subtree-push nightly-2025-10-07

[jieyouxu]

Bumping the toolchain version as part of a git subtree push.

current toolchain (nightly-2025-04-02):

• 1.88.0-nightly (e2014e876 2025-04-01)

latest toolchain (nightly-2025-10-07):

• 1.92.0-nightly (f6aa851db 2025-10-07)

Review remarks

I tried my best to mimic prior rustfmt syncs (e.g. https://github.com/rust-lang/rustfmt/pull/6531), and tried to piece this together based on clippy's subtree sync advice.

For the merge commit itself, I had to fix two merge conflicts:

1. src/types.rs (related to dyn* unstable feature removal cf. https://github.com/rust-lang/rust/pull/143036), and
2. src/closures.rs (trivial import conflict).

As far as I can tell, rustfmt repo does not have any kind of auto toolchain bump tooling, so I created the last rust-toolchain bump commit (08586bba8e66b73280d16739bb8f80a14a35e081) manually.

I also had to:

• Fix a non-trivial semantic merge conflicts related to cfg_match! changes that are present in this repo which are not present in the tree in r-l/r... So in some sense the two trees seem to have diverged a bit?
• Fix a broken test that had invalid (AFAIK) test headers (using = for the key-value separator instead of :) and a missing // rustfmt-edition: 2024 test header.
• Bump some deps in Cargo.lock (dirs update on the r-l/r side).

[jieyouxu]

~~No this isn't right, the earliest commit is too early 🤔~~ EDIT: no wait, that's commit date, not merge date.

[jieyouxu]

Ah, non-trivial merge conflicts.

[shulaoda]

Oh, I literally did the exact same thing just now! https://github.com/rust-lang/rustfmt/pull/6682 👀

[Manishearth]

• Fix a non-trivial semantic merge conflicts related to cfg_match! changes that are present in this repo which are not present in the tree in r-l/r... So in some sense the two trees seem to have diverged a bit?

Yes, and will reconverge when you sync back up. Typically you do not want there to be a lot of time to have passed between a sync and merging new PRs to avoid this problem. Clippy does a sync every 2-3 weeks which seems to work okay.

[fee1-dead]

I am also happy to do some interim reviews to help with the capacity

[jieyouxu]

I don't have bandwidth for maintaining rustfmt in addition to what I'm already maintaining, but can help with smaller cursory reviews as well in the interim if that's needed.

[ytmimi]

Kicked off the Diff-Check

[Manishearth]

Looks like something impacted import ordering in rustc.

[ytmimi]

I'll review the diff more carefully a little later to see if I can figure out what's going on similar to what I did in https://github.com/rust-lang/rustfmt/pull/6531#issuecomment-2771299604

[ytmimi]

I didn't have as much time as I thought to dig into the formatting diff yesterday. I'd like to review the changes carefully before we merge any subtree-push so that we have a clear understand on where all the changes are coming from. I likely won't have much time to look into this until early next week.

[jieyouxu]

Yeah sounds good. No rush :)

[fee1-dead]

W.r.t. Diff-Check: Did something get messed up with version="Two" and the style edition changes? All the diffs I am seeing look like stuff related to sorting and version="One" stuff like x.0.0 being reformatted back to x.0 .0

[ytmimi]

@fee1-dead It's possible that there's some sort of mismatch here. I know we need to update the Diff-Check to pass the style-edition instead of the old version config (right now version=Two should be synonymous with style-edition=2024). That might helps us get a better comparison. It's possible that we're seeing a false positive here.

We could try to update the Diff-Check job to take a style_edition parameter and pass that to each binary via the --style-edition option in the CLI

[jieyouxu]

(Let me know if you would like me to fix the merge conflict with a force-push, have not done so to avoid making it annoying to review.)

[jieyouxu]

We could try to update the Diff-Check job to take a style_edition parameter and pass that to each binary via the --style-edition option in the CLI

Two PRs (for the existing Diff Check setup):

1. #6741
2. (Stacked) #6742