Merge pull request #1095 from rylev/triage-2021-11-09

Add triage 2021-11-09

Author: rylev

triage/2021-11-09.md

@@ -0,0 +1,100 @@
+# 2021-11-09 Triage Log
+
+Largely a positive week despite taking a significant performance hit from turning on incremental compilation verification for a subsection of the total queries that the compiler does in order to more quickly catch bugs in incremental compilation. Luckily optimizations in bidi detection brought large performance improvements.
+
+Triage done by **@rylev**.
+Revision range: [6384dca100f3cedfa031a9204586f94f8612eae5..eee8b9c7bafade55981d155dae71657f1cc55a22](https://perf.rust-lang.org/?start=6384dca100f3cedfa031a9204586f94f8612eae5&end=eee8b9c7bafade55981d155dae71657f1cc55a22&absolute=false&stat=instructions%3Au)
+
+2 Regressions, 4 Improvements, 4 Mixed; 1 of them in rollups
+45 comparisons made in total
+
+#### Regressions
+
+Implement `RefUnwindSafe` for `Rc<T>` [#87467](https://github.com/rust-lang/rust/issues/87467)
+- Large regression in [instruction counts](https://perf.rust-lang.org/compare.html?start=0b4ac62ddaf9ee01b8aaf9dd7097f1f541d64551&end=4061c0407978a00c5c2518d898ad8406da28c106&stat=instructions:u) (up to 1.1% on `full` builds of `cargo`)
+- Unsurprising regression in one instance of a rustdoc run since `get_auto_trait_impls` has bad algoritmic complexity. This issue is being tracked somewhat [here](https://github.com/rust-lang/rust/issues/90512).
+
+Enable verification for 1/32th of queries loaded from disk [#90361](https://github.com/rust-lang/rust/issues/90361)
+- Very large regression in [instruction counts](https://perf.rust-lang.org/compare.html?start=3e3890c9d4064253aaa8c51f5d5458d2dc6dab77&end=495322d776fd6f679cd8cd4ca02b8fa834da654b&stat=instructions:u) (up to 6.3% on `incr-unchanged` builds of `coercions`)
+- Allows for verification of incremental compilation results to more easily catch bugs (which have unfortunately been a bit too common in the past).
+- As noted in the PR this is a regression of at most 7% on coercions opt incr-unchanged, and typically less than 0.5% on other benchmarks (largely limited to incr-unchanged).
+- The PR author and reviewer reviewed the regression impact and it was deemed acceptable.
+
+#### Improvements
+
+Optimize bidi character detection. [#90559](https://github.com/rust-lang/rust/issues/90559)
+- Very large improvement in [instruction counts](https://perf.rust-lang.org/compare.html?start=3326f19e8982ce033e04c163ddc520a76e42c737&end=5ec7d1dad6dead949a49c76c8ca0425a6e46a223&stat=instructions:u) (up to -11.6% on `incr-unchanged` builds of `coercions`)
+
+
+Fix ICE when rustdoc is scraping examples inside of a proc macro [#90583](https://github.com/rust-lang/rust/issues/90583)
+- Moderate improvement in [instruction counts](https://perf.rust-lang.org/compare.html?start=045612b8b97b344d55c13b9aad3cfb3e98405b77&end=0d1754e8bf6942b4c1d24d7c923438782129ba5a&stat=instructions:u) (up to -0.7% on `full` builds of `webrender`)
+
+
+Replace some uses of vec.drain(..) with vec.into_iter() [#90655](https://github.com/rust-lang/rust/issues/90655)
+- Moderate improvement in [instruction counts](https://perf.rust-lang.org/compare.html?start=5ec7d1dad6dead949a49c76c8ca0425a6e46a223&end=0727994435c75fdedd3e9d226cf434089b0ab585&stat=instructions:u) (up to -0.5% on `full` builds of `deeply-nested-async`)
+
+
+rustdoc: Cleanup `clean::Impl` and other parts of `clean` [#90675](https://github.com/rust-lang/rust/issues/90675)
+- Moderate improvement in [instruction counts](https://perf.rust-lang.org/compare.html?start=b3074819f615758da44a667bf3cb98c59227e2f9&end=3e3890c9d4064253aaa8c51f5d5458d2dc6dab77&stat=instructions:u) (up to -0.7% on `full` builds of `deeply-nested`)
+
+
+#### Mixed
+
+rustdoc: Add `DocVisitor` and use it where possible [#90475](https://github.com/rust-lang/rust/issues/90475)
+- Moderate improvement in [instruction counts](https://perf.rust-lang.org/compare.html?start=baba6687df3e83fdb15cc6ec239b4a1c75a30505&end=4ff90232a0c0c6adb9d2052da2206b26c3c723e4&stat=instructions:u) (up to -0.8% on `full` builds of `serde`)
+- Small regression in [instruction counts](https://perf.rust-lang.org/compare.html?start=baba6687df3e83fdb15cc6ec239b4a1c75a30505&end=4ff90232a0c0c6adb9d2052da2206b26c3c723e4&stat=instructions:u) (up to 0.5% on `full` builds of `webrender`)
+- The regressions have [already been justified](https://github.com/rust-lang/rust/pull/90475#issuecomment-959808629), but in short the improvements outweigh the regressions, and the code is much cleaner as a result.
+
+
+ Improve error when an .rlib can't be parsed [#88368](https://github.com/rust-lang/rust/issues/88368)
+- Moderate improvement in [instruction counts](https://perf.rust-lang.org/compare.html?start=46b8e7488eae116722196e8390c1bd2ea2e396cf&end=5fa94f3c57e27a339bc73336cd260cd875026bd1&stat=instructions:u) (up to -0.3% on `full` builds of `externs`)
