Re: [evolution-patches] More fixes for the evolution-data-server memory buildup [Calendar]

[Harish Krishnaswamy <kharish novell com>]

committed after omitting the icalcomponent_free on the subcomp.

I still notice a leak there but since icalcomponent_free is supposed to
operate recursively on subcomponents, i guess, the correct fix belongs
elsewhere.

thanks,
harish


On Wed, 2004-12-08 at 16:06 +0100, Rodrigo Moya wrote:
> On Wed, 2004-12-08 at 18:57 -0800, Harish Krishnaswamy wrote:
> > hi,
> > 
> >  The attached patch provides a few more fixes - basically the EList held
> > by e-cal-backend.c not getting freed and thereby not letting
> > edatacalview objects to get freed , holding the ecalbackends ....  ( a
> > classic case of domino effect)
> > 
> > kindly review the same.
> > 
> 
> > @@ -2581,7 +2581,7 @@ e_cal_get_objects_for_uid (ECal *ecal, c
> >                                         comp = e_cal_component_new ();
> >                                         e_cal_component_set_icalcomponent (comp, icalcomponent_new_clone (subcomp));
> >                                         *objects = g_list_append
> > (*objects, comp);
> > -
> > +                                       icalcomponent_free (subcomp);
> >                                         subcomp =
> > icalcomponent_get_next_component (icalcomp, kind_to_find);
> >                                 }
> this call to icalcomponent_free should not be needed, since we are
> freeing the full VCALENDAR after the while loop.
> 
> the rest looks ok