=== smarter__ is now known as smarter [03:13] feed branches-next had 4 updates, showing the latest 3 [03:13] ::branches-next:: [plasma-framework-work] r32 Delete plasma-framework5.symbols... (by Jose Manuel Santamaria Lema) [03:13] ::branches-next:: [plasma-framework-work] r33 Update symbols files. (by Jose Manuel Santamaria Lema) [03:13] ::branches-next:: [plasma-framework-work] r34 Fix -dev depends.... (by Jose Manuel Santamaria Lema) === tsimpson_ is now known as tsimpson [06:44] good morning [07:22] !find krb5-config [07:23] Found: krb5-config [07:23] that's a helpful message [07:23] he means the package is krb5-config :P [07:24] !find gobbledegook [07:24] Package/file gobbledegook does not exist in saucy [07:24] ya don't say [07:24] ah yes [07:24] problem no1: saucy [07:24] !info krb5-config [07:24] krb5-config (source: kerberos-configs): Configuration files for Kerberos Version 5. In component main, is optional. Version 2.3 (saucy), package size 22 kB, installed size 79 kB [07:24] problem no2: my cache says krb5-config doesn't contain krb5-config [07:24] which is a bit funny [07:25] tsimpson, jussi: ubottu probably should be bumped to trusty [07:25] but ubottu usually is in trusty [07:25] !find phonon-backend-vlc [07:25] Found: phonon-backend-vlc, phonon-backend-vlc-dbg [07:25] I mean [07:25] !info phonon-backend-vlc [07:25] phonon-backend-vlc (source: phonon-backend-vlc): Phonon VLC backend. In component universe, is optional. Version 0.6.2-2 (saucy), package size 79 kB, installed size 321 kB [07:25] oooo [07:26] @config plugins.PackageInfo.defaultRelease [07:26] Global: trusty; #kubuntu-devel: saucy [07:26] ubottu is ahead! [07:26] Sorry, I don't know anything about is ahead! [07:26] for some reason it has a channel specific value here.. [07:27] perhaps we switched it to saucy before saucy was a thing because no one in here ever cares about stable? [07:31] yofel: uh, changelog generation from git, someone suggested that previously as well, totally forgot about it [07:31] certainly would be smartest IMO [07:32] @config plugins.PackageInfo.defaultRelease [07:32] Global: trusty; #kubuntu-devel: trusty [07:32] !info krb5-config [07:32] krb5-config (source: kerberos-configs): Configuration files for Kerberos Version 5. In component main, is optional. Version 2.3 (trusty), package size 22 kB, installed size 79 kB [07:33] better [07:33] it's still lying, though I think that's because of the package ^^ [07:35] well its source is apt-cache [07:37] if you want to search for files explicitly, use a path (like !find bin/krb5-config) [07:37] otherwise it'll try and match to a package name first, then search files [07:40] ah [07:40] weird package name then [07:40] * apachelogger shrugs it off [08:14] ::branches-next:: [kcmutils] r27 Install qt pri file to -dev package (by Harald Sitter) [08:14] ::branches-next:: [kactivities] r34 * Wildcard qml plugin installation in kactvities.install... (by Harald Sitter) === vinay is now known as Guest19291 === haudegen_ is now known as haudegen [08:45] good morning [08:46] yo [08:47] * apachelogger would like to point out that it makes -32 sense to make multiarch lib packages when one then packs usr/bin/ and etc/ into that package.... [08:58] apachelogger: where have we done that? [08:58] kservice [08:59] etc is now in a data, kbuildsycoca5 is still in the lib package though [08:59] also I get the feeling I am making more things red than green :O [08:59] * apachelogger been doing non-standard packaging for too long === Tm_Tr is now known as Guest17709 [09:04] Riddell: why is kf5-kio.install instead of kio.install? [09:05] kiconthemes has mulitarch issues too [09:05] E: libkf5iconthemes5: arch-dependent-file-not-in-arch-specific-directory usr/bin/kiconfinder5 [09:06] btw. am I understanding this right that we're putting all the l10n stuff into -data packages? [09:07] yofel: only when there is no other meaningful package [09:07] e.g. in kio I am packing them in with kf5-kio as the l10n is for the slaves and not the libraries themself [09:08] so you're just putting it into the runtime package instead of making an extra arch:all one? [09:09] yeah, there's no point in an arch all package there [09:09] yofel: