paul@devon.UUCP (12/04/87)
After applying patch 13 to my 2.11 sources (which were stable at patch 12),
I encountered the following problem:
% postnews
Is this message in response to some other message? n
Subject: Test
Keywords: Test
Newsgroups (enter one at a time, end with a blank line):
The most relevant newsgroup should be the first, you should
add others only if your article really MUST be read by people
who choose not to read the appropriate group for your article.
But DO use multiple newsgroups rather than posting many times.
For a list of newsgroups, type ?
> local.test
>
Distribution (default='local', '?' for help) :
[ test article edited here ]
What now? [send, edit, list, quit, write, append] s
Posting article...
Article not posted - exit status 139
Exit status 139 (128 & 11) would be signal 11 (SIGSEGV).
Here is the article I was attempting to post:
% cat /tmp/post008025
Subject: Test
Newsgroups: local.test
Distribution: local
Keywords: test
This is a test
%
Using ADB, I see that the cause of the problem is a call to free() with
a 0 (or NULL) value:
% adb dist/src/inews core
$c
_ngfchec+3292: _free (0x0)
_insert+100: _ngfchec(0x139A4, 0x0, 0x0)
_main+3632: _insert ()
start+58: _main (0x0, 0x13666)
$r
d0 0x0 a0 0x1007E
d1 0x0 a1 0xFA7E
d2 0x3 a2 0x0
d3 0x0 a3 0xD01E
d4 0x1 a4 0xDAF7
d5 0x1 a5 0xFFFFFFFC
d6 0xA a6 0xF93E
d7 0x0 sp 0xF93A
ps 0x8
pc 0x8127EA _free+24: bclr #0x0,d0
$q
%
This is on a Tandy 6000 (an MC68000 machine) running Xenix 3.2. This
is the first patch I've applied to the news software that has given
me any trouble. My temporary solution is to run inews from patch 12.
Is this an old problem that has resurfaced, or did I do something
wrong?
- paul
--
Paul Sutcliffe, Jr.
UUCP (smart): paul@devon.UUCP
UUCP (dumb): ...{rutgers,ihnp4,cbosgd}!bpa!vu-vlsi!devon!paulpleasant@rutgers.rutgers.edu (Mel Pleasant) (12/04/87)
> After applying patch 13 to my 2.11 sources (which were stable at patch 12), > I encountered the following problem: > > ... output from adb ... > > Is this an old problem that has resurfaced, or did I do something wrong? You haven't done anything wrong. There is a problem that I introduced (gulp!) in patch #13 which causes it to bomb on SysV based Unix systems. Apparently, a free() call given a NULL pointer on BSD systems simply returns. On SysV based Unix Systems, a NULL pointer argument causes a core dump. Patch #14 has just been released. You'll find it in this newsgroup. This patch, which does not contain any major new features, fixes the known problems introduced in patch #13 and therefore should be as stable as patch #12 was..... -- -Mel Pleasant uucp: {ames, cbosgd, harvard, moss, seismo}!rutgers.edu!pleasant arpa: PLEASANT@RUTGERS.EDU