kcmodule and kcmoduleproxy

[Albert Astals Cid]

A Dijous 04 Desembre 2008, Albert Astals Cid va escriure:
> A Dijous 04 Desembre 2008, Matthias Kretz va escriure:
> > On Thursday 04 December 2008 20:05:59 Albert Astals Cid wrote:
> > > > IMO (not humble in this case) it is better to have a resizing dialog
> > > > than a long delay before having the dialog ready. Which is why I made
> > > > the proxy delay loading, and breaking it like this one might have
> > > > just deleted all the delayed loading logic...
> > > >
> > > > If dialogs are initially "ridicolous"ly small then the application
> > > > needs to be "fixed" to set a sensible initial size for the dialog...
> > >
> > > I disagree, i don't need to set any sensible small size, Qt does it for
> > > me, and quite good actually. You are bypassing it, and it's failing,
> > > and you say it's the developers fault?
> >
> > For your definitions of "bypassing" and "failing". I wouldn't call it
> > that way...
> >
> > It's a matter of preference:
> > - load all widgets from DSOs in a matter of ~10 seconds and have a non-
> > resizing dialog
> > vs.
> > - show the dialog immediately with at least the right size for the first
> > shown widget or a size hint given by the application and resize later
> > when needed
> >
> > If you think about it, you'll find there's no other way to achieve a non-
> > resizing dialog than to discard lazy loading of the KCMs.
> >
> > Do you really want to do that?
>
> Yes, *I* prefer taking 10 seconds to load than and absurd initial size (no
> it did not have the right size for the first kcm) and sizing jumpyness when
> other dialogs are loaded.

Not sure people understood what i wanted to mean with the "*I*", i just meant 
it was my opinion that is as worthy as others (no, it's less worthy than 
Celete's) and the change can be reverted if the majority agrees.

Albert

>
> BTW i think i discussed with someone the change but i can't find any mail
> that backups my memories :-/
>
> Albert