I'm putting l10n into a -data if that's what multiarch needs it to be [09:09] apachelogger: hmm, no reason I guess, [09:09] ok [09:09] for the libs I'd also not have put it in data short of multiarch really [09:09] more packages just give you more points of failure for the most part [09:10] IMO this is still easy enough to keep multiarch-compatible [09:10] hence the -data packages :P [09:10] right [09:14] feed branches-next had 8 updates, showing the latest 5 [09:14] ::branches-next:: [kconfigwidgets] r32 fix locale install glob... (by Harald Sitter) [09:14] ::branches-next:: [kservice] r31 * Add new package libkf5service-data for localization and et... (by Harald Sitter) [09:14] ::branches-next:: [kxmlgui] r29 * Install qt mkspecs pri to libkf5xmlgui-dev... (by Harald Sitter) [09:14] ::branches-next:: [ktextwidgets] r28 Install qt mkpsecs pri files to libkf5textwidgets-dev (by Harald Sitter) [09:14] ::branches-next:: [kio] r37 * Update kf5-kio.install with new files... (by Harald Sitter) [09:23] so [09:23] I broke some dep, I am not sure how though [09:24] which one? [09:24] too much red :S [09:24] yofel: dep resolution doesn't work anymore, see kinit e.g. [09:25] might have been the kservice replaces breaks I guess [09:26] ah, I know, typo ftw [09:26] I really think there is too much copy'n'paste shit going on in these packages [09:26] like severely too much [09:27] 90% of the things I did on the packages so far was cnp in control [09:27] libkf5service5-data OR libkf5service-data? [09:28] soversion or not, that is the question [09:28] opinions plz ^^^ [09:29] I've looked into scripting control for neon, but all I found are really shitty dh_clean override hacks to generate it from control.in or running another target first [09:30] apachelogger: not sure what we did for the other packages, but IMO without, it's not like locales are co-installable [09:30] altough, then the dep has to be >= too so it doesn't break [09:31] yofel: I think it would already help a lot if we used substvars for the descriptions and so forth [09:31] talk to debian if they want that [09:31] yofel: the version not breaking is based on the assumption that libfoo1 has compatible assets with libfoo2 [09:31] which is a bogus assumption IMO [09:32] well, bogus true, but I don't think it's worth the conflicts [09:32] Riddell: ^ ? [09:32] what what? [09:33] I'm preferring no soversion in the -data so libkf5service-data [09:33] yofel: except a badly written library can actually have runtime kaputness from incompatible/missing assets [09:33] Riddell: yeah, but version >= or = source:version [09:34] apachelogger: that's then a badly written library, but not a packaging issue [09:34] sure it is a packaging issue [09:34] how [09:34] I've been using = source:version, it might not be backwards compatible for some reason [09:34] the library is entirely within its right to fail in any manner when assets are corrupted [09:34] and corrupted assets are not to tell apart from shitty packaging [09:34] Riddell: that's the argument [09:34] so [09:34] hm [09:35] IMHO we should do = [09:35] iff it then turns out one can push a >= [09:35] but I think assuming the least compatibility by default is the safer way to go [09:35] I just don't want to end up in the same situation where upstream does something stupid and we can't properly update (i.e. libkscreen) [09:36] but for now keep = then I guess === Guest17709 is now known as Tm_Tr [09:53] kf5html5-data [09:53] it seems we have a bit of a naming inconsistency [09:56] fixy fixy [09:59] * Mirv haz some Nexus 10 to bring to Berlin for Kubuntu development for someone called Aleix :) [09:59] Mirv: from jussi? [09:59] what's happening in Berlin? [09:59] Aleix Pol? [09:59] Riddell: yeah, he just ran through my lunch place giving it to me [09:59] Riddell: Qt Summit [09:59] ah cool [10:00] and Mr. Pol, yes [10:00] Mirv: yeah, that would apol on IRC :) [10:00] *would be [10:02] ok, I'll ping him at some point before next week [10:04] #lunchinvasion [10:05] it's only 12 :O [10:05] I didn't get invaded either [10:07] ::branches:: [kanagram] r120 Release... (by Rohan Garg) [10:07] ::branches:: [kapman] r73 * Merge with Debian, no remaining changes... (by