/analysis/actions/SymlinkTreeAction.java", "src/main/java/com/google/devtools/build/lib/exec/SymlinkTreeStrategy.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaBinary.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/test/TestActionBuilder.java", "src/test/java/com/google/devtools/build/lib/analysis/SingleRunfilesSupplierTest.java", "src/test/java/com/google/devtools/build/lib/analysis/actions/SpawnActionTest.java", "src/test/java/com/google/devtools/build/lib/analysis/actions/SymlinkTreeActionTest.java", "src/test/java/com/google/devtools/build/lib/analysis/util/AnalysisTestUtil.java", "src/test/java/com/google/devtools/build/lib/exec/SymlinkTreeStrategyTest.java", "src/test/java/com/google/devtools/build/lib/rules/cpp/LtoBackendActionTest.java", "src/test/java/com/google/devtools/build/lib/starlark/StarlarkRuleImplementationFunctionsTest.java", "src/test/py/bazel/bzlmod/bazel_module_test.py" ]

[ "keyword_pr_to_issue" ]

[ "@Wyverald Could you assign me and add the appropriate labels?", "~~@comius Do you expect `java_binary` to be starlarkified in time for Bazel 6? That would simplify the changes to the Java rules.~~ Most likely no longer necessary, as the changes will apply to `buildjar` itself.", "@Wyverald Oops, I missed the use of `Closes ...` in the PR description - we should use `Work towards ...` instead. Could you reopen? For some I lack the permission to do so.", "It's the \"To fix #NNN\" line... GitHub trying to be clever here", "The most likely future seems to be that we'll standardize on Python imports _not_ containing the repository name, in which case bzlmod is a no-op.\r\n\r\nIt's not an optimal solution because it requires an entry in `$PYTHONPATH` for each repository in the transitive closure of the Python binary, but it's the status quo and I'd much prefer thinking up a solution to that (e.g. binary-specific symlink trees that do not look like runfiles trees, dunno) than to standardize on having the repository names in imports because the concept of \"repository\" is specific to Bazel.\r\n\r\nPut differently: the only way for bzlmod to cause trouble for Python is if we decide to require repository names in imports, which seems to be unlikely and it looks like the decision on that will take a while and I prefer not blocking either Bazel 6.0 or the removal of bzlmod from the experimental domain to avoid an unlikely failure mode. And should that happen, we can always paper over it with some sort of import hook.\r\n\r\n/cc @rickeylev ", "@oquenchil I'm thinking about whether we should use a macro to pass `BAZEL_CURRENT_REPOSITORY` to the `Rlocation` function automatically (e.g. do `#define Rlocation(path) RlocationReal(path, BAZEL_CURRENT_REPOSITORY)`). Do you think that would be too intrusive/magical?", "Would users have to write that line or is it part of a file that is automatically generated?\r\n\r\nI don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.", "> Would users have to write that line or is it part of a file that is automatically generated?\r\n> \r\n> I don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.\r\n\r\n@oquenchil I would want to make that magic part of the runfiles library header so that users don't have to write it themselves.\r\n\r\nWhat I am worried about is having an `Rlocation` macro mess with people's existing codebase since there doesn't seem to be a way to put it in a namespace or restrict it to member functions only.", "Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?", "> Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?\r\n\r\nWe can call it something else, we just won't be able to have existing C++ codebases use the new repo mapping-aware lookup automatically if we do so. But of course having to replace all `runfiles::create()` calls with something else won't be too bad.", "If the macro name is Rlocation and that may start clashing with people's code in C++ files then I think it would be prudent to rename it if we decide to go with the macro.\r\n\r\nI guess the call to make here is how much of a breaking change this is compared to the benefit of passing BAZEL_CURRENT_REPOSITORY automatically for users. I would say that the benefit is not too great but as far as I understood the use of Bazel runfiles in C++ code wasn't too wide-spread either?\r\n\r\n> But of course having to replace all runfiles::create() calls with something else won't be too bad.\r\n\r\nHow do you know? Is it based on search results from GH? If a release note would be enough for it not to be too disruptive then let's rename it.", "I have experimented a bit more with APIs while working on the corresponding change for Java (https://github.com/bazelbuild/bazel/pull/16549).\r\n\r\nWhat I now like more than the macro approach is to have users (optionally) pass in a canonical repository name when creating the `Runfiles` instance and default to that repo in `Rlocation` calls, while providing ways to override the default with an overload of `Rlocation` that accepts a repository name as a second parameter. In this way, users only have to pass in `BAZEL_CURRENT_REPOSITORY` only once in the common case, which should render a macro unnecessary.", "Update: All changes relevant for Bazel 6.0.0 are now in review.", "This is mostly done now, with only the bugfix https://github.com/bazelbuild/bazel/pull/16755 and a number of cherry-picks to 6.0.0 missing.", "@Wyverald You can remove this from the \"6.0.0 nice to have\" milestone as the only remaining issue is not tied to a Bazel release (Stardoc).", "Downgrading because the most important points have been addressed; the only outstanding issue is Stardoc, which is undergoing a rewrite currently by @tetromino." ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/rules/java/JavaCommon.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaInfoBuildHelper.java", "tools/java/runfiles/BUILD", "tools/java/runfiles/BUILD.tools" ]

[ "src/main/java/com/google/devtools/build/lib/rules/java/JavaCommon.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaInfoBuildHelper.java", "tools/java/runfiles/AutoBazelRepository.java", "tools/java/runfiles/AutoBazelRepositoryProcessor.java", "tools/java/runfiles/BUILD", "tools/java/runfiles/BUILD.tools" ]

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

[ "keyword_pr_to_issue" ]

[ "@Wyverald Could you assign me and add the appropriate labels?", "~~@comius Do you expect `java_binary` to be starlarkified in time for Bazel 6? That would simplify the changes to the Java rules.~~ Most likely no longer necessary, as the changes will apply to `buildjar` itself.", "@Wyverald Oops, I missed the use of `Closes ...` in the PR description - we should use `Work towards ...` instead. Could you reopen? For some I lack the permission to do so.", "It's the \"To fix #NNN\" line... GitHub trying to be clever here", "The most likely future seems to be that we'll standardize on Python imports _not_ containing the repository name, in which case bzlmod is a no-op.\r\n\r\nIt's not an optimal solution because it requires an entry in `$PYTHONPATH` for each repository in the transitive closure of the Python binary, but it's the status quo and I'd much prefer thinking up a solution to that (e.g. binary-specific symlink trees that do not look like runfiles trees, dunno) than to standardize on having the repository names in imports because the concept of \"repository\" is specific to Bazel.\r\n\r\nPut differently: the only way for bzlmod to cause trouble for Python is if we decide to require repository names in imports, which seems to be unlikely and it looks like the decision on that will take a while and I prefer not blocking either Bazel 6.0 or the removal of bzlmod from the experimental domain to avoid an unlikely failure mode. And should that happen, we can always paper over it with some sort of import hook.\r\n\r\n/cc @rickeylev ", "@oquenchil I'm thinking about whether we should use a macro to pass `BAZEL_CURRENT_REPOSITORY` to the `Rlocation` function automatically (e.g. do `#define Rlocation(path) RlocationReal(path, BAZEL_CURRENT_REPOSITORY)`). Do you think that would be too intrusive/magical?", "Would users have to write that line or is it part of a file that is automatically generated?\r\n\r\nI don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.", "> Would users have to write that line or is it part of a file that is automatically generated?\r\n> \r\n> I don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.\r\n\r\n@oquenchil I would want to make that magic part of the runfiles library header so that users don't have to write it themselves.\r\n\r\nWhat I am worried about is having an `Rlocation` macro mess with people's existing codebase since there doesn't seem to be a way to put it in a namespace or restrict it to member functions only.", "Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?", "> Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?\r\n\r\nWe can call it something else, we just won't be able to have existing C++ codebases use the new repo mapping-aware lookup automatically if we do so. But of course having to replace all `runfiles::create()` calls with something else won't be too bad.", "If the macro name is Rlocation and that may start clashing with people's code in C++ files then I think it would be prudent to rename it if we decide to go with the macro.\r\n\r\nI guess the call to make here is how much of a breaking change this is compared to the benefit of passing BAZEL_CURRENT_REPOSITORY automatically for users. I would say that the benefit is not too great but as far as I understood the use of Bazel runfiles in C++ code wasn't too wide-spread either?\r\n\r\n> But of course having to replace all runfiles::create() calls with something else won't be too bad.\r\n\r\nHow do you know? Is it based on search results from GH? If a release note would be enough for it not to be too disruptive then let's rename it.", "I have experimented a bit more with APIs while working on the corresponding change for Java (https://github.com/bazelbuild/bazel/pull/16549).\r\n\r\nWhat I now like more than the macro approach is to have users (optionally) pass in a canonical repository name when creating the `Runfiles` instance and default to that repo in `Rlocation` calls, while providing ways to override the default with an overload of `Rlocation` that accepts a repository name as a second parameter. In this way, users only have to pass in `BAZEL_CURRENT_REPOSITORY` only once in the common case, which should render a macro unnecessary.", "Update: All changes relevant for Bazel 6.0.0 are now in review.", "This is mostly done now, with only the bugfix https://github.com/bazelbuild/bazel/pull/16755 and a number of cherry-picks to 6.0.0 missing.", "@Wyverald You can remove this from the \"6.0.0 nice to have\" milestone as the only remaining issue is not tied to a Bazel release (Stardoc).", "Downgrading because the most important points have been addressed; the only outstanding issue is Stardoc, which is undergoing a rewrite currently by @tetromino." ]

[]

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

[ "tools/java/runfiles/Runfiles.java" ]

[ "tools/java/runfiles/Runfiles.java" ]

[ "src/test/py/bazel/bzlmod/bazel_module_test.py", "tools/java/runfiles/testing/BUILD", "tools/java/runfiles/testing/RunfilesTest.java" ]

[ "keyword_pr_to_issue" ]

[ "@Wyverald Could you assign me and add the appropriate labels?", "~~@comius Do you expect `java_binary` to be starlarkified in time for Bazel 6? That would simplify the changes to the Java rules.~~ Most likely no longer necessary, as the changes will apply to `buildjar` itself.", "@Wyverald Oops, I missed the use of `Closes ...` in the PR description - we should use `Work towards ...` instead. Could you reopen? For some I lack the permission to do so.", "It's the \"To fix #NNN\" line... GitHub trying to be clever here", "The most likely future seems to be that we'll standardize on Python imports _not_ containing the repository name, in which case bzlmod is a no-op.\r\n\r\nIt's not an optimal solution because it requires an entry in `$PYTHONPATH` for each repository in the transitive closure of the Python binary, but it's the status quo and I'd much prefer thinking up a solution to that (e.g. binary-specific symlink trees that do not look like runfiles trees, dunno) than to standardize on having the repository names in imports because the concept of \"repository\" is specific to Bazel.\r\n\r\nPut differently: the only way for bzlmod to cause trouble for Python is if we decide to require repository names in imports, which seems to be unlikely and it looks like the decision on that will take a while and I prefer not blocking either Bazel 6.0 or the removal of bzlmod from the experimental domain to avoid an unlikely failure mode. And should that happen, we can always paper over it with some sort of import hook.\r\n\r\n/cc @rickeylev ", "@oquenchil I'm thinking about whether we should use a macro to pass `BAZEL_CURRENT_REPOSITORY` to the `Rlocation` function automatically (e.g. do `#define Rlocation(path) RlocationReal(path, BAZEL_CURRENT_REPOSITORY)`). Do you think that would be too intrusive/magical?", "Would users have to write that line or is it part of a file that is automatically generated?\r\n\r\nI don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.", "> Would users have to write that line or is it part of a file that is automatically generated?\r\n> \r\n> I don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.\r\n\r\n@oquenchil I would want to make that magic part of the runfiles library header so that users don't have to write it themselves.\r\n\r\nWhat I am worried about is having an `Rlocation` macro mess with people's existing codebase since there doesn't seem to be a way to put it in a namespace or restrict it to member functions only.", "Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?", "> Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?\r\n\r\nWe can call it something else, we just won't be able to have existing C++ codebases use the new repo mapping-aware lookup automatically if we do so. But of course having to replace all `runfiles::create()` calls with something else won't be too bad.", "If the macro name is Rlocation and that may start clashing with people's code in C++ files then I think it would be prudent to rename it if we decide to go with the macro.\r\n\r\nI guess the call to make here is how much of a breaking change this is compared to the benefit of passing BAZEL_CURRENT_REPOSITORY automatically for users. I would say that the benefit is not too great but as far as I understood the use of Bazel runfiles in C++ code wasn't too wide-spread either?\r\n\r\n> But of course having to replace all runfiles::create() calls with something else won't be too bad.\r\n\r\nHow do you know? Is it based on search results from GH? If a release note would be enough for it not to be too disruptive then let's rename it.", "I have experimented a bit more with APIs while working on the corresponding change for Java (https://github.com/bazelbuild/bazel/pull/16549).\r\n\r\nWhat I now like more than the macro approach is to have users (optionally) pass in a canonical repository name when creating the `Runfiles` instance and default to that repo in `Rlocation` calls, while providing ways to override the default with an overload of `Rlocation` that accepts a repository name as a second parameter. In this way, users only have to pass in `BAZEL_CURRENT_REPOSITORY` only once in the common case, which should render a macro unnecessary.", "Update: All changes relevant for Bazel 6.0.0 are now in review.", "This is mostly done now, with only the bugfix https://github.com/bazelbuild/bazel/pull/16755 and a number of cherry-picks to 6.0.0 missing.", "@Wyverald You can remove this from the \"6.0.0 nice to have\" milestone as the only remaining issue is not tied to a Bazel release (Stardoc).", "Downgrading because the most important points have been addressed; the only outstanding issue is Stardoc, which is undergoing a rewrite currently by @tetromino." ]

[]

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

[ "tools/cpp/runfiles/runfiles_src.cc", "tools/cpp/runfiles/runfiles_src.h", "tools/cpp/runfiles/runfiles_test.cc" ]

[ "tools/cpp/runfiles/runfiles_src.cc", "tools/cpp/runfiles/runfiles_src.h", "tools/cpp/runfiles/runfiles_test.cc" ]

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

[ "keyword_pr_to_issue" ]

[ "@Wyverald Could you assign me and add the appropriate labels?", "~~@comius Do you expect `java_binary` to be starlarkified in time for Bazel 6? That would simplify the changes to the Java rules.~~ Most likely no longer necessary, as the changes will apply to `buildjar` itself.", "@Wyverald Oops, I missed the use of `Closes ...` in the PR description - we should use `Work towards ...` instead. Could you reopen? For some I lack the permission to do so.", "It's the \"To fix #NNN\" line... GitHub trying to be clever here", "The most likely future seems to be that we'll standardize on Python imports _not_ containing the repository name, in which case bzlmod is a no-op.\r\n\r\nIt's not an optimal solution because it requires an entry in `$PYTHONPATH` for each repository in the transitive closure of the Python binary, but it's the status quo and I'd much prefer thinking up a solution to that (e.g. binary-specific symlink trees that do not look like runfiles trees, dunno) than to standardize on having the repository names in imports because the concept of \"repository\" is specific to Bazel.\r\n\r\nPut differently: the only way for bzlmod to cause trouble for Python is if we decide to require repository names in imports, which seems to be unlikely and it looks like the decision on that will take a while and I prefer not blocking either Bazel 6.0 or the removal of bzlmod from the experimental domain to avoid an unlikely failure mode. And should that happen, we can always paper over it with some sort of import hook.\r\n\r\n/cc @rickeylev ", "@oquenchil I'm thinking about whether we should use a macro to pass `BAZEL_CURRENT_REPOSITORY` to the `Rlocation` function automatically (e.g. do `#define Rlocation(path) RlocationReal(path, BAZEL_CURRENT_REPOSITORY)`). Do you think that would be too intrusive/magical?", "Would users have to write that line or is it part of a file that is automatically generated?\r\n\r\nI don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.", "> Would users have to write that line or is it part of a file that is automatically generated?\r\n> \r\n> I don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.\r\n\r\n@oquenchil I would want to make that magic part of the runfiles library header so that users don't have to write it themselves.\r\n\r\nWhat I am worried about is having an `Rlocation` macro mess with people's existing codebase since there doesn't seem to be a way to put it in a namespace or restrict it to member functions only.", "Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?", "> Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?\r\n\r\nWe can call it something else, we just won't be able to have existing C++ codebases use the new repo mapping-aware lookup automatically if we do so. But of course having to replace all `runfiles::create()` calls with something else won't be too bad.", "If the macro name is Rlocation and that may start clashing with people's code in C++ files then I think it would be prudent to rename it if we decide to go with the macro.\r\n\r\nI guess the call to make here is how much of a breaking change this is compared to the benefit of passing BAZEL_CURRENT_REPOSITORY automatically for users. I would say that the benefit is not too great but as far as I understood the use of Bazel runfiles in C++ code wasn't too wide-spread either?\r\n\r\n> But of course having to replace all runfiles::create() calls with something else won't be too bad.\r\n\r\nHow do you know? Is it based on search results from GH? If a release note would be enough for it not to be too disruptive then let's rename it.", "I have experimented a bit more with APIs while working on the corresponding change for Java (https://github.com/bazelbuild/bazel/pull/16549).\r\n\r\nWhat I now like more than the macro approach is to have users (optionally) pass in a canonical repository name when creating the `Runfiles` instance and default to that repo in `Rlocation` calls, while providing ways to override the default with an overload of `Rlocation` that accepts a repository name as a second parameter. In this way, users only have to pass in `BAZEL_CURRENT_REPOSITORY` only once in the common case, which should render a macro unnecessary.", "Update: All changes relevant for Bazel 6.0.0 are now in review.", "This is mostly done now, with only the bugfix https://github.com/bazelbuild/bazel/pull/16755 and a number of cherry-picks to 6.0.0 missing.", "@Wyverald You can remove this from the \"6.0.0 nice to have\" milestone as the only remaining issue is not tied to a Bazel release (Stardoc).", "Downgrading because the most important points have been addressed; the only outstanding issue is Stardoc, which is undergoing a rewrite currently by @tetromino." ]

[]

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

[ "tools/bash/runfiles/runfiles.bash", "tools/bash/runfiles/runfiles_test.bash" ]

[ "tools/bash/runfiles/runfiles.bash", "tools/bash/runfiles/runfiles_test.bash" ]

[ "src/test/py/bazel/bzlmod/bazel_module_test.py", "src/test/shell/bazel/BUILD", "src/test/shell/bazel/bazel_rules_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "Dupe of https://github.com/bazelbuild/bazel/issues/15831" ]

[]

[ "more data needed", "type: support / not a bug (process)", "team-OSS" ]

[ "src/main/cpp/blaze.cc" ]

[ "src/main/cpp/blaze.cc" ]

[]

[ "keyword_pr_to_issue" ]

