gcc -Woverloaded-virtual

[David Faure]

On Monday 26 November 2007, Olivier Goffart wrote:
> Le lundi 26 novembre 2007, David Faure a écrit :
> > On Sunday 25 November 2007, Aurélien Gâteau wrote:
> 
> > > A better fix IMO would be to get rid of (1), make (2) not inherit meta
> > > data and add a method like this:
> > >
> > >     void mergeMetaDataFromJob(KJob* job)
> > >
> > > But it means all code which relies on KIO::Job::addSubjob() merging meta
> > > data needs to be fixed.
> >
> > This is out of the question IMHO. Merging was =true by default, so it
> > should happen by default, because it's the most likely case. If it's not
> > done automatically people will forget to call mergeMetaData for sure.
> 
> 
> Maybe the opposite : something like
> void addJobWithoutMerge(KJob *job)
> 
> So we will use the compile error to help us to port. 
> This is also better API than using a meaningless bool.

Yes, addSubjobNoMetadataMerge is a good solution.
We'll also find out if that's really used anywhere... I don't see any use within kio, for instance...
Hmm... so maybe we should first try to compile KDE without that bool to find out if that's really used anywhere.
Copying the metadata into the subjob gives it the right window-id and user-timestamp for instance,
I don't know of a case where we don't want to copy the metadata to the subjob.

-- 
David Faure, faure at kde.org, sponsored by Trolltech to work on KDE,
Konqueror (http://www.konqueror.org), and KOffice (http://www.koffice.org).