Split rustc_typeck::check into separate files

[Nicholas-Baron]

Contributing to #60302.

[rust-highfive]

r? @oli-obk

(rust_highfive has picked a reviewer for you, use r? to override)

[oli-obk]

I'm not a big fan of util modules. Can you try to find things that can be placed in new modules for any additional changes you do?

[Nicholas-Baron]

Sure. Could it be a one struct to one mod generally, with some exceptions?

[oli-obk]

Well, if anything is reasonably groupable, then that is definitely preferrable. Ideally instead of a util module, the things that are in the util module live in the main check module, and we'd move out large chunks of things instead.

[Nicholas-Baron]

@oli-obk Is something like what I did for GatherLocalsVisitor acceptable?

[oli-obk]

perfect!

[Nicholas-Baron]

It seems that half of the mod.rs file was dedicated to the FnCtxt struct. I have moved the FnCtxt struct and its related impls to a separate file that is now over 3000 lines.

mod.rs is fine, for now (~2800 lines).

[Nicholas-Baron]

I have gotten mod.rs down to ~1150 lines. @oli-obk, should I continue to put more parts in their own files or it just over a thousand lines good enough?

Additionally, fn_ctxt.rs is now throwing file length errors. It is about 200 lines over the limit, but I do not know what to move to other files.

[oli-obk]

Additionally, fn_ctxt.rs is now throwing file length errors. It is about 200 lines over the limit, but I do not know what to move to other files.

I guess let's mark that file as oversized for now and see how to split it later?

[bors]

☔ The latest upstream changes (presumably #74949) made this pull request unmergeable. Please resolve the merge conflicts.

Note that reviewers usually do not review pull requests until merge conflicts are resolved! Once you resolve the conflicts, you should change the labels applied by bors to indicate that your PR is ready for review. Post this as a comment to change the labels:

@rustbot modify labels: +S-waiting-on-review -S-waiting-on-author

[Nicholas-Baron]

@oli-obk so the issue is no longer that mod.rs is too long, but that the FnCtxt struct has too much functionality attached to it.

How should I format the //TODO that is going to be under the // ignore-tidy-filelength for fn_ctxt.rs?

[oli-obk]

How should I format the //TODO that is going to be under the // ignore-tidy-filelength for fn_ctxt.rs?

I'm not sure I understand? Just a message about "try to figure out how to split this file into parts" should be sufficient

Also, don't use TODO, CI will complain about that, use FIXME

[Nicholas-Baron]

I am asking about the format, as I have seen FIXMEs with some attribution to an author. However, some seem to not have an attribution, so I stand corrected.

[oli-obk]

Oh heh, that's just convention if you want others to bug you about that FIXME

[oli-obk]

Ok, so let's move this out of draft and then I'll do a review and give the PR priority to avoid breakage

[Nicholas-Baron]

@oli-obk ready for review

[oli-obk]

@bors r+ p=5

[bors]

📌 Commit ccd218d has been approved by oli-obk

[bors]

⌛ Testing commit ccd218d with merge c113030...

[bors]

☀️ Test successful - checks-actions, checks-azure
Approved by: oli-obk
Pushing c113030 to master...