[ "Duplicate of https://github.com/bazelbuild/bazel/issues/14548", "Hello @dmivankov, Can you try the above with the latest release[ 5.2.0](https://github.com/bazelbuild/bazel/releases/tag/5.2.0). Thanks! ", "I'm not dmivankov, but I tried compling v 5.2.0 with openjdk17 and am getting the same crash. Here's a full build log:\r\n\r\n```sh\r\n\r\nBuilding Bazel from scratch../opt/openjdk-bin-17.0.3_p7/bin/javac -classpath derived/jars/zstd-jni/libzstd-jni.jar:derived/jars/com_google_protobuf/java/core/liblite.jar:derived/jars/com_google_protobuf/java/core/libcore.jar:derived/jars/com_google_protobuf/java/util/libutil.jar:third_party/checker_framework_annotations/checker-qual-3.2.0-sources.jar:third_party/checker_framework_annotations/checker-qual-3.2.0.jar:third_party/javax_annotations/javax.annotation-api-1.3.2-sources.jar:third_party/javax_annotations/javax.annotation-api-1.3.2.jar:third_party/truth/truth-1.0.1.jar:third_party/bazel_bootstrap/libserialization-constant-annotation.jar:third_party/bazel_bootstrap/libserialization-constant-processor.jar:third_party/bazel_bootstrap/libserialization.jar:third_party/bazel_bootstrap/libautocodec-processor.jar:third_party/bazel_bootstrap/libunsafe-provider.jar:third_party/bazel_bootstrap/libautocodec-annotation.jar:third_party/bazel_bootstrap/libserialization-processor-util.jar:third_party/bazel_bootstrap/libregistered-singleton.jar:third_party/turbine/turbine_direct.jar:third_party/jcip_annotations/jcip-annotations-1.0-1.jar:third_party/apache_commons_pool2/commons-pool2-2.8.0.jar:third_party/apache_commons_compress/apache-commons-compress-1.19.jar:third_party/animal_sniffer/animal-sniffer-annotations-1.14.jar:third_party/flogger/flogger-system-backend-0.5.1.jar:third_party/flogger/flogger-0.5.1.jar:third_party/flogger/google-extensions-0.5.1.jar:third_party/jetbrains_annotations/annotations-13.0.jar:third_party/truth8/truth-java8-extension-1.0.1.jar:third_party/objenesis/objenesis-1_3.jar:third_party/api_client/google-api-client-1.22.0.jar:third_party/api_client/google-api-client-gson-1.22.0.jar:third_party/api_client/google-http-client-1.22.0-SNAPSHOT.jar:third_party/api_client/google-http-client-gson-1.22.0.jar:third_party/tomcat_annotations_api/tomcat-annotations-api-8.0.5.jar:third_party/jimfs/jimfs-1.1.jar:third_party/hamcrest/hamcrest-core-1.3.jar:third_party/diffutils/diffutils-1.3.0.jar:third_party/caffeine/caffeine-2.9.2.jar:third_party/checker_framework_javacutil/javacutil-3.2.0-sources.jar:third_party/checker_framework_javacutil/javacutil-3.2.0.jar:third_party/jsr330_inject/javax.inject.jar:third_party/java/jdk/langtools/jdk_compiler-src.jar:third_party/java/jdk/langtools/jdk_compiler.jar:third_party/java/jdk/langtools/java_compiler.jar:third_party/java/jdk/langtools/java_compiler-src.jar:third_party/java/jdk/langtools/javac-9+181-r4173-1.jar:third_party/java/proguard/proguard6.2.2/lib/proguard.jar:third_party/java/proguard/proguard6.2.2/lib/proguardgui.jar:third_party/java/proguard/proguard6.2.2/lib/retrace.jar:third_party/java/javapoet/javapoet-1.8.0.jar:third_party/java/android_databinding/v3_4_0/base.jar:third_party/java/android_databinding/v3_4_0/exec.jar:third_party/java/android_databinding/v3_4_0/x_common.jar:third_party/java/android_databinding/v3_4_0/compiler.jar:third_party/java/android_databinding/v3_4_0/compiler_common.jar:third_party/java/jacoco/org.jacoco.ant-0.8.7-sources.jar:third_party/java/jacoco/org.jacoco.agent-0.8.7.jar:third_party/java/jacoco/org.jacoco.ant-0.8.7-nodeps.jar:third_party/java/jacoco/org.jacoco.agent-0.8.7-sources.jar:third_party/java/jacoco/org.jacoco.core-0.8.7.jar:third_party/java/jacoco/jacocoagent-0.8.7.jar:third_party/java/jacoco/org.jacoco.report-0.8.7-sources.jar:third_party/java/jacoco/org.jacoco.ant-0.8.7.jar:third_party/java/jacoco/org.jacoco.report-0.8.7.jar:third_party/java/jacoco/org.jacoco.core-0.8.7-sources.jar:third_party/java/j2objc-annotations/j2objc-annotations-1.1.jar:third_party/java/jcommander/jcommander-1.48.jar:third_party/juniversalchardet/juniversalchardet-1.0.3.jar:third_party/guava/failureaccess-1.0.1.jar:third_party/checker_framework_dataflow/dataflow-errorprone-3.15.0.jar:third_party/checker_framework_dataflow/dataflow-shaded-3.11.0.jar:third_party/checker_framework_dataflow/dataflow-errorprone-3.15.0-sources.jar:third_party/checker_framework_dataflow/dataflow-shaded-3.11.0-sources.jar:third_party/jsr305/jsr-305.jar:third_party/asm/asm-util-9.2.jar:third_party/asm/asm-util-9.2-sources.jar:third_party/asm/asm-commons-9.2-sources.jar:third_party/asm/asm-tree-9.2.jar:third_party/asm/asm-analysis-9.2-sources.jar:third_party/asm/asm-commons-9.2.jar:third_party/asm/asm-analysis-9.2.jar:third_party/asm/asm-9.2-sources.jar:third_party/asm/asm-tree-9.2-sources.jar:third_party/asm/asm-9.2.jar:third_party/apache_velocity/velocity-1.7.jar:third_party/mockito/mockito-core-3.12.4.jar:third_party/allocation_instrumenter/java-allocation-instrumenter-3.3.0.jar:third_party/antlr/antlr4-4.5.3.jar:third_party/perfmark/perfmark-api-0.19.0.jar:third_party/error_prone/error_prone_core-2.9.0.jar:third_party/error_prone/threeten-extra-1.5.0.jar:third_party/error_prone/error_prone_type_annotations-2.9.0.jar:third_party/error_prone/error_prone_annotation-2.9.0.jar:third_party/error_prone/error_prone_annotations-2.9.0.jar:third_party/error_prone/error_prone_check_api-2.9.0.jar:third_party/ijar/test/libwrongcentraldir.jar:third_party/ijar/test/jar-without-manifest.jar:third_party/ijar/test/sealed/sealed.jar:third_party/ijar/test/nestmates/nestmates.jar:third_party/ijar/test/records/records.jar:third_party/ijar/test/jar-with-manifest.jar:third_party/ijar/test/jar-with-manifest-and-target-label.jar:third_party/javax_activation/javax.activation-api-1.2.0-sources.jar:third_party/javax_activation/javax.activation-api-1.2.0.jar:third_party/apache_commons_collections/commons-collections-3.2.2.jar:third_party/xz/xz-1.9.jar:third_party/apache_commons_lang/commons-lang-2.6.jar:third_party/junit/junit-4.13.jar:third_party/java-diff-utils/java-diff-utils-4.0.jar:third_party/rxjava3/rxjava-3.1.2.jar:third_party/apache_commons_logging/commons-logging-1.1.1.jar:third_party/jaxb/2.3.0/jaxb-core-2.3.0-sources.jar:third_party/jaxb/2.3.0/jaxb-core-2.3.0.jar:third_party/jaxb/2.3.0/jaxb-api-2.3.0-sources.jar:third_party/jaxb/2.3.0/jaxb-impl-2.3.0.jar:third_party/jaxb/2.3.0/jaxb-api-2.3.0.jar:third_party/jaxb/2.3.0/jaxb-impl-2.3.0-sources.jar:third_party/jaxb/jaxb-api-2.3.1.jar:third_party/jaxb/jaxb-api-2.3.1-patched.jar:third_party/jaxb/jaxb-api-2.3.1-sources.jar:third_party/hungarian_algorithm/software-and-algorithms-1.0.jar:third_party/hungarian_algorithm/software-and-algorithms-1.0-src.jar:third_party/kotlin_stdlib/kotlin-stdlib-jdk8-1.3.21.jar:third_party/kotlin_stdlib/kotlin-stdlib-1.3.21.jar:third_party/kotlin_stdlib/kotlin-stdlib-jdk7-1.3.21.jar:third_party/kotlin_stdlib/kotlin-stdlib-common-1.3.21.jar:third_party/apache_commons_io/commons-io-2.4.jar:third_party/reactive_streams/reactive-streams-1.0.3.jar:third_party/netty/netty-transport-native-unix-common-4.1.69.Final-linux-aarch_64.jar:third_party/netty/netty-handler-proxy-4.1.69.Final.jar:third_party/netty/netty-codec-http2-4.1.69.Final.jar:third_party/netty/netty-buffer-4.1.69.Final.jar:third_party/netty/netty-handler-4.1.69.Final.jar:third_party/netty/netty-transport-native-unix-common-4.1.69.Final-osx-x86_64.jar:third_party/netty/netty-transport-native-epoll-4.1.69.Final-linux-x86_64.jar:third_party/netty/netty-transport-native-unix-common-4.1.69.Final-osx-aarch_64.jar:third_party/netty/netty-transport-native-unix-common-4.1.69.Final.jar:third_party/netty/netty-transport-sctp-4.1.69.Final.jar:third_party/netty/netty-transport-native-kqueue-4.1.69.Final.jar:third_party/netty/netty-transport-native-unix-common-4.1.69.Final-linux-x86_64.jar:third_party/netty/netty-common-4.1.69.Final.jar:third_party/netty/netty-resolver-4.1.69.Final.jar:third_party/netty/netty-transport-native-epoll-4.1.69.Final.jar:third_party/netty/netty-resolver-dns-4.1.69.Final.jar:third_party/netty/netty-transport-4.1.69.Final.jar:third_party/netty/netty-codec-http-4.1.69.Final.jar:third_party/netty/netty-transport-native-epoll-4.1.69.Final-linux-aarch_64.jar:third_party/netty/netty-transport-native-kqueue-4.1.69.Final-osx-aarch_64.jar:third_party/netty/netty-transport-native-kqueue-4.1.69.Final-osx-x86_64.jar:third_party/netty/netty-codec-4.1.69.Final.jar:third_party/auto/auto-service-annotations-1.0-rc7.jar:third_party/auto/auto-value-1.8.2.jar:third_party/auto/auto-service-1.0-rc7.jar:third_party/auto/auto-value-annotations-1.8.2.jar:third_party/auto/auto-common-1.1.2.jar:third_party/truth_proto/truth-proto-extension-1.0.1.jar:third_party/truth_proto/truth-liteproto-extension-1.0.1.jar:third_party/pcollections/pcollections-2.1.2.jar:third_party/jetifier/jetifier-core-1.0.0-beta04.jar:third_party/compile_testing/compile-testing-0.18.jar:third_party/auth/google-auth-library-oauth2-http-1.6.0.jar:third_party/auth/google-auth-library-credentials-1.6.0.jar:third_party/android_common/com.android_annotations_25.0.0.jar:third_party/android_common/com.android.tools.build_builder-test-api_2.0.0.jar:third_party/android_common/com.android.tools_sdk-common_25.0.0-patched-stripped.jar:third_party/android_common/com.android.tools.lint_lint-checks_25.0.0.jar:third_party/android_common/com.android.tools_dvlib_25.0.0.jar:third_party/android_common/com.android.tools.build_builder-model_2.0.0.jar:third_party/android_common/com.android.tools_sdklib_25.0.0.jar:third_party/android_common/com.android.tools.build_manifest-merger_25.0.0-patched.jar:third_party/android_common/com.android.tools_common_25.0.0.jar:third_party/android_common/com.android.tools.layoutlib_layoutlib_26.1.2-stripped.jar:third_party/android_common/com.android.tools_sdk-common_25.0.0-patched.jar:third_party/android_common/com.android.tools.layoutlib_layoutlib_26.1.2.jar:third_party/android_common/com.android.tools.build_builder_2.0.0.jar:third_party/android_common/com.android.tools.external.lombok_lombok-ast_0.2.3.jar:third_party/android_common/com.android.tools.lint_lint-api_25.0.0.jar:third_party/android_common/com.android.tools_repository_25.0.0.jar:third_party/opencensus/opencensus-api-0.24.0.jar:third_party/opencensus/opencensus-contrib-grpc-metrics-0.24.0.jar:third_party/gson/gson-2.8.0.jar:third_party/apache_commons_codec/commons-codec-1.10.jar:third_party/bytebuddy/byte-buddy-agent-1.9.7.jar:third_party/bytebuddy/byte-buddy-1.9.7.jar::third_party/grpc/grpc-core-1.41.0.jar:third_party/grpc/grpc-stub-1.41.0.jar:third_party/grpc/grpc-context-1.41.0.jar:third_party/grpc/grpc-protobuf-lite-1.41.0.jar:third_party/grpc/grpc-protobuf-1.41.0.jar:third_party/grpc/grpc-api-1.41.0.jar:third_party/grpc/grpc-netty-1.41.0.jar:third_party/grpc/grpc-auth-1.41.0.jar::third_party/guava/guava-30.1.1-jre.jar:third_party/guava/guava-testlib-30.1.1-jre.jar::/var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61 -sourcepath src/java_tools/singlejar/java/com/google/devtools/build/zip:src/main/java:tools/java/runfiles:/var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/src -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes -source 1.8 -target 1.8 -encoding UTF-8 @/var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_owYrdOr7/param\r\nwarning: [options] bootstrap class path not set in conjunction with -source 8\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:17: warning: Signal is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Signal;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:18: warning: SignalHandler is internal proprietary API and may be removed in a future release\r\nimport sun.misc.SignalHandler;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/skyframe/serialization/DynamicCodec.java:30: warning: ReflectionFactory is internal proprietary API and may be removed in a future release\r\nimport sun.reflect.ReflectionFactory;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/skyframe/serialization/EnumMapCodec.java:23: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/StringUnsafe.java:21: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:18: warning: [removal] AccessController in java.security has been deprecated and marked for removal\r\nimport java.security.AccessController;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:21: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:17: warning: Signal is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Signal;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:18: warning: SignalHandler is internal proprietary API and may be removed in a future release\r\nimport sun.misc.SignalHandler;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/skyframe/serialization/DynamicCodec.java:30: warning: ReflectionFactory is internal proprietary API and may be removed in a future release\r\nimport sun.reflect.ReflectionFactory;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/skyframe/serialization/EnumMapCodec.java:23: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/StringUnsafe.java:21: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:18: warning: [removal] AccessController in java.security has been deprecated and marked for removal\r\nimport java.security.AccessController;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:21: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:17: warning: Signal is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Signal;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:18: warning: SignalHandler is internal proprietary API and may be removed in a future release\r\nimport sun.misc.SignalHandler;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/skyframe/serialization/DynamicCodec.java:30: warning: ReflectionFactory is internal proprietary API and may be removed in a future release\r\nimport sun.reflect.ReflectionFactory;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/skyframe/serialization/EnumMapCodec.java:23: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/StringUnsafe.java:21: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:18: warning: [removal] AccessController in java.security has been deprecated and marked for removal\r\nimport java.security.AccessController;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:21: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:17: warning: Signal is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Signal;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:18: warning: SignalHandler is internal proprietary API and may be removed in a future release\r\nimport sun.misc.SignalHandler;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/skyframe/serialization/DynamicCodec.java:30: warning: ReflectionFactory is internal proprietary API and may be removed in a future release\r\nimport sun.reflect.ReflectionFactory;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/skyframe/serialization/EnumMapCodec.java:23: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/StringUnsafe.java:21: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:18: warning: [removal] AccessController in java.security has been deprecated and marked for removal\r\nimport java.security.AccessController;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:21: warning: Unsafe is internal proprietary API and may be removed in a future release\r\nimport sun.misc.Unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:29: warning: Signal is internal proprietary API and may be removed in a future release\r\n private static final Signal SIGINT = new Signal(\"INT\");\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:29: warning: Signal is internal proprietary API and may be removed in a future release\r\n private static final Signal SIGINT = new Signal(\"INT\");\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:29: warning: Signal is internal proprietary API and may be removed in a future release\r\n private static final Signal SIGINT = new Signal(\"INT\");\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:29: warning: Signal is internal proprietary API and may be removed in a future release\r\n private static final Signal SIGINT = new Signal(\"INT\");\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:29: warning: Signal is internal proprietary API and may be removed in a future release\r\n private static final Signal SIGINT = new Signal(\"INT\");\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:31: warning: SignalHandler is internal proprietary API and may be removed in a future release\r\n private SignalHandler oldHandler;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:31: warning: SignalHandler is internal proprietary API and may be removed in a future release\r\n private SignalHandler oldHandler;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:31: warning: SignalHandler is internal proprietary API and may be removed in a future release\r\n private SignalHandler oldHandler;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:31: warning: SignalHandler is internal proprietary API and may be removed in a future release\r\n private SignalHandler oldHandler;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:39: warning: Signal is internal proprietary API and may be removed in a future release\r\n this.oldHandler = Signal.handle(SIGINT, signal -&gt; run());\r\n ^\r\nwarning: Signal is internal proprietary API and may be removed in a future release\r\nsrc/main/java/com/google/devtools/build/lib/server/signal/InterruptSignalHandler.java:45: warning: Signal is internal proprietary API and may be removed in a future release\r\n Signal.handle(SIGINT, oldHandler);\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/skyframe/serialization/DynamicCodec.java:276: warning: ReflectionFactory is internal proprietary API and may be removed in a future release\r\n ReflectionFactory.getReflectionFactory()\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/skyframe/serialization/EnumMapCodec.java:59: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n Unsafe unsafe = UnsafeProvider.getInstance();\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/StringUnsafe.java:38: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private final Unsafe unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/StringUnsafe.java:38: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private final Unsafe unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/StringUnsafe.java:38: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private final Unsafe unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/StringUnsafe.java:38: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private final Unsafe unsafe;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:31: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private static final Unsafe UNSAFE = getUnsafe();\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:31: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private static final Unsafe UNSAFE = getUnsafe();\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:31: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private static final Unsafe UNSAFE = getUnsafe();\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:31: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private static final Unsafe UNSAFE = getUnsafe();\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:33: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n public static Unsafe getInstance() {\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:33: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n public static Unsafe getInstance() {\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:33: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n public static Unsafe getInstance() {\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:33: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n public static Unsafe getInstance() {\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:43: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private static Unsafe getUnsafe() {\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:43: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private static Unsafe getUnsafe() {\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:43: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private static Unsafe getUnsafe() {\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:43: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n private static Unsafe getUnsafe() {\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:48: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n new PrivilegedExceptionAction&lt;Unsafe&gt;() {\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:50: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n public Unsafe run() throws Exception {\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:51: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n Class&lt;Unsafe&gt; unsafeClass = Unsafe.class;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:51: warning: Unsafe is internal proprietary API and may be removed in a future release\r\n Class&lt;Unsafe&gt; unsafeClass = Unsafe.class;\r\n ^\r\nsrc/main/java/com/google/devtools/build/lib/unsafe/UnsafeProvider.java:47: warning: [removal] AccessController in java.security has been deprecated and marked for removal\r\n return AccessController.doPrivileged(\r\n ^\r\nwarning: Implicitly compiled files were not subject to annotation processing.\r\n Use -proc:none to disable annotation processing or -implicit to specify a policy for implicit compilation.\r\nNote: Some input files use or override a deprecated API.\r\nNote: Recompile with -Xlint:deprecation for details.\r\nNote: Some input files use unchecked or unsafe operations.\r\nNote: Recompile with -Xlint:unchecked for details.\r\n65 warnings\r\n.unzip -qn derived/jars/zstd-jni/libzstd-jni.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn derived/jars/com_google_protobuf/java/core/liblite.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn derived/jars/com_google_protobuf/java/core/libcore.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn derived/jars/com_google_protobuf/java/util/libutil.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/checker_framework_annotations/checker-qual-3.2.0-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/checker_framework_annotations/checker-qual-3.2.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/javax_annotations/javax.annotation-api-1.3.2-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/javax_annotations/javax.annotation-api-1.3.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/truth/truth-1.0.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/bazel_bootstrap/libserialization-constant-annotation.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/bazel_bootstrap/libserialization-constant-processor.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/bazel_bootstrap/libserialization.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/bazel_bootstrap/libautocodec-processor.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/bazel_bootstrap/libunsafe-provider.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/bazel_bootstrap/libautocodec-annotation.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/bazel_bootstrap/libserialization-processor-util.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/bazel_bootstrap/libregistered-singleton.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/turbine/turbine_direct.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jcip_annotations/jcip-annotations-1.0-1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/apache_commons_pool2/commons-pool2-2.8.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/apache_commons_compress/apache-commons-compress-1.19.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/animal_sniffer/animal-sniffer-annotations-1.14.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/flogger/flogger-system-backend-0.5.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/flogger/flogger-0.5.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/flogger/google-extensions-0.5.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jetbrains_annotations/annotations-13.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/truth8/truth-java8-extension-1.0.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/objenesis/objenesis-1_3.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/api_client/google-api-client-1.22.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/api_client/google-api-client-gson-1.22.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/api_client/google-http-client-1.22.0-SNAPSHOT.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/api_client/google-http-client-gson-1.22.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/tomcat_annotations_api/tomcat-annotations-api-8.0.5.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jimfs/jimfs-1.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/hamcrest/hamcrest-core-1.3.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/diffutils/diffutils-1.3.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/caffeine/caffeine-2.9.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/checker_framework_javacutil/javacutil-3.2.0-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/checker_framework_javacutil/javacutil-3.2.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jsr330_inject/javax.inject.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jdk/langtools/jdk_compiler-src.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jdk/langtools/jdk_compiler.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jdk/langtools/java_compiler.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jdk/langtools/java_compiler-src.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jdk/langtools/javac-9+181-r4173-1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/proguard/proguard6.2.2/lib/proguard.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/proguard/proguard6.2.2/lib/proguardgui.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/proguard/proguard6.2.2/lib/retrace.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/javapoet/javapoet-1.8.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/android_databinding/v3_4_0/base.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/android_databinding/v3_4_0/exec.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/android_databinding/v3_4_0/x_common.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/android_databinding/v3_4_0/compiler.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/android_databinding/v3_4_0/compiler_common.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jacoco/org.jacoco.ant-0.8.7-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jacoco/org.jacoco.agent-0.8.7.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jacoco/org.jacoco.ant-0.8.7-nodeps.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jacoco/org.jacoco.agent-0.8.7-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jacoco/org.jacoco.core-0.8.7.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jacoco/jacocoagent-0.8.7.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jacoco/org.jacoco.report-0.8.7-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jacoco/org.jacoco.ant-0.8.7.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jacoco/org.jacoco.report-0.8.7.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jacoco/org.jacoco.core-0.8.7-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/j2objc-annotations/j2objc-annotations-1.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java/jcommander/jcommander-1.48.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/juniversalchardet/juniversalchardet-1.0.3.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/guava/failureaccess-1.0.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/checker_framework_dataflow/dataflow-errorprone-3.15.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/checker_framework_dataflow/dataflow-shaded-3.11.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/checker_framework_dataflow/dataflow-errorprone-3.15.0-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/checker_framework_dataflow/dataflow-shaded-3.11.0-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jsr305/jsr-305.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/asm/asm-util-9.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/asm/asm-util-9.2-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/asm/asm-commons-9.2-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/asm/asm-tree-9.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/asm/asm-analysis-9.2-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/asm/asm-commons-9.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/asm/asm-analysis-9.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/asm/asm-9.2-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/asm/asm-tree-9.2-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/asm/asm-9.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/apache_velocity/velocity-1.7.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/mockito/mockito-core-3.12.4.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/allocation_instrumenter/java-allocation-instrumenter-3.3.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/antlr/antlr4-4.5.3.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/perfmark/perfmark-api-0.19.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/error_prone/error_prone_core-2.9.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/error_prone/threeten-extra-1.5.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/error_prone/error_prone_type_annotations-2.9.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/error_prone/error_prone_annotation-2.9.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/error_prone/error_prone_annotations-2.9.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/error_prone/error_prone_check_api-2.9.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/ijar/test/libwrongcentraldir.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/ijar/test/jar-without-manifest.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/ijar/test/sealed/sealed.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/ijar/test/nestmates/nestmates.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/ijar/test/records/records.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/ijar/test/jar-with-manifest.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/ijar/test/jar-with-manifest-and-target-label.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/javax_activation/javax.activation-api-1.2.0-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/javax_activation/javax.activation-api-1.2.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/apache_commons_collections/commons-collections-3.2.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/xz/xz-1.9.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/apache_commons_lang/commons-lang-2.6.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/junit/junit-4.13.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/java-diff-utils/java-diff-utils-4.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/rxjava3/rxjava-3.1.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/apache_commons_logging/commons-logging-1.1.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jaxb/2.3.0/jaxb-core-2.3.0-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jaxb/2.3.0/jaxb-core-2.3.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jaxb/2.3.0/jaxb-api-2.3.0-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jaxb/2.3.0/jaxb-impl-2.3.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jaxb/2.3.0/jaxb-api-2.3.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jaxb/2.3.0/jaxb-impl-2.3.0-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jaxb/jaxb-api-2.3.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jaxb/jaxb-api-2.3.1-patched.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jaxb/jaxb-api-2.3.1-sources.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/hungarian_algorithm/software-and-algorithms-1.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/hungarian_algorithm/software-and-algorithms-1.0-src.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/kotlin_stdlib/kotlin-stdlib-jdk8-1.3.21.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/kotlin_stdlib/kotlin-stdlib-1.3.21.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/kotlin_stdlib/kotlin-stdlib-jdk7-1.3.21.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/kotlin_stdlib/kotlin-stdlib-common-1.3.21.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/apache_commons_io/commons-io-2.4.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/reactive_streams/reactive-streams-1.0.3.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-native-unix-common-4.1.69.Final-linux-aarch_64.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-handler-proxy-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-codec-http2-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-buffer-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-handler-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-native-unix-common-4.1.69.Final-osx-x86_64.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-native-epoll-4.1.69.Final-linux-x86_64.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-native-unix-common-4.1.69.Final-osx-aarch_64.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-native-unix-common-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-sctp-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-native-kqueue-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-native-unix-common-4.1.69.Final-linux-x86_64.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-common-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-resolver-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-native-epoll-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-resolver-dns-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-codec-http-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-native-epoll-4.1.69.Final-linux-aarch_64.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-native-kqueue-4.1.69.Final-osx-aarch_64.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-transport-native-kqueue-4.1.69.Final-osx-x86_64.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/netty/netty-codec-4.1.69.Final.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/auto/auto-service-annotations-1.0-rc7.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/auto/auto-value-1.8.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/auto/auto-service-1.0-rc7.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/auto/auto-value-annotations-1.8.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/auto/auto-common-1.1.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/truth_proto/truth-proto-extension-1.0.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/truth_proto/truth-liteproto-extension-1.0.1.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/pcollections/pcollections-2.1.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/jetifier/jetifier-core-1.0.0-beta04.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/compile_testing/compile-testing-0.18.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/auth/google-auth-library-oauth2-http-1.6.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/auth/google-auth-library-credentials-1.6.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android_annotations_25.0.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools.build_builder-test-api_2.0.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools_sdk-common_25.0.0-patched-stripped.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools.lint_lint-checks_25.0.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools_dvlib_25.0.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools.build_builder-model_2.0.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools_sdklib_25.0.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools.build_manifest-merger_25.0.0-patched.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools_common_25.0.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools.layoutlib_layoutlib_26.1.2-stripped.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools_sdk-common_25.0.0-patched.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools.layoutlib_layoutlib_26.1.2.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools.build_builder_2.0.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools.external.lombok_lombok-ast_0.2.3.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools.lint_lint-api_25.0.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/android_common/com.android.tools_repository_25.0.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/opencensus/opencensus-api-0.24.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/opencensus/opencensus-contrib-grpc-metrics-0.24.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/gson/gson-2.8.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/apache_commons_codec/commons-codec-1.10.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/bytebuddy/byte-buddy-agent-1.9.7.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/bytebuddy/byte-buddy-1.9.7.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/grpc/grpc-core-1.41.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/grpc/grpc-stub-1.41.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/grpc/grpc-context-1.41.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/grpc/grpc-protobuf-lite-1.41.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/grpc/grpc-protobuf-1.41.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/grpc/grpc-api-1.41.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/grpc/grpc-netty-1.41.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/grpc/grpc-auth-1.41.0.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/guava/guava-30.1.1-jre.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\nunzip -qn third_party/guava/guava-testlib-30.1.1-jre.jar -d /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes\r\n./opt/openjdk-bin-17.0.3_p7/bin/jar cmf /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/MANIFEST.MF /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/libblaze.jar -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes A.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes AUTHORS -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes AnnotatedClass.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes Annotations$ParametersOnlyAnnotation.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes Annotations$RuntimeInvisible.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes Annotations$RuntimeVisible.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes Annotations.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes B.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes LICENSE -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes LICENSE-junit.txt -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes LICENSE.txt -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes MANIFEST.MF -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes NOTICE -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes NOTICE.txt -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes NestTest$P.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes NestTest.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes RecordTest.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes SealedTest$Bar.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes SealedTest$Foo.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes SealedTest.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes about.html -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes android -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes androidx -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes api-versions.xml -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes api_diff.proto -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes autovalue -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes blogspot -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes build -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes com -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes data_binding_version_info.properties -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes default.config -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes default.generated.config -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes deps.proto -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes difflib -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes io -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes jacocoagent.jar -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes javax -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes jdk -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes junit -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes kotlin -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes libzstd-jni.so -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes lombok -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes manifest.proto -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes module-info.class -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes module-info.java -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes net -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes org -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes pom.xml -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes proguard -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes sun -C /var/tmp/portage/dev-util/bazel-5.2.0/temp/bazel_bf8keP61/classes textfile.txt\r\n..\r\nBuilding Bazel with Bazel.\r\n.OpenJDK 64-Bit Server VM warning: Options -Xverify:none and -noverify were deprecated in JDK 13 and will likely be removed in a future release.\r\nINFO: Analyzed target //src:bazel_nojdk (367 packages loaded, 10170 targets configured).\r\nINFO: Found 1 target...\r\n[0 / 653] [Prepa] BazelWorkspaceStatusAction stable-status.txt\r\nFATAL: bazel crashed due to an internal error. Printing stack trace:\r\njava.lang.ExceptionInInitializerError\r\n at com.google.devtools.build.lib.actions.ParameterFile.writeContent(ParameterFile.java:118)\r\n at com.google.devtools.build.lib.actions.ParameterFile.writeParameterFile(ParameterFile.java:111)\r\n at com.google.devtools.build.lib.analysis.actions.ParameterFileWriteAction$ParamFileWriter.writeOutputFile(ParameterFileWriteAction.java:170)\r\n at com.google.devtools.build.lib.exec.FileWriteStrategy.beginWriteOutputToFile(FileWriteStrategy.java:58)\r\n at com.google.devtools.build.lib.analysis.actions.FileWriteActionContext.beginWriteOutputToFile(FileWriteActionContext.java:49)\r\n at com.google.devtools.build.lib.analysis.actions.AbstractFileWriteAction.beginExecution(AbstractFileWriteAction.java:66)\r\n at com.google.devtools.build.lib.actions.Action.execute(Action.java:133)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor$5.execute(SkyframeActionExecutor.java:907)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor$ActionRunner.continueAction(SkyframeActionExecutor.java:1076)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor$ActionRunner.run(SkyframeActionExecutor.java:1031)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionState.runStateMachine(ActionExecutionState.java:152)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionState.getResultOrDependOnFuture(ActionExecutionState.java:91)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor.executeAction(SkyframeActionExecutor.java:492)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionFunction.checkCacheAndExecuteIfNeeded(ActionExecutionFunction.java:856)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionFunction.computeInternal(ActionExecutionFunction.java:349)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionFunction.compute(ActionExecutionFunction.java:169)\r\n at com.google.devtools.build.skyframe.AbstractParallelEvaluator$Evaluate.run(AbstractParallelEvaluator.java:590)\r\n at com.google.devtools.build.lib.concurrent.AbstractQueueVisitor$WrappedRunnable.run(AbstractQueueVisitor.java:382)\r\n at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136)\r\n at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635)\r\n at java.base/java.lang.Thread.run(Thread.java:833)\r\nCaused by: java.lang.reflect.InaccessibleObjectException: Unable to make java.lang.String(byte[],byte) accessible: module java.base does not \"opens java.lang\" to unnamed module @3e07d849\r\n at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:354)\r\n at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:297)\r\n at java.base/java.lang.reflect.Constructor.checkCanSetAccessible(Constructor.java:188)\r\n at java.base/java.lang.reflect.Constructor.setAccessible(Constructor.java:181)\r\n at com.google.devtools.build.lib.unsafe.StringUnsafe.&lt;init&gt;(StringUnsafe.java:75)\r\n at com.google.devtools.build.lib.unsafe.StringUnsafe.initInstance(StringUnsafe.java:56)\r\n at com.google.devtools.build.lib.unsafe.StringUnsafe.&lt;clinit&gt;(StringUnsafe.java:37)\r\n ... 21 more\r\n```", "The same issue happens for me with Bazel 5.3 and OpenJDK 17, `StringUnsafe`'s constructor fails with an `java.lang.reflect.InaccessibleObjectException: Unable to make java.lang.String(byte[],byte) accessible: module java.base does not \"opens java.lang\" to unnamed module` error:\r\n\r\n```\r\nFATAL: bazel crashed due to an internal error. Printing stack trace:\r\njava.lang.ExceptionInInitializerError\r\n at com.google.devtools.build.lib.actions.ParameterFile.writeContent(ParameterFile.java:118)\r\n at com.google.devtools.build.lib.actions.ParameterFile.writeParameterFile(ParameterFile.java:111)\r\n at com.google.devtools.build.lib.analysis.actions.ParameterFileWriteAction$ParamFileWriter.writeOutputFile(ParameterFileWriteAction.java:170)\r\n at com.google.devtools.build.lib.exec.FileWriteStrategy.beginWriteOutputToFile(FileWriteStrategy.java:58)\r\n at com.google.devtools.build.lib.analysis.actions.FileWriteActionContext.beginWriteOutputToFile(FileWriteActionContext.java:49)\r\n at com.google.devtools.build.lib.analysis.actions.AbstractFileWriteAction.beginExecution(AbstractFileWriteAction.java:66)\r\n at com.google.devtools.build.lib.actions.Action.execute(Action.java:133)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor$5.execute(SkyframeActionExecutor.java:907)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor$ActionRunner.continueAction(SkyframeActionExecutor.java:1076)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor$ActionRunner.run(SkyframeActionExecutor.java:1031)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionState.runStateMachine(ActionExecutionState.java:152)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionState.getResultOrDependOnFuture(ActionExecutionState.java:91)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor.executeAction(SkyframeActionExecutor.java:492)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionFunction.checkCacheAndExecuteIfNeeded(ActionExecutionFunction.java:856)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionFunction.computeInternal(ActionExecutionFunction.java:349)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionFunction.compute(ActionExecutionFunction.java:169)\r\n at com.google.devtools.build.skyframe.AbstractParallelEvaluator$Evaluate.run(AbstractParallelEvaluator.java:590)\r\n at com.google.devtools.build.lib.concurrent.AbstractQueueVisitor$WrappedRunnable.run(AbstractQueueVisitor.java:382)\r\n at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136)\r\n at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635)\r\n at java.base/java.lang.Thread.run(Thread.java:833)\r\nCaused by: java.lang.reflect.InaccessibleObjectException: Unable to make java.lang.String(byte[],byte) accessible: module java.base does not \"opens java.lang\" to unnamed module @3e07d849\r\n at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:354)\r\n at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:297)\r\n at java.base/java.lang.reflect.Constructor.checkCanSetAccessible(Constructor.java:188)\r\n at java.base/java.lang.reflect.Constructor.setAccessible(Constructor.java:181)\r\n at com.google.devtools.build.lib.unsafe.StringUnsafe.<init>(StringUnsafe.java:75)\r\n at com.google.devtools.build.lib.unsafe.StringUnsafe.initInstance(StringUnsafe.java:56)\r\n at com.google.devtools.build.lib.unsafe.StringUnsafe.<clinit>(StringUnsafe.java:37)\r\n ... 21 more\r\n```", "/cc @cushon ", "```\r\nCaused by: java.lang.reflect.InaccessibleObjectException: Unable to make java.lang.String(byte[],byte) accessible: module java.base does not \"opens java.lang\" to unnamed module @3e07d849\r\n at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:354)\r\n at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:297)\r\n at java.base/java.lang.reflect.Constructor.checkCanSetAccessible(Constructor.java:188)\r\n at java.base/java.lang.reflect.Constructor.setAccessible(Constructor.java:181)\r\n at com.google.devtools.build.lib.unsafe.StringUnsafe.&lt;init&gt;(StringUnsafe.java:75)\r\n```\r\n\r\nThis is https://openjdk.org/jeps/403\r\n\r\nThe workaround is to pass `--add-opens=java.base/java.nio=ALL-UNNAMED` as a server JVM flag. The bazel launcher already passes that flag, but only for embedded JDKs:\r\n\r\nhttps://github.com/bazelbuild/bazel/blob/03ad0cd1203b7b54ec02ea3f5788d1d673a9d6d6/src/main/cpp/blaze.cc#L362-L367", "> The workaround is to pass `--add-opens=java.base/java.nio=ALL-UNNAMED` as a server JVM flag\r\n\r\n@cushon \r\nHow to do that? I've tried `--jvmopt=--add-opens=java.base/java.nio=ALL-UNNAMED` unsuccessfully.", "I think you want `bazel --host_jvm_args=--add-opens=java.base/java.nio=ALL-UNNAMED build ...`, see https://bazel.build/docs/user-manual#host-jvm-args\r\n\r\nThis needs to be passed to the JVM running Bazel itself, `bazel build --jvmopt=...` is configuring JVM flags for the JVM in the target configuration (e.g. for a `java_binary` or `java_test`)", "The command executed successfully but it didn't fix my issue. Still seeing `Unable to make java.lang.String(byte[],byte) accessible: module java.base does not \"opens java.lang\" to unnamed module`.\r\n\r\nAnyway, as a workaround, I'm using `asdf` to use the `openjdk-11` in my project for now.\r\n\r\nThank you for your help.", "> java.base does not \"opens java.lang\"\r\n\r\nThat one requires `--add-opens=java.base/java.lang=ALL-UNNAMED`, not `--add-opens=java.base/java.nio=ALL-UNNAMED`, you might need to pass multiple `--add-opens=` flags", "Thank you @cushon. The command below worked:\r\n```\r\nbazel \\\r\n--host_jvm_args=\"--add-opens=java.base/java.nio=ALL-UNNAMED\" \\\r\n--host_jvm_args=\"--add-opens=java.base/java.lang=ALL-UNNAMED\" \\\r\nbuild ...\r\n```" ]

