From: G G Papazoglou (grp@med.uoc.gr)
Date: Mon Sep 09 2002 - 14:53:13 CDT
Hi. I have tried this both with versions 2.1.4 and 2.1.5 with no luck. The
following is a strange behaviour, and please forgive me if it is a common
issue but I haven't had the time to browse the previous articles.
If I include a "delete_msgnum = xxx" line in my config file to delete a
message, hypermail crashes as you can see:
---------BEGIN LOG------------------
Reading old headers...
6 messages.
Loading mailbox "/home/grp/hyper/ebmbox"...
13 messages.
Writing messages to "/usr/local/apache/htdocs/hypermail/"...
./cronHour.sh: line 12: 22042 Segmentation fault
/usr/local/bin/hypermail -p -u -m "/home/grp/hyper/ebmbox" -d
"/usr/local/apache/htdocs/hypermail/" -c "/home/grp/hyper/hypermail.cfg"
Reading old headers...
6 messages.
Loading mailbox "/home/grp/hyper/ebmbox"...
13 messages.
Writing messages to "/usr/local/apache/htdocs/hypermail/archives/"...
./cronHour.sh: line 15: 22043 Segmentation fault
/usr/local/bin/hypermail -p -u -m "/home/grp/hyper/ebmbox" -d
"/usr/local/apache/htdocs/hypermail/archives/" -c
"/home/grp/hyper/hypermail.cfg"
----------END LOG-----------------
This does not happen when the delete_msgnum line is not used. BUT, and AFTER
the crash has occured, if I remove this line, kill the stuck processes and
wait for the next call to occur, I notice that the message xxx is actually
removed !!!!! I really can't tell what's happening. Anyway this behaviour
causes me trouble because I call hypermail inside a cron command, and the
result is that the application never terminates, leaving the .lock files
existing and of course preventing all future executions.
Now I use RedHat 7.3 with Hypermail 2.1.4. Note that this whole situation
was not happening when I was using Solaris 8.
Any help would be appreciated.
Thanks
Greg