2:12 a.m.
This bug report will be sent to the XEmacs Development Team,
not to your local site managers!!
Please write in English, because the XEmacs maintainers do not have
translators to read other languages for them.
In XEmacs 21.2 (beta32) "Kastor & Polydeukes" [Lucid]
(i386-unknown-freebsdelf3.4, Mule) of Wed Apr 19 2000 on elvenbow.nc.kyushu-u.ac.jp
configured using `configure --compiler=gcc295 --with-mule --without-xim --without-canna
--without-wnn --without-wnn6 --error-checking=gc,malloc --debug --with-widgets=athena
--with-dialogs=athena --with-xfs'
Please describe exactly what actions triggered the bug
and the precise symptoms of the bug:
I tried using CVS-updated xemacs-21.2.b32 yesterday, but it crashed
(SEGV) after 5 minutes of normal use (I was typing some texts in
irchat). Next time XEmacs crashed while I was reading mails using
Mew. Still I couldn't reproduce the crash, but it occured a bit too
frequent, so I decided to report once. If there are any good methods
to isolate the problem, please tell me and I'll try them.
It hadn't occured while I was using 21.2.b31, I believe.
There were no lisp trace. I attached the stack trace of the latter
crash. I lost the former one, but it also died inside
garbage_collect_1().
--
Yoshiaki Kasahara
kasahara(a)nc.kyushu-u.ac.jp
11:25 a.m.
On Thu, 20 Apr 2000 10:12:25 +0900,
Yoshiaki Kasahara <kasahara(a)nc.kyushu-u.ac.jp> said:
I tried using CVS-updated xemacs-21.2.b32 yesterday, but it crashed
(SEGV) after 5 minutes of normal use (I was typing some texts in
irchat). Next time XEmacs crashed while I was reading mails using
Mew. Still I couldn't reproduce the crash, but it occured a bit too
frequent, so I decided to report once. If there are any good methods
to isolate the problem, please tell me and I'll try them.
It seems that after I use 'gnuclient -nw' to call a frame into a tty
terminal, XEmacs tends to die soon while garbage-collect. It doesn't
occur when I use only X frames. Also, it doesn't crash when I use tty
only (xemacs -nw). It crashes when I use X frames and tty frames
simultaneously.
It doesn't occur on xemacs -vanilla. But after I start gnuserv,
create a tty frame and load .emacs, it almost always crashes. Now I'm
trying to divide my .emacs to find which function/package is the cause
of the crash.
Also, I noticed that there is something wrong with line wrapping in a
tty frame. When I lengthen a line longer than the width of the frame,
the tail of the line is disappeared and doesn't display on the next
line until I hit ^L. It never occurs in a X frame.
--
Yoshiaki Kasahara
kasahara(a)nc.kyushu-u.ac.jp
9:19 a.m.
On Sun, 23 Apr 2000 19:25:50 +0900,
Yoshiaki Kasahara <kasahara(a)nc.kyushu-u.ac.jp> said:
It doesn't occur on xemacs -vanilla. But after I start gnuserv,
create a tty frame and load .emacs, it almost always crashes. Now I'm
trying to divide my .emacs to find which function/package is the cause
of the crash.
It seems that after (display-time) is invoked, xemacs dies soon while
garbage-collecting. I have not yet run xemacs without (display-time)
very long, but it seems more stable than it with (display-time).
The crash still occurs on 21.2-b33 (CVS updated 5/4).
Also, I ran xemacs inside gdb to find where SIGSEGV occurs. Here is
the backtrace of a crash.
(gdb) run -vanilla
Starting program: /usr/local/src/XEmacs/xemacs-21.2/xemacs/src/xemacs -vanilla
Program received signal SIGSEGV, Segmentation fault.
