kde-sig
LOGS
15:08:43 <rdieter> #startmeeting kde-sig
15:08:43 <zodbot> Meeting started Tue Sep 29 15:08:43 2015 UTC.  The chair is rdieter. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:08:43 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:08:47 <rdieter> #meetingname kde-sig
15:08:47 <zodbot> The meeting name has been set to 'kde-sig'
15:08:50 <rdieter> #topic roll call
15:08:59 <rdieter> hi all, friendly kde-sig meeting, who's present today?
15:09:10 <jgrulich> hello o/
15:09:10 <danofsatx> oh, so it's now
15:09:15 <danofsatx> .hello dmossor
15:09:16 <zodbot> danofsatx: dmossor 'Dan Mossor' <danofsatx@gmail.com>
15:09:26 <tosky> hi
15:09:29 * danofsatx checks his coffee level
15:10:53 <rdieter> #info rdieter jgrulich danofsatx tosky present
15:11:54 <than> present
15:11:59 <rdieter> #info than present
15:12:05 <rdieter> #chair jgrulich danofsatx tosky than
15:12:05 <zodbot> Current chairs: danofsatx jgrulich rdieter than tosky
15:12:09 <rdieter> #topic agenda
15:12:17 <rdieter> anything to discuss today?
15:12:37 <rdieter> danofsatx brought up topic of apper vs muon in #fedora-kde earlier, if we have nothing better to do
15:13:31 <danofsatx> yes, that I did.
15:14:07 <than> what benefit do we have in muon?
15:14:20 <danofsatx> Currently, when a user goes to the "Software Management" tool in Plasma, it defaults to Apper.
15:14:22 <than> i never used moun before
15:14:42 <danofsatx> Clicking on any of the groups in Apper will immediately crash the application.
15:14:43 <rdieter> apper cannot list software to install, due to long-standing PK backend deficiency
15:15:02 <danofsatx> Muon supports appstream, which is where the group metadata lives.
15:15:04 <rdieter> danofsatx: it doesn't (or shouldn't!) crash, but it will give an error and not show anything
15:15:17 <rdieter> #topic apper vs muon
15:15:37 <danofsatx> At one point, muon was working, but the current version of it does not.
15:15:38 <rdieter> apper can only install new software using the search box
15:16:10 <rdieter> muon worked last I tried it in f22, but doesn't seem to on my f23 box yet.  haven't had a chance to poke ximion about it yet
15:16:23 <danofsatx> oh, sorry, the pop up titled "An internal system error has occurred - Apper ?" would appear to a common user as a crash.
15:16:46 <rdieter> It's an error from the PK backend
15:17:47 <danofsatx> You and I understand that, but when I put Plasma in front of my wife or children and they go looking for an application, I'm going to get frantic calls saying it doesn't work.
15:17:55 <rdieter> .bug 1181859
15:17:57 <zodbot> rdieter: Bug 1181859 apper: missing/needed PackageKit-hif (hawkey) backend features - https://bugzilla.redhat.com/1181859
15:18:13 <rdieter> .bug 1177137
15:18:14 <zodbot> rdieter: Bug 1177137 Apper Err:SearchGroups not supported by backend - https://bugzilla.redhat.com/1177137
15:18:14 <rdieter> and ^^
15:18:21 <rdieter> that's the error that should appear
15:18:34 <danofsatx> yes, after you hit "details"
15:19:04 <rdieter> .bug 1098735
15:19:05 <zodbot> rdieter: Bug 1098735 apper: PackageKit-hif (hawkey) backend missing comps group support - https://bugzilla.redhat.com/1098735
15:19:06 <rdieter> and related ^^
15:19:25 <rdieter> looks like the first and last one should be consolidated
15:19:41 <rdieter> either way, it sucks, yes
15:20:17 <rdieter> the only gui that works currently is yumex, and we all know how Kevin will like that idea
15:20:26 * danofsatx is opening new bug for F23 muon
15:21:31 <rdieter> so I don't think there's anything constructive to do atm, besides pushing to get muon better
15:22:07 <rdieter> does any kde-sig member have contacts with ximion?  I've been having trouble finding him on irc lately
15:22:08 <danofsatx> As I said in the QA meeting yesterday, apper is dead for all intents and purposes. But Muon is currently broken.
