Libs Meeting 2022-06-08

tags: Libs Meetings Minutes

Meeting Link: https://meet.jit.si/rust-libs-meeting-ujepnbwg2lzqgt6wvrndwimi
Attendees: Amanieu, Jane, Josh Triplett, Josh Stone, Mara, The 8472, Dan Gohman, Ben Kimock

Agenda

• Triage
• Updates
  • New process for unstable features (UCP / ACP / µCP / ICP / ..)
    • First experiment: https://github.com/rust-lang/rust/pull/97629
    • We don't have a bot for this yet.
    • Jane: We can use the agenda generator for now.
    • Too many TLAs.
    • �CP.
    • Summary: a lot of "argh, bikeshedding", and ultimately "let's just use 'comment period'"
    • Jane to draft a Blog Post
  • Glob crate (https://github.com/rust-lang/glob)
    • This was moved to rust-lang. Libs has write and publish access.
    • We should add a clarification to the readme of all of our official rust-lang crates that explains what status/mode the crate is in. E.g. only accepting security bug fixes.
  • Improved locks (tracking issue)
    • Nearly done. Only issue left is dropping locked locks when leaking a guard on macOS (and other pthread platforms). We can solve that by try_lock+unlocking it in Drop, and leaking if it was locked.
  • format_args plans recap
    • Most exciting approach is to use closures. We can start with this one, as an experiment.
      -write("a"); write("b"); doesn't optimize to write("ab");, unfortunately.
• Anything else?

Triage

FCPs

1 open T-libs FCPs:

Amanieu (1), yaahc (1), m-ou-se (1), cuviper (1)

Critical

None

Backports

• 1 rust-lang/rust T-libs beta-nominated items
  • [97009] Allow `unused_macro_rules` in path tests
    • Jane: looks like this one has already been backported and there is no further work needed from us

Prioritization Requested

None

Nominated

• 1 rust-lang/rust I-libs-nominated items
  • [97467] Experiment: force debug_assertions in consteval context
    • Discussed last time.
    • With this big of a performance hit, it should be opt-in, through some mechanism.
    • Josh Triplett to reply.

Waiting on team

None

Needs decision

None

Regressions

• 1 rust-lang/rust T-libs regression-from-stable-to-stable items
  • [97549] P-high Higher memory usage on Rust 1.61.0 (compared to 1.60.0) leading to SIGKILL
    • Not a regression in std. It's a regression in cargo. Cargo's regression points at a potential issue with available_parallelism, but that's not a regression.
• 1 rust-lang/rust T-libs regression-from-stable-to-nightly items
  • [97124] Synchronous Windows stdin pipes broke tokio

Actions

• Reply to all issues/PRs discussed in this meeting, or add them to the open action items.

Generated by fully-automatic-rust-libs-team-triage-meeting-agenda-generator