Skip walking into param-env component if it has no placeholder/re-var #143500
Conversation
rustbot
added
T-compiler
|
@bors2 try @rust-timer queue |
This comment has been minimized.
This comment has been minimized.
[perf] Skip walking into param-env component if it has no placeholder/re-var
rustbot
added
the
S-waiting-on-perf
This comment has been minimized.
This comment has been minimized.
|
Finished benchmarking commit (e3ecd5f): comparison URL. Overall result: ❌✅ regressions and improvements - no action neededBenchmarking this pull request means it may be perf-sensitive – we'll automatically label it not fit for rolling up. You can override this, but we strongly advise not to, due to possible changes in compiler perf. @bors rollup=never Instruction countOur most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.
Max RSS (memory usage)This benchmark run did not return any relevant results for this metric. CyclesResults (secondary 0.2%)A less reliable metric. May be of interest, but not used to determine the overall result above.
Binary sizeThis benchmark run did not return any relevant results for this metric. Bootstrap: 461.008s -> 461.232s (0.05%) |
rustbot
removed
the
S-waiting-on-perf
compiler-errors
changed the title
|
r? lcnr |
rustbot
added
the
S-waiting-on-review
| @@ -1069,8 +1070,10 @@ where | |||
| } else { | |||
| ControlFlow::Continue(()) | |||
| } | |||
| } else { | |||
| } else if ty.has_type_flags(TypeFlags::HAS_PLACEHOLDER | TypeFlags::HAS_RE_INFER) { | |||
There was a problem hiding this comment.
why only RE_INFER instead of full infer?
There was a problem hiding this comment.
bc we only care about infer regions in this folder's logic -- it would not even be necessary if we didn't replace free regions with infer
There was a problem hiding this comment.
why would we care about ty/ct infer for this folder anyways? 🤔
There was a problem hiding this comment.
we only care as in: there should never be infer types/consts in the param_env 😁 if we end up with a ty var for whatever reason, then this folder would start incorrectly treating (): Trait<?x> as global evne though it shouldn't be. It currently detects ambiguous aliases in the env as global
There was a problem hiding this comment.
I don't see how that is the case and how this behavior would change with this fast path, given that visit_ty does not care about infer types at all and because structurally_normalize_ty just returns Ok(_) for infer types.
Can you explain in more detail, because I am obviously missing something here?
lcnr
added
S-waiting-on-author
compiler-errors
added
S-waiting-on-review
Although it only provides a minor perf improvement, it seems like it could matter in more pathological cases.