[red-knot] unify LoopState and saved_break_states #16406
Merged
+42
−52
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
41e8d59 to
01b0235
Compare
AlexWaygood
approved these changes
Feb 26, 2025
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
dcreager
added a commit
that referenced
this pull request
Feb 27, 2025
* main: [red-knot] unify LoopState and saved_break_states (#16406) [`pylint`] Also reports `case np.nan`/`case math.nan` (`PLW0177`) (#16378) [FURB156] Do not consider docstring(s) (#16391) Use `is_none_or` in `stdlib-module-shadowing` (#16402) [red-knot] Upgrade salsa to include `AtomicPtr` perf improvement (#16398) [red-knot] Fix file watching for new non-project files (#16395) document MSRV policy (#16384) [red-knot] fix non-callable reporting for unions (#16387) bump MSRV to 1.83 (#16294) Avoid unnecessary info at non-trace server log level (#16389) Expand `ruff.configuration` to allow inline config (#16296) Start detecting version-related syntax errors in the parser (#16090)
dcreager
added a commit
that referenced
this pull request
Feb 27, 2025
* dcreager/dont-have-a-cow: [red-knot] unify LoopState and saved_break_states (#16406) [`pylint`] Also reports `case np.nan`/`case math.nan` (`PLW0177`) (#16378) [FURB156] Do not consider docstring(s) (#16391) Use `is_none_or` in `stdlib-module-shadowing` (#16402) [red-knot] Upgrade salsa to include `AtomicPtr` perf improvement (#16398) [red-knot] Fix file watching for new non-project files (#16395) document MSRV policy (#16384) [red-knot] fix non-callable reporting for unions (#16387) bump MSRV to 1.83 (#16294) Avoid unnecessary info at non-trace server log level (#16389) Expand `ruff.configuration` to allow inline config (#16296) Start detecting version-related syntax errors in the parser (#16090)
MichaReiser
reviewed
Feb 27, 2025
| #[derive(Clone, Debug, Default)] | ||
| struct Loop { | ||
| /// Flow states at each `break` in the current loop. | ||
| break_states: Vec<FlowSnapshot>, |
There was a problem hiding this comment.
I haven't looked through all usages but I'm wondering if we could use a single break_states struct stored on SemanticIndexBuilder instead of allocating a Vec for each Loop.
What I'm thinking about:
Loop: stores the length of the "arena" at the time this loop was pushed- Getting all break states for the current loop is the same as
arena[start..] - When popping a loop, truncate the arena to
startto remove all snapshots belonging to the current state.
There was a problem hiding this comment.
I had originally thought I might invest more time in reducing vec allocations here, but the simple version was completely neutral on benchmarks, so I'm not sure it's worth putting more time into this right now. I think perf optimization work should be motivated by some evidence.
I suspect that in practice loops are relatively uncommon, and loops with break in them even more uncommon, so in the great majority of cases we never allocate for this vec at all. I think if we want to improve performance of semantic indexing, there will be much higher-ROI things to look at.
dcreager
added a commit
that referenced
this pull request
Feb 27, 2025
* main: [red-knot] unify LoopState and saved_break_states (#16406) [`pylint`] Also reports `case np.nan`/`case math.nan` (`PLW0177`) (#16378) [FURB156] Do not consider docstring(s) (#16391) Use `is_none_or` in `stdlib-module-shadowing` (#16402) [red-knot] Upgrade salsa to include `AtomicPtr` perf improvement (#16398) [red-knot] Fix file watching for new non-project files (#16395) document MSRV policy (#16384) [red-knot] fix non-callable reporting for unions (#16387) bump MSRV to 1.83 (#16294) Avoid unnecessary info at non-trace server log level (#16389) Expand `ruff.configuration` to allow inline config (#16296) Start detecting version-related syntax errors in the parser (#16090)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
We currently keep two separate pieces of state regarding the current loop on
SemanticIndexBuilder. One is an enum simply reflecting whether we are currently inside a loop, and the other is the saved flow states forbreakstatements found in the current loop.For adding loopy control flow, I'll need to add some additional loop state (
continuestates, for example). Prepare for this by consolidating our existing loop state into a single struct and simplifying the API for pushing and popping a loop.This is purely a refactor, so tests are not changed.