On Mon, 22 Feb 2016 at 21:18:13 +0100, Samuel Thibault wrote:
> Simon McVittie, on Mon 22 Feb 2016 07:18:28 +0000, wrote:
> > Since dbus 1.10 (in stretch), we have mitigated this differently, by
> > running dbus-launch from 75dbus_dbus-launch as you suggest,
>
> ? No, it is still prepended to $STARTUP only.
My mistake; it's started early if you have the dbus-user-session package
(dbus started by systemd --user) but not if you're relying on dbus-launch.
Nobody has reported major issues with dbus-user-session so far, so I
think we could extend this model to the dbus-launch code path in stretch.
I still think this is too disruptive a change to be suitable for stable,
though.
S
