Mask mode sometimes require internal codepage for "no reason"

This error is unavoidable, because mask ranges and placeholders are byte oriented:

../run/john -stdout -mask=m[uü]ller
Error: Mask contains UTF-8 characters; You need to set a legacy codepage
       with --internal-codepage (UTF-8 is not a codepage).

However, this one is definitely avoidable:

$ ../run/john -stdout -mask:mülle[rR]
Error: Mask contains UTF-8 characters; You need to set a legacy codepage
       with --internal-codepage (UTF-8 is not a codepage).

In the latter case, the multi-byte character is in the fixed part of the mask, so it's not actually a problem. The reason we bail with error anyway is it's slightly complicated to sort it out but it could definitely be fixed.