+- Small regression in [instruction counts](https://perf.rust-lang.org/compare.html?start=46b8e7488eae116722196e8390c1bd2ea2e396cf&end=5fa94f3c57e27a339bc73336cd260cd875026bd1&stat=instructions:u) (up to 0.5% on `incr-unchanged` builds of `helloworld`)
+- It was pointed out that this could be simply due to a different binary layout impacting performance.
+- When taking into account significance factor, the improvements outweigh the regressions, and so it's likely not worth it to investigate deeper.
+
+
+Rollup of 4 pull requests [#90695](https://github.com/rust-lang/rust/issues/90695)
+- Moderate improvement in [instruction counts](https://perf.rust-lang.org/compare.html?start=515472757d2991e280965d7c1d235cfbb6ceb006&end=c57704f3eb4319cc93513c232e9c434a73af46d2&stat=instructions:u) (up to -0.5% on `full` builds of `deeply-nested-async`)
+- Small regression in [instruction counts](https://perf.rust-lang.org/compare.html?start=515472757d2991e280965d7c1d235cfbb6ceb006&end=c57704f3eb4319cc93513c232e9c434a73af46d2&stat=instructions:u) (up to 0.3% on `incr-unchanged` builds of `html5ever`)
+- Nothing jumps out as a possible cause for this but luckily the perf changes in question are small.
+
+
+Don't destructure args tuple in format_args! [#90485](https://github.com/rust-lang/rust/issues/90485)
+- Large improvement in [instruction counts](https://perf.rust-lang.org/compare.html?start=c57704f3eb4319cc93513c232e9c434a73af46d2&end=60952bc3da442cf2ea11dcc9c80a461a659bbcd7&stat=instructions:u) (up to -4.0% on `incr-unchanged` builds of `inflate`)
+- Large regression in [instruction counts](https://perf.rust-lang.org/compare.html?start=c57704f3eb4319cc93513c232e9c434a73af46d2&end=60952bc3da442cf2ea11dcc9c80a461a659bbcd7&stat=instructions:u) (up to 1.3% on `incr-patched: println` builds of `html5ever`)
+- The regressions seem to be coming in `expand_crate` which might be impacted by this change. However, nothing stands out immediately as a definite cause for concern.
+- Left a [comment](https://github.com/rust-lang/rust/pull/90485#issuecomment-964330008) for the author and reviewer to ask if they have an ideas before we investigate more.
+
+
+#### Untriaged Pull Requests
+
+- [#90695 Rollup of 4 pull requests](https://github.com/rust-lang/rust/pull/90695)
+- [#90485 Don't destructure args tuple in format_args!](https://github.com/rust-lang/rust/pull/90485)
+- [#90462 [master] Fix CVE-2021-42574](https://github.com/rust-lang/rust/pull/90462)
+- [#90422 Rollup of 5 pull requests](https://github.com/rust-lang/rust/pull/90422)
+- [#90235 Rollup of 6 pull requests](https://github.com/rust-lang/rust/pull/90235)
+- [#90183 Show all Deref implementations recursively](https://github.com/rust-lang/rust/pull/90183)
+- [#90067 Rollup of 10 pull requests](https://github.com/rust-lang/rust/pull/90067)
+- [#89939 Rollup of 10 pull requests](https://github.com/rust-lang/rust/pull/89939)
+- [#89858 Rollup of 6 pull requests](https://github.com/rust-lang/rust/pull/89858)
+- [#89695 Move top part of print_item to Tera templates](https://github.com/rust-lang/rust/pull/89695)
+- [#89608 Rollup of 12 pull requests](https://github.com/rust-lang/rust/pull/89608)
+- [#89534 Introduce `tcx.get_diagnostic_name`](https://github.com/rust-lang/rust/pull/89534)
+- [#89435 Rollup of 6 pull requests](https://github.com/rust-lang/rust/pull/89435)
+- [#89405 Fix clippy lints](https://github.com/rust-lang/rust/pull/89405)
+- [#89263 Suggest both of immutable and mutable trait implementations](https://github.com/rust-lang/rust/pull/89263)
+- [#89165 Fix read_to_end to not grow an exact size buffer](https://github.com/rust-lang/rust/pull/89165)
+- [#89125 Don't use projection cache or candidate cache in intercrate mode](https://github.com/rust-lang/rust/pull/89125)
+- [#89124 Index and hash HIR as part of lowering](https://github.com/rust-lang/rust/pull/89124)
+- [#89103 Migrate in-tree crates to 2021](https://github.com/rust-lang/rust/pull/89103)
+- [#89100 resolve: Use `NameBinding` for local variables and generic parameters](https://github.com/rust-lang/rust/pull/89100)
+- [#89047 Rollup of 10 pull requests](https://github.com/rust-lang/rust/pull/89047)
+- [#89030 Introduce `Rvalue::ShallowInitBox`](https://github.com/rust-lang/rust/pull/89030)
+- [#88945 Remove concept of 'completion' from the projection cache](https://github.com/rust-lang/rust/pull/88945)
+- [#88880 Rework HIR API to make invocations of the hir_crate query harder.](https://github.com/rust-lang/rust/pull/88880)
+- [#88824 Rollup of 15 pull requests](https://github.com/rust-lang/rust/pull/88824)
+- [#88804 Revise never type fallback algorithm](https://github.com/rust-lang/rust/pull/88804)
+- [#88719 Point at argument instead of call for their obligations](https://github.com/rust-lang/rust/pull/88719)
+- [#88703 Gather module items after lowering.](https://github.com/rust-lang/rust/pull/88703)
+- [#88627 Do not preallocate HirIds](https://github.com/rust-lang/rust/pull/88627)
+