Re: [gtk-vnc-devel] possible cure for that crash....

["Daniel P. Berrange" <berrange redhat com>]

On Fri, May 02, 2008 at 02:45:20PM -0700, Rob Stoddard wrote:
> I changed gvnc_shutdown like so:
> 
> void gvnc_shutdown(struct gvnc *gvnc)
> {
>         gvnc->has_error = 1;
>         GVNC_DEBUG("Waking up couroutine to shutdown gracefully\n");
>         g_io_wakeup(&gvnc->wait);
>         close(gvnc->fd);
>         gvnc->fd = -1;
> }
> 
> 
> This allows the vnc coroutine to still be able to use the socket until 
> it can get back to the top of the server_messages routine.  I don't know 
> much about coroutines, so I hope somebody out there can affirm that the 
> coroutine will always reach the bottom of server_message so that it can 
> kick out (because of the has_error variable being set) and gracefully 
> shutdown the VNC widget.  Perhaps it would be wise to let the coroutine 
> close the socket and set the FD to -1, and then change gvnc_shutdown()  
> to look more like this...
> 
> void gvnc_shutdown(struct gvnc *gvnc)
> {
>         gvnc->has_error = 1;
>         GVNC_DEBUG("Waking up couroutine to shutdown gracefully\n");
>         while(gvnc->fd  != -1)
>             g_io_wakeup(&gvnc->wait);
> }
> 
> 
> This allows the coroutine to reach a cleaned up state and close the 
> resources it relies upon.   Then the vncdisplay can do whatever cleaning 
> up it needs to do.

Can you check the latest HG head code - as Jonh says, we believe we 
already fixed this.

Dan.
-- 
|: Red Hat, Engineering, Boston   -o-   http://people.redhat.com/berrange/ :|
|: http://libvirt.org  -o-  http://virt-manager.org  -o-  http://ovirt.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: GnuPG: 7D3B9505  -o-  F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|