GNOME Bugzilla – Bug 318274
Gnome-pilot won't work when pilot-link is upgraded
Last modified: 2006-03-08 08:18:42 UTC
Version details: 2.0.13 Distribution/Version: SNV 22 The current version 2.0.13 of gnome-pilot work on pilot-link-0.12.0-pre4. However,the current gnome-pilot won't work when pilot-link is upgraded. So we should develop the next version of gnome-pilot.
It happens on Solaris.
jpr updated the gnome-pilot CVS in November (IIRC) to port to the pilot-link 0.12.0 API. There hasn't been a new gnome-pilot release as yet. I don't know when this is planned. As far as I can tell, as I reported elsewhere, there is still a need to remove the final 'pi_close(listen_sd)' call in gpilotd.c:pilot_connect().
However, I encountered a problem. After building gnome-pilot(trunk codes from community CVS) successfully based on pilot-link-0.12.0-pre4 or -pre5, a error below happened when starting up "/usr/libexec/gpilotd". "(gpilotd:12950): gpilotd-WARNING **: An error occured while getting the pilot's system data". My machine OS: Linux; Os version: Suse 9; pilot-link version: 0.12.0-pre4 or 0.12.0 pre5; gnome-pilot version: trunk codes from CVS. And would you like give us some information about Jpr's machine such as which version of OS, which version of pilot-link, which version of gnome-pilot?
I have no idea what build machine jpr uses. The gnome-pilot CVS is intended to compile against either 0.11.8 or 0.12.0 APIs, but the CVS HEAD still has a couple of issues. the bug "An error occured while getting the pilot's system data" is, as far as I can tell, caused by pilot-link code. The most recent fedora package claims to be a pilot-link 'pre4' release, but is actually based on a pilot-link-0.12.0-cvs20050830.tar.bz2 tarball. See a recent thread to pilot-link-devel list: http://lists.pilot-link.org/pipermail/pilot-link-devel/2006-January/001490.html You might find success by configuring the device timeout to be zero. PLEASE: if you find this works, please send a post to the pilot-link-devel mailing list, to help get traction getting the proposed fix committed to pilot-link before 0.12.0 gets released.
Sorry.It doesn't yet work by configuring the device timeout to be zero. In fact, I filed a bug about this on bugs.pilot-link.org. Below is the comments from its owner. "I'm pretty sure gnome-pilot hasn't been updated to work with the changes from -pre4 to -pre5 of pilot-link. Do a diff of the two pilot-link trees and see what changed, and then submit a patch to the gnome-pilot folks. Unfortunately, we don't have anything to do with their project, so we can't really solve the bug on this end."
I've opened a bug against the pilot-link timeout issue: #1585 on bugs.pilot-link.org, as it is pilot-link specific and reproducible. Did you try removing the final 'pi_close(listen_sd)' call in gpilotd.c:pilot_connect() ? I have a working patched version of gnome-pilot that works for me with the pilot-link CVS HEAD. I haven't had time to break the patch down into manageable chunks (most of it is to do with netsync support), so the above fix is just off the top of my head -- I may have fixed something else critical as well...
Jerry, I've updated the gnome-pilot CVS. It would be great if you could try compiling on Solaris against the ***CVS*** version of pilot-link and update this ticket to say how you get on. I basically don't completely trust any of the packaged pilot-link 0.12.0pre* releases. There are far too many broken versions floating around for my liking. Thanks, Matt
gnome-pilot worked well by setting the device timeout to zero with gnome-pilot CVS + pilot-link CVS. However, bug below won't be fixed. http://bugzilla.gnome.org/show_bug.cgi?id=322903
marking this bug as resolved.