The previous fix introduce a critical regression - display :0 can not gain Console User profile.
I rework the patch under ConsoleKit code base. Please review it. -Halton On Thu, 2009-11-26 at 17:05 +0800, Halton Huo wrote: > Hi, > > To resolve bugster bug 6885612, we use workaround as following: > We set device value to /dev/console if display number is :0 and device > is like /dev/vt/# or " ? ". This will let user login in * on :0 gain > the permission of devices. > > Because the display :0 device is hardcoded as "/dev/console" rather than > the actual vt device "/dev/vt/#", the User Siwtch applet won't go back > to the display :0 session. This is how doo 12563 comes. > > With the attached patch, gdm will get the actual device again before > calling ConsoleKit D-Bus interface OpenSessionWithPamameters. Thus, > ConsoleKit database maintains correct VT information for Xorg session, > > For more information, please refer to > http://defect.opensolaris.org/bz/show_bug.cgi?id=12563 > > This patch committed in gnome-2-28 branch as well as trunk. > > Thanks, > Halton. -------------- next part -------------- A non-text attachment was scrubbed... Name: ConsoleKit-10-actual-vt-on-switch.diff Type: text/x-patch Size: 2189 bytes Desc: not available URL: <http://mail.opensolaris.org/pipermail/jds-review/attachments/20091130/9ae29c04/attachment.bin>
