Major async/await compiler performance regression

[Mark-Simulacrum]

Something in #67670 caused a major performance regression. Query that regressed is normalize_ty_after_erasing_regions, but not sure if that means much. I suspect that #65244 is the PR responsible, but have no proof.

[Centril]

Seems like the most prudent course of action is to create a revert PR for #65244 to see if perf is regained.

[Mark-Simulacrum]

I wouldn't be opposed to such a course. Not sure if I'll have the time to do so for a day or two, though.

[wesleywiser]

Testing a revert in #67768

[wesleywiser]

Looks like #65244 is the cause.

[pnkfelix]

triage: P-high; leaving nominated for discussion at meeting (namely discussion of revert)

[pnkfelix]

Discussed in T-compiler meeting.

Members present generally approved a revert of PR #65244 to resolve this issue in short term.

[wesleywiser]

I guess I should re-open my revert then?

[Mark-Simulacrum]

Yes, please do, and I (or someone else) can review it, I think, fairly quickly.