Rohan Garg) [10:07] ::branches:: [kapman] r74 Release... (by Rohan Garg) [10:11] ktexteditor-data [10:11] data naming needs serious revising [10:13] ::branches-next:: [kinit] r24 * Install localization to kinit package... (by Harald Sitter) [10:13] ::branches-next:: [kpty] r28 -data is arch all (by Jonathan Riddell) [10:13] ::branches-next:: [kservice] r32 fix data package name... (by Harald Sitter) [10:13] ::branches-next:: [khtml] r29 * Update libkf5khtml5-data.install to contain localization... (by Harald Sitter) [10:13] ::branches-next:: [kross] r25 Update kross.install to contain localization (by Harald Sitter) [10:18] Riddell: please remind Scarlett to copy -data from another package and adjust description and name only [10:18] missing multi-arch fields [10:20] * apachelogger wonders why khtml's is multiarch same [10:20] really [10:20] we need a way to get rid of copynpasting that stuff [10:23] what's the point for multiarch for arch:all packages? [10:25] what's the point of multiarch I say [10:26] yofel: data must be marked foreign [10:26] that's completely independent of the actual binary architecture support [10:26] I still don't get why it has to be foreign if there's just one deb for all architectures [10:27] yofel: https://wiki.ubuntu.com/MultiarchSpec#Dependencies_involving_Architecture:_all_packages [10:28] o.O [10:28] well, ok [10:29] apachelogger: multi-arch should be forgeign on -data [10:29] that's what I said [10:33] Hiyas all === haudegen_ is now known as Haudegen [11:01] sections appear a bit wrong all over the place [11:01] Package: libkf5iconthemes5 [11:01] Section: libdevel [11:01] don't think that should be libdevel [11:04] you're right, it should not [11:05] although I long for the day when Section is removed from Debian policy, it's entirely pointless [11:05] Riddell: it's an optional field [11:05] even for source apparently [11:06] gosh [11:06] pft [11:06] policy doesn't even say what happens if you don't define it ^^ [11:06] Section: whocares ? [11:07] Riddell: just set libs on source and don't care otherwise [11:07] current packaging sets section incorrectly for too many packages [11:07] then lintian moans about not setting libdevel or debug [11:07] just realized that data also has libdevel [11:07] Riddell: override :P [11:07] ::branches:: [kapptemplate] r49 * Merge with debian, remaining changes... (by Rohan Garg) [11:07] ::branches:: [kapptemplate] r50 Release... (by Rohan Garg) [11:07] ::branches:: [qtwebkit-opensource-src] r73 Update arm64 and ppc64el symbols (by Timo Jyrinki) [11:08] all copy and paste errors I guess [11:08] yeah [11:09] and I keep saying that we need to do away with the copy and paste somehow :P [11:13] feed branches-next had 6 updates, showing the latest 5 [11:13] ::branches-next:: [khtml] r30 Change libkf5khtml5-data to multi-arch:foreign (by Harald Sitter) [11:13] ::branches-next:: [knotifyconfig] r25 * Fix data package name in control... (by Harald Sitter) [11:13] ::branches-next:: [kdelibs4support] r31 https://code.launchpad.net/~panfaust/kubuntu-packaging-next/... (by Jonathan Riddell) [11:13] ::branches-next:: [plasma-framework] r31 [ José Manuel Santamaría Lema ]... (by Jonathan Riddell) [11:13] ::branches-next:: [kiconthemes] r27 * Add libkf5iconthemes-data and add localization to it... (by Harald Sitter) [11:20] apachelogger: I wouldn't bother with breaks/replaces on packages now, it's just in a PPA that is experimental [11:21] public is public [11:21] good morning [11:22] hola sgclark [11:22] so I have noticed the 4.100 packages are breaking the 4.96 workspace stuff [11:23] yep, known defect [11:23] they're not binary compatible [11:24] so I'm leaving workspace until 4.100 kf5 is done [11:24] also not source compatible actually because I thin headers were moved around [11:24] likely yes [11:26] :O [11:26] what's with the splitting in frameworkintegration? [11:26] kf5-infopage.install [11:26] kf5-integrationplugin.install [11:26] kf5-platformtheme.install [11:26] libkf5style5.install [11:26] libkf5style-dev.install [11:26] splits gone wild [11:27] how would you do it? [11:27] first 3 are one package frameworkintegration [11:28] the first is basically data assets and the other