search

sjshuck / hs-pcre2

Complete Haskell binding to PCRE2
Apache License 2.0
12 stars 2 forks source link

Options for TH regexes #3

Closed amesgen closed 2 years ago

amesgen commented 3 years ago

Thanks for this awesome package, the TH regexes for type-safe capture groups are excellent!

I migrated from lens-regex-pcre, and noted that it is not possible to customize the Options for a TH regexes (in particular, SubGlobal), in contrast to e.g. _capturesOpt.

Would something like mkRegexQQ or mkRegexTraversalQQ be feasible?

sjshuck commented 3 years ago

Thanks! Yeah, that's possible.

Although, "Opt" variants of current QQ functions would not work with SubGlobal. That's for substitutions only. Are you trying to do something like this (pseudocode):

myNonAffineTraversal = [_regexOpt SubGlobal|foo (...) baz|]

?

amesgen commented 3 years ago

Yeah exactly, I stumbled upon the current traversal being affine.

sjshuck commented 3 years ago

OK. I meant to specify SubGlobal is for PCRE2-native substitutions only, since it's free there.

I will open another issue about that since someone on Reddit needed that as well, and prioritize that one, since it's more of a blocker.

sjshuck commented 3 years ago

1.1.0 is released. Traversals in the library are now global rather than affine.

I don't want to add mkRegexQQ right now but I may in the future if there is more of a need for it. I'll keep this ticket open for now in case it becomes an issue again for you or someone else.

amesgen commented 3 years ago

Perfect, #5 fixed my use case, but I agree that this can stay open as a low-priority issue. :+1:

sjshuck commented 2 years ago

Closing due to inactivity.

Final thoughts: A mkRegexQQ helper function that takes arbitrary Options is probably undesirable because of things like callouts[^1]. If a specific Option is desired for TH regexes that cannot be achieved using (?...) syntax in the pattern, let an issue be opened for it.

[^1]: Option may in the future be refactored to distinguish between safe and unsafe options at the type level using GADTs or something, or unsafe options may be removed altogether.