httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roy T. Fielding" <>
Subject Re: segfaults on CGI's with current CVS snapshot
Date Mon, 04 Aug 1997 03:50:39 GMT
>A backtrace on the core file showed
>  (gdb) where
>  #0  0x67fdd in kill ()
>  #1  0x6753b in abort ()
>  #2  0x3500 in just_die (sig=11) at http_main.c:1482
>  #3  <signal handler called>
>  #4  copy_array (p=0xdc034, arr=0xd9074) at alloc.c:526
>  #5  0xb4b8 in sub_req_lookup_uri (new_file=0x9760c "/docs", r=0xd9074)
>      at http_request.c:660
>  #6  0x133fc in add_cgi_vars (r=0xd9074) at util_script.c:298
>  #7  0x278b8 in cgi_child (child_stuff=0xefbfbb98) at mod_cgi.c:298

I tried to find the problem but this trace does not match the current
code at all.  sub_req_lookup_uri does not call copy_array directly and
the line numbers are way off.  I also looked at a diff from 1.2.0 and
couldn't find anything that would result in a SIGSEGV in http_request.c.
Either you have local changes applied or the stack is blown.

The only thing I can think of is that get_module_config() is returning
a null pointer and none of the _walk routines are checking for it,
which might happen if m->module_index is screwed in http_config.c,
but too much has changed in http_config.c to be able to find anything.


View raw message