[ty] Improve completion rankings for raise-from/except contexts#22775
Merged
BurntSushi merged 1 commit intoastral-sh:mainfrom Jan 22, 2026
Merged
Conversation
Typing conformance resultsNo changes detected ✅ |
|
RasmusNygren
requested review from
AlexWaygood,
Gankra,
MichaReiser,
carljm,
dcreager and
sharkdp
as code owners
MichaReiser
requested review from
BurntSushi
and removed request for
AlexWaygood,
Gankra,
carljm,
dcreager and
sharkdp
BurntSushi
approved these changes
Jan 21, 2026
Member
BurntSushi
left a comment
BurntSushi
left a comment
There was a problem hiding this comment.
LGTM!
This does regress the completions when no character is typed after the raise keyword raise but I think that's a reasonable trade-off to greatly simplify the implementation with AST-traversal rather than extending the token-based approach.
Can you add a test with a comment about this? (Pretty much just saying what you've said here.)
| base_exception_ty, | ||
| Type::homogeneous_tuple(db, base_exception_ty), | ||
| ], | ||
| ); |
Member
There was a problem hiding this comment.
I kind of wonder whether it'd made sense to only constructs these if we know we'll need them. Since I'd expect the common case is that we won't. But I think that's easy enough to do if this shows up on a profile.
Contributor
Author
There was a problem hiding this comment.
Yep I believe so, it would be quite cool to have some performance tests for completions running in CI (or at least an easy local benchmark) to have a baseline to compare performance against.
Member
There was a problem hiding this comment.
There is now an ad hoc CLI for benchmarking completions that I just added. The README should say how to use it. :-) https://github.com/astral-sh/ruff/tree/main/crates/ty_completion_bench
I already did a round of plucking low hanging fruit here: #22630
From looking at profiles, my sense is that this sort of work done here (even when it isn't necessary) is marginal.
And yeah, ideally we'd have some benchmarks in CI: astral-sh/ty#2570
Contributor
Author
There was a problem hiding this comment.
Oh look at that, nice!
Completions of the correct types are now favoured in `except <CURSOR>` and `raise <EXPR> from <CURSOR>` contexts
RasmusNygren
force-pushed
the
completions-exception-context
branch
from
c6fcabb to
dacf708
Compare
BurntSushi
approved these changes
Jan 22, 2026
Member
BurntSushi
left a comment
BurntSushi
left a comment
There was a problem hiding this comment.
Thanks!
(I kinda suspect we'll end up wanting the token matching to recover raise <CURSOR> completions, but we can do that in response to user feedback.)
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
3 participants
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.
Completions of the correct types are now favoured in
except <CURSOR>andraise <EXPR> from <CURSOR>contextsFixes astral-sh/ty#1779
This does regress the completions when no character is typed after the raise keyword
raise <CURSOR>but I think that's a reasonable trade-off to greatly simplify the implementation with AST-traversal rather than extending the token-based approach.Test Plan
New completion evaluation tasks with correct rankings (that produces incorrect rankings on main)