Re: Pilot synch issues



On Wed, 2007-08-29 at 09:41 +1000, Andrew Greig wrote:
> I was not too fresh at the time, but I did wake up to it for the last
> message, I realised I was root and so switched back to $ as you see
> below.  I wrestle with this problem for hours.  Frederic Crozat says
> he has no problem using libusb under Mandriva to synch his Palm.  I
> have had no success, and only limited success when reverting to the
> visor script method. 

I think you're trying too many things at once, without determining the
cause of the core problem. If using the visor module works, stick with
that. 

Once you get the visor module working 100% of the time, you can try
converting to using libusb, but jumping from "My device doesn't sync,
let me try libusb..." is going to just result in frustration and much
lost hair as you pull it out trying to figure it out. 

Crawl, stand, walk, run. We're still at the crawl stage. 

Here are a few ground rules to help:

	1. You can't have one daemon listening for connections on the
           port, and then try to communicate across that same port with
           another tool or application (like pilot-xfer or jpilot). It
           would be like trying to receive a fax, and use a modem on the
           same telephone line. 

	2. You *ABSOLUTELY MUST* hit HotSync on the Palm device before
           you launch any client application (like pilot-xfer, jpilot,
           etc.). Hitting the HotSync button makes the electrical
           connection from Palm to desktop that causes visor to load,
           udev to wake up, and other magical things to happen. You
           can't get light out of a lamp until you pull the chain to
           begin the flow of electrons. The Palm over USB is the same
           way. 

	3. Have you identified the proper set of udev rules to
           CONSISTENTLY create the pseudo device in /dev/ every single
           time you hit HotSync on your device? 

	4. The timing between pressing HotSync and the device wakeup (to
           the point where the device is usable at the application
           level) is *VERY* device, hardware, kernel and distribution
           specific. We can't predict how long it will take, and so it
           takes anywhere from 1-10 seconds (that I've seen). 

           This means you'll have to try waiting 1 second between
           pressing HotSync and launching your client app. If 1 second
           doesn't work, wait for the device to time out, then try
           waiting 2 seconds. Then try 3 seconds, and so on... until you
           finally get an interval which works for _your hardware_. 

           I have a Thinkpad T42p here that I use every day for
           development, and on this USB hardware (uhci_hcd), my T3 takes
           1 second to connect, but my Treo 680 takes 4 seconds. Same
           desktop hardware, different Palm devices... each has its own
           specific timing requirement. 

	5. Forget trying to get gnome-pilot (gpilotd) to talk to your
           Palm device and sync, until you know you have working
           hardware and a _consistent_ connection between Palm and
           desktop. The easiest and lowest-level way to do that, is to
           use any of the tools in the pilot-link package (which I
           maintain).

           Once you know your connection is consistent and solid, THEN
           you can begin configuring gnome-pilot to communicate using
           that same port and timing interval. 

I don't mean to sound condescending or preaching, but I've helped
literally thousands of people with their Palm device connection issues
over the last decade using these tools, so I have a passing familiarity
with debugging these kinds of issues.


-- 
David A. Desrosiers
desrod gnu-designs com
setuid gmail com
http://projects.plkr.org/
Skype...: 860-967-3820

Attachment: signature.asc
Description: This is a digitally signed message part



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]