The theory in the description of this bug is that something is starting
X at the time, and that ConsoleKit is racing with it. As such, you will
see different results from the same ioctl from moment to moment.
It's not obviously that GDM is calling ConsoleKit too early - ConsoleKit
is started by D-Bus activation when GDM asks for it, no matter what
stage it asks for it, and I assert that ConsoleKit ought to be able to
back off for a while and cope even if X is starting at the time. I'll
construct a simpler test program and see what I got wrong in my patch.
The reason you're seeing odd sequences of messages is that there are
multiple threads running in parallel.
The theory in the description of this bug is that something is starting
X at the time, and that ConsoleKit is racing with it. As such, you will
see different results from the same ioctl from moment to moment.
It's not obviously that GDM is calling ConsoleKit too early - ConsoleKit
is started by D-Bus activation when GDM asks for it, no matter what
stage it asks for it, and I assert that ConsoleKit ought to be able to
back off for a while and cope even if X is starting at the time. I'll
construct a simpler test program and see what I got wrong in my patch.
The reason you're seeing odd sequences of messages is that there are
multiple threads running in parallel.