--- On Tue, 6/12/11, Werner LEMBERG <w...@gnu.org> wrote: > > If you tell me which example files it is supposed to > work on - it > > looks like it does thai, chinese big5, korean, at > least. I can try > > testing stuff along. > > This would be great. The test file which covers > almost everything is > `CJKbabel.tex'. If you load this into Emacs 23 and > inspect the > properties of a character with `C-u C-x =', you can see > that for all > non-ASCII characters a `charset' text property is > set. Maybe we > should use this information instead.
I have a slightly better idea - if you have an input file which covers everything, since I also have a small emacs 22, I can run it through emacs 22 and see what emacs 22's charset returns, and restrict emacs 23 to that list (which would exclude unicode). This idea can only work if a complete-coverage input file exists though, which you sound like it does. The current cjk-enc.el further down the charset probe, does seem to drop into assumptions of a fairly limited list, so this might be the safest option - i.e. make a known wide-coverage example work. The disadvantage of this approach is that anything not in that example input file is gauranteed *not* to work :-). But it is a start :-). Hin-Tak _______________________________________________ Cjk maillist - Cjk@ffii.org https://lists.ffii.org/mailman/listinfo/cjk
