[This post is me doing my best Strunk ("If you don't know how to say a
word, say it loud!" x 2), so please correct my ignorances and augment my
half-truths]On Mon, Oct 31, 2005 at 10:54:18AM +0800, Mat wrote: > I have currently made a mini bootable linux of mine which includes linux > kernel 2.6.11(UTF8 and the 4 code pages for CJK enabled) I think all of that enableage only affects the console interface. By the time you're in X, you're in a different world. > an NCURSES based application. That's going to potentially be a problem - there are two ncurses libs - ncurses and ncursesw. The second one has "wide" char support - which is meant in the physical sense of "wide" (as opposed to the "wide byte" meaning). If you don't have ncursesw in use, and you try to draw a rectangle around some double-wide chinese char, ncurses will miscompute the width of the rectangle. To chech to see if any particular app is linked against the right one, just do an `ldd /usr/bin/foo`. If it's not linked against ncursesw, check ./configure for such an option, or munch on the Makefile a bit. > I would like to display string resources which are CJK on my shell or the > NCURSES interface. You'll need: - a terminal that doesn't chork when it sees UTF-8 (properly compiled xterms are good for this, these days) - fonts that cover all the code points for the characters you want to display. This can be tricky. Check out: http://eyegene.ophthy.med.umich.edu/unicode/fontguide/ > What steps do i have to perform to make this CJK display possible. I am > using the stripped down version of bash which is ash(available in > busybox). I don't have handy access to ash at the moment, but bash does a great job with UTF-8 characters that I've seen. Really, the shell doesn't have to have a whole lot of smarts to display this stuff well - the only piece I can think of that fails (e.g. in my favorite shell, zsh) is the command-line editor. HTHALB, -rjk -- Linux-UTF8: i18n of Linux on all levels Archive: http://mail.nl.linux.org/linux-utf8/