0x80858bb in compact_string_chars () at alloc.c:3050
3050 assert (!(FREE_STRUCT_P (string)));
(gdb) where
#0 0x80858bb in compact_string_chars () at alloc.c:3050
#1 0x8085e83 in gc_sweep () at alloc.c:3199
#2 0x8086509 in garbage_collect_1 () at alloc.c:3519
#3 0x80c652d in Ffuncall (nargs=1, args=0xbfbfd288) at eval.c:3167
#4 0x80c787b in run_hook_with_args_in_buffer (buf=0x873c400, nargs=1,
args=0xbfbfd288, cond=RUN_HOOKS_TO_COMPLETION) at eval.c:3704
#5 0x80c78d1 in run_hook_with_args (nargs=1, args=0xbfbfd288,
cond=RUN_HOOKS_TO_COMPLETION) at eval.c:3717
#6 0x80c74ce in Frun_hooks (nargs=1, args=0xbfbfd288) at eval.c:3571
#7 0x80c7a60 in run_hook (hook=141909220) at eval.c:3818
#8 0x80c87a6 in catch_them_squirmers_run_hook (hook_symbol=137418460)
at eval.c:4262
#9 0x80c2ffd in condition_case_1 (handlers=137320548,
bfun=0x80c8794 <catch_them_squirmers_run_hook>, barg=137418460,
hfun=0x80c8554 <allow_quit_safe_run_hook_caught_a_squirmer>,
harg=145941840) at eval.c:1644
#10 0x80c8a02 in safe_run_hook_trapping_errors (
warning_string=0x8272ac0 "Error in `pre-idle-hook' (setting hook to
nil)",
hook_symbol=137418460, allow_quit=1) at eval.c:4323
#11 0x8116ffb in run_pre_idle_hook () at event-stream.c:2012
#12 0x8117459 in Fnext_event (event=141889116, prompt=137320452)
at event-stream.c:2179
#13 0x80a1b54 in Fcommand_loop_1 () at cmdloop.c:566
#14 0x80a191f in command_loop_1 (dummy=137320452) at cmdloop.c:490
#15 0x80c2ffd in condition_case_1 (handlers=137320548,
bfun=0x80a18f8 <command_loop_1>, barg=137320452,
hfun=0x80a1124 <cmd_error>, harg=137320452) at eval.c:1644
#16 0x80a1247 in command_loop_3 () at cmdloop.c:252
#17 0x80a126f in command_loop_2 (dummy=137320452) at cmdloop.c:263
#18 0x80c2b9c in internal_catch (tag=137402564,
func=0x80a1264 <command_loop_2>, arg=137320452, threw=0x0) at eval.c:1310
#19 0x80a14a8 in initial_command_loop (load_me=137320452) at cmdloop.c:301
#20 0x80be15c in xemacs_21_2_b33_i386_unknown_freebsdelf3_4 (argc=1,
argv=0xbfbfd68c, envp=0xbfbfd694, restart=0) at emacs.c:1873
#21 0x80bed9e in main (argc=1, argv=0xbfbfd68c, envp=0xbfbfd694)
at emacs.c:2292
--
Yoshiaki Kasahara
Computing and Communications Center, Kyushu University
kasahara(a)nc.kyushu-u.ac.jp
4:24 p.m.
At 05:19 PM 5/4/00 +0900, Yoshiaki Kasahara wrote:
On Sun, 23 Apr 2000 19:25:50 +0900,
Yoshiaki Kasahara <kasahara(a)nc.kyushu-u.ac.jp> said:
> It doesn't occur on xemacs -vanilla. But after I start gnuserv,
> create a tty frame and load .emacs, it almost always crashes. Now I'm
Is this still true. I can't reproduce.
andy
--------------------------------------------------------------
Dr Andy Piper
Principal Consultant, BEA Systems Ltd
4:47 a.m.
On Sat, 06 May 2000 16:24:50 +0100,
Andy Piper <andy(a)xemacs.org> said:
>> It doesn't occur on xemacs -vanilla. But after I start
gnuserv,
>> create a tty frame and load .emacs, it almost always crashes. Now I'm
Is this still true. I can't reproduce.
Yes, it still occurs on recent xemacs. I found simpler method to
reproduce this problem.
1) xemacs -vanilla on X
2) M-x gnuserv-start
3) gnuclient -nw on a TTY (I'm using kterm)
4) M-x hanoi (there is no reason why I select this)
5) M-x display-time on a X frame
6) M-x garbage-collect ... crash
If I don't do 4), it seems that xemacs won't crash. 4) and 5) seem to
be interchangable.
If I start display-time on a TTY frame and garbage-collect just after
that, xemacs doesn't crash. But after I focus on a X frame and the
display-time's display on a mode-line changes to graphical icons, next
garbage-collection crashes the xemacs.
I hope this helps.
--
Yoshiaki Kasahara
Computing and Communications Center, Kyushu University
kasahara(a)nc.kyushu-u.ac.jp
8958
days inactive
8977
days old
4 comments
2 participants
participants (2)
-
Andy Piper -
Yoshiaki Kasahara