Pavel@Janik.cz (Pavel Janík) writes:

> Can you please look at it, test it and report here the changes you need to
> make to it to suit your needs? Is this enough to support that method? I do
> not speak Vietnamese so I can not judge here :-(

I cannot build the source from CVS, the error was:
,----
| gcc -c  -D_BSD_SOURCE  -I/usr/X11R6/include     -g -O2  -DEMACS_BITMAP_FILES -DHAVE_CONFIG_H -Demacs -I. -I../src -I/mnt/hda11/tmp/emacs/lwlib -I/mnt/hda11/tmp/emacs/lwlib/../src xlwmenu.c
| In file included from /usr/X11R6/include/X11/Xos.h:284,
|                  from xlwmenu.c:38:
| /usr/X11R6/include/X11/Xarch.h:48: sys/byteorder.h: No such file or directory
| make[2]: *** [xlwmenu.o] Error 1
| make[2]: Leaving directory `/mnt/hda11/tmp/emacs/lwlib'
| make[1]: *** [really-lwlib] Error 2
| make[1]: Leaving directory `/mnt/hda11/tmp/emacs/src'
| make: *** [bootstrap-src] Error 2
`----
Just in case, I'm running GNU/Linux-2.4.18-rc4, XFree-4.2.0.

Though, I successfully built Emacs-21.2 with vntelex.el (copied from
CVS) and can type "Vietnamese" characters with telex input method.
However, Emacs pop ups the following message when saving the buffer to
file:
,----
| These default coding systems were tried:
|   vietnamese-viscii-unix
| However, none of them safely encodes the target text.
| ...
`----
and it prompts for utf-16-le-unix as the default coding system.

I don't really understand how MULE works, but I guess Werner would use
a wrong charset (It seems ISO-8859-? rather than
vietnamese-lower/upper--these character sets share many accent
characters) for translation characters in `quail-define-rules'.  I
re-edited the table and the problem has gone. Below is the changes.

Regards,
TL