From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2b) Gecko/20021029 Phoenix/0.4 Description of problem: Using serial cradle my Trio syncs fine, but using the USB cradle it does not work. JPilot works fine with the same device. I have tried gnome-pilot 0.1.65 as shipped with RH8, gnome-pilot 0.1.69 from Rawhide, pilot-link 0.11.3 from RH8 and pilot-link 0.11.5 from CVS (with gnome-pilot and j-pilot recompiled against new pilot-link) Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. set gnome-pilot to use /dev/ttyUSB1 and USB connection 2. synchronise Actual Results: Trio times out Expected Results: Trio synchronise Additional info: Older Handspring Visors work fine.
Does this work with 0.71 in rawhide?
Tried using recompiled pilot-link and gnome-pilot (0.11.5 and 0.1.71 respectively) and I get this; jpilot works as usual gpilotd-Message: gnome-pilot 0.1.71 starting... gpilotd-Message: compiled for pilot-link version 0.11.5 gpilotd-Message: compiled with [VFS] [USB] [IrDA] [Network] Gtk-WARNING **: Unable to locate image file in pixmap_path: "mcheck2.png" line 318 Gtk-WARNING **: Unable to locate image file in pixmap_path: "mcheck1.png" line 326 Gtk-WARNING **: Unable to locate image file in pixmap_path: "mcheck2.png" line 318 Gtk-WARNING **: Unable to locate image file in pixmap_path: "mcheck1.png" line 326 gpilotd-Message: Activating CORBA server gpilotd-Message: Watching Cradle (/dev/ttyS0) gpilotd-Message: Watching Cradle1 (/dev/ttyUSB1) gpilotd-Message: setting PILOTRATE=57600 Gtk-WARNING **: cannot open display:
This is the log in messages file. Apr 11 12:21:33 localhost gnome-name-server[16526]: starting Apr 11 12:21:33 localhost gnome-name-server[16526]: name server starting Apr 11 12:21:47 localhost kernel: hub.c: USB new device connect on bus1/1, assigned device number 12 Apr 11 12:21:47 localhost kernel: usbserial.c: Handspring Visor converter detected Apr 11 12:21:47 localhost kernel: visor.c: Handspring Visor: Number of ports: 2 Apr 11 12:21:47 localhost kernel: visor.c: Handspring Visor: port 1, is for Generic use and is bound to ttyUSB0 Apr 11 12:21:47 localhost kernel: visor.c: Handspring Visor: port 2, is for HotSync use and is bound to ttyUSB1 Apr 11 12:21:47 localhost kernel: usbserial.c: Handspring Visor converter now attached to ttyUSB0 (or usb/tts/0 for devfs) Apr 11 12:21:47 localhost kernel: usbserial.c: Handspring Visor converter now attached to ttyUSB1 (or usb/tts/1 for devfs) Apr 11 12:21:51 localhost /etc/hotplug/usb.agent: Setup visor for USB product 82d/100/100 Apr 11 12:21:52 localhost kernel: usb.c: USB disconnect on device 12 Apr 11 12:21:52 localhost kernel: usbserial.c: Handspring Visor converter now disconnected from ttyUSB0 Apr 11 12:21:52 localhost kernel: usbserial.c: Handspring Visor converter now disconnected from ttyUSB1
The problem here is a logic error in gpilotd/pilot-link. When you plug a TREO (not TRIO) into the USB port, it does not initiate a hot-plug connection, so no serial device is allocated. Try the following. Plug in a Treo and then cat /proc/bus/usb/devices. Now hit the sync button on the Treo and again cat /proc/bus/usb/devices. See that the serial device was connected AFTER hitting the sync button? What gpilotd/pilot-link does is imediately grab the serial line when it starts and, in the case of a Treo, finds nobody home and that's the end of it. What it needs to do is wait for the hot-plug event after the sync button is pressed and THEN grab the serial line. You can emulate this on the command line. First type "killall gpilotd", then type "gpilotd" BUT DON'T HIT ENTER. Now hit the sync button on the Treo and count to three. Hit ENTER and watch gpilotd sync just fine. You may need several tries to get the timing just right. (Too soon and there's no tty; too late and you miss the Treo's request for a sync event.) This should be an easy fix. Simply adjust the logic so that gpilotd polls for the serial line to come up instead of simply bailing as soon as it finds no connection.
Still does not work in RH9 - haven't tried Severn but suspect same case since it's not even in the changelog for gnome-pilot 2.0.10. Perhaps this should be filed with Ximian since they maintain the package.. tkeitt, can I quote from your explanation? I got my USB cradle running after searching high and low on gnome-pilot lists, missed your Bugzilla comment.. heh.
I'm sorry that this bug has gone for so long without activity. Have you been able to get this to work with more recent versions of the packages? Or are you still seeing this sort of behaviour? I've been seeing similar behaviour (with gnome-pilot-2.0.12 using a Tungsten E with a USB connector) to that described in comment #4 (and have used a similar workaround) I think the ideal way to fix this will be to integrate HAL support into gnome-pilot, so I'm marking this as a blocker for bug #133127
I no longer have the device with me, so I cannot test this further - should I close this for now?
Thanks for following this up. I'm going to resolve this bug as "deferred". Feel free to reopen it if necessary.