[ic] Status of server error generating bug in rpc mode?

Stefan Hornburg Racke interchange-users@interchange.redhat.com
Fri Nov 16 02:41:01 2001


Mike Heins <mheins@redhat.com> writes:

> Quoting interch (interch@client.paymentonline.com):
> > The problem appears to be that when you reference a database, this time in
> > a loop tag using fi=db, that table gets added to the list of databases IC
> > thinks is valid.  dummy_database loops through all of those, and dies when
> > it runs into one that isn't defined.  The only place I have found the
> > offending code so far is in lib/UI/pages/admin/template_edit.html in the
> > code I posted in the last message.  Just go in and change fi=backup to
> > fi=someothername, then go to content->template edit, edit a template, and
> > save it.  You should get the error at that point, and then after that
> > click on admin and you will see someothername listed as an IC table..
> > 
> > If not running in pre fork mode this doesn't happen.  From here I am
> > pretty lost, not knowing the internals of IC I'm not really sure where to
> > start looking next without spending a lot of time.
> > 
> 
> Thank you, Chris. Very much. Your analysis put me onto it straightaway.
> 
> Shows to go you that you can learn something new every day. I never
> thought Perl auto-vivified hash keys on a "defined" check, but now
> I know different. It must do a depth-first search, autovififying
> on the way.

Just for the record: only "exists" prevents autovivifying.

Ciao
        Racke

-- 
Die Erde bleibt keine Scheibe. --- The earth remains no disk.

For projects and other business stuff please refer to COBOLT NetServices
(URL: http://www.cobolt.net; Email: info@cobolt.net; Phone: 0041-1-3884400)