Re: [Tracker] 0.6.4 ready for final testing

[jamie <jamiemcc blueyonder co uk>]

On Tue, 2007-12-11 at 03:45 +0100, Michael Biebl wrote:
> 2007/12/10, jamie <jamiemcc blueyonder co uk>:
> > On Mon, 2007-12-10 at 17:54 +0100, Michael Biebl wrote:
> > > 2007/12/10, jamie <jamiemcc blueyonder co uk>:
> > > > Latest svn now contains the intended release for 0.6.4
> > > >
> > > > there are a few minor things I need to add to TST and the applet which I
> > > > will do tomorrow but everything else is finished and ready for testing
> > > >
> > > > I intent to release by midnight monday
> > > >
> > > >
> > > > Things to test:
> > >
> > > I still encounter the problem, that if both tracker-applet *and*
> > > trackerd are started by gnome-session (via /etc/xdg/autostart),
> > > trackerd keeps stuck in the
> > > Initializing state (that's what tracker-status reports).
> > >
> > > According to jamie this is a race condition between tracker-applet
> > > trying to access trackerd (and causing a dbus autolauched trackerd
> > > instance) and trackerd started by gnome-session.
> > > A short wait time in tracker-applet is said to circumvent this race
> > > condition somehow.
> > > Unfortunately I couldn't find this place in tracker-applet sources and
> > > see if increasing this wait time helps.
> >
> > As I said I have a few things to do tonight on the applet and TST and
> > that is one of them. FOr now pls concentrate your testing on the daemon
>
> Haven't encountered any crashes, memory leaks etc so far.
> Some basic testing (search) did work too.
>
> > I will likely increase the sleep time of the applet for a few seconds
> > before calling tracker via dbus
>
> Looks like 2 secs is not enough for me/my computer. Had to bump it to
> 10secs to work around this issue reliably.

I have moved dbus connecting of trackerd as close as possible to start to help mitigate this

Im leaving it at 2 secs though!

In future the applet might try again every few secs to reconnect if
connection fails but that is too late for 0.6.4

jamie