add regression tests for the fn sig ice

[im-lunex]

follow-up to #160628 which fixed the suggest_add_reference_to_arg fn_sig ice

adds the regression tests @teor2345 asked for in review

• closure call with a missing-field struct literal (iced without the guard)
• method call whose where-clause fails (covers the method-call branch)

the impl def_id case cant be constructed, those obligations get dropped during probing and turn into e0599 instead

[rustbot]

r? @camelid

rustbot has assigned @camelid.
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:

• Owners of files modified in this PR: compiler
• compiler expanded to 75 candidates
• Random selection from 18 candidates

[teor2345]

Cc @JohnTitor @Kivooeo who worked on who have context on #160628

[Kivooeo]

feel free to review i will r= you

[teor2345]

Thanks, these are good tests and improve the coverage of PR #160628.

the impl def_id case cant be constructed, those obligations get dropped during probing and turn into e0599 instead

I try to avoid modifying code we can't test, or code for things that can never happen. But I think this part of PR #160628 is still a good idea, because it removes an ICE that could become reachable after future code changes.

Sometimes, I add tests that are "as close as we can get" to reaching the changed code. But I think we've done a lot here already, so it's up to you if you want to add something like that in another PR. I don't mind either way.

View changes since this review

[Kivooeo]

also add a links to issue in both tests

View changes since the review

[im-lunex]

done

[im-lunex]

Thanks, these are good tests and improve the coverage of PR #160628.

the impl def_id case cant be constructed, those obligations get dropped during probing and turn into e0599 instead

I try to avoid modifying code we can't test, or code for things that can never happen. But I think this part of PR #160628 is still a good idea, because it removes an ICE that could become reachable after future code changes.

Sometimes, I add tests that are "as close as we can get" to reaching the changed code. But I think we've done a lot here already, so it's up to you if you want to add something like that in another PR. I don't mind either way.

View changes since this review

thanks letting me choose and yeah i would like to keep this much for now.. and sure im open if you say i should add more.

[teor2345]

Thanks, this looks good to me!

View changes since this review