plugin loading (was: kcmoduleinfo / Plugin linking problem)

Matthias Kretz kretz at
Tue Aug 14 15:06:31 BST 2007

On Monday 13 August 2007, David Faure wrote:
> On Monday 13 August 2007, Matthias Kretz wrote:
> > On Monday 13 August 2007, David Faure wrote:
> > > > While going over it I found one thing which I think we should change
> > > > in any case: the "const QStringList &args" parameter should be a
> > > > "const QVariantList &args" instead.
> > >
> > > I think this was discussed before, and the gain isn't big compared to
> > > the porting effort imho.
> >
> > Ok, I'll check how big the effort is. If it's too much effort should we
> > deprecate the QStringList functions and add QVariantList functions?
> This is about a pure virtual method. It's one or the other, no deprecation
> possible.

We could keep KGenericFactory as it is now and deprecate the whole class. Then 
all K_EXPORT_COMPONENT_FACTORY calls that use KGenericFactory will show the 
deprecation warning.

The new KPluginFactory OTOH will replace KLibFactory and provide a 
QVariantList and a QStringList create method where the latter calls the 
former with the QStringList as the first (and only) entry of the 
QVariantList. The QStringList create method will be marked as deprecated.

> (and it should remain pure virtual otherwise the risks of using the wrong
> signature in the reimplementation will be too high)

Actually am not sure which method you mean... you're not talking about the 
plugin ctors but the KLibFactory createObject method?

Matthias Kretz (Germany)                            <><
MatthiasKretz at, kretz at,
Matthias.Kretz at
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <>

More information about the kde-core-devel mailing list