Hi
Bisz István wrote:
>
> Could you please definitively switch in your development environment the
> release method to HBQT_RELEASE_WITH_DELETE:
>
> //HBQT_SET_RELEASE_METHOD( HBQT_RELEASE_WITH_DESTRUTOR ) //
> Exits cleanly
> HBQT_SET_RELEASE_METHOD( HBQT_RELEASE_WITH_DELETE ) //
> Exits cleanly
> //HBQT_SET_RELEASE_METHOD( HBQT_RELEASE_WITH_DELETE_LATER ) //
> Exits cleanly
>
In hbIDE or in hbQT as a whole ?
In hbIDE I have changed it as default.
Are you done with your experiments with HBQT_RELEASE_WITH_* and
have come to conclusion that HBQT_RELEASE_WITH_DELETE is the best ?
OR
We need more time to decide?
> In QUiLoader.cpp I changed temporarily the QT_QUILOADER_LOAD to:
>
> /*
> * QWidget * load ( QIODevice * device, QWidget * parentWidget = 0 )
> */
> HB_FUNC( QT_QUILOADER_LOAD )
> {
> // hb_retptr( ( QWidget* ) hbqt_par_QUiLoader( 1 )->load(
> hbqt_par_QIODevice( 2 ), hbqt_par_QWidget( 3 ) ) );
> hb_retptrGC( hbqt_gcAllocate_QWidget( ( QWidget* ) hbqt_par_QUiLoader(
> 1
> )->load( hbqt_par_QIODevice( 2 ), hbqt_par_QWidget( 3 ) ) ) );
> }
>
Done.
Will committ today.
> Maybe we should have here a general solution for the wrappers returning
> objects, some of them are already solved with inserting their pointers to
> the GC system evidence.
>
A good idea but I am at a loss how we can achieve it dynamically
and in the generator. The above change you asked is manually added in
QUiLoader.qth.
Regards
Pritpal Bedi
--
View this message in context:
http://old.nabble.com/%28no-subject%29-tp27183396p27183900.html
Sent from the Harbour - Dev mailing list archive at Nabble.com.
_______________________________________________
Harbour mailing list (attachment size limit: 40KB)
[email protected]
http://lists.harbour-project.org/mailman/listinfo/harbour
