2021-06-08 Triage Log

Some good improvements, and a few regressions. No large changes.

3 Regressions, 3 Improvements, 1 Mixed 1 of them in rollups

Always go through the expn_that_defined query. #86002

Moderate regression in instruction counts (up to 1.6% on incr-patched: println builds of webrender-check)
Necessary for correctness, regresions largely limited to incremental.

Add variance-related information to lifetime error messages #85343

Moderate regression in instruction counts (up to 4.3% on full builds of wg-grammar-check)
Seems to potentially be necessary, but difference from PR run suggests more can possibly be done to avoid some of the losses noted on the final merge.

Revert “Merge CrateDisambiguator into StableCrateId” #85891

Moderate regression in instruction counts (up to 4.9% on full builds of html5ever-opt)
Reverted due to performance concerns on the original PR (#85804), but does not seem to have recovered performance. Suggested path is to re-land but with measurements the original PR; potentially there are some performance improvements that can be done there as well.

Remove CrateNum::ReservedForIncrCompCache #85829

Very large improvement in instruction counts (up to -11.2% on full builds of deeply-nested-async-debug)
Unclear what the exact cause of the large improvement is. It may be worth investigating, as at first glance the change does not seem like it should have such strong effects.

Only compute the trait map once #85905

Moderate improvement in instruction counts (up to -1.6% on incr-unchanged builds of cranelift-codegen-check)

Miscellaneous inlining improvements #85892

Moderate improvement in instruction counts (up to -2.2% on full builds of ctfe-stress-4-check)

Rollup of 13 pull requests #85952

Moderate improvement in instruction counts (up to -4.8% on full builds of html5ever-opt)
Moderate regression in instruction counts (up to 2.9% on incr-patched: println builds of cargo-opt)
Predominantly an improvement, some incremental regressions; does not appear worthwhile to investigate in depth the exact cause.