[04:49] http://is11.snstatic.fi/img/978/1288646664471.jpg === Jacky is now known as jacky [07:29] hrm, so just read on planet kde that there will be a ported version of krunner going forward, until the replacement is ready. I dont remember what we decided, but we should stick with that as per upstream IMHO [07:30] "The future of KRunner, Plasma's mini-commandline was discussed. There is a replacement in the works, though at this early stage is it unclear when it is going to land. In order to allow an alternative to fully mature before it replaces a core component, the team decided to port the current version of KRunner, and adapt it for improvements in the desktop search area." [08:28] I was hoping for something like this [08:28] But I'm curious to what they'd replace it to be [08:28] and if all of the prior extensions would be compatable. [08:29] My hopes was mainly for the searching aspect of it to be sent to a daemon of sorts and it having a QML frontend that be used/invoked anywhere [08:29] * jacky goes to the planet [08:29] gvim [08:29] Riddell: ubiquity review is in feburary [08:30] when a card is overdue it will have a red bell :) === schmidtm_ is now known as schmidtm [09:17] morning [09:21] good morning shadeslayer [09:27] Riddell: I'm unsure what you were asking me? [09:33] InstallationMedia: Fantasy32 13.10 - Release i386 [09:33] we should just close all bugs that were not installed from an official flavor -.- [09:33] *nod* [09:35] shadeslayer: oh, do we have a PPA for qt 4.8.5 yet? [09:36] uploaded to staging [09:36] ok [09:36] I haven't followed afterwards [09:37] shadeslayer: going to look at it when I am done with mails [09:59] Riddell: ping [09:59] :'( [10:00] yofel: ping [10:00] everyone is hiding from me :'< [10:00] apachelogger: awwww [10:00] * apachelogger hugs jussi [10:00] apachelogger: I began work on the activity thing :) [10:01] weeeh [10:01] when was the deadline again? [10:01] shadeslayer: opinions on where to put the new qt? [10:01] jussi: feb 20 [10:01] apachelogger: beta backports [10:01] once it passes QA we can move to backports [10:01] apachelogger: ok, excellent. [10:01] jussi: well, before feb 20, because we have lots of quality control stuff around feb 20 [10:01] ok [10:01] apachelogger: maybe even do experimental -> backports [10:02] shadeslayer: yeah, I was thinking experimental for preliminary testing and then -> backports [10:02] apachelogger: if it isnt there already, can you add what I need to submit (which files etc) to the trello card? [10:02] apachelogger: yep, either one is fine IMHO [10:03] jussi: a snapshot with possible names of widgets and snippets of text you used [10:03] jussi: I need to manually write a script that produces such an activity then [10:03] oh. last time you wanted some config file or something... [10:03] jussi: oh and did you poke upstream about this possibly? [10:03] apachelogger: I wrote to aaron today about it, yes [10:04] jussi: if he doesn't respond within the week please send one to plasma-devel@kde.org [10:04] broder audience anyway :) [10:04] ok [10:04] although you may do that anyway, because that's what aaron will tell you [10:05] not quite sure why we concluded that a mail should be sent to him directly, I always get annoyed when people do that to me :P [10:05] shadeslayer: throwing at experimental [10:05] apachelogger: pong, kinda, not really though [10:05] yofel: nevermind [10:06] not responding in time = not getting a say :P [10:06] Riddell: also unping :P [10:06] shadeslayer: mh, or maybe a separate PPA [10:06] ah well, I'll throw it at experimental, if users want a separate one that can still be arranged [10:07] I don't think it makes much sense for a separate PPA [10:07] No one should be running experimental long term [10:08] apachelogger: I was told to "talk to aaron, he is the activities god", therefore I am doing just that :) [10:08] apachelogger: is it required to call reloadCache in the finished function? [10:08] I mean finished means it finished successfully no? [10:09] oh hmm [10:09] not really [10:09] jussi: yeah, did we say that in munich? I think we said something like that [10:09] in retrospect it still seems silly tho ^^ [10:09] yep [10:09] apachelogger: oh well, it is done now. lets see how we go [10:10] shadeslayer: not necessarily it's best tho [10:10] right [10:10] that is to say I like to foward the error signal to the finished slot [10:10] because having 2 signals is shitty API [10:10] should be finished(State), enum State { Finished, Yolo, Error }, ErrorType errorType() const [10:11] so you have one function that always gets called when the backend stops doing stuff and you can call the backend for more information on possibly errors etc. [10:11] yeah :/ [10:11] apachelogger: http://paste.kde.org/pc64cab84 [10:13] apachelogger: also, this is confusing [10:13] does finished get called even if the transaction error'd out? [10:14] if the transaction errors out, then both slots get called and I'm calling reloadCache twice at possibly the same time [10:18] shadeslayer: no [10:18] that's the thing [10:18] *facepalm* [10:18] you either get an