GNOME Bugzilla – Bug 566535
gtk_widget_get_snapshot does not work if double buffered is disabled
Last modified: 2009-01-23 05:34:24 UTC
I have looked at the issue a bit, and I understand what is happening. get_snapshot redirects the expose event using gdk_window_redirect_to_drawable. The problem with this is, that the redirection will only work if gdk_window_begin_paint_*/gdk_window_end_paint is called, as it just blits the pixel data into the snapshot pixmap. Now begin_paint and end_paint will *only* be called if the widget is double buffered. I'll attach a patch rough (and untested) patch that should fix it right away. This is a problem for Sugar, which uses the function to grab screenshots of activities.
Created attachment 125748 [details] [review] patch
(In reply to comment #0) > I have looked at the issue a bit, and I understand what is happening. > get_snapshot redirects the expose event using gdk_window_redirect_to_drawable. > The problem with this is, that the redirection will only work if > gdk_window_begin_paint_*/gdk_window_end_paint is called, as it just blits the > pixel data into the snapshot pixmap. Now begin_paint and end_paint will *only* > be called if the widget is double buffered. > > I'll attach a patch rough (and untested) patch that should fix it right away. > > This is a problem for Sugar, which uses the function to grab screenshots of > activities. We have seen that at least embedded abiword and mozilla use their own double buffering, so get_snapshot is not working there.
Ping. Need to know if this approach is viable because if not, we would need to work on Mozilla and Abiword so they use gtk's double buffer instead of their own strategies. Thanks!
I don't think that is quite right, as it affects how things work with redirected windows that are used for other things than get_snapshot. On the client-side windows branch I have a different approach to get_snapshot() code that does direct exposing instead of using invalidate/process_updates (since in that branch we can't handle invalidation of non-visible window parts). I think that approach is fundamentally more right, and furthermore its easier to handle your problems in it. I'm attaching a patch that has a gdk fix and the new gtk_widget_get_snapshot with added support for automatically enabling double buffering. Can you verify if it works for your?
Created attachment 126424 [details] [review] Another approach
(In reply to comment #4) > > I'm attaching a patch that has a gdk fix and the new gtk_widget_get_snapshot > with added support for automatically enabling double buffering. Can you verify > if it works for your? Yup, works great with both abiword and mozilla. Thanks!
Is there any issue with that second patch? Or can it be applied?
It would be good if we could get this in as soon as possible. That way we can get applications maintainers (abiword ones example) to look into issues with their rendering code.
Looks fine to me.
* gtk/gtkwidget.c (gtk_widget_get_snapshot): Use a different approach to snapshotting that is in line with what is done in the client-side windows branch, and that works for widgets regardless if they are double-buffered or not. Patch by Alex Larsson. * gdk/gdkwindow.c (_gdk_window_calculate_full_clip_region): Clip to the parent.