[]

[ "P4", "type: support / not a bug (process)", "team-Rules-Java" ]

[ "src/main/cpp/blaze.cc" ]

[ "src/main/cpp/blaze.cc" ]

[]

[ "keyword_pr_to_issue" ]

[ "Looping in @frazze-jobb who wrote the hermetic sandbox code.", "Leaving out `--experimental_use_sandboxfs` doesn't change this but reduces the problem space.", "@frazze-jobb is on another project now. I worked with him on this sandbox and answer instead.\r\n\r\nThanks for reporting this issue.\r\n\r\nI will try to find time to investigate this and planning to create test cases for a few related special cases.\r\n", "FYI, I'm encountering the same problem in the context of a hermtic cc_toolchain using a symlink.\r\nMinimal reproducible example:\r\n```\r\n#!/bin/bash\r\nset -exuo pipefail\r\n\r\n# Create files for bazel\r\ntouch src.cc\r\nmkdir a\r\nln -s a b\r\n\r\ntouch WORKSPACE\r\ncat > BUILD << EOF\r\nload(\"toolchain.bzl\", \"gcc_toolchain\")\r\n\r\ncc_binary(\r\n name = \"test\",\r\n srcs = [\"src.cc\"],\r\n)\r\n\r\ncc_toolchain_suite(\r\n name = \"my_toolchain\",\r\n toolchains = {\"k8\": \"dummy_toolchain\"},\r\n)\r\n\r\nfilegroup(\r\n name = \"compiler_files\",\r\n srcs = [\"b\"],\r\n)\r\n\r\ngcc_toolchain(\"dummy_toolchain\")\r\nEOF\r\n\r\ncat > toolchain.bzl << EOF\r\ndef _impl(ctx):\r\n crosstool_proto = cc_common.create_cc_toolchain_config_info(\r\n ctx = ctx,\r\n toolchain_identifier = \"cc-compiler-\" + ctx.label.name,\r\n host_system_name = \"x86_64-linux-gnu\",\r\n target_system_name = \"linux\",\r\n target_cpu = \"k8\",\r\n target_libc = \"unknown-glibc\",\r\n compiler = \"gcc\",\r\n )\r\n return [crosstool_proto]\r\n\r\ngcc_toolchain_configuration = rule(\r\n attrs = {},\r\n provides = [CcToolchainConfigInfo],\r\n implementation = _impl,\r\n)\r\n\r\ndef gcc_toolchain(name):\r\n gcc_toolchain_configuration(name = name + \"_configuration\")\r\n\r\n native.cc_toolchain(\r\n name = name,\r\n toolchain_config = name + \"_configuration\",\r\n all_files = \"//:compiler_files\",\r\n compiler_files = \"//:compiler_files\",\r\n dwp_files = \"//:compiler_files\",\r\n linker_files = \"//:compiler_files\",\r\n objcopy_files = \"//:compiler_files\",\r\n strip_files = \"//:compiler_files\",\r\n supports_param_files = True,\r\n supports_header_parsing = True,\r\n )\r\nEOF\r\n```\r\n\r\nTrying to build the binary via\r\n```\r\nbazel build //:test --crosstool_top=//:my_toolchain --experimental_use_hermetic_linux_sandbox\r\n```\r\n\r\nresults in\r\n```\r\nERROR: .../BUILD:3:10: Compiling src.cc failed: I/O exception during sandboxed execution: input dependency /home/.../de7a35bfd5cefbaf6a88e91201c016c4/execroot/__main__/b could not be checked for modifications during execution.\r\n```", "So the command-line output is not that useful, but looking into the server log I see this stack trace which shows a UnsupportedOperationException at FileArtifactValue.java:352 - it's trying to get a ContentsProxy of a directory, which isn't supported. Despite what FileArtifactValue may think, it's not always an actual file. So the modification check should take appropriate action on things that are not files.\r\n\r\n```\r\ncom.google.devtools.build.lib.skyframe.ActionExecutionFunction$ActionExecutionFunctionException: com.google.devtools.build.lib.actions.AlreadyReportedActionEx\r\necutionException: I/O exception during sandboxed execution: input dependency /usr/local/google/home/larsrc/.cache/bazel/_bazel_larsrc/b91bdcc9c8a35d79096b4ba9\r\nb5794356/execroot/__main__/b could not be checked for modifications during execution.\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionFunction.computeInternal(ActionExecutionFunction.java:368)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionFunction.compute(ActionExecutionFunction.java:169)\r\n at com.google.devtools.build.skyframe.AbstractParallelEvaluator$Evaluate.run(AbstractParallelEvaluator.java:590)\r\n at com.google.devtools.build.lib.concurrent.AbstractQueueVisitor$WrappedRunnable.run(AbstractQueueVisitor.java:382)\r\n at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(Unknown Source)\r\n at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)\r\n at java.base/java.lang.Thread.run(Unknown Source)\r\nCaused by: com.google.devtools.build.lib.actions.AlreadyReportedActionExecutionException: I/O exception during sandboxed execution: input dependency /usr/loca\r\nl/google/home/larsrc/.cache/bazel/_bazel_larsrc/b91bdcc9c8a35d79096b4ba9b5794356/execroot/__main__/b could not be checked for modifications during execution.\r\n ... 7 more\r\nCaused by: com.google.devtools.build.lib.actions.UserExecException: I/O exception during sandboxed execution: input dependency /usr/local/google/home/larsrc/.\r\ncache/bazel/_bazel_larsrc/b91bdcc9c8a35d79096b4ba9b5794356/execroot/__main__/b could not be checked for modifications during execution.\r\n at com.google.devtools.build.lib.sandbox.AbstractSandboxSpawnRunner.exec(AbstractSandboxSpawnRunner.java:97)\r\n at com.google.devtools.build.lib.sandbox.SandboxModule$SandboxFallbackSpawnRunner.exec(SandboxModule.java:494)\r\n at com.google.devtools.build.lib.exec.SpawnRunner.execAsync(SpawnRunner.java:245)\r\n at com.google.devtools.build.lib.exec.AbstractSpawnStrategy.exec(AbstractSpawnStrategy.java:146)\r\n at com.google.devtools.build.lib.exec.AbstractSpawnStrategy.exec(AbstractSpawnStrategy.java:108)\r\n at com.google.devtools.build.lib.actions.SpawnStrategy.beginExecution(SpawnStrategy.java:47)\r\n at com.google.devtools.build.lib.exec.SpawnStrategyResolver.beginExecution(SpawnStrategyResolver.java:68)\r\n at com.google.devtools.build.lib.rules.cpp.CppCompileAction.beginExecution(CppCompileAction.java:1454)\r\n at com.google.devtools.build.lib.actions.Action.execute(Action.java:133)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor$5.execute(SkyframeActionExecutor.java:907)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor$ActionRunner.continueAction(SkyframeActionExecutor.java:1076)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor$ActionR\r\nunner.run(SkyframeActionExecutor.java:1031)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionState.runStateMachine(ActionExecutionState.java:152)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionState.getResultOrDependOnFuture(ActionExecutionState.java:91)\r\n at com.google.devtools.build.lib.skyframe.SkyframeActionExecutor.executeAction(SkyframeActionExecutor.java:492)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionFunction.checkCacheAndExecuteIfNeeded(ActionExecutionFunction.java:856)\r\n at com.google.devtools.build.lib.skyframe.ActionExecutionFunction.computeInternal(ActionExecutionFunction.java:349)\r\n ... 6 more\r\nCaused by: java.io.IOException: input dependency /usr/local/google/home/larsrc/.cache/bazel/_bazel_larsrc/b91bdcc9c8a35d79096b4ba9b5794356/execroot/__main__/b could not be checked for modifications during execution.\r\n at com.google.devtools.build.lib.sandbox.LinuxSandboxedSpawnRunner.checkForConcurrentModifications(LinuxSandboxedSpawnRunner.java:399)\r\n at com.google.devtools.build.lib.sandbox.LinuxSandboxedSpawnRunner.verifyPostCondition(LinuxSandboxedSpawnRunner.java:380)\r\n at com.google.devtools.build.lib.sandbox.AbstractSandboxSpawnRunner.runSpawn(AbstractSandboxSpawnRunner.java:136)\r\n at com.google.devtools.build.lib.sandbox.AbstractSandboxSpawnRunner.exec(AbstractSandboxSpawnRunner.java:92)\r\n ... 22 more\r\nCaused by: java.lang.UnsupportedOperationException\r\n at com.google.devtools.build.lib.actions.FileArtifactValue$DirectoryArtifactValue.getContentsProxy(FileArtifactValue.java:352)\r\n at com.google.devtools.build.lib.sandbox.LinuxSandboxedSpawnRunner.checkForConcurrentModifications(LinuxSandboxedSpawnRunner.java:395)\r\n ... 25 more\r\n```", "Looking at the possible subclasses of FileArtifactValue, they mostly don't support the operations needed for the modification check. @frazze-jobb we'll have to figure out which kinds of inputs to ignore and which to handle properly. Here are the direct subclasses:\r\n\r\n```\r\ngoogle3/third_party/bazel/src/main/java/com/google/devtools/build/lib/actions/ActionCacheChecker.java\r\n851: private static final class ConstantMetadataValue extends FileArtifactValue\r\ngoogle3/third_party/bazel/src/main/java/com/google/devtools/build/lib/actions/FileArtifactValue.java\r\n332: private static final class DirectoryArtifactValue extends FileArtifactValue {\r\n399: private static final class HashedDirectoryArtifactValue extends FileArtifactValue {\r\n462: private static final class RegularFileArtifactValue extends FileArtifactValue {\r\n549: public static class RemoteFileArtifactValue extends FileArtifactValue {\r\n718: public static final class UnresolvedSymlinkArtifactValue extends FileArtifactValue {\r\n758: public static final class InlineFileArtifactValue extends FileArtifactValue {\r\n832: public static final class SourceFileArtifactValue extends FileArtifactValue {\r\n895: private static final class SingletonMarkerValue extends FileArtifactValue implements Singleton {\r\n940: private static final class OmittedFileValue extends FileArtifactValue implements Singleton {\r\ngoogle3/third_party/bazel/src/main/java/com/google/devtools/build/lib/skyframe/TreeArtifactValue.java\r\n210: private static final class TreeArtifactCompositeFileArtifactValue extends FileArtifactValue {\r\n```", "Thanks for looking into this. Not that it would help much, but leaving this smaller minimal example here for sake of simplicity\r\n```\r\n#!/bin/bash\r\ntouch WORKSPACE\r\nmkdir a\r\nln -s a b\r\n\r\ncat > BUILD << EOF\r\ngenrule(\r\n name = \"target\",\r\n srcs = [\":b\"],\r\n outs = [\"dummy\"],\r\n cmd = \"ls $< > \\$@\",\r\n)\r\nEOF\r\n```\r\n\r\n```\r\nbazel build //:target --experimental_use_hermetic_linux_sandbox\r\n```", "That's a nice small repro. It can also be done with just a directory as an input, so both directory and symlink handling needs to be fixed.", "The purpose of `checkForConcurrentModifications` is to check that an action has not changed the content of hardlinked files outside of the sandbox. `HardlinkedSandboxedSpawn.hardLinkRecursive` is resolving symbolic links and are only creating hardlinks for files, not directories. Therefor it should be safe to skip the concurrent modification check when `metadata.getType().isFile()` is false. That should address the use cases in this ticket reliably.\r\n\r\nHowever, then the following subclasses of FileArtifactValue, mentioned by @larsrc-google, remains:\r\n\r\n- ConstantMetadataValue\r\n- InlineFileArtifactValue\r\n- RemoteFileArtifactValue\r\n- SourceFileArtifactValue\r\n\r\nI don't see any good way to handle them. Are they rare? As workaround, should we catch UnsupportedOperationException and log a warning each time it occurs? (That could be up to one warning per action.)", "Thanks for the fix!\r\n\r\nI don't know anything about Bazel release policies - is this fix likely to make it to a 6.x release?", "@lunch-glide-pepper You can use the `@bazel-io flag` command in a GitHub comment to mark it for cherry-picking and then create a PR with the commit cherry-picked on `release-6.0.0`." ]

[]

[ "type: bug", "P2", "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_hermetic_sandboxing_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "@bazel-io flag", "Submitted a fix in https://github.com/bazelbuild/bazel/pull/16851", "@bazel-io fork 6.0.0" ]

[]

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

[ "tools/java/runfiles/Runfiles.java" ]

[ "tools/java/runfiles/Runfiles.java" ]

[]

[ "keyword_pr_to_issue" ]

