Re: Unicode-capable srfi-14 charsets (f49dbcadf...)

[Andy Wingo]

On Thu 27 Aug 2009 17:20, Mike Gran <address@hidden> writes:

> I pushed the commit for Unicode-capable srfi-14.  Of all parts necessary
> for Unicode, this patch has the largest amount of new code and is most
> likely to cause compilation problems.

On Linux, this now works fine for me; make succeeds, one test fails for
me (utf8 r6rs ports).

On Mac 10.5, I get an exception running the test suite, which actually
stops the test suite: bad argument to setlocale. I don't recall which
file that was in. Apologies for the imprecision, that machine is at
work.

> Since the default charsets such as char-set:lower-case are much larger
> in the Unicode scenario, I had to change both the underlying
> structures that hold them as well as the print format. I also had to
> pregenerate them, since it was too slow to create them at startup. So
> this is also the part of this project where I've taken the greatest
> liberties.

Looks good to me. Ideally over time these will migrate to being
implemented in Scheme. For now, please make that static table const, so
it can be shared between processes, and avoid some relocations at link
time.

Cheers,

Andy
-- 
http://wingolog.org/