Ali merge upstream 0.8.1#1
Merged
Ali-Piccioni merged 46 commits intomainfrom Aug 4, 2022
Merged
Conversation
To use config.string with allow_multiple
* WIP coverage support * Make it work * Regenerate documentation * Gate env vars on coverage being enabled * Add coverage targets * Allow 2 actions when generating coverage Co-authored-by: Krasimir Georgiev <krasimir@google.com>
…d#1426) * crate_universe support for individually updating packages. * Regenerated crate_universe outputs * Regenerate documentation
* Added Rust 1.62.0 * Regenerate documentation * Updated crate_universe examples * Addressed new clippy warnings
….llvm-cov is None (bazelbuild#1432) Do not attempt to instrument rust code for coverage if toolchain.llvm-cov is None.
…ld#1433) * update crate_universe `--repin` args to not require values. * Regenerate documentation
* Updated crate_universe version to `0.4.0`. * Regenerated crate_universe outputs * Regenerated crate_universe example outputs
…ild#1434) This is supposed to fix the long standing Windows flakiness as described in bazelbuild#1427 (comment) Initially I thought it's an issue with `rustdoc_test`, however the actual issue is that `rust_binary` and its `rust_test` have the same crate name and generate the same intermediary `.o` files. For sandboxed builds this is not an issue, as the actions are isolated, however, we have a race condition in non-sandboxed builds (Windows): Let's consider the following build: ``` bazel build --spawn_strategy=standalone \ //test/unit/rustdoc:bin_with_transitive_proc_macro \ //test/unit/rustdoc:bin_with_transitive_proc_macro_test ``` The `bin_with_transitive_proc_macro` compile action will create the following intermediate file: `bazel-out/k8-fastbuild/bin/test/unit/rustdoc/bin_with_transitive_proc_macro.bin_with_transitive_proc_macro.573369dc-cgu.2.rcgu.o`, as will the `bin_with_transitive_proc_macro_test` action. These two files differ slightly (as the second action is for a test), namely the `bin_with_transitive_proc_macro` output has the following symbols: ``` 0000000000000000 T main 0000000000000000 t _ZN30bin_with_transitive_proc_macro4main17hfc292cc42314e131E U _ZN3std2rt10lang_start17h1dbc829c47cd61d9E ``` while the `bin_with_transitive_proc_macro_test` `.o` output looks like this: ``` 0000000000000000 T main 0000000000000000 t _ZN30bin_with_transitive_proc_macro4main17h28726504dc060f8dE U _ZN3std2rt10lang_start17h1dbc829c47cd61d9E U _ZN4test16test_main_static17h37e3d88407f5b40fE ``` Now, when the two actions run in parallel, it can happen that `bin_with_transitive_proc_macro` creates the output, and `bin_with_transitive_proc_macro_test` overwrites it. Then, at linking time for `bin_with_transitive_proc_macro`, `rustc` will complain: ``` note: ld.lld: error: undefined symbol: test::test_main_static::h37e3d88407f5b40f ``` This is because `bin_with_transitive_proc_macro` is not a test and as such is not linked against `libtest`. This PR fixes the issue by directing the compilation outputs of `rust_test` to be under a `test-{hash}` directory.
…argets (bazelbuild#1437) * exclude `BUILD` and `WORKSPACE` files from generated crate_universe targets * Regenerated crate_universe outputs
v0.2.0 does not have `rust_bindgen_dependencies`
bazelbuild#1442) * Updated toolchain repository rules to represent one toolchain per repo * Regenerate documentation
* Upgrade stardoc I am eagerly awaiting bazelbuild/stardoc#124 so the anchor links work. * Regenerate documentation Co-authored-by: UebelAndre <github@uebelandre.com>
…uild#1454) * shorten `crate_universe_crate_index` to `cui` * refresh crate_universe/3rdparty * Update crate_universe/repositories.bzl Co-authored-by: UebelAndre <github@uebelandre.com> Co-authored-by: UebelAndre <github@uebelandre.com>
* Add extra_rustc_flag and extra_exec_rustc_flag * Restore Label wrappers * Update docs/BUILD.bazel and docs/symbols.bzl * Regenerate documentation * Revert "Regenerate documentation" This reverts commit 37a39e1. * Regenerate documentation
**Context** When building [rustdocs](https://doc.rust-lang.org/rustdoc/what-is-rustdoc.html) it is very useful to pass `-Dwarnings`. This allows you to easily assert things like [intra-doc-links](https://doc.rust-lang.org/rustdoc/write-documentation/linking-to-items-by-name.html) aren't broken. As far as I can tell, passing flags to `rustdoc` isn't possible today. **Solution** Add the `rustc_flags` attr to the `rust_doc` rule. `rustc_flags` is already parsed in [`construct_arguments(...)`](https://github.com/bazelbuild/rules_rust/blob/main/rust/private/rustc.bzl#L601) which `rust_doc` calls into.