[ "@Wyverald Could you assign me and add the appropriate labels?", "~~@comius Do you expect `java_binary` to be starlarkified in time for Bazel 6? That would simplify the changes to the Java rules.~~ Most likely no longer necessary, as the changes will apply to `buildjar` itself.", "@Wyverald Oops, I missed the use of `Closes ...` in the PR description - we should use `Work towards ...` instead. Could you reopen? For some I lack the permission to do so.", "It's the \"To fix #NNN\" line... GitHub trying to be clever here", "The most likely future seems to be that we'll standardize on Python imports _not_ containing the repository name, in which case bzlmod is a no-op.\r\n\r\nIt's not an optimal solution because it requires an entry in `$PYTHONPATH` for each repository in the transitive closure of the Python binary, but it's the status quo and I'd much prefer thinking up a solution to that (e.g. binary-specific symlink trees that do not look like runfiles trees, dunno) than to standardize on having the repository names in imports because the concept of \"repository\" is specific to Bazel.\r\n\r\nPut differently: the only way for bzlmod to cause trouble for Python is if we decide to require repository names in imports, which seems to be unlikely and it looks like the decision on that will take a while and I prefer not blocking either Bazel 6.0 or the removal of bzlmod from the experimental domain to avoid an unlikely failure mode. And should that happen, we can always paper over it with some sort of import hook.\r\n\r\n/cc @rickeylev ", "@oquenchil I'm thinking about whether we should use a macro to pass `BAZEL_CURRENT_REPOSITORY` to the `Rlocation` function automatically (e.g. do `#define Rlocation(path) RlocationReal(path, BAZEL_CURRENT_REPOSITORY)`). Do you think that would be too intrusive/magical?", "Would users have to write that line or is it part of a file that is automatically generated?\r\n\r\nI don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.", "> Would users have to write that line or is it part of a file that is automatically generated?\r\n> \r\n> I don't mind adding whatever will make it easier but if it gets too magical it should maybe be wrapped so that the details aren't visible to the user unless they want to look.\r\n\r\n@oquenchil I would want to make that magic part of the runfiles library header so that users don't have to write it themselves.\r\n\r\nWhat I am worried about is having an `Rlocation` macro mess with people's existing codebase since there doesn't seem to be a way to put it in a namespace or restrict it to member functions only.", "Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?", "> Aren't we in control of the name of the macro? Can we rename it to something that would be very unlikely to clash with anything?\r\n\r\nWe can call it something else, we just won't be able to have existing C++ codebases use the new repo mapping-aware lookup automatically if we do so. But of course having to replace all `runfiles::create()` calls with something else won't be too bad.", "If the macro name is Rlocation and that may start clashing with people's code in C++ files then I think it would be prudent to rename it if we decide to go with the macro.\r\n\r\nI guess the call to make here is how much of a breaking change this is compared to the benefit of passing BAZEL_CURRENT_REPOSITORY automatically for users. I would say that the benefit is not too great but as far as I understood the use of Bazel runfiles in C++ code wasn't too wide-spread either?\r\n\r\n> But of course having to replace all runfiles::create() calls with something else won't be too bad.\r\n\r\nHow do you know? Is it based on search results from GH? If a release note would be enough for it not to be too disruptive then let's rename it.", "I have experimented a bit more with APIs while working on the corresponding change for Java (https://github.com/bazelbuild/bazel/pull/16549).\r\n\r\nWhat I now like more than the macro approach is to have users (optionally) pass in a canonical repository name when creating the `Runfiles` instance and default to that repo in `Rlocation` calls, while providing ways to override the default with an overload of `Rlocation` that accepts a repository name as a second parameter. In this way, users only have to pass in `BAZEL_CURRENT_REPOSITORY` only once in the common case, which should render a macro unnecessary.", "Update: All changes relevant for Bazel 6.0.0 are now in review.", "This is mostly done now, with only the bugfix https://github.com/bazelbuild/bazel/pull/16755 and a number of cherry-picks to 6.0.0 missing.", "@Wyverald You can remove this from the \"6.0.0 nice to have\" milestone as the only remaining issue is not tied to a Bazel release (Stardoc).", "Downgrading because the most important points have been addressed; the only outstanding issue is Stardoc, which is undergoing a rewrite currently by @tetromino." ]

[]

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

[ "tools/java/runfiles/Runfiles.java" ]

[ "tools/java/runfiles/Runfiles.java" ]

[]

[ "keyword_pr_to_issue" ]

[ "@bazel-io flag", "@jwnimmer-tri Thanks for the report. Does #16801 look reasonable to you?", "@bazel-io fork 6.0.0" ]

[]

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

[ "tools/cpp/runfiles/runfiles_src.h" ]

[ "tools/cpp/runfiles/runfiles_src.h" ]

[]

[ "keyword_issue_to_pr" ]

[ "@kshyanashree Please cherry pick https://github.com/bazelbuild/bazel/commit/cd40666001e8d599bb61735898c195c6d2fae55b to close this one, thanks! ", "After cherry pick this one, we can create rc3 ;)", "> @kshyanashree Please cherry pick [cd40666](https://github.com/bazelbuild/bazel/commit/cd40666001e8d599bb61735898c195c6d2fae55b) to close this one, thanks!\r\n\r\nHi @meteorcloudy ! I have already cherry-picked this one here , but check is failing due to missing CLA #16886 . Wanted to check with you. Thanks!", "For some reason the original commit was showing up as being authored by \"[email protected]\", which can't pass the CLA check. I think we can just force merge that PR.", "Merged in #16886\r\n\r\nLet's create an rc3!", "Sure, Thanks @Wyverald!" ]

[]

[]

[ "src/main/starlark/builtins_bzl/common/cc/cc_binary.bzl" ]

[ "src/main/starlark/builtins_bzl/common/cc/cc_binary.bzl" ]

[]

[ "keyword_pr_to_issue" ]

[ "This issue also happens with Bazel@HEAD: https://buildkite.com/bazel/bazel-at-head-plus-downstream/builds/2766#0184e535-8df3-4114-9f65-8777f5bc7fe4, but the Bazel Auto Sheriff points to a [commit](https://github.com/bazelbuild/bazel/commit/fb2324620a4adee19bb4e8a030b6fba58f672dc0) that's not in Bazel 6.0.0rc4: https://buildkite.com/bazel/bazel-auto-sheriff-face-with-cowboy-hat/builds/1118\r\n\r\n**Update**: culprit finder failed due to not being able to download the Bazel binary.\r\n```\r\n:bazel: Test with Bazel built at fb2324620a4adee19bb4e8a030b6fba58f672dc0\r\n:googlecloud: Downloading Bazel built at fb2324620a4adee19bb4e8a030b6fba58f672dc0\r\n \r\ngsutil cp gs://bazel-builds/artifacts/centos7/fb2324620a4adee19bb4e8a030b6fba58f672dc0/bazel /tmp/tmp09usce0c/bazel\r\nCommandException: No URLs matched: gs://bazel-builds/artifacts/centos7/fb2324620a4adee19bb4e8a030b6fba58f672dc0/bazel\r\nFailed to download Bazel binary at fb2324620a4adee19bb4e8a030b6fba58f672dc0, error message:\r\nCommand '['gsutil', 'cp', 'gs://bazel-builds/artifacts/centos7/fb2324620a4adee19bb4e8a030b6fba58f672dc0/bazel', '/tmp/tmp09usce0c/bazel']' returned non-zero exit status 1.\r\n\r\n```\r\n\r\nWe should fix culprit finder\r\n", "/cc @coeuvre Since this only happens with RBE enabled. https://github.com/bazelbuild/bazel/blob/2a88c8da7783490ef6192336b091910cd662cc67/.bazelci/postsubmit.yml#L266", "@coeuvre <del>Is https://github.com/bazelbuild/bazel/pull/16749 the possible culprit?</del> probably not since the original change was submitted in Oct, we have only been seen the failure with Bazel@HEAD in Nov.", "Last green: https://buildkite.com/bazel/bazel-at-head-plus-downstream/builds/2729\r\nFirst fail: https://buildkite.com/bazel/bazel-at-head-plus-downstream/builds/2730\r\n\r\nCulprit should fall into (6e945e8ab422ca0a195d09899cce61ba6c0cfa59, 1500195986dc5095a21f32633cb4e3aabe037f1e]", "Bisecting the rc4 branch at https://buildkite.com/bazel/culprit-finder/builds/3876 and it points to https://github.com/bazelbuild/bazel/commit/0b645254b41edc738c6641fd192fca86203ff2e2", "Bisecting again on master branch after fixing the CI issue: https://buildkite.com/bazel/culprit-finder/builds/3877", "While waiting for the bisect result, I can also reproduce at https://github.com/bazelbuild/bazel/commit/839ce7f5c40240d8b6f49c416c3769e226f43fee (original commit of the 6.0.0 cherrypick https://github.com/bazelbuild/bazel/commit/0b645254b41edc738c6641fd192fca86203ff2e2)\r\n```\r\nUSE_BAZEL_VERSION=839ce7f5c40240d8b6f49c416c3769e226f43fee bazel test //src/test/shell/bazel:bazel_example_test --config=ubuntu1804_java11 --remote_executor=grpcs://remotebuildexecution.googleapis.com\r\n```\r\nBut not at it's parent commit.\r\n\r\n/cc @fmeum @Wyverald ", "@meteorcloudy Could you run these tests with `--test_env=RUNFILES_LIB_DEBUG=1`? That should provide us with very helpful information.", "@fmeum Here is the test log:\r\n```\r\nexec ${PAGER:-/usr/bin/less} \"$0\" || exit 1\r\nExecuting tests from //src/test/shell/bazel:bazel_example_test\r\n-----------------------------------------------------------------------------\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/integration_test_setup.sh): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/integration_test_setup.sh): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/integration_test_setup.sh): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/unittest.bash): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/unittest.bash): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/unittest.bash): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nWARNING: Passing test names in arguments (--test_arg) is deprecated, please use --test_filter='/usr/lib/jvm/java-11-openjdk-amd64' instead.\r\nWARNING: Arguments do not specify tests!\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/testenv.sh): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/testenv.sh): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/testenv.sh): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/bazel): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/bazel): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/bazel): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bin/bazel): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bin/bazel): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bin/bazel): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[runfiles.bash]: rlocation(io_bazel/tools/BUILD): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/tools/BUILD): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/tools/BUILD): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bazel/testing_server.py): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bazel/testing_server.py): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bazel/testing_server.py): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[bazel_example_test 2022-12-07 09:02:51 (+0000)] bazel binary is at /b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles/io_bazel/src/test/shell/bin\r\nINFO[bazel_example_test 2022-12-07 09:02:51 (+0000)] setting up client in /b/f/w/_tmp/bfbf694ed20a8c587876e28f8f9820b1/workspace\r\nINFO[runfiles.bash]: rlocation(/usr/lib/jvm/java-11-openjdk-amd64/bin/java): start\r\nINFO[runfiles.bash]: rlocation(/usr/lib/jvm/java-11-openjdk-amd64/bin/java): absolute path, return\r\nINFO[runfiles.bash]: rlocation(/usr/lib/jvm/java-11-openjdk-amd64/bin/java): not using repository mapping () since it does not exist\r\nERROR[runfiles.bash]: cannot look up runfile \"/usr/lib/jvm/java-11-openjdk-amd64/bin/java\" (RUNFILES_DIR=\"/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles\", RUNFILES_MANIFEST_FILE=\"\")\r\n```", "@coeuvre I suspected a different issue, so I would like to confirm that isn't a problem either: The Bash runfiles library currently relies on absolute (!) paths of generated files containing an `execroot` segment. Is that guaranteed to be the case with remote execution, just like it is with local execution?", "The logic for absolute paths passed to `rlocation` broke. I am investigating why the existing test didn't catch that.", "> @coeuvre I suspected a different issue, so I would like to confirm that isn't a problem either: The Bash runfiles library currently relies on absolute (!) paths of generated files containing an `execroot` segment. Is that guaranteed to be the case with remote execution, just like it is with local execution?\r\n\r\nMy gut feeling is that `execroot` is not included in the path when actions inspect the path inside remote worker but I may be wrong. We use `execroot` as input root when uploading inputs for actions. Remote worker is free to move the input root to any local directories so a file `/.../execroot/bazel-out/a` in bazel host is probably become `/..../bazel-out/a` in remote worker.", "@coeuvre Thanks, I will remove that assumption in a second PR.\r\n\r\n@meteorcloudy The Bash runfiles test didn't cover the exit codes of `rlocation`, only whatever it printed, and thus missed some corner cases. I have reworked the tests to check the exit codes and am now fixing these cases.", "@fmeum @coeuvre Thanks!!", "I submitted two PRs with fixes, #16945 and #16946.", "I think I know how to fix the RBE test failures, but now that I studied this more closely, there is some underlying inconsistent behavior that makes the library difficult to test and potentially also hard to use: `$(rlocation does/not/exist)` has exit code 0 if and only if the runfiles manifest exists. This means that the behavior differs between local and remote execution on Windows, which is probably why we only noticed this in downstream tests.\r\n\r\n@laszlocsomor Sorry for the sideline mention, but can you recall how you intended the exit code of Bash `$(rlocation ...)` to behave?", "> `$(rlocationpath does/not/exist)` has exit code 0 if and only if the runfiles manifest exists.\r\n\r\nDid you mean `$(rlocation does/not/exist)`?", "@Wyverald Yes, sorry, fixed.", "Fixing this inconsistency isn't required to get the tests to pass, but if we want to do it, 6.0.0 is probably the right time. Since `$(failing_command)` doesn't usually fail the outer shell script, having `$(rlocation does/not/exist)` exit with exit code 1 in all cases doesn't seem to be *that* breaking.\r\n\r\n@Wyverald @meteorcloudy What do you think?" ]

[]

[ "type: bug", "P1" ]

[ "tools/bash/runfiles/runfiles.bash" ]

[ "tools/bash/runfiles/runfiles.bash" ]

[]

[ "keyword_pr_to_issue" ]

[ "This issue also happens with Bazel@HEAD: https://buildkite.com/bazel/bazel-at-head-plus-downstream/builds/2766#0184e535-8df3-4114-9f65-8777f5bc7fe4, but the Bazel Auto Sheriff points to a [commit](https://github.com/bazelbuild/bazel/commit/fb2324620a4adee19bb4e8a030b6fba58f672dc0) that's not in Bazel 6.0.0rc4: https://buildkite.com/bazel/bazel-auto-sheriff-face-with-cowboy-hat/builds/1118\r\n\r\n**Update**: culprit finder failed due to not being able to download the Bazel binary.\r\n```\r\n:bazel: Test with Bazel built at fb2324620a4adee19bb4e8a030b6fba58f672dc0\r\n:googlecloud: Downloading Bazel built at fb2324620a4adee19bb4e8a030b6fba58f672dc0\r\n \r\ngsutil cp gs://bazel-builds/artifacts/centos7/fb2324620a4adee19bb4e8a030b6fba58f672dc0/bazel /tmp/tmp09usce0c/bazel\r\nCommandException: No URLs matched: gs://bazel-builds/artifacts/centos7/fb2324620a4adee19bb4e8a030b6fba58f672dc0/bazel\r\nFailed to download Bazel binary at fb2324620a4adee19bb4e8a030b6fba58f672dc0, error message:\r\nCommand '['gsutil', 'cp', 'gs://bazel-builds/artifacts/centos7/fb2324620a4adee19bb4e8a030b6fba58f672dc0/bazel', '/tmp/tmp09usce0c/bazel']' returned non-zero exit status 1.\r\n\r\n```\r\n\r\nWe should fix culprit finder\r\n", "/cc @coeuvre Since this only happens with RBE enabled. https://github.com/bazelbuild/bazel/blob/2a88c8da7783490ef6192336b091910cd662cc67/.bazelci/postsubmit.yml#L266", "@coeuvre <del>Is https://github.com/bazelbuild/bazel/pull/16749 the possible culprit?</del> probably not since the original change was submitted in Oct, we have only been seen the failure with Bazel@HEAD in Nov.", "Last green: https://buildkite.com/bazel/bazel-at-head-plus-downstream/builds/2729\r\nFirst fail: https://buildkite.com/bazel/bazel-at-head-plus-downstream/builds/2730\r\n\r\nCulprit should fall into (6e945e8ab422ca0a195d09899cce61ba6c0cfa59, 1500195986dc5095a21f32633cb4e3aabe037f1e]", "Bisecting the rc4 branch at https://buildkite.com/bazel/culprit-finder/builds/3876 and it points to https://github.com/bazelbuild/bazel/commit/0b645254b41edc738c6641fd192fca86203ff2e2", "Bisecting again on master branch after fixing the CI issue: https://buildkite.com/bazel/culprit-finder/builds/3877", "While waiting for the bisect result, I can also reproduce at https://github.com/bazelbuild/bazel/commit/839ce7f5c40240d8b6f49c416c3769e226f43fee (original commit of the 6.0.0 cherrypick https://github.com/bazelbuild/bazel/commit/0b645254b41edc738c6641fd192fca86203ff2e2)\r\n```\r\nUSE_BAZEL_VERSION=839ce7f5c40240d8b6f49c416c3769e226f43fee bazel test //src/test/shell/bazel:bazel_example_test --config=ubuntu1804_java11 --remote_executor=grpcs://remotebuildexecution.googleapis.com\r\n```\r\nBut not at it's parent commit.\r\n\r\n/cc @fmeum @Wyverald ", "@meteorcloudy Could you run these tests with `--test_env=RUNFILES_LIB_DEBUG=1`? That should provide us with very helpful information.", "@fmeum Here is the test log:\r\n```\r\nexec ${PAGER:-/usr/bin/less} \"$0\" || exit 1\r\nExecuting tests from //src/test/shell/bazel:bazel_example_test\r\n-----------------------------------------------------------------------------\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/integration_test_setup.sh): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/integration_test_setup.sh): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/integration_test_setup.sh): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/unittest.bash): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/unittest.bash): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/unittest.bash): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nWARNING: Passing test names in arguments (--test_arg) is deprecated, please use --test_filter='/usr/lib/jvm/java-11-openjdk-amd64' instead.\r\nWARNING: Arguments do not specify tests!\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/testenv.sh): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/testenv.sh): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/testenv.sh): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/bazel): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/bazel): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/bazel): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bin/bazel): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bin/bazel): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bin/bazel): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[runfiles.bash]: rlocation(io_bazel/tools/BUILD): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/tools/BUILD): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/tools/BUILD): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bazel/testing_server.py): start\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bazel/testing_server.py): not using repository mapping () since it does not exist\r\nINFO[runfiles.bash]: rlocation(io_bazel/src/test/shell/bazel/testing_server.py): found under RUNFILES_DIR (/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles), return\r\nINFO[bazel_example_test 2022-12-07 09:02:51 (+0000)] bazel binary is at /b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles/io_bazel/src/test/shell/bin\r\nINFO[bazel_example_test 2022-12-07 09:02:51 (+0000)] setting up client in /b/f/w/_tmp/bfbf694ed20a8c587876e28f8f9820b1/workspace\r\nINFO[runfiles.bash]: rlocation(/usr/lib/jvm/java-11-openjdk-amd64/bin/java): start\r\nINFO[runfiles.bash]: rlocation(/usr/lib/jvm/java-11-openjdk-amd64/bin/java): absolute path, return\r\nINFO[runfiles.bash]: rlocation(/usr/lib/jvm/java-11-openjdk-amd64/bin/java): not using repository mapping () since it does not exist\r\nERROR[runfiles.bash]: cannot look up runfile \"/usr/lib/jvm/java-11-openjdk-amd64/bin/java\" (RUNFILES_DIR=\"/b/f/w/bazel-out/k8-fastbuild/bin/src/test/shell/bazel/bazel_example_test.runfiles\", RUNFILES_MANIFEST_FILE=\"\")\r\n```", "@coeuvre I suspected a different issue, so I would like to confirm that isn't a problem either: The Bash runfiles library currently relies on absolute (!) paths of generated files containing an `execroot` segment. Is that guaranteed to be the case with remote execution, just like it is with local execution?", "The logic for absolute paths passed to `rlocation` broke. I am investigating why the existing test didn't catch that.", "> @coeuvre I suspected a different issue, so I would like to confirm that isn't a problem either: The Bash runfiles library currently relies on absolute (!) paths of generated files containing an `execroot` segment. Is that guaranteed to be the case with remote execution, just like it is with local execution?\r\n\r\nMy gut feeling is that `execroot` is not included in the path when actions inspect the path inside remote worker but I may be wrong. We use `execroot` as input root when uploading inputs for actions. Remote worker is free to move the input root to any local directories so a file `/.../execroot/bazel-out/a` in bazel host is probably become `/..../bazel-out/a` in remote worker.", "@coeuvre Thanks, I will remove that assumption in a second PR.\r\n\r\n@meteorcloudy The Bash runfiles test didn't cover the exit codes of `rlocation`, only whatever it printed, and thus missed some corner cases. I have reworked the tests to check the exit codes and am now fixing these cases.", "@fmeum @coeuvre Thanks!!", "I submitted two PRs with fixes, #16945 and #16946.", "I think I know how to fix the RBE test failures, but now that I studied this more closely, there is some underlying inconsistent behavior that makes the library difficult to test and potentially also hard to use: `$(rlocation does/not/exist)` has exit code 0 if and only if the runfiles manifest exists. This means that the behavior differs between local and remote execution on Windows, which is probably why we only noticed this in downstream tests.\r\n\r\n@laszlocsomor Sorry for the sideline mention, but can you recall how you intended the exit code of Bash `$(rlocation ...)` to behave?", "> `$(rlocationpath does/not/exist)` has exit code 0 if and only if the runfiles manifest exists.\r\n\r\nDid you mean `$(rlocation does/not/exist)`?", "@Wyverald Yes, sorry, fixed.", "Fixing this inconsistency isn't required to get the tests to pass, but if we want to do it, 6.0.0 is probably the right time. Since `$(failing_command)` doesn't usually fail the outer shell script, having `$(rlocation does/not/exist)` exit with exit code 1 in all cases doesn't seem to be *that* breaking.\r\n\r\n@Wyverald @meteorcloudy What do you think?" ]

[]

[ "type: bug", "P1" ]

[ "tools/bash/runfiles/runfiles.bash", "tools/bash/runfiles/runfiles_test.bash" ]

[ "tools/bash/runfiles/runfiles.bash", "tools/bash/runfiles/runfiles_test.bash" ]

[]

[ "keyword_pr_to_issue" ]

[ "Fix has been cherry-picked at https://github.com/bazelbuild/bazel/pull/17004" ]

[]

[]

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

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

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

[ "keyword_pr_to_issue" ]

[ "In addition, it's possible for an action to consume only one of the files in a tree artifact (see `SpawnActionTemplate`), in which case the prefetcher will try to fetch the entire tree artifact; this currently fails because the metadata for the entire tree artifact isn't present (see https://github.com/bazelbuild/bazel/issues/16987).", "Reopening because https://github.com/bazelbuild/bazel/commit/25ba76c1b8d420ebae43c1f59890ae46153c2240 was just a workaround, and we have yet to fix it properly.", "@tjgq Should this be cherry-picked?", "Yes, but I'm not sure 6.1.0 is still accepting cherry-picks. If we decide to do an rc3, I'll cherry-pick it there. Otherwise, it will have to go into 6.2.0.", "@bazel-io fork 6.2.0" ]

[]

[ "type: bug", "P2", "team-Remote-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/actions/ActionInputPrefetcher.java", "src/main/java/com/google/devtools/build/lib/remote/RemoteActionInputFetcher.java", "src/main/java/com/google/devtools/build/lib/skyframe/ActionExecutionFunction.java", "src/main/java/com/google/devtools/build/lib/skyframe/ActionInputMapHelper.java", "src/main/java/com/google/devtools/build/lib/skyframe/CompletionFunction.java", "src/main/java/com/google/devtools/build/lib/skyframe/SkyframeActionExecutor.java" ]

[ "src/main/java/com/google/devtools/build/lib/actions/ActionInputPrefetcher.java", "src/main/java/com/google/devtools/build/lib/remote/RemoteActionInputFetcher.java", "src/main/java/com/google/devtools/build/lib/skyframe/ActionExecutionFunction.java", "src/main/java/com/google/devtools/build/lib/skyframe/ActionInputMapHelper.java", "src/main/java/com/google/devtools/build/lib/skyframe/CompletionFunction.java", "src/main/java/com/google/devtools/build/lib/skyframe/SkyframeActionExecutor.java" ]

