|> So am I correct in gathering that, while the actual right-to-left |> functionality of emacs-bidi is sugject to crashes, I can still (via |> the leim/quail/arabic.el code in the current emacs-bidi download) |> input Arabic from left to right (forget about shaping for the moment) |> and, upon downloading a fixed-width Arabic (UTF-8) font such as |> console-data-1999.08.29, have zero-width diacritics appear in "their |> own space" as it were, without stacking on top of each other?
I haven't been following the bidi branch recently, only the unicode-2 branch. The arabic quail method allows input -- following the layout used by the X11 Arabic keyboard -- of the Arabic characters. If the bidi branch is following the unicode-2 branch then that much should work there, too. Since you mention console-data-1999.08.29 I presume you are running emacs at the Linux or BSD text console. I know the unicode-2 branch can display Arabic characters on the text console when run in a UTF-8 locale. Again, if the bidi branch is based on that, it should have similar support. But, as you ask, in the unicode-2 branch the stacking diacritics do fail to stack. Once Handa-san adds the shaping code derived from the http://www.m17n.org/ code (which is also in lisp), I suspect that both the combining diacritics and full Arabic shaping will work. You can see the lack of combining diacritics by viewing the arabic.el file itself. The characters at positions QWERASX all combine when viewed in eg rxvt-unicode or gvim but do not when viewed in (the unicode-2 branch of) emacs. I'm sure Eli can be more definitive on the bidi branch. -JimC -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6 _______________________________________________ emacs-bidi mailing list [email protected] http://lists.gnu.org/mailman/listinfo/emacs-bidi
