[feature] consider using fancy-regex to support look-around and backtracking

[eugenesvk]

⭐ Suggestion

Currently ast-grep regex support is missing a couple of features

The regular expression is written in Rust syntax, not the popular PCRE like syntax. So some features are not available like arbitrary look-ahead and back references."

However, there is this fancy regex crate that implements "look-around and backtracking" https://github.com/fancy-regex/fancy-regex, so it might be better to use that instead of the default regex

💻 Use Cases

Same use cases where you'd use regex, but richer and allowing for better matches

[HerringtonDarkholme]

Hi, thanks for the suggestion.

fancy-regex looks like not as mature as regex. I would await more feedback and benchmarks about fancy-regex for its integration.

[HerringtonDarkholme]

From fancy-regex's README

Still in development, though the basic ideas are in place. Currently, the following features are missing:

The status was committed 7 years ago, telling from git blame...

Looking at its issue list, I can see these opening bugs:

• Invalid back reference error fancy-regex/fancy-regex#112
• Fuzzing results in index out of bounds fancy-regex/fancy-regex#122
• (|a)* matches more text than (|a)+ does fancy-regex/fancy-regex#70 (this has been open for more than two years)

I sincerely don't think fancy-regex is production-ready now.

[eugenesvk]

Well, you could get feedback and benchmarks in the most important use case - AST-grepping - by adding that as an optional engine :)

Though from my "searching end user" perspective the issues linked are dwarfed by the lack of look-ahead and back-references, those I use often in alternative engines, and it's a big limitation of the "production ready" Rust's default one (even if the back-references,bugs in cases like [\042\223\224\262\263\271]{2}\S{0,16}[\042\223\224\262\263\271]{2} that I'll never encounter and some fuzzed incantation I'll also never use fails)

[HerringtonDarkholme]

@eugenesvk Would you like to give me an example where you have to use lookaround operator? That will help me understand your use case.

Also, you can use JS API or Python API if lookahead/lookbehind is critical for you.