index: measure regexp optimizer thresholds in runes - #1149
Merged
Conversation
The trigram optimizer operates on rune trigrams, but its eligibility check measured UTF-8 bytes. This sent short non-ASCII literals through an inconsistent planning path and could produce surprising composite trees. Use the parsed rune count so the optimizer and substring planner apply the same threshold. Amp-Thread-ID: https://ampcode.com/threads/T-01a03d06-0149-77e8-9156-33171018a250
keegancsmith
changed the title
keegancsmith
marked this pull request as ready for review
August 26, 2026 10:24
burmudar
approved these changes
Aug 26, 2026
eseliger
approved these changes
Aug 26, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Zoekt indexes rune trigrams, but the regexp optimizer currently decides whether a literal is large enough by measuring its UTF-8 byte length. Non-ASCII literals can therefore cross the optimizer threshold despite containing fewer than three runes, only to fall back to a different matcher when the substring planner counts runes correctly.
This makes the optimizer use the parsed regexp rune count and adds coverage around two-rune and three-rune Unicode literals, including the alternation shape from #1147. This is intentionally independent hardening rather than the complete symbol-alternation fix: the equivalent ASCII alternation still requires the separate symbol-planning change.
Validated with
go test ./... -short -count=1andgo build ./cmd/....