Re: Initial sync pickup

[Nigel Metheringham <Nigel Metheringham dev intechnology co uk>]

On Wed, 2004-05-26 at 22:59, Lars Weissflog wrote:
> as Nigel has the very same problem with a Tungsten C and Philip  with
> another type of Clie, I still think it is _not_ entirely hardware
> related; at least _not_ limited to my combination. 

I actually spent a while chasing this yesterday.  It appears to be a
full scale heisenbug - as soon as you look for the actual bug it appears
to vanish :-)

If I strace gpilotd (running under the control of gnome-pilot applet)
then sync pickup works 100% of the time.

Similarly if I get rid of the gnome-pilot applet and run gpilotd on the
command line (gives more debugging information that way), it again works
100% of the time.

My original thought was that gpilotd was failing to pick up the change
in /proc/bus/usb/devices but having looked at the code I don't see how
that can happen so it must be hanging in the very start of the sync
code.   I'll have a closer look at how I can get a bit more debugging
data out the system running as a whole without scaring the bug back into
hiding.

I may also add a one second sleep after the detection of the device and
before starting the sync - I suspect that will make things work,
although there should be a better answer.   Interestingly the first sync
of the day normally works, later ones don't (even if I kill/restart all
palm related stuff on the box).  

Has anyone got FC2 without major changes around the pilot stuff running
with a USB palmOS device which picks up every time?

	Nigel.