Gtk--, set_name()object types, and those warning messages
- From: Paul Barton-Davis <pbd Op Net>
- To: gtk-list redhat com
- Subject: Gtk--, set_name()object types, and those warning messages
- Date: Sat, 20 Mar 1999 11:18:23 -0500
Well, there seems to be a small bug in Gtk--.
In widget.gen_h, we have:
void Gtk_Widget_Class::style_set_c_callback(GtkWidget *o, GtkStyle *p1)
{
CppObjectType *obj=static_cast<CppObjectType *>(gtk_object_get_data(GTK_OBJECT(o),"cpp"));
g_return_if_fail(obj);
if (obj) {
// try to use signal with Gtk_Widget argument if available, else with GtkWidget
if(p1) {
Gtk_Style *warg = (Gtk_Style*)gtk_object_get_data(GTK_OBJECT(p1),"cpp");
if(warg)
obj->style_set(warg);
else
obj->style_set_c_impl(p1);
} else
obj->style_set_c_impl(p1);
} //else {
// sigs_bk->style_set_c(o,p1);
//}
return;
}
If you end up calling this from this:
string someWidgetName;
Gtk_Widget->set_name (someWidgetName);
then sometimes, for reasons I don't understand (*), the p1 argument is of
type GTK_TYPE_CHAR, and is therefore not an object. This means that
the lookup in gtk_object_get_data() will produce a warning, even
though the result of the call to set_name() will be successful,
presumably because we use the C level style set function.
Any comments anyone ?
--p
(*) although I'll hazard a guess. This seems to happen whenever an object
is having its style reset from the default widget style, which it
picked up c/o being packed into a container. I'm guessing that something
about the way the default widget style is defined by me in an rc file,
or used/set up in GTK+ is responsible. This only seems to happen
after I've packed an object into a container, and then I reset its
name.
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]