markdownError look-behind is slower than necessary - #232
Open
jvoisin wants to merge 1 commit into
Open
Conversation
Problem: The markdownError syntax rule uses an unbounded look-behind
"\w\@<=_\w\@=" which fires on every underscore and rescans
backward further than needed, making it the highest-cost inline
rule on snake_case-heavy text.
Solution: Bound the look-behind to one byte and use "\ze" for the trailing
word-character check: "\w\@1<=_\ze\w". Byte-identical highlighting,
~27% faster for that rule on underscore-heavy input.
The rule flags an underscore between two word characters (e.g. foo_bar).
"\w\@1<=" checks exactly the preceding byte instead of the unbounded
"\w\@<=", and "\ze" ends the match before the following word character
instead of a separate look-ahead. The match boundaries are unchanged,
including for consecutive underscores such as "a__b" and "a_b_c".
Signed-off-by: Julien Voisin <julien.voisin@dustri.org
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem: The markdownError syntax rule uses an unbounded look-behind
"\w@<=\w@=" which fires on every underscore and rescans
backward further than needed, making it the highest-cost inline
rule on snake_case-heavy text.
Solution: Bound the look-behind to one byte and use "\ze" for the trailing
word-character check: "\w@1<=\ze\w". Byte-identical highlighting,
~27% faster for that rule on underscore-heavy input.
The rule flags an underscore between two word characters (e.g. foo_bar). "\w@1<=" checks exactly the preceding byte instead of the unbounded "\w@<=", and "\ze" ends the match before the following word character instead of a separate look-ahead. The match boundaries are unchanged, including for consecutive underscores such as "a__b" and "a_b_c".
Signed-off-by: Julien Voisin <julien.voisin@dustri.org