Fix hygiene regression in patterns

[petrochenkov]

Closes #29746

[rust-highfive]

r? @nikomatsakis

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

[petrochenkov]

Hm, no, this breaks other things.
It looks like the change in 40ce804, causing this regression, at the same time fixed some ICEs, like #23891 and #27320.
It needs further investigation.

[petrochenkov]

Ok, now all tests pass.
It's strange, however, that mtwt::resolve is needed during type checking.

[brson]

Nominating for backport. Linked issues says this is on beta. cc @rust-lang/compiler

[petrochenkov]

It looks like there are couple of very similar issues in rustc_trans and rustc_mir, in places where "the first of bindings with the same name" is needed in match.

match x {
    pat1(a) | pat2(a) | pat3(a) => {}
         ^ the first `a` is here, `a`s have to be compared hygienically
}

[petrochenkov]

Updated with fixes for mir and trans.
Ideally, resolve should build some special table for these cases and save it for later, so mtwt tables could be cleared immediately after name resolution.

[bors]

☔️ The latest upstream changes (presumably #29763) made this pull request unmergeable. Please resolve the merge conflicts.

[nrc]

r+ with a rebase.

I am really unhappy about keeping the mtwt tables around longer than we have to - this will cause memory use and possibly compile time regressions. But I guess it is better to be slow and correct. I'll file an issue for a better fix.

[nikomatsakis]

r? @nrc (since he already reviewed)

[petrochenkov]

Rebased.

[nrc]

@bors: r+

[bors]

📌 Commit fba1926 has been approved by nrc

[bors]

⌛️ Testing commit fba1926 with merge cc30948...

[bors]

☀️ Test successful - auto-linux-32-nopt-t, auto-linux-32-opt, auto-linux-64-debug-opt, auto-linux-64-nopt-t, auto-linux-64-opt, auto-linux-64-x-android-t, auto-linux-cross-opt, auto-linux-musl-64-opt, auto-mac-32-opt, auto-mac-64-nopt-t, auto-mac-64-opt, auto-win-gnu-32-nopt-t, auto-win-gnu-32-opt, auto-win-gnu-64-nopt-t, auto-win-gnu-64-opt, auto-win-msvc-32-opt, auto-win-msvc-64-opt

[nikomatsakis]

Accepted for beta because this is a regression.

[brson]

This doesn't cherry-pick cleanly at all. Can somebody else do it in the next week or so? @petrochenkov perhaps?