Improve performance by doing an early filter

[flaviut]

Is your feature request related to a problem? Please describe.

I'm always frustrated when I type in a query and it takes a while to return results.

Describe the solution you'd like

It'd be nice if performance was optimized. For example, the following command:

comby -o -matcher .c -jobs 8 ':[_] << :[_] << endl' . 2>/dev/null $(rg -H '<<|endl' | cut -d: -f 1 | uniq)

is way faster (30s) than this command:

comby -o -matcher .c -jobs 8 ':[_] << :[_] << endl' . 2>/dev/null

(3min).

I'm sure there's other clever ways to improve performance as well.

Additional context
Add any other context or screenshots about the feature request here.

[rvantonder]

Hi 👋

This is almost certainly because the pattern starts with :[_], which unlike grep or line-based matching, will match across newlines (i.e., entire chunks of the file) before hitting the << part. This is a known behavior (there's a documentation tip for about it) and a warning was added in a recent release #147 to warn on this type of pattern.

My guess is that if your use case is somewhat line based, starting the pattern with something like :[_.] would work better (see reference, but it may not be sufficient for your use case depending on what the expression looks like--if so, let me know, this is still an active area of improvement).

[rvantonder]

@flaviut closing this for now. Future syntax improvements will help avoid the slow behavior.

[flaviut]

Understood--apologies for not getting back to you earlier, I read your message but didn't get around to replying.

Very cool project, thank you for starting it and all the time you put into it :)