Re: [evolution-patches] NNTP one-liners

[Meilof <meilof wanadoo nl>]

Not Zed wrote:

> > The point is not that it is an invalid input, the point is that later 
> > on, the code in the folder tree does this:
> >
> > >     /* FIXME: camel's IMAP code is totally on crack here, @top's
> > >      * folder info should be @fi and fi->child should be what we
> > >      * want to fill our tree with... *sigh* */
> > >     if (m->top && m->fi && !strcmp (m->fi->full_name, m->top)) {
> > >         if (!(fi = m->fi->child))
> > >             fi = m->fi->sibling;
> > >     } else
> > >         fi = m->fi;
> >
> > As I understand it, this code is meant to check whether the first folder 
> > info item it receives is for the top node we requested (so in that case, 
> > it moves on to the children, and, according to the comment above, this 
> > should be the case). This works if m->fi is not NULL, but if it's NULL, 
> > the check is obviously not done. In this case, you'd see a hierarchy like:
> >
> > news.wanadoo.nl
> >  -> .
> >    -> foo
> >
> > I did find it kind of odd that other stores shouldn't have this problem 
> > though...
> >    
> Well that code checks for m->top being null ...
But if it's null, it doesn't accept the fact that it's a root item that 
should be ignored, or am I something missing here?

> FWIW, IIRC, the imap (or one of the others) code just does:
> if top == NULL
> top = "";
>
> anyway, so you can just do the same in nntp.  But the interface is ...
> NULL top is ok input.
The NNTP does this.

> Jeff really should've just fixed imap, if its returning a funny return,
> rather than hacking around it in the mailer!
Umm.. yes, I think the "totally on crack" comment speaks for itself here 
:) I currently use the crack-method myself, but changing this caused 
this problem for me...



Meilof
-- meilof wanadoo nl