[ "src/test/java/com/google/devtools/build/lib/skyframe/SequencedSkyframeExecutorTest.java", "src/test/shell/bazel/remote/build_without_the_bytes_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "As an add-on: Seems like all these variables are missing from aquery output, too.", "It seems like maybe there's something much more general broken with environment variable reporting for actions.\r\n\r\nI ran back into an equivalent problem on Windows and with aquery, and it really does complicate and embrittle tooling.\r\n\r\nI've updated the original report to reflect that.\r\n\r\n@gregestren @oquenchil, might this merit different labels/priorities if its a more general, non-apple-specific, non-extra-action-specific bug?\r\n\r\n", "I don't have expertise on this part of the code base (aquery, extra actions, environment variable consumption). I'm going to defer to @oquenchil .", "This is blocking us too. Is there any meaningful progress here?", "(Sadly not that I know of.) But I'm curious what you're building on top of aquery/extra_actions, @sualehasif!", "I think this may be an _even_ more general bug where toolchain environment variables are missing from aquery output.\r\n\r\n(I've been trying to help a user over at https://github.com/hedronvision/bazel-compile-commands-extractor/issues/83#issuecomment-1312067222 and noticing that EM_BIN_PATH and the other emscripten environment variables are missing from aquery output, despite being set in [the toolchain](https://github.com/emscripten-core/emsdk/blob/005063304bb4e575f97b7c5f50368f937c06cb54/bazel/emscripten_toolchain/toolchain.bzl))\r\n\r\n@oquenchil, any chance of a higher prioritization, given it's now clear it's a more general issue (not platform-specific), present across the action querying mechanisms (not just extra action), and causing problems for others, too?", "Is this just for C++ rules and it works for others like Java?", "I'm pretty sure it also affects Objective-C, at least, but I'm not sure whether it affects other language toolchains/Java.\r\n[It's been a problem everywhere I've looked, but I haven't needed to dive into the Bazel toolchains for other languages just yet.] ", "This is complicating tooling for me as well. I've implemented a toolchain to use the EWDK to build windows drivers, and it is a real bummer to not be able to query the INCLUDE environment variable for this.", "@meisterT is this something that aquery could get to?", "Regarding `DEVELOPER_DIR` and `SDKROOT`: I suspect that these are missing from `aquery` output as the real values for these variable are only determined at execution time by running `xcode-locator` and are then injected into spawns that need them (see https://cs.opensource.google/bazel/bazel/+/abe26b84df03eac149f731f1e71fd9a5f0138f0a:src/main/java/com/google/devtools/build/lib/exec/local/XcodeLocalEnvProvider.java;l=68). Ultimately that seems to be necessary as the values have to be host-dependent absolute paths and thus shouldn't be part of the action key.", "Regarding the missing toolchain variables on all OSes: Based on a cursory glance, it looks to me as if `aquery` only emits environment variables for `SpawnAction`s: https://cs.opensource.google/bazel/bazel/+/abe26b84df03eac149f731f1e71fd9a5f0138f0a:src/main/java/com/google/devtools/build/lib/skyframe/actiongraph/v2/ActionGraphDump.java;l=177. But the C++ compilation and linking actions aren't `SpawnAction`s. Maybe that `if` could just check for `AbstractAction` instead?", "Thanks so much for diving in, guys.", "Can confirm that @fmeum's good work in https://github.com/bazelbuild/bazel/pull/17108 surfaced the emscripten environment variables--but the Apple (and I presume the Windows) variables from the original issue are still missing.\r\n\r\nStill think this might be worth fixing (at least by starlarkificiation time). " ]

[ "@kshyanashree This test made it in by mistake, only the one after it is part of the original PR.", "Hi @fmeum! I removed this test and ran the presubmits, but still checks are failing. Could you please look into it?" ]

[ "type: bug", "P4", "platform: apple", "team-Rules-CPP", "team-Rules-ObjC" ]

[ "src/main/java/com/google/devtools/build/lib/skyframe/actiongraph/v2/ActionGraphDump.java" ]

[ "src/main/java/com/google/devtools/build/lib/skyframe/actiongraph/v2/ActionGraphDump.java" ]

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

[ "keyword_pr_to_issue" ]

[ "I have a PR that should fix this: https://github.com/bazelbuild/bazel/pull/17045\r\n" ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/actions/AbstractAction.java", "src/main/java/com/google/devtools/build/lib/actions/ActionExecutionMetadata.java", "src/main/java/com/google/devtools/build/lib/pkgcache/LoadingPhaseCompleteEvent.java", "src/main/java/com/google/devtools/build/lib/runtime/SkymeldUiStateTracker.java", "src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java", "src/main/java/com/google/devtools/build/lib/skyframe/TargetPatternPhaseFunction.java" ]

[ "src/main/java/com/google/devtools/build/lib/actions/AbstractAction.java", "src/main/java/com/google/devtools/build/lib/actions/ActionExecutionMetadata.java", "src/main/java/com/google/devtools/build/lib/pkgcache/LoadingPhaseCompleteEvent.java", "src/main/java/com/google/devtools/build/lib/runtime/SkymeldUiStateTracker.java", "src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java", "src/main/java/com/google/devtools/build/lib/skyframe/TargetPatternPhaseFunction.java" ]

[ "src/test/java/com/google/devtools/build/lib/runtime/SkymeldUiStateTrackerTest.java", "src/test/java/com/google/devtools/build/lib/runtime/UiStateTrackerTest.java" ]

[ "keyword_pr_to_issue" ]

[]

[]

[ "type: bug", "team-Configurability", "untriaged" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/config/BuildConfigurationValue.java", "src/main/java/com/google/devtools/build/lib/analysis/config/ExecutionTransitionFactory.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/config/BuildConfigurationValue.java", "src/main/java/com/google/devtools/build/lib/analysis/config/ExecutionTransitionFactory.java" ]

[]

[ "keyword_pr_to_issue" ]

[ "cc @comius @c-mita ", "It's seems the issue is that cc_tests unconditionally pulls those coverage tools dependencies.\r\nhttps://cs.opensource.google/bazel/bazel/+/master:src/main/java/com/google/devtools/build/lib/analysis/test/TestConfiguration.java;l=255-270" ]

[]

[ "type: bug", "P2", "team-Rules-CPP" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/BUILD", "src/main/java/com/google/devtools/build/lib/rules/BUILD" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/BUILD", "src/main/java/com/google/devtools/build/lib/rules/BUILD" ]

[]

[ "keyword_pr_to_issue" ]

[ "Forgot to add the place in code that seems to be the writing the broken entry to the disk\r\nhttps://github.com/bazelbuild/bazel/blob/3748bcc8bf0c45e34295583fafaf1f52449114b8/src/main/java/com/google/devtools/build/lib/remote/disk/DiskAndRemoteCacheClient.java#L202", "@coeuvre @tjgq ", "@bazel-io flag", "Thanks for the report!\r\n\r\nThe issue is tricky. When building without the bytes with disk cache alone, we want to check ac entry integrity for disk cache. Otherwise the build could fail if the any blob is missing when we are reading it later.\r\n\r\nHowever, when combined with remote cache, if we hit remote cache, we only download ac and skip downloading of blobs. Hence the AC in the disk cache is always incomplete until all the referenced blobs are downloaded later. The current code didn't check the integrity before deciding using the disk part of a combined cache.\r\nhttps://github.com/bazelbuild/bazel/blob/3748bcc8bf0c45e34295583fafaf1f52449114b8/src/main/java/com/google/devtools/build/lib/remote/disk/DiskAndRemoteCacheClient.java#L190\r\n\r\nWe could fix that by checking the integrity when deciding which part to use of a combined cache. The problem is we will never hit the disk cache until all the referenced blobs are downloaded. The fix could reduce the cache-hit rate for disk cache. Considering 2 consecutive builds:\r\n\r\n```\r\nbazel build --config=remote_exec_with_disk --remote_download_minimal ...\r\nbazel build --config=remote_exec_with_disk --remote_download_minimal ...\r\n```\r\n\r\nThe second build will miss the disk cache because blobs referenced by AC are not downloaded (but can still hit the remote cache). On the other hand, this should the correct behaviour. Since remote cache could evict blobs. We shouldn't accept AC from disk cache if any referenced blob is missing, hoping we could get it from remote cache -- it could also be missing there.\r\n\r\nAnother fix is drop the integrity check for AC from disk cache when using combined cache. This will increase the disk-cache hit rate, but the chance to trigger \"missing blobs\" error is higher if the TTL of blobs on remote cache is low.\r\n\r\nI will go with the second fix for now (for performance reason) and see how can we improve it with lease service.", ">Another fix is drop the integrity check for AC from disk cache when using combined cache. This will increase the disk-cache hit rate, but the chance to trigger \"missing blobs\" error is higher if the TTL of blobs on remote cache is low.\r\n\r\nCan I vote that we _don't_ do this? At least until the lease service is up and running. The change to have the AC check if it's entries are valid is a very good one, in particular with remote caching. If you need to disable the check, and you do it only when disk cache + remote exec, but not disk cache + remote cache?", "I agree AC check is a good one. But if we enable it for a combined cache (either disk cache + remote exec, or disk cache + remote cache), we will probably never hit the disk cache when building without the bytes.\r\n\r\nTaking the repro as an example, since we are using minimal mode, the outputs of `//:main` will nerve be downloaded which means we will never hit the disk cache if we check the integrity.\r\n\r\nIn real world builds, I believe there are actions whose outputs will never be downloaded. If we do AC check, these actions will never hit the disk cache.", "I'm fine with the AC result being used in the combined cache, as long as Bazel calls FindMissingBlobs on the remote cache to verify that the result is valid. I think it would be cheaper if Bazel instead just asked for the AC result from the remote cache in that case.\r\n\r\nThe main benefit of the combined disk cache is the storing of the CAS items. There might be some local-only actions that do end up storing themselves, even with BwtB. I feel that BwtB + disk cache only is going to be very rare. The most common case will be the combined cache. And if we want the BwtB experience to be as nice as possible, the combined cache shouldn't degrade the experience. By that, I mean that if one was using just the remote cache, and `bazel clean --expunge` was run (until the lease service lands), Bazel wouldn't run into missing blobs errors because the remote cache wouldn't return AC results if they had missing CAS blobs. This change would degrade that because the disk cache could incorrectly return an invalid AC result.", "I get you. From this perspective, not throwing an error is definitely a better experience than slowing the build. And yes, just check remote AC is probably not slower. Reusing CAS is the main benefit." ]

[]

[ "type: bug", "P1", "team-Remote-Exec" ]

[ "src/main/java/com/google/devtools/build/lib/remote/disk/DiskAndRemoteCacheClient.java", "src/main/java/com/google/devtools/build/lib/remote/disk/DiskCacheClient.java" ]

[ "src/main/java/com/google/devtools/build/lib/remote/disk/DiskAndRemoteCacheClient.java", "src/main/java/com/google/devtools/build/lib/remote/disk/DiskCacheClient.java" ]

[ "src/test/java/com/google/devtools/build/lib/remote/BUILD", "src/test/java/com/google/devtools/build/lib/remote/DiskCacheIntegrationTest.java" ]

[ "keyword_pr_to_issue" ]

[ "CC @katre @illicitonion ", "I prototyped a PR that realizes the desired behavior (but isn't polished): https://github.com/bazelbuild/bazel/pull/16968\r\n\r\n@katre Given that 5.4.0 hasn't been released yet, do you see a way to add no-op parameters to `find_cpp_toolchain` and/or targets to `@bazel_tools//tools/cpp` that would allow rulesets to make use of optional toolchains sooner? \r\n\r\nFor example, if we added an alias of `current_cc_toolchain` called `optional_current_cc_toolchain` to Bazel 5.4.0, it could become the needed \"optional\" form in Bazel 6.X and rulesets could refer to it once their minimum supported Bazel version is 5.", "Notes, no particular order:\r\n\r\n1. Please note that this is specifically for CC toolchains. Can you update the issue name and description to clarify that?\r\n2. I see in your example that you get the error text \"Unable to find a CC toolchain using toolchain resolution. Did you properly set --platforms\". Are you actually using your modified version of `find_cpp_toolchain`?\r\n\r\nI can see backporting a small fix to `find_cpp_toolchain` to make it more clear, but looking at your proposed #16968, it looks overly complicated to backport to an LTS release, and not needed at HEAD (because in HEAD optional toolchains exist).\r\n\r\nWhat specific versions do you want to support, and which flags with each version? Part of the appeal of the LTS model for the Bazel team was, honestly, not doing this sort of backporting and incremental rollout of features, because it gets very complicated (and risky) to support the older versions. The optional toolchains API was added a version before the actual implementation in order to allow projects to use the new API, and then get the new features when users upgrade to a Bazel that contained them. Was this the wrong approach? What would have been a better approach, so that we can handle this next time?", "> 1. Please note that this is specifically for CC toolchains. Can you update the issue name and description to clarify that?\r\n\r\nGood point, done.\r\n\r\n> 2. I see in your example that you get the error text \"Unable to find a CC toolchain using toolchain resolution. Did you properly set --platforms\". Are you actually using your modified version of `find_cpp_toolchain`?\r\n\r\nYes, this text exists in two places:\r\n1. In `@bazel_tools//tools/cpp:toolchain_utils.bzl%find_cpp_toolchain`, which I thus had to inline and modify slightly (simply replaced `fail` with `return None`) here in order to make it work with optional toolchains.\r\n2. In `com.google.devtools.build.lib.rules.cpp.CppHelper#getToolchainFromPlatformConstraints`, which is used by `cc_toolchain_alias` to get the toolchain. This is the root cause of this issue: Compatibility with `--noincompatible_enable_cc_toolchain_resolution` means that the dependency on `cc_toolchain_alias` is needed, but if it is there, this method is always called and always throws this error, even if the toolchain dependency is marked optional.\r\n\r\nThe error you see in my example comes from 2., not 1.\r\n\r\n> I can see backporting a small fix to `find_cpp_toolchain` to make it more clear, but looking at your proposed #16968, it looks overly complicated to backport to an LTS release, and not needed at HEAD (because in HEAD optional toolchains exist).\r\n\r\nI am not proposing to backport #16968 - the change we could backport could be as simple as adding `alias(name = \"optional_current_cc_toolchain\", actual = \"current_cc_toolchain\")` to `@bazel_tools//tools/cpp`. The backport is not meant to add any functionality or make optional toolchains available, it just makes it so that rulesets can expect this target to exist earlier (that is, already when Bazel 5 is their minimum supported version).\r\n\r\nWhile optional toolchains exist at HEAD, I don't see a way for a ruleset to use them for C++ even at HEAD while remaining compatible with scenarios in which `--incompatible_enable_cc_toolchain_resolution` is disabled - this is why I created this issue. It is not about preserving compatibility with an earlier version of Bazel, it is just about compatibility with both values of the flag for any recent Bazel version, including HEAD. #16968 is a draft of a change that would make this possible at HEAD.\r\n\r\n> What specific versions do you want to support, and which flags with each version?\r\n\r\nSummarizing the above, I want to allow a Bazel ruleset to declare an optional C++ toolchain dependency and simultaneously:\r\n* make use of the toolchain actually being optional with Bazel@HEAD and `--incompatible_enable_cc_toolchain_resolution=true`\r\n* do not break with Bazel@HEAD and `--incompatible_enable_cc_toolchain_resolution=false`\r\n\r\nThe example I gave is meant to demonstrate that this doesn't seem to be possible today, so even if we do not care about backwards compatibility, there is something that needs to be fixed.\r\n\r\nIf we do care about allowing a ruleset to do this and remain compatible with a previous version of Bazel (e.g. 5 or 6), I think that this would require a backport even if https://github.com/bazelbuild/rules_go/pull/3390#discussion_r1043539582 is used to only request an optional toolchain dependency if the Bazel version supports it. That is because I don't see a way to fix the example without introducing a new `cc_toolchain_alias`, as is done in #16968.\r\n\r\n> Part of the appeal of the LTS model for the Bazel team was, honestly, not doing this sort of backporting and incremental rollout of features, because it gets very complicated (and risky) to support the older versions. The optional toolchains API was added a version before the actual implementation in order to allow projects to use the new API, and then get the new features when users upgrade to a Bazel that contained them. Was this the wrong approach? What would have been a better approach, so that we can handle this next time?\r\n\r\nI think that the approach was correct and adding the API a version before the feature is makes it so that rulesets can use it earlier, gracefully falling back to the old behavior but not breaking with versions of Bazel that don't have the feature. In fact, the proposal for a new `cc_toolchain_alias` target above is exactly of this type: It would add API to Bazel 5 and/or 6, with the actual \"feature\"/bug fix coming in (likely) Bazel 6.1.\r\n\r\nThe only thing that seems to have been missed is compatibility with `--incompatible_enable_cc_toolchain_resolution=false`. Given that this flag is toggleable even at HEAD makes it even more important for rulesets to remain compatible with both values and, as I intend to demonstrate with this issue, that doesn't seem to be possible at all with any version of Bazel.\r\n", "If I understand correctly this is about the old mechanism for toolchain resolution used by C++ rules and the new mechanism not being compatible? If the old one is deprecated, we should focus on removing it instead, no need to support the flag --noincompatible_enable_cc_toolchain_resolution. Every incompatible flag is temporary and not meant for long term support.\r\n\r\nIf the problem is that without fixing the issue described here it becomes impossible to migrate, then that'd be a different story.\r\n", "@oquenchil The problem is that `--noincompatible_enable_cc_toolchain_resolution` is still the default. Even if it is flipped in Bazel 7 (not even sure that is planned), rulesets will only be able to assume it is flipped when they stop supporting Bazel 6, which will happen when Bazel 8 is released at the earliest. As a result, rulesets may not be able to use optional toolchains until the release of Bazel 8, which is probably a year away.\r\n\r\nAdding a single alias to `@bazel_tools` now (that is, for Bazel 6.1), could speed this up by at least half a year.", "Understood. Thanks for the clarification.", "@katre For the reasons given by Fabian I think we should backport to 6.1 the simple change adding an alias to a BUILD file (not necessary for 5.4, although doing it for 5.4 wouldn't hurt either). https://github.com/bazelbuild/bazel/pull/16968 can be checked in at HEAD and that's it, no need to backport that PR. It's only the alias that would be needed to speed up progress by a lot as explained by Fabian.\r\n\r\nI know that this falls in the domain of C++ rules but you have a lot of expertise in this area. With those clarifications, would you be ok with backporting the alias?", "This looks very reasonable. I took a quick look at the change and it makes sense, but you should do the detailed review." ]

[]

[ "type: feature request", "P2", "team-Rules-CPP" ]

[ "tools/cpp/BUILD.tools", "tools/cpp/toolchain_utils.bzl" ]

[ "tools/cpp/BUILD.tools", "tools/cpp/toolchain_utils.bzl" ]

[]

[ "keyword_pr_to_issue" ]

[ "Cherry-picked in #17360 " ]

[]

[]

[ "src/main/java/com/google/devtools/build/lib/analysis/actions/LazyWritePathsFileAction.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaCompilationHelper.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/actions/LazyWritePathsFileAction.java", "src/main/java/com/google/devtools/build/lib/rules/java/JavaCompilationHelper.java" ]

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

[ "keyword_pr_to_issue" ]

[ "I guess this is done?" ]

[]

[ "P1", "type: process", "team-Rules-CPP", "potential 5.x cherry-picks" ]

[ "src/main/starlark/builtins_bzl/common/cc/cc_binary.bzl", "src/main/starlark/builtins_bzl/common/cc/experimental_cc_shared_library.bzl" ]

[ "src/main/starlark/builtins_bzl/common/cc/cc_binary.bzl", "src/main/starlark/builtins_bzl/common/cc/experimental_cc_shared_library.bzl" ]

[ "src/main/starlark/tests/builtins_bzl/cc/cc_shared_library/test_cc_shared_library/BUILD.builtin_test", "src/main/starlark/tests/builtins_bzl/cc/cc_shared_library/test_cc_shared_library/failing_targets/BUILD.builtin_test", "src/main/starlark/tests/builtins_bzl/cc/cc_shared_library/test_cc_shared_library/foo.cc", "src/main/starlark/tests/builtins_bzl/cc/cc_shared_library/test_cc_shared_library/starlark_tests.bzl", "src/main/starlark/tests/builtins_bzl/cc/cc_shared_library/test_cc_shared_library3/BUILD.builtin_test", "src/main/starlark/tests/builtins_bzl/cc/cc_shared_library/test_cc_shared_library3/diff_pkg.cc", "src/main/starlark/tests/builtins_bzl/cc/cc_shared_library/test_cc_shared_library3/diff_pkg.h" ]

[ "keyword_pr_to_issue" ]

[ "We got the same issue on macOS, but it seems to be a problem with git (tried with `git clone --shallow-since`).", "Yeah, seems like it. Though I'd say Bazel is still responsible for avoiding recommending `shallow_since` unless/until this is fixed (a version check on git might make sense). Otherwise it's really confusing for the user to be recommended an option that's broken.", "@aehlig Do you know what causes it? ", "I did some digging here, and reported [an upstream bug to `git`](https://public-inbox.org/git/CAPSG9dZV2EPpVKkOMcjv5z+NF7rUu=V-ZkZNx47rCv122HsiKg@mail.gmail.com/T/#u)\r\n\r\nThe upshot is that `--shallow-since` is subtly incorrect when the provided timestamp would cause us to include _some_ parents of a merge, but not others. In the case provided here, nelhage/rules_boost@df90835 has a commit timestamp of 1569511515. That commit is the second part of a merge, nelhage/rules_boost@76f3064; that merge's first parent is nelhage/rules_boost@a83197e, with an older commit timestamp of 1569487118\r\n\r\nThe bug I reported results in `git` discarding _all_ parents of the merge (nelhage/rules_boost@76f3064), including the desired commit.\r\n\r\nFor now, the workaround for bazel _users_ -- where available -- is to switch your `commit=` line to point to a commit on the first-parent spine, e.g. nelhage/rules_boost@76f3064. If you add the suggested `shallow_since` line for that commit, things work fine.\r\n\r\nFor Bazel itself, it might be nice to recognize this situation and adjust the `shallow_since` recommendation, at least until the upstream bug can be sorted out. I'm not certain offhand what the right/easiest heuristic to detect it is, though.", "It turns out that the workaround proposed here doesn't work either.\r\n\r\nThe shallow_since bug is even worse than I expected. It stops at the first commit which has any parents before the cutoff time. This means we have the following horrible scenario:\r\n\r\n1. Latest commit on main branch is aaaa, with parents baaa and bbbb. aaaa is at 8am, baaa is at 6am and bbbb is at 4am\r\n2. I create a git_repository rule with commit = \"aaaa\", shallow_since = \"4am\"\r\n3. This works fine when I build it\r\n4. Somebody adds a new commit caaa to main branch, with parents aaaa and cbbb. caaa is at 9am, cbbb is at 2am\r\n5. git clone --shallow_since=4am will now fetch only caaa because it has a parent before 4am, and aaaa will not be fetched\r\n\r\nThis is completely unusable and I don't see a workaround.", "Yeah, I agree with @asuffield's analysis -- I ran into that situation since posting https://github.com/bazelbuild/bazel/issues/10292#issuecomment-680205748. In light of this and of #12857 it really seems like the right path forward is to stop suggesting `--shallow-since`. It might be appropriate in cases where the developers know neither of these problems will apply -- or are able to debug their way out if necessary -- but it's clearly too broken to suggest as a default.", "I'm piecing together an alternative solution - for cases where a tag or commit hash is given, we can use point fetch with --depth=1 as long as we verify the results and correctly fallback to deep fetch if we don't get what we needed. It doesn't work on all git versions, but I don't believe there is a solution other than deep fetch for older git.", "If you auto-detect GitHub URLs you can instead just suggest HTTPS cloning. It's probably a good suggestion in any case; we switched to that as it's faster too.", "> If you auto-detect GitHub URLs you can instead just suggest HTTPS cloning. It's probably a good suggestion in any case; we switched to that as it's faster too.\r\n\r\nThis is an interesting idea which would likely be best implemented in git itself - the protocol's rich enough to pull it off.", "`git archive` already exists in git. GitHub has refused to support it.", "I would like to get some progress on this issue. Would it be fine if I create a PR to remove the debug log saying:\r\n\r\n> indicated that a canonical reproducible form can be obtained by modifying arguments shallow_since = \r\n\r\nThe issue with this log is that one a user see it the first thing will do is to set up the shallow_since, but because of this bug it is something that should consider twice.\r\n" ]

