🚀 go-pugleaf

GNU Emacs version: 18.55 + interruptable IO patches as in 18.57/src/s-sunos4.h
Hardware:          Sun-4/60
Operating system:  SunOS 4.1.1
GNUS version:	   3.13

I recently posted this to gnu.emacs.gnus as a followup to another
message. It really belongs here. I never got this problem under SunOS
4.1 or previous version of SunOS.

The following debugging information was generated by starting GNU
Emacs and then issuing the M-x gnus RET command. An error(6) followed
by an error(11) resulted. The core file from that dump was useless as
the stack had been corrupted. A subsequent 'gdb emacs' was started
which resulted in the following backtrace. Attempts to repeat the
problem after that failed.

-- 
Mark D. Baushke
mdb@ESD.3Com.COM

------- Forwarded Article from gnu.emacs.gnus

From: mdb@ESD.3Com.COM (Mark D. Baushke)
Newsgroups: gnu.emacs.gnus
Subject: Re: Too many groups, anyone else seen this?
Date: 18 Feb 91 08:48:38 GMT
Organization: 3Com Corp., Santa Clara, CA.
In-reply-to: gobbel@cogsci.edu's message of 17 Feb 91 05:06:48 GMT

On 17 Feb 91 05:06:48 GMT, gobbel@cogsci.edu (Randy Gobbel) said:

Randy> I was trying to set up a friend's gnuemacs to use gnus, and it
Randy> kept dumping core, after getting "error 6".  Setting him up
Randy> with a copy of my .newsrc and .newsrc.el took care of it.
Randy> Apparently there are now so many usenet groups that something
Randy> in the initial setup processing causes emacs to run out of
Randy> memory (I'm assuming).  Has anyone else run into this problem,
Randy> and is there a fix?

Randy> -Randy

I have seen this problem too. I have no fix, but I did manage to save
a traceback of what was going on. While I was debugging, my server
apparently got over some threshold on our news server and it stopped
getting the error. I did not get very far in that debugging round.

If anyone has a fix let me know.

It appears that mark_object is using the default: abort() at the
switch statement. At that time my objptr had a value of 0xf7ffec28
VALBITS is 24 on a Sun-4/60 running SunOS 4.1.1.

Here is the backtrace I got. The problem is hard to debug since it
appears to vary depending on how many newsgroups or unread articles
are around for your particular .newsrc file.

(gdb) bt
#0  0x100630 in kill ()
#1  0x41010 in mark_object (objptr=(int *) 0xf7ffec28) (alloc.c line 1079)
#2  0x407a4 in Fgarbage_collect  () (alloc.c line 790)
#3  0x4ed7c in Ffuncall (nargs=2, args=(int *) 0xf7ffe070) (eval.c line 1617)
#4  0x5f3d0 in Fbyte_code (bytestr=-2145040476, vector=-2077880312, maxdepth=28) (bytecode.c line 271)
#5  0x4e414 in Feval (form=86329332) (eval.c line 1398)
#6  0x4b800 in Fprogn (args=86305828) (eval.c line 305)
#7  0x4f98c in funcall_lambda (fun=86330948, nargs=0, arg_vector=(int *) 0xf7fff57c) (eval.c line 1823)
#8  0x4f3a8 in Ffuncall (nargs=1, args=(int *) 0xf7ffe3b0) (eval.c line 1713)
#9  0x5f3d0 in Fbyte_code (bytestr=-2145040596, vector=-2077880440, maxdepth=7) (bytecode.c line 271)
#10 0x4e414 in Feval (form=86328940) (eval.c line 1398)
#11 0x4b800 in Fprogn (args=86328892) (eval.c line 305)
#12 0x4f98c in funcall_lambda (fun=86305860, nargs=0, arg_vector=(int *) 0xc) (eval.c line 1823)
#13 0x4f3a8 in Ffuncall (nargs=1, args=(int *) 0xf7ffe6b0) (eval.c line 1713)
#14 0x5f3d0 in Fbyte_code (bytestr=-2145097956, vector=-2077906296, maxdepth=9) (bytecode.c line 271)
#15 0x4e414 in Feval (form=86271780) (eval.c line 1398)
#16 0x4b800 in Fprogn (args=86271884) (eval.c line 305)
#17 0x4f98c in funcall_lambda (fun=86271132, nargs=0, arg_vector=(int *) 0x0) (eval.c line 1823)
#18 0x4f3a8 in Ffuncall (nargs=1, args=(int *) 0xf7ffe9b8) (eval.c line 1713)
#19 0x5f3d0 in Fbyte_code (bytestr=-2145166148, vector=-2078055416, maxdepth=11) (bytecode.c line 271)
#20 0x4e414 in Feval (form=86203572) (eval.c line 1398)
#21 0x4b800 in Fprogn (args=86203708) (eval.c line 305)
#22 0x4fdb4 in unbind_to (count=3) (eval.c line 1899)
#23 0x5f488 in Fbyte_code (bytestr=52601192, vector=69428360, maxdepth=5) (bytecode.c line 287)
#24 0x4e414 in Feval (form=86203460) (eval.c line 1398)
#25 0x4b800 in Fprogn (args=86203492) (eval.c line 305)
#26 0x4f98c in funcall_lambda (fun=86203692, nargs=1, arg_vector=(int *) 0x1ef84) (eval.c line 1823)
#27 0x4f3a8 in Ffuncall (nargs=2, args=(int *) 0xf7ffef30) (eval.c line 1713)
#28 0x4af58 in Fcall_interactively (function=18594424, record=18388020) (callint.c line 448)
#29 0x21310 in Fcommand_execute (cmd=18594424, record=18388020) (keyboard.c line 1482)
#30 0x2167c in Fexecute_extended_command (prefixarg=86584492) (keyboard.c line 1540)
#31 0x4f1b0 in Ffuncall (nargs=2, args=(int *) 0xf7fff2b8) (eval.c line 1685)
#32 0x4af58 in Fcall_interactively (function=18451528, record=18387980) (callint.c line 448)
#33 0x21310 in Fcommand_execute (cmd=18451528, record=18387980) (keyboard.c line 1482)
#34 0x1f8e0 in command_loop_1 () (keyboard.c line 666)
#35 0x4d1a8 in internal_condition_case (bfun=(int (*)()) 0x1ef84 <command_loop_1>, handlers=18388220, hfun=(int (*)()) 0x1e974 <cmd_error>) (eval.c line 996)
#36 0x1eda4 in command_loop_2 () (keyboard.c line 431)
#37 0x4cba4 in internal_catch (tag=18388200, func=(int (*)()) 0x1ed78 <command_loop_2>, arg=18387980) (eval.c line 814)
#38 0x1ed40 in command_loop () (keyboard.c line 413)
#39 0x1e8a0 in Frecursive_edit () (keyboard.c line 301)
#40 0x1dfc0 in main (argc=1, argv=(char **) 0xf7fff84c, envp=(char **) 0xf7fff854) (emacs.c line 461)

-- 
Mark D. Baushke
mdb@ESD.3Com.COM

------- End of Forwarded Article from gnu.emacs.gnus