"Philipp Rotmann, Linksystem Muenchen" wrote:
> 
>      Hi guys,
> 
> On 20.01.01 (13:41), Emiliano wrote:
> > > >One specific page?
> 
> Now I've got nearly the same problem, also with one specific page
> -- Apache sends part of the page and then segfaults. It doesn't
> always die, only often enough to make it a problem.
> 
> A request to http://www.wochenanzeiger.de/ works perfectly well,
> as well as to all other URIs at that host I know about. Only
> when serving http://www.wochenanzeiger.de/termine/, Midgard
> (Apache/1.3.12 (Unix) PHP/3.0.18+Midgard/1.4/SG Midgard/1.4/SG)
> segfaults:
> 
> [Sun Jan 28 22:11:11 2001] [debug] mod_midgard.c(1178): [client
> 195.30.2.138] midgard_translate_handler(latin1,
> http://www.wochenanzeiger.de:80/termine/)
> [Sun Jan 28 22:11:11 2001] [debug]
> /usr/local/include/mgd_apache.h(123): [client 195.30.2.138]
> Midgard: select_database kept current (newmidgard)
> [Sun Jan 28 22:11:11 2001] [debug] mod_midgard.c(1263): [client
> 195.30.2.138] Midgard: using style #9 (14 elements)
> [Sun Jan 28 22:11:11 2001] [debug]
> /usr/local/include/mgd_apache.h(123): [client 195.30.2.138]
> Midgard: select_database kept current (newmidgard)
> [Sun Jan 28 22:11:11 2001] [info] [client 195.30.2.138] Midgard:
> Template not found code-compat
> [Sun Jan 28 22:11:11 2001] [debug] midgard/mgd_main.c(135):
> [client 195.30.2.138] Midgard: snippet parse common path
> [Sun Jan 28 22:11:12 2001] [notice] child pid 9335 exit signal
> Segmentation fault (11)

David? Do snippets do anything outlandish that could be causing
this? I'm guessing no, I couldn't find anything.

> The HTML code sent until the segfault is still part of the style
> elements executed before the page content, and the page does at
> least not contain Ingo's friend mgd_is_article_in_topic_tree().

At what point does it break? What was the last midgard function called,
if known of course.

> Another observation which may be helpful: with a normal request
> including the same style elements I observe lots of blobs and
> filetemplates lines in the debugging output, as every image
> request is examined for being a blob or filetemplate my
> mod_midgard -- but they're simply files in the fs, and at the
> end they get served in those pages.

Hmyes, in this setup mod_midgard would check for each.

Emile

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to