GNOME Bugzilla – Bug 592604
Get rid of Bonobo
Last modified: 2011-03-20 22:53:00 UTC
...and the hopefully last clean-up ticket filed against gnome-pilot: Bonobo will be removed for GNOME 3. D-Bus (and other stuff) is available nowadays.
See also http://live.gnome.org/GnomePilotRemoveBonobo and the gnome-pilot dbus branch: http://git.gnome.org/browse/gnome-pilot/?h=dbus
Created attachment 156708 [details] [review] replace bonobo with dbus This is draft patch, I'm sending out to Matt for review. The whole log are on http://git.gnome.org/browse/gnome-pilot/log/?h=dbus&showmsg=1, start from 6a5e8b4a2ce608fe3e34b5c892cb3ec81a430324. 2010-02-26 gnome-pilot.idl -> gpilotd-daemon.xml.
Matt: Can this get a review please? GNOME3 is coming close...
Hi Halton, Great work on this patch. Apologies for the delay in reviewing. I've done some initial testing, and it seems that most things seem to be working. I would like to resolve a lot of compiler warnings. Try 'make CFLAGS="-g -Wall -Werror" ' and start squashing warnings. Some of these are serious. For example: I did find that if I cleared my configuration, causing the wizard to run, that gpilotd crashes reliably. This only happens if I compile with optimization, suggesting uninitialised memory. valgrind confirms there are problems. I'm not sure of the cause yet, but in nearby code it seems that "GNOME_Pilot_SysInfo" disagrees with the "struct SysInfo" from pilot-link, which shows up as an incompatible pointer warning when building. This looks like an old bug, so not sure why it would start causing problems now. Also, I noticed a make problem with genmarshal not getting run after gp-marshal.list changes. Probably makes most sense to continue this review and cleanup on the dbus branch.
[update] I've fixed the mis-usage of GNOME_Pilot_SysInfo, and this seems to fix my crasher. I've committed this change to the dbus branch. Quite a lot more testing is going to be needed, but we're on the right track.
Continuing to make progress. Added some "gtk_init" calls to gpilotd and other programs. Working through the main functionality methodically now. One question: can someone summarise the status of bonobo/panel? When should/must we migrate the gnome-pilot panel applet from bonobo to dbus? This hasn't been done yet.
(In reply to comment #6) > can someone summarise the status of bonobo/panel? When > should/must we migrate the gnome-pilot panel applet from bonobo to dbus? This > hasn't been done yet. Latest gnome-panel 2.31 can already be build without bonobo (there is a HAVE_BONOBO parameter in configure.ac). Also see http://live.gnome.org/GnomeGoals/AppletsDbusMigration
It's hard to keep up with the pace of change when you don't live and breathe all things gnome. That GnomeGoals page still leads with a banner saying "This is only a proposal at this time, and not approved for project-wide application. This should not be applied before being officially turned into a GNOME Goal!". Thanks for the info - appreciated.
Review of attachment 156708 [details] [review]: Many thanks for this patch, Halton. This is now committed, with some minor fixes applied. I've been testing for a little while, and it's looking pretty solid now. I've merged the dbus branch back to master.
Cool and many thanks for reviewing.
(In reply to comment #9) > I've merged the dbus branch back to master. Can this be closed then?
I think we can close this issue. gnome-pilot 2.32.0 removes the direct libbonobo dependency. I'll open a new issue to track the applet migration from bonobo to dbus, as per comment 7.
Opened #630725 to track applet migration.
Reopening, as this bug is being used to track the gnome-pilot bonobo dependency. http://live.gnome.org/GnomeGoals/AppletsDbusMigration
Applet now migrated to lipanel4. Fixed on master.