15:22:21 <rdieter> danofsatx: accurate statement +1
15:24:01 * jreznik_ is around, sorry for being late
15:25:37 <adamw> so the result's going to be stick with apper, but it's not going to work with groups?
15:27:24 <rdieter> adamw: correct, the situation with PK backend hasn't improved since f21 when it was first introduced unfortunately
15:27:28 <than> adamw:  Muon is currently broken. we should stick with apper
15:27:31 <adamw> alrighty
15:27:32 <tosky> rdieter: isn't ximion connected right now on #fedora-kde ?
15:27:48 <rdieter> tosky: if he is, I'll poke after meeting
15:27:49 <rdieter> thx
15:28:12 <danofsatx> Final Release Criteria: http://fedoraproject.org/wiki/Fedora_23_Final_Release_Criteria#Default_application_functionality
15:28:25 <danofsatx> I'd argue that neither option meets that criteria.
15:29:04 <rdieter> danofsatx: yeah, we'd pushed for that back in f21, but folks eventually caved anyway :(
15:29:48 <rdieter> (I think the idea was that it *could* be fixed post-release, except it never was)
15:30:55 * rdieter poked ximion, filed issue https://github.com/ximion/appstream/issues/6
15:31:04 <rdieter> ximion: hi
15:31:54 <ximion> hi!
15:31:54 <rdieter> ximion: in short, we'd love to ship muon with fedora 23, but we're stuck on https://github.com/ximion/appstream/issues/6
15:32:10 <rdieter> any help to figure out what's going wrong there would be greatly appreciated
15:32:56 <ximion> okay, do you have a link to your appstream-data RPM package?
15:33:50 <rdieter> ximion: https://koji.fedoraproject.org/koji/buildinfo?buildID=678448 is the one I'm currently testing
15:33:57 <ximion> "some metadata was ignored due to errors" is likely the problem, but it should tell you the reason in the verbose output
15:34:00 <rdieter> I can mush it into a tarball or something if that's easier for you
15:34:14 <danofsatx> .bug 1267312
15:34:16 <zodbot> danofsatx: Bug 1267312 muon-iscover not showing any packages - https://bugzilla.redhat.com/1267312
15:34:35 <ximion> I'll just extract it from the rpm
15:34:40 <rdieter> k thanks!
15:34:49 <ximion> rdieter: did you talk to apol about the state of Muon aleady?
15:34:55 <rdieter> no
15:34:57 * danofsatx fixed the typo in the title
15:35:11 <rdieter> ximion: should I?
15:35:21 <ximion> yes
15:35:25 <rdieter> k
15:35:46 <rdieter> ximion: why?  are there big changes coming or something?
15:35:59 <ximion> rdieter: since Muon Disover is currently not working great with PackageKit, and the "Muon" part is unmaintained, our plan is to split Discover out of Muon as KDE Discover
15:36:02 <rdieter> I vaguely recall reading a blog about UI design
15:36:13 <ximion> to make maintaining it easier
15:36:31 <rdieter> there's only muon-discover and muon-update, what else is there?
15:36:37 <ximion> Apper would then become the "package-manager" while muon would be the Application center
15:36:41 <rdieter> or is that the point, split them ?
15:37:00 <rdieter> ah, sounds good to me
15:37:11 <danofsatx> this sounds like an F24 goal
15:37:18 <ximion> yes - you don't build the full thing because you're not Ubuntu ;-) (look at the toplevel CMakeLists.txt file ^^)
15:37:28 <ximion> jup, probably
15:37:37 <rdieter> anyting distro specific doesn't exist , yes :)
15:37:58 <ximion> but it also means that PackageKit support in Muon Discover currently is not that awesome - we have a few problems with it on Tanglu
15:38:20 <rdieter> <nod>, can't be any worse than (kde4) apper right now
15:38:22 <ximion> also, Discover doesn't implement some AppStream features, e.g. last time I checked, it wasn't using AppStream to fecth screenshots
15:38:30 <ximion> jup
15:38:41 * ximion needs help to finish the KF5 port of Apper
15:39:03 <danofsatx> but the question is, can the PackageKit support be fixed?
