-
Notifications
You must be signed in to change notification settings - Fork 13.4k
Only insert nodes which changes lint levels in the LintLevelMap #58176
New issue
Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? # to your account
Conversation
The job Click to expand the log.
I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact |
@bors try |
⌛ Trying commit 76b688633cd55ff5ee282a70fc1790d229f539b1 with merge 0677dc2b768296023b6cb338d6afbdc5ab0be878... |
☀️ Test successful - checks-travis |
@rust-timer build 0677dc2b768296023b6cb338d6afbdc5ab0be878 |
Success: Queued 0677dc2b768296023b6cb338d6afbdc5ab0be878 with parent b2c6b8c, comparison URL. |
Finished benchmarking try commit 0677dc2b768296023b6cb338d6afbdc5ab0be878 |
1d97ce4
to
59c6c5a
Compare
@bors try |
[WIP] Only insert nodes which changes lint levels in the LintLevelMap r? @eddyb
The job Click to expand the log.
I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact |
@bors try |
Only insert nodes which changes lint levels in the LintLevelMap r? @eddyb
☀️ Test successful - checks-travis |
@rust-timer build f94acd0 |
Success: Queued f94acd0 with parent 65e647c, comparison URL. |
Finished benchmarking try commit f94acd0 |
@bors try |
⌛ Trying commit 776e2eacdd50eae29db3d72949bada4ea43659fe with merge 74c0b1340d42de21d4d37f561338cbab04a9e6e2... |
@oli-obk I don't know how to run clippy locally. I just get |
@bors try |
Only insert nodes which changes lint levels in the LintLevelMap r? @eddyb
How did you run it? |
I set the |
☀️ Try build successful - checks-travis |
@rust-timer build f2f65c8 |
Success: Queued f2f65c8 with parent f22dca0, comparison URL. |
Finished benchmarking try commit f2f65c8 |
perf is mixed, but max-rss seems mostly positive |
Btw, one change I'd like to eventually make is to allow lints to declare that they're pure, i.e. they don't cache any state between
|
filed #59024 |
cc @estebank in case you know stuff about lints |
@oli-obk Are you comfortable approving this? |
Implementation wise yes, but I'm not happy with the perf regressions. Since this enables actually making lints incremental, which again will be a perf bonus outstripping most of the regressions, I'm going to r+ @bors r+ |
📌 Commit f07ce55 has been approved by |
@oli-obk We might be able to get rid of the regressions by storing |
In order to do that properly we'd need to have different diagnostic systems for early lint passes and late lint passes (as early lint passes don't have |
I meant to say storing the HIR lint location in the serialized lints, so emitting a lint won't depend on the lint level map. |
Only insert nodes which changes lint levels in the LintLevelMap r? @eddyb
☀️ Test successful - checks-travis, status-appveyor |
r? @eddyb