This patch avoids the segfault. Basically, the core_module hook needs to run first.
Any problems
with this?
Index: core.c
===================================================================
RCS file: /home/cvs/httpd-2.0/server/core.c,v
retrieving revision 1.8
diff -u -u -r1.8 core.c
--- core.c 2001/04/11 23:37:16 1.8
+++ core.c 2001/04/20 20:35:44
@@ -3274,7 +3274,7 @@
/* FIXME: I suspect we can eliminate the need for these - Ben */
ap_hook_type_checker(do_nothing,NULL,NULL,APR_HOOK_REALLY_LAST);
ap_hook_access_checker(do_nothing,NULL,NULL,APR_HOOK_REALLY_LAST);
- ap_hook_create_request(core_create_req, NULL, NULL, APR_HOOK_MIDDLE);
+ ap_hook_create_request(core_create_req, NULL, NULL, APR_HOOK_FIRST);
/* register the core's insert_filter hook and register core-provided
* filters
> I am fairly sure this patch is at least a contributing factor to a seg fault I am
>trying to debug.
> The seg fault, which happens on the first request to the server, is in
>ap_getline(). req_cfg is
> NULL after the call to ap_get_module_config(). We blow up trying to reference
>req_cgg->bb. with a
> NULL pointer. Here is the code snip...
>
> req_cfg = (core_request_config *)
> ap_get_module_config(r->request_config, &core_module);
> b = req_cfg->bb
>
> The problem is that ap_set_module_config() has not been called on core_module yet.
>Ryan, did this
> code serve pages for you before you committed it?
>
> Bill
>
>
>
> > rbb 01/04/17 20:53:35
> >
> > Modified: include httpd.h
> > modules/http config.m4 http_core.c http_protocol.c
> > mod_core.h
> > modules/loggers mod_log_config.c
> > server protocol.c
> > Log:
> > Move the keepalives field out of the conn_rec and into an HTTP specific
> > connection record. This also moves some HTTP specific back out of the
> > core and into the HTTP module.
> >
> > Revision Changes Path
> > 1.147 +0 -4 httpd-2.0/include/httpd.h
> >
>
>
