Stream: t-compiler/wg-prioritization/alerts

Topic: I-prioritize #77477 simd_insert and simd_extract allow garb…


triagebot (Oct 03 2020 at 05:16, on Zulip):

@WG-prioritization/alerts issue #77477 has been requested for prioritization.

Procedure

Joshua Nelson (Oct 03 2020 at 05:21, on Zulip):

I think we need to confirm this is a bug before prioritizing it

Joshua Nelson (Oct 03 2020 at 05:22, on Zulip):

(clearly the miri ICE is a bug, but not sure about the unsafe code)

lcnr (Oct 03 2020 at 07:13, on Zulip):

looks like we don't actually valide this somewhere and miri expects it to be validated

lcnr (Oct 03 2020 at 07:14, on Zulip):

https://github.com/rust-lang/rust/blob/8c54cf67c1bdae93ecef9791dc07c0982bb76041/compiler/rustc_mir/src/interpret/intrinsics.rs#L393-L399

lcnr (Oct 03 2020 at 07:17, on Zulip):

I wasn't able to find if an incorrect index is supposed to be an error or UB :shrug:

lcnr (Oct 03 2020 at 07:17, on Zulip):

anyways, I guess this looks P-medium to me

Joshua Nelson (Oct 03 2020 at 07:18, on Zulip):

seems fine to me, anywhere that has unsafe I think it's reasonable to put the onus on the caller

lcnr (Oct 03 2020 at 07:21, on Zulip):

seems fine to me, anywhere that has unsafe I think it's reasonable to put the onus on the caller

a bit unrelated but I actually disagree there :smile: even for unsafe code we should prevent as much "certainly wrong" unsafe code as possible.
i.e. with CString::as_ptr adding the lint to the compiler is extremely valuable imo

Joshua Nelson (Oct 03 2020 at 07:22, on Zulip):

yes, that's fair enough

triagebot (Oct 03 2020 at 18:05, on Zulip):

Issue #77477's prioritization request has been removed.

Camelid (Oct 03 2020 at 18:05, on Zulip):

Assigned P-medium

Jubilee (Oct 03 2020 at 18:06, on Zulip):

Yeah you're supposed to generally have a bounded set of invariants to uphold, not literally anything.

Last update: Apr 10 2021 at 22:45UTC