Edward Cherlin wrote: >On Sunday 30 March 2003 06:29 pm, Jungshik Shin wrote: > > >>Edward Cherlin wrote: >> >> >>>On Sunday 30 March 2003 03:26 am, Jungshik Shin wrote: >>> >>> >>> >>>I >>>can't test some of the others myself, and haven't heard any >>>detailed information on them. I have not found any problems >>>with diacritics in Latin and Cyrillic. >>> >>> >> Well, you do have problems with characters with diacritics >>in Latin,Greek and Cyrillic for which >>Unicode does NOT have assigned and will NEVER assign separate >>codepoints. That's >>what I was talking about. There are tens , if not hundreds, >> >> > >thousands, if not tens of thousands. I'm a mathematician. > >
I know how to multiply, too. It doesn't take a mathematician to multiply, does it? :-) The reason I wrote tens/ hundreds instead of thousands/tens of thousands was that I like to give the number of combinations that have turned up in existing documents rather than the number of all possible combinations. > > >>of combinations >>(base character + one or more diacritic mark(s)) that can ONLY >>be represented by combining character sequences. >> >> > >Like this? >à̀ >It's an a with two accents, and it composes and displays >correctly in kwrite and kmail, with one accent above the other. > >Let's try some more. >á̀ế̀̀î́̀ổ́̀̀û̀̀n̂́̀x̂̉́̀̀ >Not too bad, except that only the first three accents on each >letter are actually displayed, and the dot on the i isn't >removed. Curiously, Yudit doesn't handle multiple accents as >well as these simple-minded apps do. > Yudit needs the same change as I proposed for Pango in this mail and a couple of others. Yudit supports opentype layout table for several Indic scripts and it needs to do the same for Latin/Greek/Cyrillic alphabets. SIL has one such font. Unfortunately, the last time I downloaded it, there's something wrong with zip and I couldn't try it. (http://www.sil.org/~gaultney/gentium/index.html) > >What do you see in your mail? > > I can't tell without knowing what I'm supposed to see. Anyway, what I see is two diacritics overlapped over each other instead of taking disjoint 'spaces' alongside or on top of /below each other. See http://www.columbia.edu/kermit/st-erkenwald.html for a real life example. Didn't I specifically write that Pango does not support diacrtic marks combined with base characters while Uniscribe does (although it didn't until very recently)? I know that xterm and vim support up to two combining characters and that's how pre-1933 Korean script and Latin/Greek/Cyrillic diacritic marks are supported by xterm/vim. I guess kmail/kwrite do likewise. However, that's a kind of the last resort when you don't have a better way to do it properly. Eventually, what we need is support in Pango and that's filed as bug 101079 (see http://bugzilla.gnome.org/show_bug.cgi?id=101079) Other pango bugs I filed (excluding Korean-specific ones) include : http://bugzilla.gnome.org/show_bug.cgi?id=101081 http://bugzilla.gnome.org/show_bug.cgi?id=106624 >The starting point of this discussion was the inability to use >Chinese, Korean, and Japanese IMEs in the same locale. I write >documents in all three languages, and I would do it more often >if it were actually convenient. > This is becoming rather frustrating. How many times do I have to write that it IS possible right now to install all of them and switch between them in a *single* application (session) running under any UTF-8 locale of your choice? Why don't you try installing all three of them (im-ja, imhangul and wenju ) and fire up gedit and right-click on the text input area to see what you have? The very same information was given in last Decemeber and this thread doesn't add any new information except for im-ja in place of other less advanced Japanese gtk2 input modules. Jungshik -- Linux-UTF8: i18n of Linux on all levels Archive: http://mail.nl.linux.org/linux-utf8/
