Hi, On 2 September 2016 at 09:30, Giulio Camuffo <giuliocamu...@gmail.com> wrote: > 2016-09-02 10:27 GMT+02:00 Olivier Fourdan <ofour...@redhat.com>: >>> 2016-09-01 11:52 GMT+02:00 Olivier Fourdan <ofour...@redhat.com>: >>> >> I’m ok with that, but why not updating all three "grab_*" members? >>> > >>> > While I think the key itself would be OK, I suspect updating the grab_time >>> > on all states (ie like key release) might have additional side effects. >>> > >>> > So I took the most conservative approach which was to update the serial >>> > only, as this one is compared against the popup serial. >>> >>> I can't tell if it's right or not, but so maybe it would be good to >>> add a comment about it, unless someone (daniels?) can confirm it's >>> right. >> >> Well, this is needed when traversing submenus using the keyboard. >> >> Once you navigate to a submenu with the keyboard, the corresponding popup is >> created by the toolkit using the serial provided by the compositor, which >> will compare against the last known serial. If the grab serial is updated on >> key press only, the serial won't match and the submenu will be dismissed. > > Yes, i meant if it's right or not to leave the grab_time and grab_key > as they are.
Yes, leaving those as-is makes things work as intended. What's unpleasant about this patch isn't that though, but that clients which attempt to use the key _press_ to trigger something are now going to fail. I don't like that, but only triggering on release makes much much more sense (hence why GTK+ does it), and maintaining multiple possible serials is a lot more work. So, pushed with review, thanks! To ssh://git.freedesktop.org/git/wayland/weston be8a6d3..df84dbe push -> master Cheers, Daniel _______________________________________________ wayland-devel mailing list wayland-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/wayland-devel
