Re: gnome-pilot can't see my palm V (jpilot can)

[Matt Davey <mcdavey mrao cam ac uk>]

On Thu, 2006-05-25 at 10:14 -0400, matt price utoronto ca wrote:
> > try starting gpilotd with a
> > debugger:
> >         killall gpilotd
> >         gdb /usr/libexec/gpilotd
> >
> > When it crashes, type 'bt' to get a 'backtrace'.
> > If you can send the complete gpilotd output, from when it starts
> > to the backtrace, that would be great.  This output will also
> > confirm which version of pilot-link you're using.
> >
> 
> ok, I've pasted that in to the bottom of this em.

Thanks for that.  Another thing that would be handy would be to turn on
pilot-link debugging by setting the environment variables:
        PILOT_DEBUG="DEV SLP CMP PADP NET SOCK"
        PILOT_DEBUG_LEVEL="DEBUG"

So, try rerunning with all that stuff:
  env PILOT_DEBUG="DEV SLP CMP PADP NET SOCK" \
    PILOT_DEBUG_LEVEL="DEBUG" gdb /usr/libexec/gpilotd.debug


> > The other simple thing worth trying (with pilot-link 0.11.8) is
> > increasing the 'timeout' value configured in gnome-pilot. 
> 
> found that value after some searching in
> ~/.gnome2/gnome-pilot.d/gpilotd.  Changing it didn't seem to have much
> effect.

It's also available in the configuration applet, under the 'devices'
tab.

> > Would you be interesting in trying to compile from a tarball?
> > If so, I can give you some pointers to the most recent gnome-pilot
> > development source.
> 
> sure, I'd try that.  But I guess the better scenario would be to
> figure out where Dapper's going wrong & fix THAT too, since there are
> zillions of Ubuntu users out there...

true

> OK, here is the gpilotd output followed by the stacktrace.

I've had a look at the output.  My current theory is that gpilotd tries
to connect, fails for some unknown reason, and then tries again before
the palm has given up.  This causes gp to read half-consumed content
from the palm, and processing this unexpected garbage causes a crash in
pilot-link.  So, the crash we're seeing may not be the start of our
troubles.  I'm hopeful that the extra pilot-link output might show an
earlier problem.

One thing worth trying: if you are running the gnome-pilot panel applet
(that shows the enigmatic progress bar on sync, and the 'sync' icon that
changes colour) you can right click on the applet to 'pause' the daemon.
Try 'pausing' the daemon and then unpausing just after starting a sync
on the palm.  This might help if the problem is caused by kernel/udev
slowness in producing the ttyUSB device.

Matt

Matt Davey        Why do they only eat one egg in France?
mcdavey mrao cam ac uk 	Because in France, one egg is an oeuf.