Fix hypothetical ICE in variances_of

[zetanumbers]

A cleanup based on #153694.

This PR changes panic messages in fn_sig's value_from_cycle_error to more fitting ones. Also this PR makes variances_of's value_from_cycle_error to utilize new query key argument instead of extracting it from CycleError.

Closes #127971

r? @nnethercote

[nnethercote]

Seems plausible, but please add a short description that explains the current situation and how/why this PR improves things. Currently the only descriptive word is "cleanup" which tells the reader very little.

cc @Zoxc

[zetanumbers]

Seems plausible, but please add a short description that explains the current situation and how/why this PR improves things. Currently the only descriptive word is "cleanup" which tells the reader very little.

Updated the top comment. Is it good?

[nnethercote]

It's an improvement, but it still doesn't explain what was going wrong in #127971 and how that has been fixed. The new fn_sig calls default, is that important?

[zetanumbers]

It's an improvement, but it still doesn't explain what was going wrong in #127971 and how that has been fixed.

This PR basically eliminates panic that have caused #127971 and avoids fn_sig's issue like #153391 to happen to variances_of.

The new fn_sig calls default, is that important?

Previously there was:

tcx.dcx().abort_if_errors();
unreachable!()

Now compare it to default's definition:

let Some(guar) = tcx.sess.dcx().has_errors() else {
    bug!(
        "`from_cycle_error_default` on query `{query_name}` called without errors: {:#?}",
        cycle_error.cycle,
    );
};
guar.raise_fatal()

These are essentially the same. I've only changed unreachable! to more descriptive bug!.

[nnethercote]

A better PR title would be something like "Fix ICE in variances_of." That makes it clear that it's not just a cleanup, but also fixing a bug.

A better commit message would be something like this:

variances_of currently used search_for_cycle_permutation, which can fail and abort when <... describe circumstances... >. This commit changes variances_of to receive a def_id which means it can compute a value without using search_for_cycle_permutation, avoiding the possible abort.

Fixes #127971

This explains the current situation and the change in a way that is meaningful to a reader who isn't already familiar with #127971.

[nnethercote]

Needs rebasing over #153694, and the title/description/commit messages updated as described above.

[zetanumbers]

@rustbot ready

[nnethercote]

@bors r+ rollup

[rust-bors]

📌 Commit 567f8bd has been approved by nnethercote

It is now in the queue for this repository.

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[rustbot]

⚠️ Warning ⚠️

• There are issue links (such as #123) in the commit messages of the following commits.
  Please move them to the PR description, to avoid spamming the issues with references to the commit, and so this bot can automatically canonicalize them to avoid issues with subtree.
  • b5e53dc

[nnethercote]

@bors r+

[rust-bors]

📌 Commit b5e53dc has been approved by nnethercote

It is now in the queue for this repository.