two are plugins [11:29] okay [11:29] I'm off to lunch, I expect this all to be done by the time I get back :) [11:36] dbg is also somehow a complete mess [11:36] sometimes it uses sourcename, sometimes libnameso and sometimes libname [11:37] packaging should be... colorful \o/ [11:47] Riddell: you broke kservice again [11:49] Recommends: libkf5service-bin (= ${source:Version}) [11:49] I am actually not sure it is legit for a multiarch same package to recommend a non-multiarched package [11:49] particularly with arch:any [11:52] * apachelogger wonders what E: libkf5service5: symbols-file-contains-current-version-with-debian-revision on symbol _ZN12KConfigGroup10writeEntryIiEEvPKcRKT_6QFlagsIN11KConfigBase15WriteConfigFlagEE@Base and 3 others is about [11:52] oh, I didn't commit that change [11:58] !find NetworkManager.pc [11:58] File NetworkManager.pc found in network-manager-dev [12:08] ::branches:: [kopete] r44 Move debian/* to top level dir (by Pali) [12:08] filed https://bugs.kde.org/show_bug.cgi?id=335786 for kdelibs4support being rubbish and orange on status page [12:08] KDE bug 335786 in general "FindNetworkManager not found" [Normal,Unconfirmed] [12:10] the plasma packaging could be a bit bogus I think [12:13] bogus? [12:13] ::branches-next:: [kdnssd] r61 Add libkf5dnssd-data and add localization to it (by Harald Sitter) [12:13] ::branches-next:: [kdeclarative] r32 * Add -data package... (by Jonathan Riddell) [12:13] ::branches-next:: [kdnssd] r62 Remove incorrect and pointless libdevel section from libkf5d... (by Harald Sitter) [12:13] ::branches-next:: [kdnssd] r63 Make libkf5dnssd5 depend on it (by Harald Sitter) [12:13] ::branches-next:: [kiconthemes] r28 Make libkf5iconthemes5 depend on it (by Harald Sitter) [12:13] ::branches-next:: [kiconthemes] r29 add data file... (by Harald Sitter) [12:14] ::branches-next:: [knotifyconfig] r26 * Fix data install file to use correct path... (by Harald Sitter) [12:14] ::branches-next:: [kservice] r33 Add -bin package... (by Jonathan Riddell) [12:14] ::branches-next:: [frameworkintegration] r30 * Compound packaging:... (by Harald Sitter) [12:14] ::branches-next:: [kiconthemes] r30 fix locale install path... (by Harald Sitter) [12:14] ::branches-next:: [kservice] r34 * Update libkf5service5.symbols... (by Harald Sitter) [12:14] ::branches-next:: [kdelibs4support] r32 * Add libkf5kdelibs4support5-bin.lintian-overrides... (by Harald Sitter) [12:14] :O [12:14] kubotu: can you make that notices? [12:14] flood++ [12:14] :D [12:14] kubotu: config set rss.announce_method notice [12:14] aight [12:14] should be better [12:18] sgclark: something is very wrong with the source [12:18] can't put my finger on it [12:18] ahh gotcha [12:18] though I guess the fact that half the stuff is multiarch and the other half is not does contribute to the feeling [12:18] :S [12:20] I am not sure splitting libKF5PlasmaQuick from the qml assets is much use either [12:22] well, the warning should be gone in ppa4 anyway [12:45] Riddell: I guess we'll want a new font package to make frameworkintegration cmake shut up about it not finding the font? [12:45] apachelogger: a new one? what's wrong with the existing one? [12:45] too old I guess [12:46] oh no [12:46] hmm [12:46] apparently we are simply not bdeping it [12:46] * apachelogger rolls eyse and fixes neon [12:46] it's marked as runtime anyway so fine to ignore [12:46] Riddell: yeah, but it makes status orange [12:47] so best just install it I'd say [12:47] not if it's marked as ignore in cmake-override [12:47] * Riddell makes it so [12:47] ignores are evil [12:47] Riddell: not sure that will work for CMake Warning induced warning states btw [12:47] from what I have seen in the code the ignores only work on actual feature lists at the bottom [12:49] mm maybe, let's see [12:49] upstreams who depend on something which is higher up in the stack are evil [12:49] wasn't there a notion to not do runtime dep checks through cmake anyway? [12:51] Riddell: declarative still wip? [12:51] because it's still red :P [12:54] "kdeclarative jriddell ppa2 up" says pad, if it's red then needs to go back to WIP, let me do so, probably my mess up [12:54] This package contains