Re: gtk--: Really weird thing!!
- From: Owen Taylor <owt1 cornell edu>
- To: Tero Pulkkinen <terop students cc tut fi>
- Cc: gtk-list redhat com
- Subject: Re: gtk--: Really weird thing!!
- Date: 08 Mar 1998 13:49:19 -0500
Tero Pulkkinen <terop@students.cc.tut.fi> writes:
> > class ColorButton : public Gtk_Button {
> > public:
> > ColorButton(gchar * label):Gtk_Button(){
> > GdkWindow * win = GTK_WIDGET(gtkobject)->window;
>
> This window -variable was changed in 0.99.4 and has caused quite many
> sigsegv's. I think its not updated until the widget has been shown to
> screen anymore. (i.e. it needs to be in the screen before you can do
> this)
>
> I couldnt test the programs yet (no X11 here), but that looks a good
> ccandidate of what new problems people have with 0.99.4.
There has is no change I know of to widget->window in 0.99.4.
As before, widget->window will be NULL until the window is realized,
(which occurs before the widget is shown on the screens, and
can be forceably triggered earlier with gtk_widget_realize(),
at any point after the widget has been added into the widget
heirarchy)
What has changed in 0.99.4, is that it is no longer permissable
to create pixmaps without either specifying the window or
specifying the depth. But this will produce easily identifiable
warnings before any segfaults can occur.
I'd guess it is more likely that gtk-- is not handling reference
counting in a proper manner with the changes in 0.99.4.
Regards,
Owen
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]