T Compiler Meeting Agenda 2026 03 19

T-compiler Meeting Agenda 2026-03-19

Announcements

  • Reminder: if you see a PR/issue that seems like there might be legal implications due to copyright/IP/etc, please let us know (or at least message @davidtwco or @Wesley Wiser so we can pass it along).

Other WG meetings

  • None

MCPs/FCPs

Backport nominations

T-compiler beta / T-compiler stable

  • :beta: “Don’t look for non-type-level assoc consts when checking trait object types” rust#153738
    • Authored by fmease
    • Closes #153731
    • Voting Zulip topic, in favor
    • If approved, an additional PR #153958 will need to be backported (thanks @fmease!)
  • :beta: “Don’t add empty target features for target-cpu=native on macOS” rust#153763
    • Authored by nikic (thanks!!)
    • Fixes #153397 (P-critical issue breaking compilation on Apple M5/M4)
    • Voting Zulip topic, def. in favor
  • :beta: “Do not shallow resolve to root var while fudging” rust#153869
    • Authored by ShoyuVanilla
    • Meant to address #153816 and #153849, still in review
    • Voting Zulip topic, seems a revert of #151380 looks preferable
    • left a comment: is this approved for merge? See comment
  • :beta: “Fix some suggestions of the for-loops-over-fallibles lint” rust#153913
    • Authored by JohnTitor
    • Only FIY: approved but not yet merged, we can revisit next week
    • Fixes Fix #148114, #147973 and rust-lang/rust-clippy#16133
    • Voting Zulip topic
  • No stable nominations for T-compiler this time.

T-types beta / T-types stable

  • No beta nominations for T-types this time.
  • No stable nominations for T-types this time.

PRs S-waiting-on-t-compiler

T-compiler

Issues of Note

Short Summary

P-critical

T-compiler

  • “[ICE]: LLVM ERROR “Broken module” with -C target-cpu=native on aarch64-apple-darwin (Apple M5, LLVM 22.1.0)” rust#153397
    • Addressed by #153763

T-types

  • No P-critical issues for T-types this time.

P-high regressions

P-high beta regressions

  • “1.95 beta regression affecting old versions of bitvec: “… is not dyn compatible”” rust#153731
    • Addressed in #153738
  • “Significant compilation time regression starting in v1.95.0-nightly (bevy)” rust#153910
    • Being investigated (if anyone can help, see comment), seems originating from #149904. An MCVE could also be helpful
  • “Complex EH flow miscompiled on wasm” rust#153948
    • being investigated, @_hoodmane points to an LLVM upgrade (comment), opened a fix upstream llvm-project#187302
  • “1.95 beta regression: “no method named accept found for struct…”” rust#153849
    • Could be fixed by #153869 (or by a revert)

Note: See here there is a number of other beta regressions from the crater run. They are investigated, unsure about their impact.

  • “1.95 beta regression: “displacement … is not within [-2147483648, 2147483647]”” rust#153852
    • Caused by llvm-project#75747 (which could be “working as intended”)

Unassigned P-high nightly regressions

  • No unassigned P-high nightly regressions this time.

Performance logs

2026-03-16 Triage Log

Another fairly quiet week, with few changes and overall neutral performance.

Triage done by @simulacrum. Revision range: 3945997a..5b61449e

Summary:

(instructions:u)meanrangecount
Regressions (primary)0.2%[0.1%, 0.4%]53
Regressions (secondary)0.2%[0.0%, 0.6%]45
Improvements (primary)-1.8%[-6.7%, -0.2%]4
Improvements (secondary)-1.2%[-8.5%, -0.2%]10
All (primary)0.1%[-6.7%, 0.4%]57

1 Regression, 1 Improvement, 2 Mixed; 3 of them in rollups 35 artifact comparisons made in total

Regressions

Rollup of 13 pull requests #153672 (Comparison Link)

(instructions:u)meanrangecount
Regressions (primary)0.2%[0.1%, 0.4%]10
Regressions (secondary)0.3%[0.0%, 0.5%]44
Improvements (primary)--0
Improvements (secondary)--0
All (primary)0.2%[0.1%, 0.4%]10

Due to Implement coercions between &pin (mut|const) T and &(mut) T when T: Unpin and Refactor ActiveJobGuard.

The first of those seems like an unavoidable regression (doing more work) and regressions are limited to secondary benchmarks which stress coercions and the second had some investigation done but didn’t lead to success for CI-built artifacts: see this comment for more details.

Improvements

Rollup of 4 pull requests #153642 (Comparison Link)

(instructions:u)meanrangecount
Regressions (primary)--0
Regressions (secondary)--0
Improvements (primary)-0.2%[-0.2%, -0.1%]3
Improvements (secondary)-0.2%[-0.3%, -0.1%]15
All (primary)-0.2%[-0.2%, -0.1%]3

Affects benchmarks that have been known to be spurious… they also regressed in the next commit (rollup just above) so it’s hard to say how real this improvement is.

Mixed

Rollup of 12 pull requests #153741 (Comparison Link)

(instructions:u)meanrangecount
Regressions (primary)0.1%[0.1%, 0.2%]4
Regressions (secondary)0.0%[0.0%, 0.0%]3
Improvements (primary)-0.2%[-0.2%, -0.2%]1
Improvements (secondary)-0.2%[-0.3%, -0.1%]13
All (primary)0.1%[-0.2%, 0.2%]5

Unclear what the root cause of the regression is, which is mostly affecting just diesel. Secondary benchmarks have a bunch of improvements. Marking as triaged since the regression seems fairly limited in scope and is probably due to a cleanup that’s worth doing even with the small perf cost. See comment for cachegrind diff.

rustdoc-search: update to stringdex 0.0.6 #153799 (Comparison Link)

(instructions:u)meanrangecount
Regressions (primary)0.4%[0.4%, 0.4%]1
Regressions (secondary)0.1%[0.1%, 0.1%]1
Improvements (primary)-3.5%[-6.8%, -0.2%]2
Improvements (secondary)-8.6%[-8.6%, -8.6%]1
All (primary)-2.2%[-6.8%, 0.4%]3

Fairly expected to be a mixed result (as most heuristic/optimization tuning things are). Improvements seem to outweigh regressions in practice, though few benchmarks are affected.

Nominated Issues

T-compiler

  • “Clean up crate type names to fix dylib vs staticlib confusion” rust#153863
    • Proposal to “[…] rename dylib to rdylib and staticlib to cstaticlib” (see comment)
    • Cargo raised concerns about the impact on cargo users (comment)
    • Any comment from T-compiler or should we be waiting on the cargo concerna to be first clarified?

RFC

  • No I-compiler-nominated RFCs this time.

Oldest PRs waiting for review

T-compiler

  • “Add support for xray in aarch64 unknown none target” rust#148666 (last review activity: 4 months ago)
    • cc: @Wesley Wiser
  • “Stabilize stack-protector” rust#146369 (last review activity: 3 months ago)
    • IIUC this is/was waiting on rfcs#3855, now in FCP, see comment cc: @Jieyou Xu @Santiago Pastorino @Wesley Wiser
  • “Add AddrspacePtr for pointers to non-0 addrspaces” rust#150452 (last review activity: 2 months ago)
    • This ended up being “blocked” on creating a notification group for the “GPU target” (see comment). The proposal for creating the group was approved
    • I’ve just opened a PR on team#2337 (feel free to have a look)
  • “naked functions: respect function-sections” rust#147811 (last review activity: 2 months ago)
    • cc: @Amanieu d’Antras (last ping 2 weeks ago)

Next meetings’ agenda draft: hackmd link