Re: MidCOM: error editing an article (solved)

On man, 2006-01-16 at 18:48 +0100, Ingo Herz wrote:

There is no debug-switch in lib/rcs_functions, but I added some=20
debug-statements around the reported lines and ...

the problem was that the creator of the article has somehow been removed.=20
After fixing the creator manually in mysql to an existing person-id all run=
s=20
fine now with aegir and MidCOM.

Ingo

> At 12:10 15.01.2006, you wrote:
> >On l=F8r, 2006-01-14 at 17:27 +0100, Ingo Herz wrote:
> > > Am Samstag, 14. Januar 2006 15:58 schrieb Tarjei Huse:
> > > > On tor, 2006-01-12 at 11:15 +0100, Ingo Herz wrote:
> > > > > Fatal error: Call to a member function on a non-object
> > > > > in /var/cache/midgard/xxx/1-10-5-0.php(991) : eval()'d code on li=
ne
> > > > > 531
> > > > >
> > > > > This is line 991 in .../1-10-5-0.php:
> > > > > 991:
> > > > > eval('
> > > > >=20
> > ?>'.mgd_preparse(mgd_snippet($set["snippetpath"]."/lib/rcs_functions"))=
=2E'
> > > > ><?php ');
> > > >
> > > > Hi, have you checked if the rcs directory is writeable by the apache
> > > > process?
> > >
> > > #/var/lib/aegir/cvs# ls -ld .
> > > drwxr-xr-x  2 www-data www-data 8192 2006-01-14 17:22 .
> > >
> > > and there are lots of files like
> > >
> > > -r--r--r--  1 www-data www-data  7447 2006-01-13 09:52=20
> > fa577112f2a481c19897c8df44f6d3f6,v
> >   ^^^^^ <- are you sure? This looks wrong to me.
>=20
> One should never be shure... Of course this is wrong. Changed all 444 to=
=20
> 644, tried again but the error is still the same. What else could I do to=
=20
> debug this? I'm quite shure that it will work if I recreate the topic but=
 I=20
> really would like to know what is going wrong here.

Take a look at the snippet in question (lib/rcs_functions) I think it
has a debug paramet you can set to see what happens inside it.
T