error signal [10:18] or you get a finished signal [10:18] that's why it is shitty API [10:18] fix it fix it fix it [10:18] filed a bug, can't do anything about that at this point [10:18] would be a rather grave API change [10:19] seeing as every application is forced to pick up both signals (further evidence of the shittyness .... :P) [10:19] shadeslayer: note that you can connect to a slot without params [10:20] connect(d->transaction, SIGNAL(errorOccurred(QApt::ErrorCode)), [10:20] this, SLOT(transactionError())); [10:20] so since you have the same logic you can just connect the errorOccured signal to SLOT(finished())) [10:21] (because you don't evaluate the ExitStatus you don't need the param really) [10:21] [10:21] apachelogger: right, but I'm adding code that will show the user a error message widget when there's an error with the transaction [10:21] so I do need two functions now? [10:22] 3 [10:22] error, finished, cleanup [10:22] error calls cleanup, finished calls cleanup [10:22] hmm, could rename restoreUI to cleanu [10:22] *cleanup [10:22] yeah, that thing would reload the backend then [10:23] general idea: cleanup does everything that needs to be done on transaction end [10:23] *nod* [10:27] hurrat [10:28] libqapt doesn't provide error details when you have an auth error [10:37] apachelogger: might want to look at 1182784 [10:38] apachelogger: and does this look right to you https://projects.kde.org/projects/extragear/sysadmin/libqapt/repository/revisions/master/entry/utils/qapt-batch/qaptbatch.cpp#L212 [10:39] I mean the string says "The following packages have not been verified by their authors. " , and then passes the size [10:42] and regarding http://techbase.kde.org/Development/Tutorials/Localization/i18n_Semantics#Structure_tags , do you mean ask tsdegos? [10:43] because it won't work without i18nc as mentioned yesterday [10:44] you may be able to use the structure tags without i18n somehow, I meant that you should ask whether you should use i18n for that string in general (for the sake of localization) [10:44] becasue I am not sure whether that may be necessary [10:48] I don't see how you can localize something like "Radeon HD XXXX" [10:48] or "Broadcom wireless FOO" [10:59] shadeslayer: that's why you ask a specialist [11:02] k gonna ask tsdegos [11:02] who isn't around at the moment [11:02] huh, he's here, but not in #kde-devel :) [11:03] tsdgeos: poke poke [11:05] shadeslayer: he's at work work that's why he isn't in kde-devel ;) [11:12] Good afternoon. [11:19] apachelogger: http://paste.kde.org/p28371404 [11:19] shadeslayer: "is hardware vendor", "is hardware model" maybe? [11:22] k [11:25] apachelogger: experimental PPA should have pacakges soon [11:28] time to figure out layouting issues [11:35] hm [11:36] * yofel realized why he didn't see apachelogger's ping earlier [11:36] plasma isn't updating o.O [11:36] it's not 12:08 now... [11:38] lol [11:38] deadlock? [11:39] shadeslayer: cool [11:40] some timer probably forgot to fire ;) [11:44] Hiyas all [11:50] apachelogger: so the driverOptionsVLayout QVBoxLayout is the member causing layouting issues, but I have no clue how to delete the layout so that all the child widgets get deleted [11:51] atleast not without writing a while loop that iterates over the child widgets and deleting them one by one [11:57] maybe it would be beneficial to migrate to a model-view system :) [11:57] shadeslayer: layout must not have a parent [11:57] not a qwidget one anyway [11:57] otherwise what happens is parent.setlayout(this) [11:57] which may be why things fall apart [11:58] I don't follow [11:58] oh [11:58] Oo [11:59] ah, oh [11:59] shadeslayer: you don't delete the layout [11:59] apachelogger: I delete the entire ui [11:59] oh [11:59] oh [11:59] QList m_driverwidgets [11:59] qDeleteAll(m_driverwidgets) [11:59] for example [12:00] there's no call to find driverwidgets :/ [12:00] shadeslayer: ^ things like that are why mode/view would be better [12:00] I'll have to store them at the time of calling shit [12:00] shadeslayer: you create those widgets.... [12:00] apachelogger: yeah, I'm thinking of migrating to model/view now :/ [12:00] maybe tomorrow [12:00] well [12:00] I don't have a clue how to work on that tbh [12:00] you could just do the cheapshot widget solution [12:01] QListWidgetView or what its called [12:01] see usb-creator [12:01] it's the cheap shot solution where you use an existing view without a model [12:01] not sure whether that will be enough for your layout concept though [12:02] though TBH, a less complicated layout working application would be much preferred over a well layed out one [12:02] apachelogger: http://paste.kde.org/p8fddf42d works [12:02] that's shit [12:02] I know [12:02] you do not ever call setupUi [12:02] whaaaa [12:03] just create a list of widgets to delete [12:03] and qdeleteall on that [12:03] yeah, I'm going to port this to model/view tomorrow [12:03] so we don't need such shitty things [12:03] shadeslayer: maybe ask for opinion in the office [12:03] hopefully tomorrow is the day I can understand model/view [12:04] apachelogger: ok [12:04] perhaps