15:39:07 <rdieter> unfortunately, that wouldn't help *us*, our PackageKit backend still doesn't support groups
15:39:09 * pino|work is late, sorry :/
15:39:17 <rdieter> #info pino|work present
15:39:19 <rdieter> pino|work: hi!
15:39:43 <ximion> rdieter: that's probably not too relevant in the AppStream world
15:39:56 <rdieter> danofsatx: it's between heliocastro/Kevin_Kofler and hughsie to work out.  As it is, there's been a *lot back-n-forth without much action
15:40:11 <danofsatx> I meant in muon-discover
15:40:12 <rdieter> ximion: it is if we want to use apper as *package* manager
15:40:24 <ximion> the current main issue from an architecture point-of-view is that Discover fetches packages with PackageKit, and then resolves them to AppStream components - while it should actually be the other way round
15:40:38 <danofsatx> I understand/accept that PK will never come to apper
15:41:00 <ximion> Apper is *the* PK frontend for KDE
15:41:06 <ximion> you probably mean AppStream ;-)
15:41:33 <rdieter> ximion: I'm confused
15:41:38 <rdieter> you said, [10:36] <ximion> Apper would then become the "package-manager" while muon would be the Application center
15:41:53 <danofsatx> yes, I did mean appstream in regards to apper, but my original question was whether PK support could be fixed in muon-discover.
15:42:22 <rdieter> (though I could be confusing myself, since there's multiple conversations going on)
15:42:56 <rdieter> danofsatx: *appstream* support could be fixed in muon-discover
15:43:00 <rdieter> PK there works fine
15:43:18 <ximion> anything can be fixed, of course ;-)
15:43:38 <danofsatx> [9/29/15 10:35] <ximion> rdieter: since Muon Disover is currently not working great with PackageKit,
15:43:48 <rdieter> so anyway, we'll (seriously) consider shipping muon-discover if it works by final freeze
15:44:01 <ximion> rdieter: Discover would provide the "application-centric" view on the package repositories, while Apper would provide a package-centric view for technical users who want to use a GUI to browse their packages
15:44:09 <rdieter> danofsatx: I think he's only talking about the internal architecture there
15:44:13 <danofsatx> ok
15:44:32 <rdieter> ximion: <nod>, except apper doesn't work *for us* due to missing PK backend feature(s)
15:44:46 <rdieter> which is a separate issue
15:44:59 <danofsatx> but, as it stands, neither application can pass a "Basic expected functionality" test. One, or both, need fixed and/or documented.
15:45:03 <ximion> huh? So you used it in broken form the past years?
15:45:18 <danofsatx> ximion: apper has been broken since F21
15:45:32 <ximion> wow
15:45:44 <rdieter> ximion: fedora's PK backend doesn't support groups
15:46:01 <rdieter> so apper only works via searching for stuff, browsing is a fail
15:46:16 * rdieter shakes fist menacingly
15:46:22 <danofsatx> unless you browse the full package list.
15:46:44 <rdieter> ximion: but that's not your problem
15:47:05 <apol> hi
15:47:24 <ximion> patches would of course be welcome :-P
15:47:32 <ximion> rdieter> I asked apol to join
15:47:44 <rdieter> apol: hola, fedora is seriously considering shipping muon discover (by default), and ximion said we should talk to you first
15:47:47 <ximion> he knows about the development plans for Muon much better than I do
15:48:12 <rdieter> for fedora 23 kde spin, in particular
15:48:35 <apol> sure
15:48:37 <ximion> &me can|t reproduce https://github.com/ximion/appstream/issues/6
15:48:42 <apol> well, I'm happy you're considering it
15:48:59 <rdieter> ximion: fwiw, we've submitted a patch, but hughsie has been rejecting it for *reasons* (some pedantic silliness)
15:49:17 <rdieter> ximion: oh fun
15:49:19 <apol> I'm not sure how can I help you take a conclusion, but if you have any question or concern, please tell me
15:50:25 <rdieter> apol: ok, I don't have any questions yet, I generally like muon discover
15:50:34 <apol> hurray!
