Re: [Sbcl-devel] unicode support

[William H. N. <wil...@ai...>]

On Wed, Dec 19, 2001 at 08:32:14PM +1100, Brian Spilsbury wrote:
> Christophe Rhodes wrote:
> 
> >Just to clarify, you want a utf-8-encoded string. This is a mapping
> >between indices and elements, so it's an array
> ><http://localhost/doc/hyperspec/Body/syscla_array.html#array>, but
> >it's slightly weird in that it has linear, not constant, access time
> >to arbitrary elements.
[...]
> >Hmm. Does anyone have any idea how to square this circle?
> >
> I don't think that constant access time is a significant issue for the 
> places that these get used.
> 
> utf8-encoded simple-strings must be immutable, since mutation can change 
> the length,
> and this more or less limits them to symbol-names, and string-literals.
> 
> Given that we can use fast methods for iteration (ie, not aref on each 
> element, but dovector, etc),
> in the standard/optimised functions then we don't pay for this in 
> concatenate, or print, etc.

I'm concerned that this might cause the implementation of
things like DOVECTOR to become unreasonably baroque. 

> >>You can have a patch upon 0.6.13 if you like, although it isn't ready 
> >>for release yet and will probably be about 6 megs uncompressed.
> >>
> >
> >Can we have something that doesn't include all the character data --
> >just the logic, in other words? It might be worth it so we can see
> >what's going on...
> >
> Sure, but you'll still be looking at a really big patch,  I'll have a 
> look at how big it gets. 

I don't want to try to merge the patch in 0.7.0, but I and probably
others wouldfind it interesting to see what you've done and start
thinking about it. If you don't have a place to distribute it, I
can probably figure out some way to make it available on SourceForge.
(Anonymous FTP is gone, but I think it's supposed to be possible
to put arbitrary binary stuff up for HTTP access.)

-- 
William Harold Newman <wil...@ai...>
"Programming should be fun, programs should be beautiful."
  -- P. (Paul?) Graham, quoted in comp.lang.lisp by David E. Young
PGP key fingerprint 85 CE 1C BA 79 8D 51 8C  B9 25 FB EE E0 C3 E5 7C