Re: [Chicken-users] Re: Problem w/ Utf8 egg: couldn't load case-map-1.dat

[Alex Shinn]

At Fri, 16 Sep 2005 09:00:39 -0000, Thomas Chust wrote:
> 
> Am 16.09.2005, 01:36 Uhr, schrieb Alex Shinn <address@hidden>:
> 
> > [...]
> > I'm also considering breaking the utf8 module into 3 modules:
> >
> >   utf8
> >   utf8-srfi-13
> >   utf8-srfi-14
> 
> I think this would be a good idea. I regularly use the utf-8 egg, but  
> nearly never I need all the procedures provided by it. Also I rarely use  
> the utf-8 character classes. Splitting the egg into separate units will  
> probably increase both maintainability and usability!

OK, I've made this change and uploaded a new egg (which passes all old
and new tests on x86_64 Linux and on OSX).  For the time being the
char-set and case-map libraries are still included.

[Turns out arithmetic-shift was a red herring - there *is* a bug in it
for bit patterns where the high bit gets set, but there should be no
numbers that large in any of the Unicode-related eggs (the largest
char value is 2^21).  It was an another stupid mistake in a different
direction - I was shipping static little-endian files but still
detecting the platform endianess, so OSX got out of range numbers.]

-- 
Alex