15:02:12 <rdieter> #startmeeting kde-sig
15:02:12 <zodbot> Meeting started Tue Sep 30 15:02:12 2014 UTC.  The chair is rdieter. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:02:12 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:02:16 <rdieter> #meetingname kde-sig
15:02:16 <zodbot> The meeting name has been set to 'kde-sig'
15:02:18 <rdieter> #topic roll call
15:02:27 <Kevin_Kofler> Present.
15:02:31 <rdieter> hi all, friendly kde-sig meeting, who's present today?
15:02:33 <tosky> hi
15:03:13 <dvratil> hi
15:03:16 <than> present
15:03:25 * jreznik is here
15:03:44 <jgrulich> hi
15:04:23 <rdieter> #info rdieter Kevin_Kofler tosky dvratil than jreznik jgrulich present
15:04:32 <rdieter> #chair Kevin_Kofler tosky dvratil than jreznik jgrulich
15:04:32 <zodbot> Current chairs: Kevin_Kofler dvratil jgrulich jreznik rdieter than tosky
15:04:35 <rdieter> #topic agenda
15:04:44 <rdieter> okie dokie, what should we discuss today?
15:04:57 * rdieter can give update on kde4-EL copr
15:05:39 * Kevin_Kofler wishes he could give an update on PackageKit-hif comps support, but can't really. :-(
15:06:01 <Kevin_Kofler> All the free time so far spent setting up things so I can do stuff.
15:06:03 <pino|work> mo
15:06:05 <pino|work> *me
15:06:37 <Kevin_Kofler> One thing we can discuss is making kimpanel just work. We've been punting that from release to release.
15:07:00 <rdieter> ok, let's start
15:07:08 <rdieter> #topic kde4-EL7 copr
15:07:27 <rdieter> I've been working on building kde-4.14.1 for el7 in copr, https://copr.fedoraproject.org/coprs/rdieter/kde4
15:07:41 <rdieter> it's *mostly* done, looks like I have 1 broken dep to fix
15:07:55 <rdieter> anyone who wants to help is free to request acls
15:09:19 <rdieter> #info pino|work present
15:10:05 <rdieter> #topic ibus/kimpanel love
15:10:13 <rdieter> Kevin_Kofler: ^^ go ahead
15:11:19 <Kevin_Kofler> So the thing is, for input methods, we still rely on the GTK+ stuff.
15:11:48 <Kevin_Kofler> We don't ship input methods by default at all, but when users install them, ibus automagically fires up the GTK+ stuff for them and kimpanel gets ignored.
15:12:26 <Kevin_Kofler> In fact, I think that even if the user adds kimpanel to the panel manually, it won't work, because ibus is not starting the D-Bus service that kimpanel uses, but its own built-in GTK+ applet that doesn't use D-Bus.
15:12:47 <Kevin_Kofler> (IBus uses D-Bus at lower levels, but for the applet, they don't use D-Bus by default.)
15:13:11 <Kevin_Kofler> Sho_ can probably help us test things, but AIUI, he doesn't want to get involved with the packaging work.
15:13:26 * rdieter sees mention of fcitx in #kde-devel
15:13:27 <Kevin_Kofler> Still, we should try to finally get this working.
15:13:45 <Kevin_Kofler> fcitx is another input method framework that competes with IBus.
15:13:55 <Kevin_Kofler> It's less well supported in Fedora.
15:14:17 <Kevin_Kofler> Whether it can be made to work more easily with kimpanel than IBus, I don't know.
15:14:18 <rdieter> fellow in #kde-devel working on kimpanel is fcitx developer too, so it may actually work better *for us*
15:15:01 <Kevin_Kofler> There's one developer working on IBus stuff in kimpanel, but then again, IIRC that's also the guy who thinks editing the ibus script is an acceptable way to set up kimpanel support for ibus…
15:15:17 <Kevin_Kofler> (having the user edit a script owned by a package is a no-go)
15:15:57 * rdieter agress, ideally stuff should 'just work'
15:16:18 <rdieter> so, is there any specific steps/tasks to take to make this all work better?
15:17:35 <Kevin_Kofler> xuetianweng is the developer who has apparently created both the xuetianweng/kimpanel-ibus-1.5 and xuetianweng/kimpanel-frameworks branches.