…ommand (bazelbuild#1461) * Updated header of crate_universe generated files to include a regen command * Regenerate documentation * Regenerated crate_universe outputs
Currently we pass `bazel-out/volatile-status.txt` to all the `Rustc` actions, which is bad for remote caching with `--stamp`enabled. This PR makes only `rust_binary` be affected by the `--stamp` flag by default. The rest of the rust rules have `stamp = 0` as a default. Should you want to force stamp you can set `stamp = 1`. If you want to make your target stamp behavior depend on the `--stamp` command line flag, use `stamp = -1`. Additionally, this PR makes `rules_rust` also use the stable keys for stamping, by passing `bazel-out/stable-status.txt` to the relevant `Rustc` actions when stamping is enabled. This makes `rules_rust` adhere to the contract described here: https://docs.bazel.build/versions/main/user-manual.html#flag--workspace_status_command
Bazel provides COVERAGE_MANIFEST which is a file we can pass through to llvm-cov to limit the files that we produce coverage for. Using this revealed the path-equivalence was backwards because we're using the coverage prefix map to disable absolute paths in code coverage info, but llvm-cov operates on absolute paths, so we have to remap the `.` in the coverage data to the full path so that llvm-cov understands the relative paths (which it absolutizes internal) from the manifest file. This feels a bit weird and ideally it could operate entirely on these relative paths instead. This has the side effect of excluding files that aren't part of bazel sources, which fixes bazelbuild#1466 This should also improve interaction with `--instrumentation_filter`. Co-authored-by: UebelAndre <github@uebelandre.com>
* Updated `wasm_bindgen` rules dependencies. * Regenerated crate_universe outputs.
* Updated `tools/rust_analyzer` to use `crate_universe`. * Regenerate documentation * Updated generated files. * Deleted tools/rust_analyzer/raze * Add missing bzl_lib target
* Update `//bindgen` to use crate_universe * Updated generated files. * Deleted bindgen/raze
…build#1473) bazelbuild#1452 set `rust_*.stamp=0` as a default for all rules except `rust_binary`, which has -1 as a default. The `rust_binary_without_process_wrapper` rule is used to build the process wrapper itself. It needs to override the `stamp` attribute back to 0, as we do not support stamping without the process wrapper. bazelbuild#1452 introduced a bug where the default value for this rule became `-1`, thus building `//util/process_wrapper` with `--stamp` is now broken. This PR fixes the issue by routing all the attributes through the [_common_attrs_for_binary_without_process_wrapper](https://source.corp.google.com/piper///depot/google3/third_party/bazel_rules/rules_rust/rust/private/rust.bzl;rcl=461281665;l=1069?q=rust_binary_without_process_wrapper&ct=os&sq=package:piper%20file:%2F%2Fdepot%2Fgoogle3%20-file:google3%2Fexperimental) function, thus ensuring that the `stamp` attribute is once again set to 0.
* Fixed fetch_shas.sh * Added Rust version 1.62.1 * Regenerate documentation * Updated crate_universe examples
…bazelbuild#1455) * Added `rust_analyzer_toolchain` and repository rules for creating one * Regenerate documentation * Update rust/repositories.bzl Co-authored-by: Daniel Wagner-Hall <dawagner@gmail.com> * Updated rust_analyzer_toolchain export * Fixed rust_analyzer_toolchain_repository rule Co-authored-by: Daniel Wagner-Hall <dawagner@gmail.com>
Previously if the workspace root manifest was missing, a panic would be triggered due to unwrapping a `None` option. Now, we suggest the missing manifest if possible, or direct you to work it out yourself if we can't.
…ild#1483) * Fix rust-analyzer being unable to find rust sysroot sources. * Update tools/rust_analyzer/main.rs Co-authored-by: Daniel Wagner-Hall <dawagner@gmail.com> Co-authored-by: Daniel Wagner-Hall <dawagner@gmail.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
11 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.