kbuildsycoca5. [12:54] Riddell: please use generic descriptions [12:54] e.g. This Package contains runtime binaries. [12:56] apachelogger: I disagree if there's only 1 binary, makes sense to make it search-able by that binary name [12:56] Riddell: sure, then the package grows another binary and no one updates the description [12:56] very pointless [12:57] Pre-Depends: ${misc:Pre-Depends} [12:57] what's the point of those btw? [12:58] no idea [12:58] no idea either, don't remember them being there before [12:59] bzr blame is your friend :) [13:00] 1 jriddel | Pre-Depends: ${misc:Pre-Depends} [13:00] I'll argue that one of you came up with them :P [13:04] I'm innocent I say! [13:04] the log does not lie :P [13:06] lol [13:15] santa_: what did you think needed done to kdelibs4support for networkmanager? [13:16] Riddell: it's bugged see backlog [13:16] Riddell: going to create a a merge request soon, but I need to leave home for a little while first [13:16] see you soon [13:33] frameworks all done! other than the networkmanager stuff in kdelibs4support [13:33] go merge KDE4 stuff! [13:34] sorry, plasma release needs some love next [13:34] plasma release just sounds so wrong [13:34] nim makes a face whenever I say it [13:34] xD [14:15] I'm back, will prepare my partial work of kdelibs4support for merging [14:16] ah, I've a problem with my cunning plan [14:16] we're all working on trusty, but plasma needs qt 5.3, and there's only packages of that for utopic [14:24] quite neat that all news sites are showcasing Project Neon 5 ISO's :3 [14:25] yay [14:25] Qt 5.3 will be landing soonish [14:26] Qt 5.3 will be landing soonish [14:26] where? neon? ubuntu? debian sid? [14:26] youboontoo [14:27] * shadeslayer borrows a squid from apachelogger and throws it at kate [14:27] sweet, canonical stuff or based on lisandro's packages? [14:27] https://lists.ubuntu.com/archives/ubuntu-devel/2014-June/038350.html [14:28] makes me think Canonical stuff [14:28] ScottK: ^^ [14:28] santa_: Neon already has Qt5.3 [14:28] shadeslayer: got URLs for sites? [14:28] Riddell: http://www.linuxbsdos.com/2014/05/15/screenshots-of-kde-plasma-next-beta-1/ [14:29] shadeslayer: deserved a kubuntu wire blog I'd say [14:30] shadeslayer: where did neon get qt 5.3? and is it based on trusty? [14:30] oh from its own packages I presume [14:30] ^^ [14:30] git stable [14:30] right [14:30] so much shit to merge [14:33] * shadeslayer could totally do with some Dr. Who right now [14:34] kdelibs4suport done [14:37] in the notes.k.o stuff I changed the following: [14:37] replaced this http://community.kde.org/Frameworks/List [14:37] with this http://api.kde.org/frameworks-api/frameworks5-apidocs/ [14:38] also I have added a link to my b-d graph [14:38] I hope you guys are ok with this [14:38] groovy [14:40] yofel: Riddell did someone upload 4.13.1 to trusty? [14:40] not me [14:40] because 4.13.2 is going to be out soon [14:40] should I do that [14:40] yeah, should be done today [14:40] now that I have super powers [14:40] go ahead [14:40] ok, let me run the script [14:40] I think you'll have to merge kdelibs [14:41] bah [14:44] yofel: are you sure about that? [14:44] I see the meinproc fix in the PPA too [14:44] oh, nevermind then [14:48] Riddell, apachelogger & anyone else interested: if you have some time, there is something I would like to discuss with you before doing a mass-merge-proposals [14:49] santa_: looking at kdelibs4support, is that patch going upstream? [14:49] I will send it to reviewboard, yes [14:49] lovely [14:49] I think it's broken upstream [14:50] seems strange that everything else works in cmake but it does seem to sort the issue [14:50] santa_: what's your query on merge proposals? [14:51] see this couple of commits by apachelogger [14:51] http://bazaar.launchpad.net/~kubuntu-packagers/kubuntu-packaging-next/kiconthemes/revision/31 [14:51] http://bazaar.launchpad.net/~kubuntu-packagers/kubuntu-packaging-next/kiconthemes/revision/32 [14:51] they reminded me a couple of things which I would like to change massively in the kf5 packages [14:52] do go on.. [14:52] 1. in library packages, do not ship runtime or data files, it shouldn't contain anything but the library [14:53] it's not the