15:50:48 <apol> I'm working on a redesign, as proposed by the KDE VDG
15:50:59 <apol> but it's only graphical
15:51:03 <rdieter> ximion: mind helping me debug more after the meeting? (or some other time?)
15:51:52 <ximion> rdieter: yes - I also have a few more things to try before giving up ;-P
15:52:37 <rdieter> #topic open discussion
15:52:40 <ximion> apol: do we have a specific time for splitting KDE Discover out of Muon?
15:52:42 <rdieter> anything else to discuss today?
15:53:09 <apol> ximion: somewhere betweem the last release of Plasma 5.4 and before 5.5.0
15:54:25 <rdieter> ok, with that pending big change, maybe shipping muon discover now may not be wise
15:54:36 <apol> why not?
15:54:53 <apol> rdieter: the reason why we're splitting them is because the codebases aren't related at all
15:54:56 <rdieter> apol: because muon discover future is ending soon, it'll be changed to kde discover
15:55:31 <apol> I see it as a small detail, but sure, it makes sense I guess
15:55:32 <rdieter> and shipping any *bugfix* updates with such a big change could be bad for *reasons*
15:55:59 <ximion> apol: we should also resolve AppStream components to PK packages - doing it the other way round is hurting performance a lot, at least with the aptcc backend
15:56:05 <rdieter> apol: in general, we avoid shipping any .desktop/application name changes, that can affect users' shortcuts
15:56:28 <apol> rdieter: ok
15:56:34 <rdieter> (during a release anyway, between disto releases, that is ok of course)
15:56:50 <danofsatx> can we rename it now, knowing it's coming?
15:57:11 <danofsatx> or is that a bad idea?
15:57:25 <apol> ximion: send me an email about that, I cannot really think about it now
15:57:31 <rdieter> I guess we could theoretically rename it's .desktop file
15:57:49 <rdieter> apol: assuming you have a good idea what 'kde discover' .desktop file will be named
15:58:15 <ximion> apol: should't we do the split now? - I must admit, I didn't really get the reason for waiting for the last Plasma 5.4 point release before splitting it out
15:58:15 <rdieter> something like org.kde.discover.desktop ?
15:58:27 * danofsatx wonders if it should be plasma-discover
15:58:32 <apol> rdieter: it's already called org.kde.discover in master
15:58:46 <rdieter> apol: excellent
15:58:53 <ximion> apol: sure, I'll file a bug report about this as work item - unless there is a Trello board already
15:59:11 <apol> ximion: because when we split we won't be able to ship bugfix patches, because distros will be expecting 1 tarball instead of 2
15:59:21 <apol> ximion: there's no trello
16:00:09 <ximion> apol; we could still split it and leave the branches at Muon intact until after the release for tarball-builds
16:01:04 <apol> ximion: well, you see backporting fixes would be hard and weird anyway
16:01:13 * ximion thinks kde-discover or plasma-discover is a better technical name too, since discover is really a bit too generic
16:01:38 <ximion> well, not if people are aware of the split
16:02:12 <tosky> diskover
16:02:13 * tosky hides
16:02:23 <danofsatx> heh
16:02:30 <apol> ximion: or we can just wait, there's nobody really benefiting from hurrying it
16:02:34 <apol> AFAIU
16:02:57 <ximion> yes, true
16:03:21 <rdieter> boo, kde repos don't have v5.4.1 tags yet
16:03:49 <ximion> rdieter: can zou point me to the patch hughsie rejected?
16:04:04 * ximion thinks he knows that thing already, but...
16:05:32 <rdieter> ximion: https://github.com/hughsie/PackageKit/pull/49
16:06:24 <rdieter> admittedly the folks on our end haven't exactly been great at being responsive either
16:06:58 <ximion> ah, well, that doesn't look like it's rejected - fix the patch, and everything should be good...
16:07:05 <rdieter> anyway, any final comments?   our hour is about up
16:07:26 <rdieter> ximion: <nod> our contributors working on that got a bit frustrated
16:08:08 <rdieter> ximion: first attempt, https://github.com/hughsie/PackageKit/pull/20
16:09:02 <rdieter> thanks everyone!
16:09:04 <rdieter> #endmeeting