zachs18

\ Replaced empty-string-producing regexes with non-(empty-string-producing) regexes that still cause the issue. \ One of the simpler regexes I've found that causes a stack overflow (via infinite recursion of two...

Yes, I should have time this weekend to look into that.

I rebased, but I didn't change to use the new machinery yet. I should be able to look into that sometime this week.

searched nightlies: from nightly-2023-11-11 to nightly-2024-02-01 regressed nightly: nightly-2024-01-27 searched commit range: https://github.com/rust-lang/rust/compare/5bd5d214effd494f4bafb29b3a7a2f6c2070ca5c...e7bbe8ce933123a8e327c79e38bcb790595e6a65 regressed commit: https://github.com/rust-lang/rust/commit/0c1fb2a1e65a57073dc62dd502319a00cf753371 bisected with cargo-bisect-rustc v0.6.5 Host triple: x86_64-unknown-linux-gnu Reproduce with: ```bash cargo bisect-rustc --start=1.75.0 --regress...

@ijackson > At the very least this needs to be discussed in the RFC. But I fear that the semantics are sufficiently dangerous that these functions ought not to be...

@ijackson > In principle, some other code using freeze might arbitrarily happen to read those cryptographic keys while they are in use (eg by another thread), or are still live...

Seems useful! One thing I'd mention is that seeing `#[patchable_function_entry()]` on a function at first glance seems to me like it should be *enabling* "patchability", but since the defaults are...

@rustbot label +A-allocators

:facepalm: Oops, this depends on #124980, which isn't merged yet. Converted to draft, sorry for the noise.

@rustbot author