Toolstate: Don't block beta week on already broken tools.#69624
Toolstate: Don't block beta week on already broken tools.#69624bors merged 1 commit intorust-lang:masterfrom
Conversation
rust-highfive
added
the
S-waiting-on-review
|
No PRs will pass starting Wednesday without this if all tools aren't fixed by then. I'll try to get the books fixed before then, but that is cutting it close. I think Clippy should be good to update, but there is some risk because it will require a coordinated update with Cargo. |
|
This looks correct to me, and does seem like the better behavior: realistically, we have 6 weeks to get tools working on beta branch. @bors r+ |
|
📌 Commit 22d840e has been approved by |
bors
added
S-waiting-on-bors
Rollup of 6 pull requests Successful merges: - #68682 (Add documentation to compiler intrinsics) - #69544 (Unrevert "Remove `checked_add` in `Layout::repeat`") - #69617 (constify mem::forget) - #69622 (Rename `syntax` in librustc_ast/README.md) - #69623 (stash API: remove panic to fix ICE.) - #69624 (Toolstate: Don't block beta week on already broken tools.) Failed merges: - #69626 (Toolstate: don't duplicate nightly tool list.) r? @ghost
Was this meant to link to something else? That PR doesn't touch |
|
Yea, it was supposed to be #66681. GitHub's auto-complete for issue numbers is very annoying, and too often "helpfully" picks a random number. |
| did_error = true; | ||
| eprintln!( | ||
| "error: Tool `{}` has regressed from {} to {} during beta week.", | ||
| tool, old_state, state | ||
| ); |
There was a problem hiding this comment.
Reading this code, something very odd is going on here... there are two places where the code does "if regressed and in_beta_weak, then fail": it does that here, and it does that again and seemingly independently in change_toolstate.
What is going on here? Seems like duplication of the same logic, which usually shouldn't happen.
There was a problem hiding this comment.
Yea, I had noticed that, too. I didn't want to make too many changes in one PR. The check in change_toolstate can be removed, that only runs on master. I'm not even sure if anyone monitors if master is failing?
I'll submit a PR to clean it up.
EDIT: I was wrong about master-only.
…ta, r=Mark-Simulacrum Toolstate: remove redundant beta-week check. I made a bit of a mistake in rust-lang#69624. The "beta regression" doesn't need to be checked twice. I also rolled up rust-lang#69693 to avoid merge conflicts.
…ta, r=Mark-Simulacrum Toolstate: remove redundant beta-week check. I made a bit of a mistake in rust-lang#69624. The "beta regression" doesn't need to be checked twice. I also rolled up rust-lang#69693 to avoid merge conflicts.
…ta, r=Mark-Simulacrum Toolstate: remove redundant beta-week check. I made a bit of a mistake in rust-lang#69624. The "beta regression" doesn't need to be checked twice. I also rolled up rust-lang#69693 to avoid merge conflicts.
…ta, r=Mark-Simulacrum Toolstate: remove redundant beta-week check. I made a bit of a mistake in rust-lang#69624. The "beta regression" doesn't need to be checked twice. I also rolled up rust-lang#69693 to avoid merge conflicts.
…ta, r=Mark-Simulacrum Toolstate: remove redundant beta-week check. I made a bit of a mistake in rust-lang#69624. The "beta regression" doesn't need to be checked twice. I also rolled up rust-lang#69693 to avoid merge conflicts.
6 participants
This changes it so that tools are allowed to be broken entering the beta week if they are already broken. This restores the original behavior before the changes in #66681.
Closes #68458