[imp] Apache segfaults due to error in Horde.php logMessage()

[Gonzalo Servat]

On 2/08/2003 2:22 AM +0200, Jan Schneider wrote:

>> The minimal patch attatched seems to correct the problem.
>
> This has been fixed in CVS already.

Hi Jan,

I'm still experiencing occasional child process segfaults in Apache and the 
cause still appears to be IMP/Horde. Could it be because of this Log bug? I 
switched to Log version 1.6.3 as suggested by Victorrio (sp?) but I still 
get them.

I've tried debugging Apache by using -X and also by starting minimal Apache 
child processes and attaching gdb to them to no avail. I also tried 
modifying "Logformat" for Access Log entries in Apache to include the child 
process that served the request so I can try and match the child process 
pid & date+time of the segfault and the request to find out which page was 
being accessed at the time when it happened, but that didn't help either. 
There were no requests that matched the same sort of time frame (+/- 20-30 
seconds) and the segfault time. As you can see, I'm very very keen to stop 
the segfaults :)

Should I try the CVS version?

Thanks in advance.

Best regards,
Gonzalo