Harold, It was posted in discussion board. http://matsu-www.is.titech.ac.jp/~sohda/cygwin/treebbs/treebbs.cgi?kako=1&all=10&s=10 He said that "This fix following two problem. 1. X freeze when we have 2 byte string after 1 byte string using ASTEX-X's XIM. 2. We can't use XIM on Cygwin/XFree86 4.1.0 etc."
Second problem is fixed http://cvsweb.xfree86.org/cvsweb/xc/lib/X11/lcSjis.c.diff?r1=3.9&r2=3.10 and http://cvsweb.xfree86.org/cvsweb/xc/lib/X11/lcEuc.c.diff?r1=3.11&r2=3.12 that Allow UTF8 conversion to work for Japanese locales (#A.1527, Etsushi Kato). So I delete code that related 2nd problem. I write about 1st problem. CTEXT has GL and GR two graphic character sets. But CTEXT to multi-byte string/wide string converter in lcSjis.c and lcEuc.c have only signle set. So text use GR convertion fail. Those fix patch was wrote by TAKABE. I added following code. if (charset == ct_state.GR_charset) { clen = charset->length; do { (*(Uchar *)(ctptr-clen)) = BIT8ON(*(Uchar *)(ctptr-clen)); } while (--clen); } Because multi-byte text to CTEXT converter in lcSjis.c doesn't set 8-bit on even if it is in GR. I don't know his mail address. At Sun, 12 Oct 2003 23:05:57 -0400, Harold L Hunt II wrote: > > Kensuke, > > > > Kensuke Matsuzaki wrote: > > And it seems tha libX11 has some CTEXT convertion bug, I attach > > patch that based on TAKABE's work. > > http://www.ff.iij4u.or.jp/~t-takabe/xf410_xim_fix.diff > > Could you send me more information on this? Please describe what the > current problem is. Please send any links to a message or post that > describes what the patch does. Do you have Takabe's email address as well? > > If you can send me those things then I can make sure that this patch > makes it into XFree86, as long as it is a valid patch. Right now, I > don't know how to describe what it does. :) > > Harold Kensuke Matsuzaki