15:19:27 <Kevin_Kofler> Specific task: If the user installs kimpanel and adds it to the panel, it should just work.
15:19:50 <Kevin_Kofler> The required ibus stuff should be there, and ibus should be starting the D-Bus service kimpanel needs and NOT be starting the GTK+ stuff.
15:20:24 <Kevin_Kofler> This probably means the scripts the ibus package ships need to be smarter at figuring out what to start under what conditions.
15:20:50 <Kevin_Kofler> As for "The required ibus stuff should be there", that means proper package Requires.
15:21:24 <Kevin_Kofler> Now if we want to support multiple kimpanel backends, that needs to be done with subpackages, and then we actually need to test each of the subpackages, they all ought to work out of the box.
15:21:26 <rdieter> from Sho_'s ongoing discussion in #kde-devel, looks like kimpanel+fcitx should 'just work', and does handle the dbus-activation automatically (ie kimpanel should switch into fcitx mode then too)
15:21:47 <Kevin_Kofler> Testing all the backends of course needs even more work.
15:21:52 <rdieter> <nod>
15:22:26 <Kevin_Kofler> So to sum up, the current situation is that we have one backend (ibus) that has crappy support from the KDE side, and one backend (fcitx) that has crappy support from the Fedora side, sigh…
15:22:59 <Kevin_Kofler> (To detail the "crappy support from the Fedora side" thing: e.g., all the comps groups are dragging in the ibus stuff, not the fcitx stuff.)
15:23:16 <rdieter> it's "bad(tm)" if both ibus and fcitx are installed?
15:23:37 * rdieter wonders
15:23:41 <Kevin_Kofler> I think it can make for an interesting mess if users have both installed…
15:24:02 <Kevin_Kofler> I think both will want to start up, own the input methods and show their own applet.
15:24:08 <rdieter> ok, not sure where/how best to track this, open another tracker bug ?
15:24:30 <Kevin_Kofler> I think we even had one, not sure what has become of it.
15:24:38 <Kevin_Kofler> We haven't dared touching that for a long time.
15:24:47 <rdieter> I recall and old one, but I think we closed it
15:24:49 * rdieter looks
15:25:25 <Kevin_Kofler> We may have "closed" it, but it wasn't solved in a really satisfactory way.
15:25:26 <rdieter> .bug 583545 ?
15:25:29 <zodbot> rdieter: Bug 583545 Enable building Kimpanel and make a subpackage for it - https://bugzilla.redhat.com/583545 ?
15:25:44 <rdieter> actually, we've already fulfilled that bug topic
15:26:21 <Kevin_Kofler> Yes, we enabled building it, but making it work out of the box is what we didn't get done. I think there was another bug for that, might have been closed INSUFFICIENT_DATA or something though.
15:26:24 <rdieter> 583716.bug
15:26:28 <rdieter> .bug 583716
15:26:30 <zodbot> rdieter: Bug 583716 Need additional runtime detecting in ibus.conf for Kimpanel - https://bugzilla.redhat.com/583716
15:26:34 <rdieter> there we go
15:27:21 <Kevin_Kofler> Right, the ibus maintainers dropped the ball there.
15:27:39 <Kevin_Kofler> And we also didn't care enough to prod them.
15:28:13 <Kevin_Kofler> Unless this actually was solved somehow without anybody realizing that it fixes that old bug.
15:28:57 <Kevin_Kofler> What I'm also not sure the attached patch actually does not do is disable the GTK+ applet.
15:29:19 <Kevin_Kofler> *What I'm also not sure the attached patch actually does is disable the GTK+ applet.
15:29:46 <Kevin_Kofler> (Does passing --panel= also have that effect? If not, it doesn't do it.)
15:30:16 * rdieter added that to our F22Target-kde tracker bug
15:31:15 <rdieter> in the meantime, anyone interested in testing fcitx ?
15:32:58 <rdieter> moving on...
15:33:01 <rdieter> #topic open discussion
15:33:04 <rdieter> anything else for today?
15:34:01 <tosky> I know it's a bit later, but I can confirm that 4.14.1 upgrade worked
15:34:04 <tosky> (F20)
15:35:11 <rdieter> good to hear
15:35:21 * ltinkl is back
15:39:47 * rdieter will close meeting in 1 minute
15:50:03 * rdieter lied, but thanks everyone!
15:50:06 <rdieter> #endmeeting