Combining characters and MacOSX

Combining characters and MacOSX

[Jun T.]

Since all the filenames are decomposed in MacOSX, I guess
many Mac users (including me) have been waiting for the
combining character support eagerly.

But the recent "combining character patch" does not work
on MacOSX, just because wcwidth() on MacOSX is broken.
For example,

wchar_t w = 0x0301; /* COMBINING ACUTE ACCENT */
printf("%d %d\n", wcwidth(w), iswpunct(w) );

this will print "0 1" on Linux but "1 0" on MacOSX. I think
this is a serious bug of MacOSX, but I don't know whether Apple
will fix this or not; maybe they won't at least in a near future.

Is it possible to include a working version of wcwidth()
in the zsh source tree and use it if the system's wcwidth()
is broken?
For example, a version written by Markus Kuhn can be found in:
http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c
Gnulib also has wcwidth(), but I guess the above one is much
easier to include into the zsh source.

As I wrote in another post, use of iswpunct() can be avoided
if the combining character is defined solely by wcwidth(wc)==0.

Jun

Re: Combining characters and MacOSX

[Peter Stephenson]

"Jun T." wrote:
> Since all the filenames are decomposed in MacOSX, I guess
> many Mac users (including me) have been waiting for the
> combining character support eagerly.
> 
> But the recent "combining character patch" does not work
> on MacOSX, just because wcwidth() on MacOSX is broken.
>
> Is it possible to include a working version of wcwidth()
> in the zsh source tree and use it if the system's wcwidth()
> is broken?
> For example, a version written by Markus Kuhn can be found in:
> http://www.cl.cam.ac.uk/~mgk25/ucs/wcwidth.c

Yes, that should be usable.  Presumably it shouldn't be too difficult to
craft a configure test for whether the system's wcwidth() returns
something sensible on a standard zero-width character.

-- 
Peter Stephenson <pws@csr.com>                  Software Engineer
CSR PLC, Churchill House, Cambridge Business Park, Cowley Road
Cambridge, CB4 0WZ, UK                          Tel: +44 (0)1223 692070

Re: Combining characters and MacOSX

[Vincent Lefevre]

On 2008-04-19 00:52:19 +0900, Jun T. wrote:
> Is it possible to include a working version of wcwidth()
> in the zsh source tree and use it if the system's wcwidth()
> is broken?

If you replace the whole wcwidth() function, make sure that it doesn't
become inconsistant with the other wc functions. For instance, the LRM
behavior is system-dependent:

  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=457634#13

-- 
Vincent Lefèvre <vincent@vinc17.org> - Web: <http://www.vinc17.org/>
100% accessible validated (X)HTML - Blog: <http://www.vinc17.org/blog/>
Work: CR INRIA - computer arithmetic / Arenaire project (LIP, ENS-Lyon)