Can we run Lisp function in encode_terminal_code of term.c?

Can we run Lisp function in encode_terminal_code of term.c?

[Kenichi Handa]

In article <v97joddgi4.fsf@marauder.physik.uni-ulm.de>, Reiner Steib <reinersteib+gmane@imap.cc> writes:
> gb2312 characters are not displayed correctly in an UTF-8 enabled
> xterm when started with `emacs -nw'.

This is because encode_terminal_code (that is called from
write_glyphs) doesn't run pre-write-conversion of utf-8.  I
don't remember why I didn't do that.  There's a possibility
that running Lisp here is not safe.  As write_glyphs is
called from various places in dispnew.c, I can't tell easily
whether it is safe or not.  Does anybody know?

---
Ken'ichi HANDA
handa@m17n.org

Re: Can we run Lisp function in encode_terminal_code of term.c?

[Richard Stallman]

    This is because encode_terminal_code (that is called from
    write_glyphs) doesn't run pre-write-conversion of utf-8.  I
    don't remember why I didn't do that.  There's a possibility
    that running Lisp here is not safe.  As write_glyphs is
    called from various places in dispnew.c, I can't tell easily
    whether it is safe or not.

In general, running Lisp code from redisplay is safe nowadays.
I looked at the callers of write_glyphs, and it seems to be safe
to run Lisp code there.  I am not quite certain it is ok
to get an error from there, but it looks safe.

Anyway, if you make sure this particular Lisp code will
never get an error unless Emacs is badly screwed, then
we don't have to worry about the issue of what happens if it
gets an error.