UTF8 RegexFix doesn't correct capture's text

[chekoopa]

convertMatchText @ Text.RE.ZeInternals.Types.Match does perfectly correct captures' offsets and lengths, but capturedText is left intact (at this very line you can see, it's put straight from input), which may provoke more issues with using the library, mostly Text.RE.PCRE.Text.

The workaround is take (captureLength c) $ drop (captureOffset c) $ (captureSource c), but it's kind of lame. Incorporating similar code into RegexFix would make it more transparent but may impact on performance.

[cdornan]

@chekoopa thanks for the clear analysis. I am far too busy to be able to work on this at the moment but will be amenable to carving out some time. The more demand there is the sooner I am likely to get to this so please shout if anybody needs this fixed.

[kt0d]

>matchedText $ "żX" ?=~ [re|ż|]
Just "\380X"

It's may be easy to work around if you just want one match, but I originally encountered this problem using (*=~/) (search and replace).