Crystal: Regex can't match utf-8 characters

[Nephos]

Hello,

I was trying to match some utf-8 characters but it doesn't seem to work.

/[[:alnum:]]/x.match "à" # => nil

It works in Ruby and seems compliant with PCRE, is there a bug with crystal or is there a subtlety I didn't notice here ?

Note: I also wrote an error which did not raise any alarm, a /[class]]/. Maybe it should be mandatory to escape the ']' inside or outside the class

[akzhan]

This behavior related to absent option:

class Regex
  @[Flags]
  enum Options
    PCRE_UCP = 0x20000000

But this option requires PCRE build with UCP support.

pcre brew formula on MacOS contains '--enable-unicode-properties' option, but for some reason PCRE_UCP failed. Needs more investigation.

See also https://gist.github.com/jweyrich/9803969

P.S.: Ruby works fine because built with special Oniguruma branch (not pcre).

[akzhan]

Looks like somebody should replace PCRE with Onigmo.

[straight-shoota]

I think we could at least provide Regex::Option::UCP so people can use it if necessary. Just need to document that this option requires a compatible libpcre.