On Fri, Mar 30, 2007 at 03:45:31PM +0900, Stephen J. Turnbull wrote:
Olivier Galibert writes:
> +separate_textual_runs_xft_mule (unsigned char *text_storage,
> + struct textual_run *run_storage,
> + const Ichar *str, Charcount len,
> + struct face_cachel *cachel)
Except for non-Xft Mule, `cachel' should be declared UNUSED, it
seems. (This shuts up a warning, it's not an error.)
Didn't know about this syntax. Should have looked.
> + Lisp_Object charset = ichar_charset(ch);
[...]
> + if (prev_charset != charset)
Tsk, tsk. EQ is your friend here.<wink>
I guess I have something against EQ. No idea why...
Builds, passes "make check" (ie, only the KNOWN BUGs fail),
and runs
(not used extensively yet). Interestingly, it doesn't show the
missing line/column info bug reported yesterday.
Neat. Simplifications tend to help, in practice.
Once you've answered about the str/memcpypart, I'll do a new, probably
final patch integrating your remarks.
The only part I'm really doubtful about is the ucs->byte split in the
-xft+mule translate_to_ics2 case. I don't even know how to have the
code pass through there.
OG.
_______________________________________________
XEmacs-Patches mailing list
XEmacs-Patches(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-patches