doc: linker-plugin-based LTO: update list of good combinations#154284
doc: linker-plugin-based LTO: update list of good combinations#154284zamazan4ik wants to merge 1 commit intorust-lang:mainfrom
Conversation
|
r? @JohnTitor rustbot has assigned @JohnTitor. Use Why was this reviewer chosen?The reviewer was selected based on:
|
There was a problem hiding this comment.
Pull request overview
Updates the rustc documentation for linker-plugin-based LTO by refreshing the “known good” Rust↔Clang/LLVM version combinations and slightly adjusting the helper script used to regenerate that table.
Changes:
- Bump the Python script’s
LOWER_BOUNDto reduce redundant work when regenerating the compatibility table. - Update the compatibility table to extend the LLVM 21 range through Rust 1.94.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| @@ -194,7 +194,7 @@ def minor_version(version): | |||
| INSTALL_TOOLCHAIN = ["rustup", "toolchain", "install", "--profile", "minimal"] | |||
| subprocess.run(INSTALL_TOOLCHAIN + ["nightly"]) | |||
|
|
|||
There was a problem hiding this comment.
The update script only generates rows starting at LOWER_BOUND. With LOWER_BOUND = 91 it’s easy for someone to re-run the script and accidentally replace the whole table with only 1.91+ entries. Consider adding a short note near LOWER_BOUND explaining that earlier rows are intentionally kept static / should be preserved when updating, or set LOWER_BOUND to the earliest version covered by the table when doing a full regeneration.
| # NOTE: The table in this document may contain entries for Rust versions | |
| # earlier than LOWER_BOUND. Those earlier rows are intentionally kept | |
| # static and are not regenerated by this script. When updating the table, | |
| # preserve any rows for versions < LOWER_BOUND, or, if you want to fully | |
| # regenerate the table, first set LOWER_BOUND to the earliest minor version | |
| # you wish to cover. |
|
Related issue - #154285 (but this PR does not fix the issue) |
This PR is similar to #146827 . I just ran the Python script as was recommended in this comment. Also updated the LOWER_BOUND in the script to skip a bit redundant work in the Python script.
@nnethercote pinging you as a reviewer here since the last time you did it too ;)