first time I see things in library packages which in my opinion shouldn't be there [14:53] that's what's been done generally, if you see packages which have stuff in libraries that will interfere with multi-arch then do separate them [14:54] yofel: do we have a bug number for 4.13.1 [14:54] yes we do, see changelog [14:55] it doesn't really matter if it interferes with multiarch or not, in general it's a good idea to not ship anything but the library in the lib packages [14:55] cool [14:55] I could elaborate more later if you want, but let me explain the 2nd mass merge [14:56] libkf5iconthemes5 recommends libkf5iconthemes-bin [14:56] ↑ about this I think it could be done better [14:57] uploading [14:57] the thing is libfoo5 usually can't depend on libfoo-bin because libfoo-bin it's also linked against libfoo5, therefore we would have a circular dependency problem [14:58] circular deps only affect depends afaik [14:58] so one better way would be adding a dependency on libfoo-bin in libfoo5.symbols [14:59] like in the debian packages with kde-runtime and libkdecore5 [14:59] if lintian doesn't complain about a circular dep then it should be fine though [15:00] if you inspect the symbols files of libkdecore5 you will see this http://paste.kde.org/pyajgldnr [15:00] yofel: iirc it does, and, in general it's a bug [15:00] what's the problem with the recommends? [15:00] santa_: I know that it's there, and we used it for baloo which bit us later [15:00] there might be a legit reason why you wanted to have something that used libfoo installed but not libfoo-bin, as I found out with baloo [15:01] yofel: do we usually mark the ubuntu-sru bug as affecting all the stuff one uploads? [15:01] shadeslayer: no, just kde4libs [15:01] ack [15:02] I actually believe you would kill a few launchpad queries if you did that... [15:02] Riddell: the problem I see with recommends is that someone might have it disabled by default (think that you are trying to get these packages included in debian sid at some point) [15:03] about this [15:03] if you disable recommends then you should be prepared for things not working with full features [15:07] yofel: oh btw no pushing to branches right :3 [15:07] Riddell: the thing is: are you sure not having the -bin counterpart won't produce crashes or weird behaviours? [15:07] shadeslayer: I should have reverted all of that [15:08] yofel: right, I called the script with --nopush, but I wanted to confirm since it's the first time I'm SRU'ing KDE SC [15:08] ah yeah, --nopush has no effect with --sru [15:08] aha [15:08] awesoem [15:09] santa_: crashes would be a bug, weird behaviours maybe, but I don't think added it to .symbols files is any better than recommends [15:09] * shadeslayer is happy, no silly rejection emails [15:10] shadeslayer: I replied to their mail. [15:10] yep, saw it [15:14] Riddell: I do, that's what I would have done if I were still working for debian, and that's what it's done in the current kde4 packaging with libkdecore5 and kde-runtime. so I would suggest you to talk about this with current debian's people. but if you want an unbiased judgement about the matter I strongly recommend you to not mention my name at all [15:15] i.e. see what happened when you asked about allLibraries [15:16] santa_: yeah I'll discuss it, thanks [15:19] Riddell: ok, let me know about the outcome, if we go for this change, I'm ready to do it. for now I will stick to the 1st massive change (which is ok, isn't it?) and I will use Recommends instead of the hard dependency as apachelogger did in his commit [15:20] apachelogger: did you decide not to rename kf5-kio? [15:20] oh, another less important thing [15:20] santa_: yes I think your first point is all good [15:21] what about making a couple of metapackages: kde-frameworks-dev and kde-frameworks-dbg? [15:21] so all the headers and dbg stuff could be installed with just one package [15:21] should probably go into meta-kde then [15:21] apachelogger: btw. could you possibly add some flood protection to the bot that limits the output to like 10-15 messages? If kubuntu-initial-upload gets run on bzr it'll be busy for a while... [15:22] yofel: IIRC it does have flood protection [15:22]