[]

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

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

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

[]

[ "keyword_pr_to_issue" ]

[ "+1\r\n\r\nAlso git's `--shallow-since` has a bug that can sometimes make git impossible to fetch the desired commit using its suggested timestamp: https://github.com/bazelbuild/bazel/issues/10292.", "There is no way to persuade git to fetch a specific commit with --depth=1. If you try fetching a commit hash, most servers will just give you \"server does not allow request for unadvertised object\", and if you fetch a ref name then it can return anything and is not reproducible.\r\n\r\nshallow_since is the only feature git offers to do this. Unfortunately, it's completely broken.", "Correction: --depth=1 on a single commit only works with protocol v2, see https://github.com/bazelbuild/bazel/pull/12199 ", "@asuffield I don't think we're on protocol v2 (although I'm not completely sure how to verify that) and we've been using the commit only since I posted the issue without any apparent problems or non-reproducibility. As far as I can tell, the actions bazel is taking are equivalent to the following commands, which appear to work for me.\r\n\r\n```\r\nmkdir ruby ; cd ruby\r\ngit init\r\ngit remote add origin [email protected]:ruby/ruby.git\r\n# A random old commit\r\ngit fetch origin 5de38c41ae7bf17ae599fdfa9f8face87f16d8bb --depth 1\r\ngit checkout 5de38c41ae7bf17ae599fdfa9f8face87f16d8bb\r\n```\r\n\r\n Do you mean that it doesn't work reliably with all git servers, so maybe it works on github but not elsewhere?\r\n\r\n```\r\n$ git --version\r\ngit version 2.25.1\r\n$ echo $GIT_PROTOCOL\r\n\r\n```", "It looks like maybe the server config option `uploadpack.allowReachableSHA1InWant` needs to be set to enable this, but the repository rules already seem to handle falling back when the server doesn't support this option anyway. `shallow_since` is perhaps better than the fallback option, but it's probably worse than fetching the commit directly when supported, so it shouldn't be recommended when fetching by commit is supported.\r\n\r\nhttps://github.com/bazelbuild/bazel/blob/8346ea4cfdd9fbd170d51a528fee26f912dad2d5/tools/build_defs/repo/git_worker.bzl#L129", "It definitely fails with git 2.18 (ugh, upgrading that right now) against gitlab, I haven't tested exhaustively. But the main problem in the PR linked above is that one of bazel's target platforms has an old git version, and that's going to sabotage us here.", "Are there ways to improve this situation, even given bazel's support for ubuntu 18.04 and its associated git version?\r\n\r\nOne option is simply to drop the warning from this rule (by making the rule output a canonical version that omits `shallow_since` if the caller did not specify it).\r\n\r\nAnother is to add something like an explicit `depth` arg to the `git_repository` rules: I don't like that from an API perspective, but it could allow callers to explicitly use a depth-based canonical form (rather than a `shallow_since`-based form) if that works for their environments.\r\n\r\nI think that any change to the canonical form of this rule would involve changing what's returned from the [`git_repo` function](https://github.com/bazelbuild/bazel/blob/master/tools/build_defs/repo/git_worker.bzl#L94-L96).\r\n\r\nAre there other ideas that could remove warnings from a correct repository setup that's relying on `depth=1` rather than a `shallow_since` date?" ]

[]

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

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

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

[]

[ "keyword_pr_to_issue" ]

[ "Hello @cjohnstoniv, Can you provide an example code to reproduce the above error. Thanks!", "Unfortunately not easily as the code base we build this on is private and I cannot share it. Also it's built using a localjdk workspace config so none of that would be repeatable on anyone else's machine (unless they have java in the same path).\r\n\r\nThe issue is quite straightforward / obvious though. The `Optional.isEmpty` method was added in Java 9+ so it would (as expected) fail in any Java 8 JVMs. My ask is can we just make this 1 line Java 8 compliant by making it `!optional.isPresent()` instead.\r\n\r\nIf Bazel still intends to support Java 8+ then this would be a requirement to make sure that the coverage tools remain Java 8 compliant. Unless the intention is Java 8+ compilation and not runtime.\r\n\r\nThis all works just fine on Bazel 5.4.0 and if you look at the diffs between 5.4.0 and 6.0.0, you can see this class updated to Java 9+ only complaint code.", "I personally don't see a reason not to make this one line compatible with Java 8. But even if you are *targeting* Java 8, I would expect the coverage tool to be built in the `exec` configuration and (by default) use a remote JDK 11. Using a more modern JDK just for tools could help you in the meantime. ", "We want to run our runtime on Java 8 too as unfortunately our org still has a lot of software still running on Oracle Java 8 JDKs (gross, I know). So from a Bazel build/test perspective we are trying to keep it all as consistent as possible. Testing with Java 11 runtime but with Java 8 target, isn't the same as compiling and testing on the same JDK version. There are plenty of things like needing to add-opens/exports to make Java 8 compiled code run in Java 11 runtimes.\r\n\r\nIf the answer is the devs don't feel making this Java 8 compatible is the right thing to do, then we should update the docs to make it clear that Java 8 is supported purely from compilation and not from runtime.", "I don't mean \"compile with Java 11, run with Java 8\". What I am referring to is using something like `--java_runtime_version=local_jdk_8 --tool_java_runtime_version=remotejdk_11` to:\r\n* build and test all regular targets with JDK 8\r\n* build and execute tools run as part of the build itself (for example the coverage tools) with JDK 11\r\n\r\n@cjohnstoniv Could you share at least the parts of your `.bazelrc` and `WORKSPACE` that configure Java toolchains?\r\n\r\n", ".bazelrc\r\n\r\n```\r\nbuild --embed_label=9.0.LTS-SNAPSHOT\r\nbuild --java_language_version=8\r\nbuild --tool_java_language_version=11\r\nbuild --tool_java_runtime_version=remotejdk_11\r\nbuild --java_runtime_version=localjdk_8\r\nbuild --verbose_failures\r\nbuild --experimental_ui_max_stdouterr_bytes=100048576\r\nstartup --max_idle_secs=300\r\ntest --test_output=errors\r\ntest --test_verbose_timeout_warnings\r\ncoverage --collect_code_coverage\r\ncoverage --combined_report=lcov\r\ncoverage --coverage_report_generator=@bazel_aladdin//sonarqube:SonarQubeCoverageGenerator\r\nbuild --remote_cache=grpc://[SOME IP]:8980\r\nbuild --remote_max_connections=200\r\ncoverage --experimental_split_coverage_postprocessing\r\ncoverage --experimental_fetch_all_coverage_outputs\r\nbuild --disk_cache=/Users/chjohnst/mazel/.cache\r\n```\r\n\r\nWORKSPACE toolchains:\r\n\r\n```\r\nload(\"@bazel_tools//tools/jdk:local_java_repository.bzl\", \"local_java_repository\")\r\n\r\nlocal_java_repository(\r\n name = \"localjdk\",\r\n version = '8',\r\n java_home = \"/Library/Java/JavaVirtualMachines/jdk1.8.0_202.jdk/Contents/Home\"\r\n)\r\n```\r\n\r\n@fmeum based on your comments above, I was under the impression that using `--tool_java_language_version=11 --tool_java_runtime_version=remotejdk_11` would have solved the issue (before we had --tool_java_language_version=8 --tool_java_runtime_version=remotejdk_11). But I still see the same error in the test logs:\r\n\r\nException in thread \"Thread-0\" java.lang.NoSuchMethodError: java.util.Optional.isEmpty()Z\r\n at com.google.testing.coverage.JacocoLCOVFormatter$1.getExecPathForEntryName(JacocoLCOVFormatter.java:73)\r\n at com.google.testing.coverage.JacocoLCOVFormatter$1.visitBundle(JacocoLCOVFormatter.java:117)\r\n at com.google.testing.coverage.JacocoCoverageRunner.createReport(JacocoCoverageRunner.java:156)\r\n at com.google.testing.coverage.JacocoCoverageRunner.create(JacocoCoverageRunner.java:132)\r\n at com.google.testing.coverage.JacocoCoverageRunner$2.run(JacocoCoverageRunner.java:568)\r\nGCov does not exist at the given path: ''\r\n", "If that also doesn't work I am personally out of ideas why you are seeing this issue. team-Rules-Java can probably help here.", "I'll see if I can find some time later this week from my personal PC to upload a small Java project that recreates the issue so others can test.\r\n\r\nOddly when I think about it, I would've expected the tools compilation stage to fail when I had set `--tool_java_language_version=8` as it would have tried compiling Java 11 code on Java 8; but instead it fails with a runtime NoSuchMethodError which suggests the tool is compiling on Java 11+ but for whatever reason is not running on Java 11+.", "[broken-java8-coverage.zip](https://github.com/bazelbuild/bazel/files/10460205/broken-java8-coverage.zip)\r\n\r\nAttached is a sample project that recreates the issue following the following steps (assumes you downloaded the zip to an empty dir, also assumes linux OS, modify the less location for other OS's, you may also have to modify your Java 8 localjdk location in WORKSPACE file):\r\n\r\n`unzip broken-java8-coverage.zip; bazelisk coverage //...; less bazel-out/k8-fastbuild/testlogs/unit-tests/test.log`\r\n\r\nYou will see that despite the tool java level and runtime being 11, the test runner still fails to calculate coverage due to the Java 9+ added Optional method.", "Also not sure if anyone has any experience with the Bazel distribution build but I tried checking out the 6.0 tag and compiling my own version of Bazel with changes to the JacocoLCOVFormatter class; however I found that seemingly when I run `compile.sh` and/or `bazel build //src:bazel` my changes to the class aren't picked up. I've confirmed this at least by making non-compilable changes to the class (i.e bad characters) and the bazel build runs just fine.\r\n\r\nIs there someway I can rebuild/retest the JacocoLCOVFormatter class? I was trying to test what would happen if we made the `isEmpty()` call `!isPresent()` instead.", "@cushon This may he another instance of https://github.com/bazelbuild/bazel/issues/17281: The tool Java language level/runtime version not behaving as one may expect. Could you also provide some insight for this particular situation?", "@comius \r\n\r\nI can reproduce (thanks for https://github.com/bazelbuild/bazel/issues/17165#issuecomment-1397472443)!\r\n\r\n```\r\nException in thread \"Thread-0\" java.lang.NoSuchMethodError: java.util.Optional.isEmpty()Z\r\n at com.google.testing.coverage.JacocoLCOVFormatter$1.getExecPathForEntryName(JacocoLCOVFormatter.java:73)\r\n...\r\n at com.google.testing.coverage.JacocoCoverageRunner$2.run(JacocoCoverageRunner.java:568)\r\n```\r\n\r\nIt looks like `JacocoCoverageRunner` is being built in the exec configuration, and then apparently being run in the target configuration as part of test execution:\r\n\r\nhttps://github.com/bazelbuild/bazel/blob/834ba08f9e780810743ff8eb39b125c0bf9e36b6/src/main/java/com/google/devtools/build/lib/rules/java/JavaToolchainRule.java#L279-L287\r\n\r\nI'm not sure what the intent there was, but either it shouldn't be running on the target JDK, or it should be built in the target configuration.", "In Bazel `JacocoCoverageRunner` appears to be taken from `remote_coverage_tools` in precompiled form, but the packaging process probably uses JDK 11 with `-target 8`, which doesn't catch this.", "Is there any thoughts to a solution for this? Can we not just make the coverage tool's code Java 8 compliant so we know it will run in Java 8 JVM? This is assuming the isEmpty() usage is the only place we use Java 11+ code, but I also imagine there aren't any real needs for Java 11+ code other than \"nice-to-haves\".", "https://github.com/bazelbuild/bazel/pull/17338 is in review and fixes this. @cushon Can you say whether Ivo's input is truly needed? He appears to be quite busy.\n\nIt's unfortunately hard to cover with tests as that would require maintaining a JDK 8 on all CI machine. But if this breaks in the future, it can probably be fixed again as long as Java 8 is supported.", "I still get this issue on 6.1.0, the project above still recreates this issue.", "Rules requires a Java tools update and it's possible that didn't make it into the release.\n\n@hvadehra Is there already a release with the change?", "https://github.com/bazelbuild/java_tools/releases/tag/java_v11.12 was released a few days ago which should contain https://github.com/bazelbuild/bazel/commit/c26665102d0d524bb42be8cd205c7b31a22a75c9", "So can we reopen this issue then? It seems that even if the patch did go out, it's not actually fixing the issue defined here/demonstrated by the example.", "The fix went into the Java tools, but the commit bumping the pinned version of these tools didn't go into 6.1.0. It does look like it will be cherry-picked into 6.2.0.", "@cjohnstoniv You should be able to manually update remote_java_tools to v11.12 in your WORKSPACE too." ]

[]

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

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

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

[ "src/java_tools/junitrunner/java/com/google/testing/coverage/JacocoLCOVFormatter.java" ]

[ "keyword_pr_to_issue" ]

[ "Thanks for the detailed report, I prepared a fix at https://github.com/bazelbuild/bazel/issues/17458 that can be cherry-picked into Bazel 6.1.", "I'm seeing this issue with bazel 6.1.2. (It worked in 5.3.2.) Looking at https://github.com/bazelbuild/bazel/commit/420659a9ad2a98f57e057d8c22eb621e3b12803e, it looks like it made it into various `7.0.0-pre.*` builds, but not into any `6.*` build. Is it possible to cherry-pick it?" ]

[]

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

[ "tools/java/runfiles/AutoBazelRepositoryProcessor.java" ]

[ "tools/java/runfiles/AutoBazelRepositoryProcessor.java" ]

[]

[ "keyword_pr_to_issue" ]

[ "I think that this could be fixed by appropriately merging the logic in https://cs.opensource.google/bazel/bazel/+/47c93b7e38cf9498617777fcb28928179e7f9013:src/main/java/com/google/devtools/build/lib/rules/cpp/LibrariesToLinkCollector.java;l=113-194 with that in https://cs.opensource.google/bazel/bazel/+/47c93b7e38cf9498617777fcb28928179e7f9013:src/main/java/com/google/devtools/build/lib/rules/cpp/LibrariesToLinkCollector.java;l=264-273.", "Fixed by commit 773d232f528276338098578a28c19c742e3b4e7e" ]

[]

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

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

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

[ "src/test/java/com/google/devtools/build/lib/rules/cpp/BUILD", "src/test/java/com/google/devtools/build/lib/rules/cpp/LibrariesToLinkCollectorTest.java" ]

[ "keyword_pr_to_issue" ]

[ "cc @comius @c-mita ", "It's seems the issue is that cc_tests unconditionally pulls those coverage tools dependencies.\r\nhttps://cs.opensource.google/bazel/bazel/+/master:src/main/java/com/google/devtools/build/lib/analysis/test/TestConfiguration.java;l=255-270" ]

[]

[ "type: bug", "P2", "team-Rules-CPP" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/BUILD", "src/main/java/com/google/devtools/build/lib/analysis/BaseRuleClasses.java", "src/main/java/com/google/devtools/build/lib/rules/BUILD" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/BUILD", "src/main/java/com/google/devtools/build/lib/analysis/BaseRuleClasses.java", "src/main/java/com/google/devtools/build/lib/rules/BUILD" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/test/CoverageConfiguration.java", "src/main/java/com/google/devtools/build/lib/analysis/test/TestConfiguration.java", "src/test/shell/bazel/cc_integration_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "Hi @FahriBilici, Above issue related to [Tensorflow](https://github.com/tensorflow/tensorflow) lite. Can you please check the above issue with the specific repo of [tensorflow](https://github.com/tensorflow/tensorflow). Please reach us back if the above issue is still relevant to Bazel, with complete details. Thanks!", "@sgowroji This is indeed a bug in Bazel itself. Note how the error message says that `ios` is not a valid Apple platform type but then goes on to list it as valid - but with a Turkish lowercase i. \n\nI believe that this can be fixed by using `Ascii.toLowerCase` and `Ascii.equalsIgnoreCase` in the methods of https://cs.opensource.google/bazel/bazel/+/669c298f71b5f8fbef8f62c85543cda3eab4dfcc:src/main/java/com/google/devtools/build/lib/rules/apple/ApplePlatform.java;drc=669c298f71b5f8fbef8f62c85543cda3eab4dfcc;l=246. Can't fix it myself right now, could you assign it (probably to the Apple team)?", "Yes I truly believe the issue displayed because of case sensitivity.", "I submitted a fix: https://github.com/bazelbuild/bazel/pull/17687\r\n\r\nBazel didn't really work at all with a Turkish locale (e.g. it fails when it encounters a `size = \"medium\"` on a test). The workaround is to specify a different locale with:\r\n```\r\nbazel --host_jvm_args=-Duser.country=US --host_jvm_args=-Duser.language=en build ...\r\n```" ]

[]

[ "type: bug", "platform: apple", "untriaged", "team-Rules-ObjC" ]

[ "src/main/cpp/blaze.cc" ]

[ "src/main/cpp/blaze.cc" ]

[]

[ "keyword_pr_to_issue" ]

[ "Pinging this issue. @rickeylev I noticed you've been making changes around the Python rules here. Would you be able to take a look? This issue is coming up more and more as more projects are upgraded to Bazel 6.", "What is this \"RUN_UNDER_RUNFILES\" environment variable? Is there some doc about this? As near as I can tell it's only set when Bazel runs a target as a test?\r\n\r\nAnyways, yeah, bug. I was able to repro.\r\n\r\nPosterity/notes to self: This was PR #15590", "@rickeylev thank you for getting this done so quickly. Can you merge this into the Bazel 6.1 branch so that it can go out as soon as possible?", "I suppose so. I've never done or requested a backport, so I'll have to ask around how to do that.", "@rickeylev this mentions how: https://github.com/bazelbuild/bazel/issues/17212#issuecomment-1439198224", "Hi all, can this wait till the next release (6.2)? We've already created the third release candidate for 6.1 and plan to do the final release on Monday, so we won't be able to include this.", "Hi,\r\n\r\nYes, if it's too late for 6.1, then 6.2 is fine.", "@bazel-io fork 6.2.0", "I was seeing this on 6.3.2 and tried upgrading to 7.1.1 to see if it fixed it, but it still happens.\r\n\r\nAnyone else still seeing this?" ]

[]

[ "type: bug", "team-Rules-Python" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/rules/python/python_stub_template.txt" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/rules/python/python_stub_template.txt" ]

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

[ "keyword_pr_to_issue" ]

[ "Seems to be a much more general issue:\r\n```\r\n$ bazel cquery --output=starlark --starlark:expr=\"providers(target)\" :alias\r\nNone\r\n```" ]

[ "The original change at https://github.com/bazelbuild/bazel/pull/17753/files doesn't impact lines 99-111. Does that matter?", "Yeah, that was the change I made. I removed this extra code that shouldn't have been included:) Looks like the file here is outdated." ]

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

[ "src/main/java/com/google/devtools/build/lib/analysis/ConfiguredTarget.java", "src/main/java/com/google/devtools/build/lib/analysis/configuredtargets/AbstractConfiguredTarget.java", "src/main/java/com/google/devtools/build/lib/query2/cquery/StarlarkOutputFormatterCallback.java", "src/main/java/com/google/devtools/build/lib/rules/AliasConfiguredTarget.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/ConfiguredTarget.java", "src/main/java/com/google/devtools/build/lib/analysis/configuredtargets/AbstractConfiguredTarget.java", "src/main/java/com/google/devtools/build/lib/query2/cquery/StarlarkOutputFormatterCallback.java", "src/main/java/com/google/devtools/build/lib/rules/AliasConfiguredTarget.java" ]

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

[ "keyword_pr_to_issue" ]

[ "What is the most useful and still rational behavior here? You could plausibly argue four approaches:\r\n\r\n1. keep the value of the first key seen\r\n2. keep the value of the last key seen\r\n3. merge the values of duplicates if the value is an object\r\n4. error\r\n", "experiment:\r\n```\r\n$ node -e 'console.log(JSON.parse(\"{\\\"a\\\": 1, \\\"a\\\": 2}\"))'\r\n{ a: 2 }\r\n```\r\n\r\nSo the answer I'd most like coming from a NodeJS ecosystem is to take your option 2.\r\n\r\nI dunno if that's formally documented in the nodejs spec.", "Taking the last entry seems preferable over taking the first, since it's more consistent with `dict.update`, the `dict()` constructor, etc. But it's not clear whether that's better or worse than failing. We could also make it an option, i.e. a named arg to decode, but I'm not sure I want to complicate that API.", "ping! @brandjon will you take PRs for this?", "[ECMA-262](https://262.ecma-international.org/13.0/#sec-json.parse) says under 25.5.1.1\r\n> Note 2 In the case where there are duplicate name Strings within an object, lexically preceding values for the same key shall be overwritten.\r\n\r\nThis aligns with option 2.", "ping! @brandjon would you accept a PR? Another user on Slack tripped over this.", "Different JSON parsers diverge on this point: https://stackoverflow.com/questions/21832701/does-json-syntax-allow-duplicate-keys-in-an-object\r\n\r\nJavascript, and anything that tries to be compatible with Javascript, of course allows duplicate keys and keeps the last value.\r\n\r\nC and C++ based JSON parsers typically don't allow duplicate keys.\r\n\r\nI have no strong opinion on which way to go, but given the need to integrate with npm, we may as well follow Javascript's behavior.", "I think the critical observation is that Bazel is meant to be the orchestration for a bunch of third-party package managers and tools, so it should follow the https://en.wikipedia.org/wiki/Robustness_principle\r\nMeaning, even if parsers typically reject some input, Bazel's job is to \"be permissive in what it accepts\" to maximize compatibility.\r\nSame argument I made for https://github.com/bazelbuild/bazel/commit/c9e2be52a067dd9abf5efa4f5f55bb5b98cf5d3b", "Rationale for accepting the change:\r\n\r\n* The JSON RFC apparently leaves this unspecified, so we're not prohibited from doing it.\r\n* JavaScript will allow it, with the unsurprising (and consistent with `dict`) semantics that last key wins.\r\n* It supports this Node-related use case.\r\n\r\n(The \"Robustness principle\" can be problematic -- think HTML. But we're maintaining a parser, not writing a new standard.)", "@brandjon @tetromino I don't suppose you guys know which release this will hit? ", "Given that this doesn't change the behavior on inputs that didn't hard fail before this change, I would propose to cherry-pick it into 6.2.0. Otherwise it would naturally make its way into 7.0.0.", "I am asking for a cherry-pick at https://github.com/bazelbuild/bazel/pull/17868" ]

[]

[ "type: bug", "P4", "team-Starlark-Interpreter" ]

[ "src/main/java/net/starlark/java/lib/json/Json.java" ]

[ "src/main/java/net/starlark/java/lib/json/Json.java" ]

[ "src/test/java/net/starlark/java/eval/testdata/json.star" ]

[ "keyword_pr_to_issue" ]

