Add regression test for #135287

[blackms]

Fixes #135287

The ICE (adt_sized_constraint called on non-struct type) triggered when changing a struct to an enum under incremental compilation. It was fixed indirectly by the removal of the adt_sized_constraint query in #122493, but no regression test was added.

This adds one using the minimized reproducer from the issue.

r? incremental

[rustbot]

The review request incremental corresponds to 1 reviewer(s).
However, none of them are available for a review at this time.
Use r? to specify someone else to assign.

[rustbot]

r? @Mark-Simulacrum

rustbot has assigned @Mark-Simulacrum.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Fallback group: @Mark-Simulacrum, @jieyouxu
• @Mark-Simulacrum, @jieyouxu expanded to Mark-Simulacrum, jieyouxu

[Kivooeo]

View changes since the review

use a link to an issue instead of just issue number

[Kivooeo]

also please move it to one of tests/ui directories and remove "issue-n" from test name and remove issue number from commit message

View changes since this review

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[blackms]

Thanks for the review @Kivooeo! Addressed in the latest push:

• Renamed file to adt-sized-constraint-on-enum.rs (dropped the issue-N- prefix).
• Inline comment now uses the full issue URL instead of #135287.
• Commit message rewritten to drop the #135287 / #122493 references, keeping only a description of what the test exercises.

One thing I'd like to push back on gently: I'd prefer to keep the test in tests/incremental/ rather than move it under tests/ui/. The bug is specifically an ICE that happens when the same crate is recompiled across two builds that toggle a type definition between struct and enum — it requires the incremental cache from cfail1 to be reused by cfail2, which is exactly what tests/incremental/ is wired up to do. In tests/ui/ with //@ incremental, each revision recompiles from a clean cache, so the offending query wouldn't be hit in the stale-on-disk state that originally triggered the ICE. Happy to move it if you disagree — just wanted to flag the reasoning first.

@rustbot ready

[jieyouxu]

See #155296 (comment)