Continuing the discussion from [ANN] termcolor 1.0 is out and moved to its own repository : @BurntSushi : I don't want to polute your termcolor announcement too much :slight_smile: [image] [ANN] termcolor 1.0 is out and moved to its own repository That’s incredible and actually kind of unexpected…

Ah. I didn't realize the query was complex. Yes, that is likely the source of the problem. Without being able to run this on my own (which is fine), I can't say for sure, but I can guess: Unicode aware word boundaries can be a sore point (which are used for --word-regexp). ripgrep has a few tricks…

Thanks for the ideas! [image] BurntSushi: ripgrep could read git’s index as well if it wanted to, but I’ve mostly resisted this temptation because That seems like a very sensible temptation to resist; Just the fact that it is impractical to make yourself depend on implementation details of …

@juleskers Thanks for running those! Based on that, it seems to me like the Unicode word boundary (added as part of --word-regexp) is probably preventing the DFA from running at all, and therefore falls back to the slower NFA. The DFA can't handle Unicode word boundaries sadly, so if the DFA sees a …

More benchmark results :tada: first of all: --dfa-size-limit 2G is our new hero: a blazing 13.4 seconds instead of well over an hour of my CPU-fan running at full tilt while my computer can't even keep up with my typing :smile: I'm glad you mentioned it before we did our 32K sample identifiers se…

[image] juleskers: Would it make sense to raise the default limit up from 10MB? Indeed. I expect this limit to be increased in the next release. 10MB is too stingy. But there will always be a point at which the limit is too small. :slight_smile: [image] juleskers: You mention there are “…

Unexpectedly good ripgrep performance vs. git grep

radix July 19, 2018, 12:38pm 6

Perhaps ripgrep could print a warning to stderr when it is falling back to a nfa for size reasons, which points to the flag?

7 Likes

Topic		Replies	Views
Wow, ripgrep is fast	5	1261	February 15, 2021
Questions about performance in regards to my generic regular expressions learning project help	13	1420	January 12, 2023
Ripgrep - Confessions of a father	4	1043	January 12, 2023
Help me reduce overhead of regex matching	62	10973	January 12, 2023
What's everyone working on this week (6/2016)?	23	4190	January 12, 2023

Unexpectedly good ripgrep performance vs. git grep

Related topics