[ "@meteorcloudy @Wyverald I would be interested in hearing your thoughts on whether this should be possible.", "Can you explain a bit why having `dev_dependency` for the module extension proxy call isn't enough? Under which use case, do we have to share dev and non-dev dependencies with the same module extension proxy call? ", "Having the `dev_dependency` parameter on `use_extension` is indeed enough as users can just create a new proxy when they want to add a different type of dependency.\n\nThis feature request is only about making the information that a certain proxy (in particular, the set of tags created with it) was created with `dev_dependency = True` available to the module extension. \n\nThe very concrete use case I have for this information is the `auto_use_repo` idea mentioned in #17048. If a module extension wants to emit the correct `use_repo` calls that should be made in the root module's `MODULE.bazel`, it has to be able to distinguish tags from dev and non-dev proxies.", "I understand the idea of `auto_use_repo`, but have no idea how it should work (neither on how Gazelle's `update-repos` rep o works). I thought since the tags can only be used via the extension proxy created by `use_extension`, we shouldn't need the `dev_dependency = True` on tags?", "Let's consider the following example `MODULE.bazel` of a root module:\r\n\r\n```\r\nfoo_deps = use_extension(\"@rules_foo//foo:extensions.bzl\", \"foo_deps\")\r\nfoo_deps.package(name = \"fmt\")\r\nfoo_deps.package(name = \"json\")\r\n# Managed by auto_use_repo\r\nuse_repo(foo_deps, \"fmt\", \"json\")\r\n\r\nfoo_dev_deps = use_extension(\"@rules_foo//foo:extensions.bzl\", \"foo_deps\", dev_dependency = True)\r\nfoo_dev_deps.package(name = \"testing\")\r\n# Managed by auto_use_repo\r\nuse_repo(foo_dev_deps, \"testing\")\r\n```\r\n\r\nIn this situation, the `foo_deps` extension would create a repository with a file that contains a list of all repos added due to tags defined by the root module. `auto_use_repo` could consume the list and ensure that the `use_repo` calls are up to date.\r\n\r\nHowever, with the currently available features, `foo_deps` can't know that `testing` is a dev dependency and should thus be added to the `use_repo(foo_dev_deps, ...)` call whereas `fmt` is a regular dependency and thus has to be added to the `use_repo(foo_deps, ...)` call.\r\n\r\nOne way to fulfill this feature request would be to add `module_ctx.is_dev_dependency(module_tag)`, which would return `True` if and only if the argument is a module tag created by a module extension proxy in turn created by a call to `use_extension` with `dev_dependency = True`. If that were available, `foo_deps` could create the following `repositories.json` file that `auto_use_repo` could use to add the used repositories to the correct `use_repo` call:\r\n\r\n```json\r\n[{\"name\": \"fmt\"}, {\"name\": \"json\"}, {\"name\": \"testing\", \"dev_dependency\": true}]\r\n```\r\n\r\nI expect this to be useful in other situation, such as if `foo_deps` wants to add usage or license metadata to the repositories it creates.", "Thanks for the example -- that helped me understand the use case better. I'd like to think a bit more about how the API should look like.", "Thanks!! Do I understand it correctly that we probably don't need to add `dev_dependency` to the `tag` API, but implementing `module_ctx.is_dev_dependency` is enough?", "> Thanks!! Do I understand it correctly that we probably don't need to add `dev_dependency` to the `tag` API, but implementing `module_ctx.is_dev_dependency` is enough?\r\n\r\nYes, exactly. Tags are already implicitly dev dependencies if they come from a proxy that is marked with `dev_dependency`, so I don't think adding `dev_dependency` to tags would increase expressivity.\r\n", "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 30 days. It will be closed in the next 7 days unless any other activity occurs or one of the following labels is added: \"not stale\", \"awaiting-bazeler\".", "Not stale", "@bazel-io fork 6.2.0" ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionContext.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleFileGlobals.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/Tag.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/TypeCheckedTag.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionContext.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleFileGlobals.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/Tag.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/TypeCheckedTag.java" ]

[ "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/BzlmodTestUtil.java", "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionResolutionTest.java", "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleFileFunctionTest.java", "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/TypeCheckedTagTest.java" ]

[ "keyword_pr_to_issue" ]

[]

[]

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

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionContext.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleFileGlobals.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/Tag.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/TypeCheckedTag.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionContext.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleFileGlobals.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/Tag.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/TypeCheckedTag.java" ]

[ "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/BzlmodTestUtil.java", "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionResolutionTest.java", "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleFileFunctionTest.java", "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/TypeCheckedTagTest.java" ]

[ "keyword_pr_to_issue" ]

[ "I just skimmed this, but at first glance that surprises me. Bazel should, generally, be correct for any changes. `team-Core` should be able to give you better insight.", "Are you invoking Bazel with any special flags? `--cache_computed_file_digests` could cause this problem, although I would expect the node id to still catch that the file has changed. I used your test case on both Mac and Bazel, but couldn't reproduce:\r\n\r\n```\r\n$ bazel build //:expand1\r\nStarting local Bazel server and connecting to it...\r\nINFO: Analyzed target //:expand1 (5 packages loaded, 10 targets configured).\r\nINFO: Found 1 target...\r\nTarget //:expand1 up-to-date:\r\n bazel-bin/expanded1\r\nINFO: Elapsed time: 2.257s, Critical Path: 0.03s\r\nINFO: 3 processes: 2 internal, 1 linux-sandbox.\r\nINFO: Build completed successfully, 3 total actions\r\n$ cat bazel-bin/expanded1\r\nfoo : alpha\r\n$ cp template_archive_delta.tar template_archive.tar\r\n$ bazel build //:expand1\r\nINFO: Analyzed target //:expand1 (0 packages loaded, 0 targets configured).\r\nINFO: Found 1 target...\r\nTarget //:expand1 up-to-date:\r\n bazel-bin/expanded1\r\nINFO: Elapsed time: 0.124s, Critical Path: 0.01s\r\nINFO: 3 processes: 2 internal, 1 linux-sandbox.\r\nINFO: Build completed successfully, 3 total actions\r\n$ cat bazel-bin/expanded1\r\nfoo : delta\r\n$ ls -l bazel-bin/\r\ntotal 8\r\n-r-xr-xr-x 1 janakr primarygroup 12 Mar 8 12:25 expanded1\r\n-r-xr-xr-x 1 janakr primarygroup 13 Jan 1 1970 template1\r\n```", "@janakdr On Linux, I was able to reproduce the issue with Bazel 4.1.0 and 4.2.0, but not 5.0.0.", "Thanks @fmeum . Your original comment seems to be deleted, but you also correctly pointed out there that `--cache_computed_file_digests` has a default non-zero value in Bazel, which would mostly explain the issue (I just verified that the inodes are the same). @miscott2 is upgrading to 5.0 a reasonable solution for you? Separately we should understand what the difference between 4.2 and 5.0 is, to make sure that we're doing the right thing now.\r\n\r\nI see a potential area for improvement regardless: in most of Bazel, we use the [ctime](https://cs.github.com/bazelbuild/bazel/blob/3fc17f3c482592fbda95789d5bf5e8f7b8afbe6a/src/main/java/com/google/devtools/build/lib/actions/FileContentsProxy.java#L35) as part of our file data but for the digest cache we only use [mtime](https://cs.github.com/bazelbuild/bazel/blob/3fc17f3c482592fbda95789d5bf5e8f7b8afbe6a/src/main/java/com/google/devtools/build/lib/vfs/DigestUtils.java#L58). Adding ctime should make this cache more robust.", "@janakdr I deleted my original comment because I noticed that setting `--cache_computed_file_digests=0` did *not* seem to reliably fix the issue for me - I didn't want to drop a red herring. This makes me all the more interested in finding out which change between 4.2 and 5.0 ended up fixing this. Do you plan to run a bisect?", "Thanks everyone for your comments/efforts. Yes upgrading to Bazel 5.0 is absolutely fine as a solution for us as we were planning to do so imminently anyway.", "Since I won't be able to work on this, I'm going to close, since a workaround exists. It would be nice to understand what went wrong in case it's still lurking, but I can't say the team should prioritize it.\r\n\r\ncc @alexjski for interest", "I ran into a correctness issue under similar circumstances as the original issue with bazel 7.0.0-pre.20230316.2 + Linux 6.2.8 + sandbox_base pointing to tmpfs + ext4 . I built a tarball in a rule with a fixed mtime and then extracted it in another rule which lead to this issue when changing constants in a cpp file. Sometimes it would take 2 or 3 changes of the constant before the issue occurred. In my case, ```touch```ing the extracted output files was enough to fix the correctness issue. I tested with ```--cache_computed_file_digests=0``` and this also fixed the issue. Bazel relying on the mtime in this way is extremely surprising behaviour and should really be better documented (#14401) and honestly probably disabled by default. \r\n\r\n", "I looked into this again and was able to reproduce the issue with Bazel 6.1.1 and last_green, but *not* with the original reproducer: In order to reproduce, I need to `mv` instead of `cp` the tar files between builds.\r\n\r\nIt looks like this is fixed by adding ctime to the cache key as suggested by @janakdr in https://github.com/bazelbuild/bazel/issues/14723#issuecomment-1062060278. I submitted #18003, which includes an integration test, but will have to investigate other test failures this causes.\r\n\r\n@meteorcloudy As this is an incremental correctness issue, a fix should probably go into the planned 5.4.1 and 6.1.2 releases.", "@bazel-io flag", "@bazel-io fork 6.2.0" ]

[]

[ "type: bug", "P1", "more data needed", "team-Core" ]

[ "src/main/java/com/google/devtools/build/lib/skyframe/ActionMetadataHandler.java", "src/main/java/com/google/devtools/build/lib/unix/UnixFileSystem.java", "src/main/java/com/google/devtools/build/lib/vfs/DigestUtils.java", "src/main/java/com/google/devtools/build/lib/vfs/FileStatus.java", "src/main/java/com/google/devtools/build/lib/vfs/inmemoryfs/InMemoryContentInfo.java", "src/main/java/com/google/devtools/build/lib/windows/WindowsFileOperations.java", "src/main/java/com/google/devtools/build/lib/windows/WindowsFileSystem.java", "src/main/native/windows/file-jni.cc", "src/main/native/windows/file.cc", "src/main/native/windows/file.h" ]

[ "src/main/java/com/google/devtools/build/lib/skyframe/ActionMetadataHandler.java", "src/main/java/com/google/devtools/build/lib/unix/UnixFileSystem.java", "src/main/java/com/google/devtools/build/lib/vfs/DigestUtils.java", "src/main/java/com/google/devtools/build/lib/vfs/FileStatus.java", "src/main/java/com/google/devtools/build/lib/vfs/inmemoryfs/InMemoryContentInfo.java", "src/main/java/com/google/devtools/build/lib/windows/WindowsFileOperations.java", "src/main/java/com/google/devtools/build/lib/windows/WindowsFileSystem.java", "src/main/native/windows/file-jni.cc", "src/main/native/windows/file.cc", "src/main/native/windows/file.h" ]

[ "src/test/java/com/google/devtools/build/lib/skyframe/ActionMetadataHandlerTest.java", "src/test/shell/integration/execution_phase_tests.sh", "src/test/shell/integration/loading_phase_test.sh" ]

[ "keyword_pr_to_issue" ]

[ "Hi @mafanasyev-tri, Can you provide an sample example to test the same. Thanks! ", "Updated ticket description to include test case." ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/bazel/rules/python/python_stub_template.txt" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/rules/python/python_stub_template.txt" ]

[]

[ "keyword_pr_to_issue" ]

[]

[]

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

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/BUILD", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/DelegateTypeAdapterFactory.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/GsonTypeAdapterUtil.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/Module.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionContext.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionUsage.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleFileGlobals.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/SingleExtensionEvalFunction.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/BUILD", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/DelegateTypeAdapterFactory.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/GsonTypeAdapterUtil.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/Module.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionContext.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionMetadata.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionUsage.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleFileGlobals.java", "src/main/java/com/google/devtools/build/lib/bazel/bzlmod/SingleExtensionEvalFunction.java" ]

[ "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/BUILD", "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/BazelDepGraphFunctionTest.java", "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleExtensionResolutionTest.java", "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/ModuleFileFunctionTest.java", "src/test/java/com/google/devtools/build/lib/bazel/bzlmod/StarlarkBazelModuleTest.java" ]

[ "keyword_pr_to_issue" ]

[ "We saw such bugs before; we refined the file deletion logic to deal with ever more situations, alas it's still incomplete apparently.\r\n\r\nDo you know anything else about those files? Are they marked read-only?\r\n\r\n@buchgr -- do you think those files might be held open?", "Well, on the server this is quite frequent. It affects all binaries (.exe and .dll), and my best guess is that it is connected to anti-virus or Windows defender is scanning the file, and therefore locking it.\r\n\r\nWhen the server is entered, there is no problem to remove the binary, and nothing that is locking it.", "I have some more info, most likely it is Windows Defender that locks the files, and the locking can last from minutes up to 30+ min. (so the unlocking period is not permanent, it is only very long)\r\n\r\nOnly specific files are affected, so far: .dll, .exe, .bin, .a and maybe .exp\r\n\r\nMost of the time the locking period is short, and then it only affects repeated jobs on the server side.\r\nHowever, when the locking period is long, also local users are affected.\r\n\r\n\"Unlocker\" can unlock the files in all cases, so the current solution is to walk through the output tree, and delete any locked binaries before the build (which takes quite some time on Windows.)", "Is there any update on this issue? I'm still facing the same issue on Bazel 4.1.0 on Windows 10.", "Having this issue sometimes. It'd be great if it printed out the path for the file so I could use some other way to delete it.\r\n\r\nedit: I found that `bazel clean` will list the file, but I'll have trouble deleting it normally even with sudo. doing `sudo rm -rf /home/$USER/.cache/bazel` seems to delete them on wsl, but also nukes bazel completely, requiring it to rebuild and reacquire everything.", "I'm seeing this a lot when building android targets with `bazel mobile-install`. It's incredibly irritating because it doesn't tell me which files to remove and in order to continue I have to `bazel clean`, which then forces me to rebuild *everything* (and re-sync IntelliJ) even though I suspect only some outputs related to the `android_binary` are the problem. It would be nice if there was a way to `bazel clean` a single target.", "I just encountered this bug on a Windows machine in GitHub Actions while building a jar file that does not contain any native code.", "I debugged another instance in which this happend and traced it back to being caused by an attempted rebuild of the rules_kotlin worker while it was still running due to a previous build request.\r\n\r\n@larsrc-google Is the case of using non-prebuilt workers fully supported? It looks like all worker processes would need to be shut down before the worker dependencies can be rebuilt - at least on Windows, where deleting a file that is held open by a process isn't possible.", "That does make sense. That is a Windows-specific problem, and I don't have a good immediate solution. ", "@larsrc-google A workaround that shouldn't be too costly could be to restart all workers on Windows when the analysis cache is discarded. What do you think?" ]

[]

[ "type: bug", "P2", "area-Windows", "team-OSS" ]

[ "src/main/java/com/google/devtools/build/lib/actions/ActionExecutionException.java" ]

[ "src/main/java/com/google/devtools/build/lib/actions/ActionExecutionException.java" ]

[ "src/test/java/com/google/devtools/build/lib/actions/ActionExecutionExceptionTest.java", "src/test/java/com/google/devtools/build/lib/actions/BUILD" ]

[ "keyword_pr_to_issue" ]

[ "I'd also like to note this is rather hard to work around.\r\n\r\nFabian suggested creating an extension that creates a \"middleman\" repo. The extension detects if its the root module, and, if so, populates the middleman with references to the real toolchains. If not, then it puts a fake toolchain in so downstream users inheriting the register_toolchains() call don't break. Implementing this, however, is...well I haven't really figured out a good way, yet. You can't alias `:all`; alias() doesn't work with toolchains; symlinking the underlying build file introduces more complications with repo visibility. Maybe I'm missing something obvious, but it's turning into a headache." ]

[]

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

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

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

[ "src/test/java/com/google/devtools/build/lib/skyframe/RegisteredExecutionPlatformsFunctionTest.java", "src/test/java/com/google/devtools/build/lib/skyframe/RegisteredToolchainsFunctionTest.java" ]

[ "keyword_pr_to_issue" ]

[ "@Wyverald Just noticed that this could also be relevant to your proposal." ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkRepositoryContext.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkRepositoryFunction.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkRepositoryContext.java", "src/main/java/com/google/devtools/build/lib/bazel/repository/starlark/StarlarkRepositoryFunction.java" ]

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

[ "keyword_pr_to_issue" ]

[]

[]

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

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/Checksum.java" ]

[ "src/main/java/com/google/devtools/build/lib/bazel/repository/downloader/Checksum.java" ]

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

[ "keyword_pr_to_issue" ]

[ "@comius @cushon Could you take a look? It may end up just being a doc issue, but it does seem likely to break users of the Java runfiles library updating to Bazel 6.", "I think the issue might be that the JDK version that's used to run JavaBuilder isn't controlled by `--tool_java_runtime=` version, it's configured with the `java_toolchain.java_runtime` attribute.\r\n\r\nPart of the motivation for that is that it lets us ensure JavaBuilder is running on a relative recent JDK version, even if a build wants to use an older `--tool_java_runtime=`. But it also makes it harder to use a newer `--tool_java_runtime=`.\r\n\r\nInternally we're using the latest available JDK as the `java_toolchain.java_runtime`. One option is to just update the default to a newer version, but the latest javac versions have dropped support for `-source 7 -target 7`, so that's probably an aggressive default for Bazel.\r\n\r\nAnother option here might just be to document how to configure a custom `java_toolchain` that sets a matching `java_runtime`.", "What I find surprising is that setting `--java_runtime_version=remotejdk_17 --tool_java_runtime_version=remotejdk_17` doesn't fix this, only `--java_language_version=17` does. Is that expected? I would have assume that `--java_language_version` only affects the `-source` and `-target` values, but not the runtime used to run `JavaBuilder`.\r\n\r\nIs there a way to set up the toolchain so that a recent JDK is used for compilation (both target and tool) but target compilation outputs remain compatible with old JDKs and are also tested against these old JDKs?", "> Is there a way to set up the toolchain so that a recent JDK is used for compilation (both target and tool) but target compilation outputs remain compatible with old JDKs and are also tested against these old JDKs?\r\n\r\nI remembered that this is eventually going to be improved by https://github.com/bazelbuild/bazel/commit/75e0f7dfea71a7bd40b4bdb601f64217650e8d47 making it into a `java_tools` release, which will update the default `java_toolchain.java_runtime` to JDK 17\r\n\r\nThe issue you encountered in this bug could still come up for `--tool_java_language_version`s newer than 17, though. It would be nice if a bleeding-edge `--tool_java_language_version` just worked without having to define a custom `java_toolchain`.\r\n\r\n> I would have assume that `--java_language_version` only affects the `-source` and `-target` values, but not the runtime used to run JavaBuilder.\r\n\r\nI agree this is having surprising results here, it's the result of all three of those being configured with `java_toolchain` attributes. The default java_toolchain for the target platform is still `@bazel_tools//tools/jdk:toolchain_java8`, which with Bazel 6.0.0 is setting a JDK 11 `java_runtime`:\r\n\r\n```\r\n$ bazel shutdown; bazel build --toolchain_resolution_debug=//tools/jdk:toolchain_type //:Main --tool_java_language_version=17 --tool_java_runtime_version=remotejdk_17 |& grep 'Selected toolchain'\r\nINFO: ToolchainResolution: Type @bazel_tools//tools/jdk:toolchain_type: target platform @local_config_platform//:host: execution @local_config_platform//:host: Selected toolchain @bazel_tools//tools/jdk:toolchain_java8\r\nINFO: ToolchainResolution: Type @bazel_tools//tools/jdk:toolchain_type: target platform @local_config_platform//:host: execution @local_config_platform//:host: Selected toolchain @bazel_tools//tools/jdk:toolchain_jdk_17\r\n```\r\n\r\n```\r\n$ bazel query --output build @bazel_tools//tools/jdk:toolchain_java8\r\n...\r\n java_runtime = \"@bazel_tools//tools/jdk:remote_jdk11\",\r\n```", "Thanks for the explanations. Does that mean that the language level of annotation processors is limited by the Java runtime used in the target configuration? If that's the case, would it help if processors were compiled for the target rather than the exec configuration?", "The language level of annotations processors is limited by the `java_toolchain.java_runtime`.\r\n\r\nIdeally we'd want to ensure that the provided configurations satisfy the following constraints.\r\n\r\n| | | |\r\n| -- | -- | -- |\r\n`--java_runtime_version=` | `>=` | `--java_language_level=`\r\n`--tool_java_runtime_version=` | `>=` | `--tool_java_language_level=`\r\n`java_toolchain.java_runtime` | `>=` | `--tool_java_language_level=`\r\n`java_toolchain.java_runtime` | `>=` | `--java_language_level=`\r\n\r\ni.e. to build code at the Java N language level, you need a JavaBuilder running on JDK >= N to compile it, and a JDK >= N to execute it", "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 30 days. It will be closed in the next 7 days unless any other activity occurs or one of the following labels is added: \"not stale\", \"awaiting-bazeler\".", "Possibly related: https://github.com/bazelbuild/rules_jvm_external/issues/895", "One more note on that is that another way it satisfy the `java_toolchain.java_runtime >=` constraint is to keep that version upgraded to the latest available JDK. We've been doing that at Google, but it might be too aggressive for Bazel when there are breaking changes like [JDK-8173605](https://bugs.openjdk.org/browse/JDK-8173605)", "I also ran into this bug building bazel project on Mac. Can we bump the priority?", "Warnings have been added in Bazel 7 or anything using a recent rules_java release." ]

[]

[ "type: bug", "P2", "team-Rules-Java", "not stale" ]

[ "tools/jdk/local_java_repository.bzl" ]

[ "tools/jdk/local_java_repository.bzl" ]

[]

[ "keyword_pr_to_issue" ]

