[Qt-creator] feature request: have qt-creator set a parent for promoted classes on instantiation

[Steven Bakhtiari]

Thanks for the response!

I just found this out earlier when going through the code here:
http://qt.gitorious.org/qt/qt/blobs/4.7/src/gui/kernel/qlayout.cpp and had
it confirmed on the QtCentre forums.

I feel a bit daft actually.

Thanks again,
Steven

On 14 March 2010 22:40, Coda Highland <coda at bobandgeorge.com> wrote:

> When the widget is added to the layout, it is automatically reparented
> to the widget that owns the layout. There's no need for an explicit
> setParent call because it's handled implicitly.
>
> /s/ Adam
>
> On Sun, Mar 14, 2010 at 1:16 PM, Steven Bakhtiari <steven.bakh at gmail.com>
> wrote:
> > Hello!
> >
> > I was wondering if it would be possible to have the generated code from
> > qt-creator pass a parent object into the constructor of promoted classes,
> or
> > rather, use the setParent method on the object, after instantiation?
> >
> > Scenario:
> >
> > I have a QStackedWidget object and each "page", or "layer", of the stack
> has
> > been promoted to a custom class, derived from QWidget. When adding
> promoted
> > classes through qt-creator, you define the superclass.
> >
> > On compilation some code is generated that builds the UI, however, the
> > pages, or layers, of the QStackedWidget are not passed a parent object.
> The
> > code reads:
> >
> > QStackedWidget *sw = new QStackedWidget(mainWindowPointer);
> > StackedLayer *foo = new StackedLayer();
> >
> > instead of the preferable
> >
> > QStackedWidget *sw = new QStackedWidget(mainWindowPointer);
> > StackedLayer *foo = new StackedLayer(stackedWidgetPointer);
> >
> > I feel this may cause some complications, say for example, if the custom
> > class has a different constructor. For that reason, perhaps we should
> call
> > setParent after instantiation? For example:
> >
> > StackedLayer *foo = new StackedLayer();
> > foo->setParent(stackedWidgetPointer);
> >
> > The downside to not setting the parent means the developer needs to
> > explicitly delete the object (please correct me if I'm wrong here, I'm
> > pretty new to C++ & Qt) but they're not even explicitly instantiating the
> > object, it all takes place in generated code, so this is confusing. It
> means
> > an app will likely have memory leaks. Also, it means objects that ideally
> > should have a parent, don't.
> >
> > Perhaps there's a good reason that I'm unaware of for not doing this?
> >
> > Thanks for reading,
> >
> > Steven
> >
> > _______________________________________________
> > Qt-creator mailing list
> > Qt-creator at trolltech.com
> > http://lists.trolltech.com/mailman/listinfo/qt-creator
> >
> >
> _______________________________________________
> Qt-creator mailing list
> Qt-creator at trolltech.com
> http://lists.trolltech.com/mailman/listinfo/qt-creator
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.qt-project.org/pipermail/qt-creator-old/attachments/20100314/262b5875/attachment.html