GNOME Bugzilla – Bug 577580
Local mouse cursor hidden on second entry to remote view
Last modified: 2010-03-18 13:05:42 UTC
From: https://bugs.launchpad.net/ubuntu/+source/vinagre/+bug/278489 "When you first connect to a VNC session with vinagre, you have both a 'local' mouse pointer, which moves smoothly in concert with your mouse movements (compare with xtightvnc where this local mouse pointer is a circle rather than an arrow), and the 'remote' mouse pointer for the remote computer, which lags behind the local pointer depending on the bandwidth of the connection. This is all as expected. But when you move the local mouse outside the window displaying the remote computer (eg move the mouse into Vinagre's bookmarks column or title bar) and then back into the window displaying the remote computer, the 'local' mouse pointer vanishes. This makes using vinagre quite frustrating because the 'lagging' mouse pointer gives poor feedback and you continuously overshoot menus, etc." Tested using SVN head.
I grabbed the source for vinagre 2.24.0 and it looks like the problem is with vnc_pointer_grab_cb() in vinagre-tab.c. When the mouse leaves the window, vnc_pointer_ungrab_cb() is called, but vnc_pointer_grab_cb() is not called when the mouse re-enters the window. I don't know anything about libgtk-vnc, so it's not clear to me if the problem is with the way vinagre sets up the callback or with libgtk-vnc not calling the "vnc-pointer-grab" callback when it should.
This bug does not seem to affect remote sessions to Macs. The cursor changes but appears to remain local.
Created attachment 148702 [details] [review] proposed patch Everytime the widget is left the gtk-vnc leave_event function ungrabs the pointer, one of the side-effects of this is that the local point gets unset. This patch ensure the local pointer is shown again if the widget is entered and local pointer is enabled
Sjoerd's patch (against gtk-vnc) fixes the bug for me, thanks. So I guess the bug should be repoted against gtk-vnc instead of vinagre.
Is there a way to vote for this patch to be applied? Or to reassign this bug to the gtk-vnc package? Or to change the status to confirmed?
I've applied this upstream http://git.gnome.org/browse/gtk-vnc/commit/?id=61651e0485ae3334f047f08d080ac62d4b848fae