[ "Looks like the Python rules do not add `symlinks` and `root_symlinks` to the ZIP file, which causes the `_repo_mapping` file to be missing. I will try to fix this.", "https://github.com/bazelbuild/bazel/pull/17942 is the easy part, but this is unfortunately complicated on a deeper level: \r\n\r\nThe `_repo_mapping` file is currently injected directly in https://cs.opensource.google/bazel/bazel/+/6f12510e39153e316f35ab0a1ec376ece7f1b0a9:src/main/java/com/google/devtools/build/lib/analysis/Runfiles.java;l=421, which makes it invisible to the `runfiles` Starlark API.\r\n\r\nWe could instead add it in a way similar to https://cs.opensource.google/bazel/bazel/+/master:src/main/java/com/google/devtools/build/lib/analysis/RunfilesSupport.java;l=125, but that would construct a new `Runfiles` object for every target.\r\n\r\n~An alternative could be to add the artifact to every `Runfiles` object directly upon construction (as long as Bzlmod is enabled) and then only register the generating action in `RunfilesSupport`.~ Edit: This doesn't work as those `Runfiles` objects wouldn't be mergeable anymore.\r\n\r\n@Wyverald @lberki I could use your help here as I don't have a good grasp on the potential performance implications of these approaches.", "~Looked into this some more. Even if `RunfilesSupport` adds the repo mapping artifact to root symlinks, the Python rules still wouldn't see it as it is added only after the rule implementation has run. Maybe we need to expose it on `RuleContext`?~\r\n\r\n*Edit*: Cleared up some confusion I had, which led me to this insight: Computing the repo mapping requires the runfiles, so it simply isn't known before the rule commits to its runfiles. Getting it in the context of the rule only seems to be necessary if the rule wants to build a runfiles structure manually for itself - if it did this for some other rule, that rule's repo mapping would be in its runfiles. So unless the Python rules can build the ZIP as part of a separate rule using a macro, maybe `RepoMappingManifestAction` would need to be exposed to Starlark so that users can build their own manifests for arbitrary `runfiles` objects?", "Let me see if I understand this correctly: the current way Python rules build ZIP files is that they dissect a `Runfiles` object and put the path to artifact mapping on a command line. But repo mappings make it impossible to determine that mapping solely from the `Runfiles` object. Is that correct?\r\n\r\nShooting from the hip, with the current toolbox, I could imagine creating a \"fake\" Python binary which is then put on the inputs of an action and then that action does not execute it, but simply extracts its runfiles tree. Would that work?\r\n", "@lberki That's exactly correct. If forcing such rules to become macros that obtain the runfiles from an actual target is acceptable, then that is certainly the easiest solution.\r\n\r\n@rickeylev Could the executable Python rules go through a separate target to generate the runfiles that are then packaged up into the executable ZIP file?", "Is the limitation that makes a macro necessary that `FilesToRunProvider` can only be created once per rule and outside the control of Starlark code and that the only way one can currently add a runfiles tree to the inputs of an action is by using `ctx.actions.run(tools=)` ?", "> Is the limitation that makes a macro necessary that `FilesToRunProvider` can only be created once per rule and outside the control of Starlark code and that the only way one can currently add a runfiles tree to the inputs of an action is by using `ctx.actions.run(tools=)` ?\r\n\r\nThat is part of the reason. But even if there were a way to turn `Runfiles` into a `FilesToRunProvider` from Starlark, the logic that adds the repository mapping manifest is part of `RunfilesSupport` and `RunfilesSupplier` and may thus not invoked correctly in that way. This may require additional changes.\r\n\r\n", "> Could the executable Python rules go through a separate target to generate the runfiles that are then packaged up into the executable ZIP file?\r\n\r\nMaybe? That last mile logic to generate the executable for the zip-file case is a bit hairy.\r\n\r\nDoes every target have the repo mapping? Or just certain ones? I'm thinking just have an extra target generated along side with the necessary info, e.g.\r\n\r\n```\r\ndef py_binary(name, ...\r\n repo_mapping(\"_\" + name + \"_repo_map\")\r\n py_binary_rule(..., _INTERNAL_ONLY_REPO_MAP = \"_\" + name + \"_repo_map\")\r\n\r\ndef repo_mapping_impl(ctx):\r\n return [DefaultInfo()] # repo mapping automatically added, right?\r\n\r\ndef py_binary_impl(ctx):\r\n repo_map = ctx.attr._INTERNAL_ONLY_REPO_MAP.runfiles\r\n```\r\n\r\nCouple other thoughts:\r\n* PyBuiltins.java is where various Java-hooks can go. If the necessary info can be easily gotten via Java, that's the place to put such hooks.\r\n* The `input_manifests` arg _might_ be of help? There's a (now unused) helper in PyBuiltins that converts a runfiles object into a RunfilesSupplier.", "@rickeylev With Bzlmod enabled, every target has the repo mapping as a root symlinks that is injected in `getRunfilesInputs`: https://cs.opensource.google/bazel/bazel/+/master:src/main/java/com/google/devtools/build/lib/analysis/Runfiles.java;drc=0ce17480390cdced2df8d59249561613e81f446f;l=421\nIt may thus not be accessible by rule implementations directly.\n\nThe artifact itself is created in `RunfilesSupport`: https://cs.opensource.google/bazel/bazel/+/master:src/main/java/com/google/devtools/build/lib/analysis/RunfilesSupport.java;l=135;bpv=1;bpt=1\n\nWe may have to make it available from there.", "Exposing the RepoMappingManifestAction is part of the problem, but most of what it does is computed beforehand and are inputs into that action. Manually doing that logic does look feasible, though.\r\n\r\nLooking at how RunfilesSupport builds the repo mapping, the two pieces we need are `ruleContext.getTransitivePackagesForRunfileRepoMappingManifest()` and something to get `Package.getPackageIdentifier().getRepository()` (or an equivalent).\r\n\r\nUnfortunately, these both use the `Package` object, which I don't think has a Starlark equivalent. Can an object that doesn't implement the Starlark classes be passed from Java to Starlark? If so, we can just pass them onto another helper in PyBuiltins to operate on. If not, we'll have to convert Package to something else (probably a string or struct), or move more logic to Java (the map building is pretty simple, so this looks quite feasible).\r\n\r\nMaking a Starlark implemented rule's RunfilesSupport available to it during its execution is somewhere near impossible, in particular for the Python rules, so that isn't really an option. Problem 1 is RunfilesSupport is created _after_ the rule function completes. If you try to manually create a RunfilesSupport object during execution (`PyBuiltins.new_empty_runfiles_with_middleman` does such), then it creates the same actions multiple times, which triggers duplicate action conflict detection logic later. Problem 2 is, for the Python rules, the duplicate detection ends up triggering the empty runfiles supplier code multiple times during the analysis phase, which requires flattening the runfiles, and the net effect is a huge CPU and memory regression. If there is some sort of solution involving Starlark+RunfilesSupport, I don't know it :).\r\n\r\nEr, hm. Is that repo mapping code flattening the runfiles during analysis phase? That doesn't sound good, well, a problem for another time.\r\n\r\nIf the action did more of the heavy lifting (computing the map), it would make this easier overall _and_ probably also solve the runfiles-being-flattened-during-analysis issue. We could probably reduce it to a simple helper in PyBuiltins that looked something like:\r\n\r\n```\r\npublic void createRepoMappingManifest(StarlarkRuleContext starlarkCtx, Runfiles runfiles) {\r\n starlarkCtx.registerAction(new RepoMappingAction(\r\n starlarkCtx.getRuleContext().getTransitivePackagesForRunfileRepoMappingManifest(),\r\n runfiles\r\n ));\r\n}\r\n```", "As far as I remember the repo mapping manifest action isn't really doing much work since actions must be serializable, but `Package`s are not. That's also why the `NestedSet` is indeed flattened during analysis time. \n\nWe only need a tiny fraction of the Package information, so maybe that could be solved with some kind of custom serialization logic for `RepoMappingManifestAction` instead? I don't know much about this area though.", "I was playing around with this tonight and may have come up with a somewhat simple work around: declare the repo mapping file and add it to the zip, but rely on the subsequent RunfilesSupport code to actually create the file. I think this might also have a bonus: because the same file isn't created twice, duplicate detection doesn't have to run. We can make it a bit less hacky by factoring out a \"get file name\" function from Runfiles.createRepoMappingManifestAction. I'll poke this some more tomorrow to see how promising it is.\r\n\r\nAs a side note, this is the second time I've encountered Starlark code wanting to get at some of the things in RunfilesSupport (the other case was par files, which wanted RunfilesSupport for a similar reason), so maybe there is some missing functionality somewhere.", "Alright, today's experiment was trying to use a separate target, which looks iffy and unappealing.\r\n\r\n* The repo mapping is only generated and added to the runfiles of executable/test targets. This is due to StarlarkRuleConfiguredTargetUtil#addSimpleProviders (see the RunfilesSupport.withExecutable() call, guarded by the \"is executable or test\" check\r\n* When it is added, its added to the middleman artifact. See RunfilesSupport.createRunfilesMiddleman, where the repoMappingManifest is added to the nested set fed into the middleman\r\n* The code in Runfiles.getRunfilesInputs adds repoMappingManifest to the runfiles manifest, but not the runfiles itself. (and I think this is the manifest that may not be present, depending on a flag?) (Ah, I seem to recall that the manifest and inputs work in tandem: the manifest tells what path to materialize a file under when the runfiles tree is created, and the inputs tell what inputs are actually available)\r\n\r\nThe net effect is, when you look at a target's runfiles, you don't see the repo mapping because it's in the middleman artifact, which I think is a secret dependency of the executable somehow.\r\n\r\nSo, I ran the target through `ctx.resolve_tools()`. This returns the middle man artifact (contained within a depset of all necessary files) and an Starlark-opaque RunfilesSupplier object. Well, we have the RunfilesSupplier now, so I fed that into a java hook to call RunfilesSuppler.getRepoMappingManifest(). This sort of works. It did give a repo mapping file.\r\n\r\nHowever, the resulting repo mapping is only correct if the outer target and inner target have the same dependencies (including implicit deps). In my experiments, I used a stub executable with no dependencies as the \"inner\" target. It's repo map was just the current workspace. The actual repo map includes the various implicit dependencies of py_binary itself (rules_python, platforms, bazel_tools, some others).\r\n\r\nSo to make this work, the inner target needs to copy all the passed in user values and the implicit deps of py_binary, forward along the transitive runfiles (if it doesn't, the repo map won't be correct) (and this particular runfiles object is unused), produce an unused executable file (to satisfy some starlark validations), and then the outer rule needs to dig out the repo mapping somehow. Note that all the attributes have to continue to exist on the outer rule so that things like aspects can still traverse them. I think there might be a more direct route to get the repo mapping than going through the middleman artifact (I think a java hook to go Target->FilesToRunProvider#getRunfilesSupport is possible), but that doesn't help with the generation of its content.\r\n\r\nSo, technically it's possible, but it is pretty wasteful. I'm very wary of generating essentially a copy of py_binary() -- the extra target isn't free; we already know this has a real cost. Within Google, we were forced to do this to make PAR files work and saw the effects in the performance metrics, and only deemed it OK because we didn't really have a choice and plan to make it an opt-in behavior instead of always-on behavior.\r\n\r\nAh -- I see the RepoMappingManifestAction PR got merged! I'll experiment with that -- looks pretty straight forward.", "@bazel-io flag", "@bazel-io fork 6.3.0", "@fmeum @rickeylev should this be now cherry-picked and included in the release-6.3.0?\r\n", "That would be great, yes.\r\n\r\nNote that this fix (https://github.com/bazelbuild/bazel/commit/4e60992432a49dda255eb005544ae8c03b30e243) requires https://github.com/bazelbuild/bazel/pull/18419/commits/ee46d1a6e1a3ce65384d87e9f6cc5d939e631018", "> That would be great, yes.\r\n> \r\n> Note that this fix ([4e60992](https://github.com/bazelbuild/bazel/commit/4e60992432a49dda255eb005544ae8c03b30e243)) requires [ee46d1a](https://github.com/bazelbuild/bazel/commit/ee46d1a6e1a3ce65384d87e9f6cc5d939e631018)\r\n\r\n@rickeylev @fmeum I did make sure the commit https://github.com/bazelbuild/bazel/commit/ee46d1a6e1a3ce65384d87e9f6cc5d939e631018 was merged. However, it seems like additional commits needs to be cherry-picked first in order for the commit to be cherrypicked in the release branch. Could you please submit a PR or let us know what commits to include? Currently, there are 3 files that are missing in the release-6.3.0 branch (src/main/java/com/google/devtools/build/lib/rules/python/PyBuiltins.java, src/main/starlark/builtins_bzl/common/python/py_executable_bazel.bzl, src/main/starlark/builtins_bzl/common/python/py_internal.bzl).", "Ahhh, I forgot most of the Starlark rewrite work didn't make the 6.0 cut. It won't be feasible to port this to 6.x. Even if we did, it'd require re-implementing the fix based on the (now deleted) Java implementation of the rules.", "> Ahhh, I forgot most of the Starlark rewrite work didn't make the 6.0 cut. It won't be feasible to port this to 6.x. Even if we did, it'd require re-implementing the fix based on the (now deleted) Java implementation of the rules.\r\n\r\ncc: @Wyverald @meteorcloudy " ]

[]

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

[ "src/main/java/com/google/devtools/build/lib/analysis/BUILD", "src/main/java/com/google/devtools/build/lib/analysis/RepoMappingManifestAction.java", "src/main/java/com/google/devtools/build/lib/analysis/RunfilesSupport.java" ]

[ "src/main/java/com/google/devtools/build/lib/analysis/BUILD", "src/main/java/com/google/devtools/build/lib/analysis/RepoMappingManifestAction.java", "src/main/java/com/google/devtools/build/lib/analysis/RunfilesSupport.java" ]

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

[ "keyword_pr_to_issue" ]

[ "Maybe this is a similar bug that should be fixed like https://github.com/bazelbuild/bazel/commit/edfe2a17e3434cce660757f59b14f2e9d6ab944e?", "/cc @buildbreaker2021", "We also encounter this problem. Unfortunately this causes us to not be able to compile Android on windows. Which is a major block for our project. Is there a rough estimate of when this will be fixed? Or is there any way we can workaround this? Thanks.", "The worst thing here is that this issue is not even particularly complicated, even with `WindowsOsPathPolicy` rework included (the right thing to do: support for case sensitive directories on Windows 10+ will solve a lot of problems). Yet, it breaks a whole lot of plain C libraries on Windows putting a big question mark on Bazel's viability in general.\r\n\r\nWhich is a pity.\r\n", "> The worst thing here is that this issue is not even particularly complicated, even with `WindowsOsPathPolicy` rework included (the right thing to do: support for case sensitive directories on Windows 10+ will solve a lot of problems). Yet, it breaks a whole lot of plain C libraries on Windows putting a big question mark on Bazel's viability in general.\r\n> \r\n> Which is a pity.\r\n\r\nCurrently, you can workaround this by replacing `--cxxopt` with `--per_file_copt=.*\\.cpp\\$@`.\r\n\r\nBut I agree that it's very confusing why this issue is a feature request and is P4.", "It won't work for flags which are set in the actual BUILD files. The primary issue here, is that bazel offers no way to say \"this file is plain C\" explicitly. So, in a project with several dozens of third party dependencies a lot of patching and testing is needed to get anywhere with Bazel v5.\r\n\r\nBazel v4 sort of works, but then it can not be used with VS2022 due to another open issue.\r\n", "> Bazel v4 sort of works, but then it can not be used with VS2022 due to another open issue.\r\n\r\nI thought this is not a problem for a msbuild based toolchain because msbuild only gives a warning? For us, it's a block issue because we're doing android cross compile on Windows and pass C++ flag to C compile simply failed.\r\n\r\nAnother interesting workaround is that we found somehow the C++ flags defined in toolchain won't be passed to C targets. So we end up moving the C++ flags to toolchain. But yeah it's lucky for us to be able to workaround this bug. For some complex project this may not work.", "VS2022 is broken pretty badly, we had to go back to VS2019/bazel v4 for now. We may figure something out going forward; possibly by making all sources C++ compatible (this appears to be the best option on the cost/benefit scale).\r\n\r\nStill, not nice.\r\n", "@oquenchil This issue seems to have wide user impact on Windows. Is it caused by starlarkfifying the cc rules? Do you know which exact change caused this and how hard is it to fix this?", "@meteorcloudy Looks like is caused by #14005. The `.c` extension is matched with `.C` cause bazel treat it as a C++ source. I think this can be fixed like https://github.com/bazelbuild/bazel/commit/edfe2a17e3434cce660757f59b14f2e9d6ab944e . I understand bazel team has many things to do and may not have the capacity to fix this. If you have the bandwidth to review a PR I can give this a try.", "Not sure why `.c` and `.C` should mean different file types on Windows because file names are already case insensitive. Is this a Linux convention? Can you just change files with `.C` to `.cpp`? \r\nBut given we are already making an exception in https://github.com/bazelbuild/bazel/commit/edfe2a17e3434cce660757f59b14f2e9d6ab944e, a PR to fix this sounds plausible, @oquenchil @comius should decide eventually.", "> Can you just change files with `.C` to `.cpp`?\r\n\r\nWe don't actually have any `.C` files. I think it's bazel that treated all `.c` files as `.C` files causing the problem.\r\n\r\n", "@meteorcloudy Also see #15550 for some further info.", "Any progress on this issue? Bazel cannot compile C targets under Windows anymore since version 5. I don't think this warrants classification as a \"feature request\" or \"P4\".", "Bazel can compile things on Windows, it just requires jumping through ugly loops with `--per_file_copt=` and possibly some patching.\r\n\r\nBut yes, this issue is bad and the lack of proper support for case sensitive abilities of NTFS is also unfortunate - that is, bazel would have worked fine on Windows with some extra configs if not for the rather stupid case forcing inside the Windows path resolver.\r\n\r\n", "From the [documentation](https://bazel.build/docs/user-manual): `--cxxopts: This is similar to --copt, but only applies to C++ compilation, not to C compilation or linking.` In obvious contradiction bazel 5.3.2 passes `--cxxopts` arguments to C files. This is clearly a bug and not a \"feature request\". @oquenchil\r\n\r\nI find if concerning that something that elementary is broken in bazel and swiped under the rug.", "This bug is still causing significant problems in a project, forcing us to use an old bazel version which causes further compatibility issues and surfaces other bugs without fixes. Will it ever be looked at?", "It's almost like Windows is not an important OS for developers any more. :-D\r\n\r\nYet, Windows is still the most used developer OS by a huge margin.\r\n", "@oquenchil I think this shouldn't be a P4 issue, can you take a look?\r\n\r\nIf anyone has a fix for this, a PR is also very welcome!", "Hi, I've created a PR. It's basically learned from https://github.com/bazelbuild/bazel/commit/edfe2a17e3434cce660757f59b14f2e9d6ab944e . I'm not very familiar to java actually, so please tell me if I did anything wrong. I've tested it on my environment and added a test.", "Created the referenced Cherrypick @kkpattern Any objections agains the attempt?", "> Created the referenced Cherrypick @kkpattern Any objections agains the attempt?\r\n\r\nNot at all😃.", "Why is this closed? Observed in the latest 5.x and 6.x builds.\r\n\r\nThis is C code! `sign.c` is a C source file! `--host_cxxopt=-std=c++17` causes `-std=c++17` to be passed to C compilation!!!\r\n\r\n```\r\n/home/.../.cache/bazel/_bazel_build/install/a09dbb90c658248f08f9aa0eba11997d/process-wrapper '--timeout=0' '--kill_delay=15' '--stats=/home/.../.cache/bazel/_bazel_build/36f2a9693014a18078cf75940b43707b/sandbox/processwrapper-sandbox/114/stats.out' \\\r\n\r\n/usr/bin/gcc-9 -MD -MF bazel-out/k8-dbg-gcc/bin/external/boringssl/_objs/crypto/sign.pic.d \\\r\n '-frandom-seed=bazel-out/k8-dbg-gcc/bin/external/boringssl/_objs/crypto/sign.pic.o' \\\r\n '-DBAZEL_CURRENT_REPOSITORY=\"boringssl\"' \\\r\n -iquote external/boringssl -iquote bazel-out/k8-dbg-gcc/bin/external/boringssl \\\r\n -isystem external/boringssl/src/include \\\r\n -isystem bazel-out/k8-dbg-gcc/bin/external/boringssl/src/include \\\r\n '-std=c++17' \\ # WUHT?????\r\n -g -O0 -fno-omit-frame-pointer -fPIC -fno-canonical-system-headers \\\r\n '-fdiagnostics-color=always' -Wa,--noexecstack -Wall -Werror '-Wformat=2' \\\r\n -Wsign-compare -Wmissing-field-initializers -Wwrite-strings -Wshadow \\\r\n -fno-common '-D_XOPEN_SOURCE=700' '-std=c11' -Wmissing-prototypes \\\r\n -Wold-style-definition -Wstrict-prototypes -c external/boringssl/src/crypto/evp/sign.c \\\r\n -o bazel-out/k8-dbg-gcc/bin/external/boringssl/_objs/crypto/sign.pic.o\r\n```", "6.3.0 should contain the fix? https://github.com/bazelbuild/bazel/pull/18552", "I did a simple test. When compile with MSVC with ` --cxxopt=/std:c++17`, bazel 5.1.1 added the flag to a C source file:\r\n\r\n```\r\nC:\\...\\cl.exe ... -D__TIME__=\"redacted\" /std:c++17 ... /c hello.c\r\n ^^^^^^^^^\r\n```\r\n\r\nwhile bazel 6.3.2 didn't:\r\n\r\n```\r\nC:\\...\\cl.exe ... -D__TIME__=\"redacted\" ... /c hello.c\r\n```", "I'm using Ubuntu 22.04... so I'm seeing the problem on Linux. Given that we're not using Bzlmod (because of incompatibilities), it's possible that I have a dependency of a dependency that's getting an old version of cc_tools or something. It's frustrating because it my build is broken because of two rather trivial bugs that _look_ like they're in bazel. (Where does @local_config_cc come from, after all???)" ]

[]

[ "type: bug", "P2", "team-Rules-CPP", "help wanted" ]

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

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

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

[ "keyword_pr_to_issue" ]

[ "the 1s timeout was introduced for #15373", "I am slightly confused here:\r\n\r\n1. Why is the system under load? Bazel only runs this code upon startup (SandboxModule registration), before any action is executed. So the expectation here is the host system should have enough resources dedicated to running Bazel and downstream actions. Are you suffering from the noisy neighbor problem here? (i.e. running multiple Bazel JVM, isolated by containers without dedicated resources, on the same host?).\r\n\r\n2. If the system does not have adequate resources to execute `/bin/true` inside a `linux-sandbox`, it's a pretty good sign to fail the build right there immediately instead of waiting and letting downstream actions also suffer the same faith?\r\n\r\n3. Do you restart Bazel JVM a lot? Bazel JVM should be long-lived to retain the in-memory analysis cache. Restarting Bazel JVM from cold will make your build slower and increase the chance of this problem happening.\r\n\r\nFrom the point above, I would imagine your current setup is something like this: 1 VM/baremetal host running Bazel in multiple containers. After each build, you are discarding your Bazel container, killing the Bazel JVM inside. You are overfitting the host system to saturate the CPU consumption which resulted in \"noisy neighbor\": builds from a few busier containers are consuming too many CPU, stopping new containers from being spun up successfully.\r\n\r\nIf that's indeed the case, my advice would be: to switch to using a set of \"persistent containers\" and re-use the containers + Bazel JVM inside between runs. Each container should be assigned a fixed set of CPU/RAM, enforced by the container runtime to ensure that they don't use more resources than they should.\r\n\r\nI don't think the mitigation in #18151 will solve your issue at all. You are only delaying the noisy neighbor issue from Bazel's startup phase down to the action execution phase. The 5% failure rate would still retain as all of your action execution will be delayed in scheduling by the same wait time." ]

[]

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

[]

[ "keyword_pr_to_issue" ]

[ "This feels to be Working As Intended, since modules specifically have a different way to register toolchains, but I'll leave it for @Wyverald to clarify.", "I passed this by @Wyverald on Bazel Slack and he asked to create an issue and cc him. Was just trying to find his GitHub handle :)\r\n\r\nIf WORKSPACE files can't register toolchains when bzlmod is enabled then it will be hard to use rulesets that have bzlmod support alongside rulesets that don't (while they all get up to speed) if they both register toolchains.", "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 30 days. It will be closed in the next 7 days unless any other activity occurs or one of the following labels is added: \"not stale\", \"awaiting-bazeler\".", "Not stale. ", "Ah, I think I figured out what's wrong here. It's that damn `platforms` repo again...\r\n\r\nWhen you don't enable Bzlmod, `@platforms` for everyone is the `@@platforms` repo defined in a WORKSPACE prefix/suffix (don't remember which). All good.\r\n\r\nWhen you enable Bzlmod, `@platforms` is still the `@@platforms` repo defined in a WORKSPACE prefix/suffix for everything that is *not* in the Bzlmod system (i.e. all the WORKSPACE repos including `@@rules_nodejs`). However, for things in the Bzlmod system (i.e. only built-in modules like `bazel_tools` and `local_config_platform` because your MODULE.bazel file is empty), `@platforms` actually refers to `@@platforms~0.0.5`. This is important because when we search for an appropriate execution platform (most likely your host machine), `@local_config_platform//:host` actually uses the platforms defined in `@@platforms~0.0.5`, but `rules_nodejs` defines toolchains for platforms defined in `@@platforms`.\r\n\r\nThe immediate workaround would be to add a single line in your MODULE.bazel file (`bazel_dep(name=\"platforms\", version=\"0.0.5\")`). I'll have to ponder how we could resolve this situation better, if it's even possible.", "Closing this as things are technically working as intended here. Hopefully the existence of this issue will serve as documentation for whoever's encountering this problem in the future.", "> Closing this as things are technically working as intended here. Hopefully the existence of this issue will serve as documentation for whoever's encountering this problem in the future.\r\n\r\nUhm, the module system not resolving a dependency, and the user having to manually depend on a specific version of a deep dependency is \"working as intended\"?\r\n\r\nIn my view this completely invalidates bzlmod.\r\n\r\nIt might be working \"as designed\", but definitely not \"as intended\".", "Agreed, letting the user add a dependency to make it work isn't really a solution to me.\r\nTo anyone that don't want to read the whole thing: \r\nAdd to `MODULE.bazel`:\r\n```\r\nbazel_dep(name=\"platforms\", version=\"0.0.5\")\r\n```", "if anyone has an idea how to fix this, please leave a suggestion.", "Maybe making `platforms` v0.0.5 the default on Bazel 7 would be a solution ?" ]

[]

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