Re: More character matching bits

[Jarkko Hietaniemi]

> I think, following my line of thought, that [a-\N{KATAKANA LETTER KI}]
> should be equivalent to [\x{0061}-\x{30AD}], which would match any of

I think it should be an error.  If you mean the code points write the
code points.  Mixing symbolic names (KATAKANA LETTER KI) and native
characters (that supposedly are converted into Unicode code points by
magical means) is meaningless and one's wrists should be slapped.

> Code-point order at least allows us to collate 'a' and KATAKANA LETTER
> KI, which I can't think of any other sensible way to do it.

The fewer 'sensible ways' we create for ourselves the clearer and simpler
our definition is and the easier it is to both implement and explain.

-- 
$jhi++; # http://www.iki.fi/jhi/
        # There is this special biologist word we use for 'stable'.
        # It is 'dead'. -- Jack Cohen