John Levon wrote:
> On Sun, Feb 13, 2005 at 12:43:22PM +0000, Angus Leeming wrote:
>
>> Pretty dumm ;-) At a quick look, it looks to me like wa_ptr (work area
>> pointer?) is a hack. Also at a quick look, it means that QWorkArea is
>
> Right, it's a horrible hack because we don't have access to a work-area
> in the context we need it. Nor is it clear what we could do about it.
>
> john
What about something that maps between an X11 window ID and a QWorkArea?
F.ex below. Of course, that requires a way to find the X11 window ID
associated with a particular QWorkArea and I've no idea if that is
possible...
namespace {
typedef std::map<Window, QWorkArea *> work_area_map;
work_area_map registered_work_areas;
}
QWorkArea::QWorkArea() {
// Add entry to the map
registered_work_areas[x11_window_id] = this;
}
QWorkArea::~QWorkArea() {
// Remove 'this' from the work_area_map
}
bool lyxX11EventFilter(XEvent * xev)
{
work_area_map::const_iterator it =
registered_work_areas.find(xev->window);
if (it == registered_work_areas.end())
return;
QWorkArea * wa_ptr = it->second;
...
}
--
Angus
