segv from repl

[Sandro Hawke]

Binary package hint: nmh

<fresh install of Ubuntu 7.04 Feisty>

$ sudo apt-get instal nmh mh-e
$ PATH=/usr/bin/mh:$PATH
$ comp
     send a message to yourself
$ inc
     see that message arrive - in my case it's message 7081
$ repl 7081
mhl: signal 11 (core dumped)

For some other message in my inbox I get slightly different output:
$ repl 1
Segmentation fault (core dumped)

I've gotten rid of my replcomps and replgroupcomps, and trimmed my .mh_profile down to just "Path: Mail"

[Sandro Hawke (sandro-w3)]

To be clear -- I tried it on lots of messages in my inbox, and on a new user account, and it never works.

[Larry Daffner (vizzie)]

I am seeing this problem too. It is happening in the mhl code. However, if I change -O2 to -O in debian/rules, it runs fine. Backtrace of the crash below.

vizzie@belew ~/nmh-1.2 $ gdb uip/mhl
GNU gdb 6.6-debian
Copyright (C) 2006 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you are
welcome to change it and/or distribute copies of it under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB. Type "show warranty" for details.
This GDB was configured as "x86_64-linux-gnu"...
Using host libthread_db library "/lib/libthread_db.so.1".
(gdb) run `mhpath cur`
Starting program: /home/vizzie/nmh-1.2/uip/mhl `mhpath cur`
[Thread debugging using libthread_db enabled]
[New Thread 47213561985792 (LWP 10062)]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 47213561985792 (LWP 10062)]
strcasecmp (s1=<value optimized out>, s2=<value optimized out>)
    at strcasecmp.c:33
33 while (tolower(*us1) == tolower(*us2++))
(gdb) bt
#0 strcasecmp (s1=<value optimized out>, s2=<value optimized out>)
    at strcasecmp.c:33
#1 0x0000000000404ae7 in process (folder=0x0,
    fname=0x7fffe857c85a "/home/vizzie/Mail/inbox/1952", ofilen=1, ofilec=1)
    at mhlsbr.c:915
#2 0x000000000040544a in mhl (argc=<value optimized out>,
    argv=<value optimized out>) at mhlsbr.c:459
#3 0x0000000000402cfe in main (argc=2, argv=0x7fffe857ab38) at mhl.c:24
(gdb) print us1
$1 = (const unsigned char *) 0x0
(gdb) print us2
$2 = (const unsigned char *) 0x7fffe8574890 "X-Spam-Checker-Version"

[JoshW (josh-bugs)]

I have experienced the same problem, and also found that it is corrected by changing -O2 to -O.

I would recommend that, at a minimum, this fix be incorporated.

[John Jendro (tacliat)]

I have experienced this too.

[Allan Bailey (zirpu)]

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=399271

This seems to be the same or at least similar bug.

[Allan Bailey (zirpu)]

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=420653

Is also related, stating that the -O2 switch was the problem.

The patch for the ultimate cause of the bug caused by the -O2 compile switch problem was
listed in http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=399271

[Brian Murray (brian-murray)]

Based off the upstream bug report it appears that this is resolved in package version 1.2-3 which is included with Gutsy and Hardy. In the event that someone tests it with Gutsy or Hardy and it works as expected the bug's status should be changed to Fix Released.

[Allan Bailey (zirpu)]

I can confirm that nmh works fine on Gutsy x86.

[Wouter Stomp (wouterstomp-deactivatedaccount)]

Fixed in 1.2-3