[ty] Collapse never paths in constraint set BDDs#21880
Merged
Conversation
Diagnostic diff on typing conformance testsNo changes detected when running ty on typing conformance tests ✅ |
|
carljm
approved these changes
Dec 9, 2025
dcreager
added a commit
that referenced
this pull request
Dec 10, 2025
* origin/main: (33 commits) [ty] Simplify union lower bounds and intersection upper bounds in constraint sets (#21871) [ty] Collapse `never` paths in constraint set BDDs (#21880) Fix leading comment formatting for lambdas with multiple parameters (#21879) [ty] Type inference for `@asynccontextmanager` (#21876) Fix comment placement in lambda parameters (#21868) [`pylint`] Detect subclasses of builtin exceptions (`PLW0133`) (#21382) Fix stack overflow with recursive generic protocols (depth limit) (#21858) New diagnostics for unused range suppressions (#21783) [ty] Use default settings in completion tests [ty] Infer type variables within generic unions (#21862) [ty] Fix overload filtering to prefer more "precise" match (#21859) [ty] Stabilize auto-import [ty] Fix reveal-type E2E test (#21865) [ty] Use concise message for LSP clients not supporting related diagnostic information (#21850) Include more details in Tokens 'offset is inside token' panic message (#21860) apply range suppressions to filter diagnostics (#21623) [ty] followup: add-import action for `reveal_type` too (#21668) [ty] Enrich function argument auto-complete suggestions with annotated types [ty] Add autocomplete suggestions for function arguments [`flake8-bugbear`] Accept immutable slice default arguments (`B008`) (#21823) ...
This file contains hidden or 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
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.
#21744 fixed some non-determinism in our constraint set implementation by switching our BDD representation from being "fully reduced" to being "quasi-reduced". We still deduplicate identical nodes (via salsa interning), but we removed the logic to prune redundant nodes (one with identical outgoing true and false edges). This ensures that the BDD "remembers" all of the individual constraints that it was created with.
However, that comes at the cost of creating larger BDDs, and on #21551 that was causing performance issues.
scikit-learnwas producing a function signature with dozens of overloads, and we were trying to create a constraint set that would map a return type typevar to any of those overload's return types. This created a combinatorial explosion in the BDD, with by far most of the BDD paths leading to theneverterminal.This change updates the quasi-reduction logic to prune nodes that are redundant because both edges lead to the
neverterminal. In this case, we don't need to "remember" that constraint, since no assignment to it can lead to a valid specialization. So we keep the "memory" of our quasi-reduced structure, while still pruning large unneeded portions of the BDD structure.Pulling this out of #21551 for separate review.