Re: [Chicken-users] UTF8 egg case conversion problem

[Alex Shinn]

On Tue, Jan 4, 2011 at 8:24 AM, Mehmet Köse <address@hidden> wrote:
>
> utf8-string-downcase cannot convert "Latin Letter With Dot Above"
> correctly. (U-0130, only Turkish has this letter afaik.) This one-line
> patch fixes the problem for me:

Teşekkür ederim!  I think this was an error from using the
same code for upcase and downcase.

> -     ((#x0130) (if (lang? opt "tr" "az") #\I dotted-small-i))
> +     ((#x0130) (if (lang? opt "tr" "az") #\i dotted-capital-i))

This is char-downcase, so both #\I and dotted-capital-i
are wrong.  I think it should be

  ((#x0130) (if (lang? opt "tr" "az") #\i dotted-small-i))

That is, in Turkish the dotted capital I always downcases
to normal i.  In other locales, dotted capital I isn't even a
character, so it should follow the rules for the Unicode
locale independent lowercasing, which require "dotted-small-i"
to preserve canonical equivalence.  "dotted-small-i"
is just small i with a combining dot above, which when
uppercased becomes capital I with combining dot above
(without the combining dot it uppercases to normal capital I,
losing the dot).

I'll double check some of these cases and release a new
version shortly.

-- 
Alex