Re: Emacs crash CVS 2003-10-17

[Kenichi Handa <handa@m17n.org>]

In article <20031111.062904.179201862.wl@gnu.org>, Werner LEMBERG <wl@gnu.org> writes:
>>  Please show the result of these:
>>  (gdb) p *f->face_cache
>>  (gdb) p *f->face_cache->faces_by_id[26]

>   (gdb) p *f->face_cache
>   $6 = {
>     buckets = 0x8660de0, 
>     f = 0x85af0c0, 
>     faces_by_id = 0x8ba02f0, 
>     size = 200, 
>     used = 21, 
>     menu_face_changed_p = 0
>   }
>   (gdb) p *f->face_cache->faces_by_id[26]
>   Cannot access memory at address 0x0

> The last non-zero index is 20:

Then, please see all
f->face_cache->faces_by_id[N]->font_name and check if any of
them shows a big5 font.

>>  By the way, xbacktrace shows that Emacs crashed while calling
>>  message.  I remember you are using a separate frame for the
>>  minibuffer, so the message of what-cursor-position is shown in the
>>  minibuffer-only frame, right?

> No.  I only use a separate minibuffer-only frame if I access the
> 40x40 bitmap fonts for CNS-11643 which isn't the case currently.

Hmmm.  The bug occured when you called what-cursor-position
and Emacs tried to display the big5 character under the
cursor in the echo area.  That means that the big5 charater
was already displayed on the current frame.  Is it right?

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