they have an idea for simpler thing than model/view but still better than that ^^ [12:04] true [12:04] enough driver stuff for today [12:04] I mean, at the end of the day that is just a mushed together model/view without neatly encapsulated logic, but I already complained about latter anyway :P [12:05] *nod* [12:28] 1202 e-mails to process, I will get through them today! [12:34] hi Sick_Rimmit! [12:34] Riddell: good luck with that :P [12:45] * Peace- got a 8 core phone [12:48] ::qt-bugs:: [1198806] osgviewerQt crashes on startup @ https://bugs.launchpad.net/bugs/1198806 (by Thomas Bleher) [13:16] hello world [13:17] on http://www.kubuntu.org/news/trusty-alpha-2-available-testing it should say "More information on Kubuntu *Trusty* Alpha 2 [13:19] ovidiu-florin: do you have or want the access details to change it? [13:20] Riddell: I don't have access. [13:20] I'll make the change if you give me access [13:30] apachelogger: d_ed and vHanda_ say that implementing a MVC with things like radio buttons is a PITA [13:32] and d_ed says my solution to the problem by deleting the driver layout is fine [13:34] shadeslayer: no, make a list [13:35] you are recreating the entire UI, because you did not put the layout there in the first place [13:35] and recreating the entire UI is just wrong [13:35] /o\ [13:36] kdepimlibs git needs akonadi git [13:36] I think that's it though [13:36] apachelogger: hm, so keep list of radiobuttons, call qdeleteall on the list? [13:36] shadeslayer: yeah [13:37] shadeslayer: but what you want to do anyway is create a class to contain the visual prototype of an entry in your pseudo list [13:38] so you can do QWidget *ptr = MewoClass(data); m_kittehList.append(ptr); layout->addWidget(ptr); [13:38] and on refresh qDeleteAll(m_kittehList); [13:38] hmm [13:38] gotcha [13:38] tomorrow [13:38] as for the deps... mtime on my clones suggest milou->baloo->kdepimlibs->akonadi [13:39] unless I started on day and finished the other day, then knewstuff would be in the list, although that doesn't make much sense I guess ^^ [13:39] s/on/one [13:39] good to know I don't have to go further than akonadi [14:01] apachelogger: any clue if these translations http://pastebin.kde.org/pgnw8pcja can be used by plasma-widget-homerun-kicker? [14:01] or since the file name does not have kicker, it can only be used by homerun [14:04] shadeslayer: what does homerun say it uses? [14:04] should specify a catalog somewhere [14:06] Riddell: Hi, sorry not ignoring you, just messing round with my machine setups [14:06] fullscreen/applet/homerunlauncher.cpp: KGlobal::locale()->insertCatalog("plasma_applet_org.kde.homerun"); [14:06] shadeslayer: so yes looks good [14:07] Riddell: `yeah [14:07] no catalog for kicker [14:15] sgclark: "Multi-Arch: allowed" where did you get that from? [14:18] Riddell: Umm, looked up the Arch binary in non arch error [14:18] Riddell: I don't remember sorry [14:18] Riddell: it fixes the error though [14:20] Riddell: I'm really tempted to ship homerun kicker as the default kickoff :P [14:21] shadeslayer: ug, why? [14:22] it's like going back to windows 95 isn't it? [14:22] I quite like it :) [14:23] sgclark: where did you get the rename patches from? upstream? [14:23] Riddell: either way, homerun-kicker in binary new soon [14:28] shadeslayer: does it need a separate package? [14:28] Riddell: IMHO yes [14:28] because they're 2 different plasmoids [14:28] 2 completely different things [14:28] * Riddell accepts [14:29] * shadeslayer rages at bzr [14:31] Riddell: no I made those myself [14:32] sgclark: you're good :) [14:32] sgclark: have you sent them upstream? [14:33] Riddell: no I have not, want me to? [14:35] sgclark: let me look closer to check [14:35] Riddell: good plan :) [14:40] sgclark: looks like kf5rename kf5renametelnetservice already have equivalents upstream but the .desktop files haven't been renamed for some reason [14:41] on they install to /usr/share/applications/kde4/ anyway so not a problem [14:41] oh they install to /usr/share/applications/kde4/ anyway so not a problem [14:42] for renamedbus I've been keeping the diff simple and not renaming the dbus interface file in the sources but just adding a RENAME kf5_foo.xml in cmakelists [14:42] also I've been using an underscore rather than a dash to show it's a namespace not part of the name [14:45] Riddell: the desktop files as is would launch the kde4 version, wouldn't they? [14:47] hmm, I'm not sure [14:47] I assume whoever did the upstream patch thought about it a bit [14:58] shadeslayer: yes, sure, if that name is correct [14:59] apachelogger: so I can't build release-upgrader on precise [14:59] it doesn't have things like ubuntu-drivers common [14:59] or python3-update-manager [15:02] apparently first arrives in Quantal [15:04] sgclark: kf5 meeting in #kde-devel incase you're interested [15:05] ok [15:07] shadeslayer: you build it on trusty [15:08] the point of the entire thing is that you download the release upgrader for trusty from trusty and it will work on precise [15:08] i.e. why it actually makes sense that this thing is pyth0rn [15:08] so it copies libs that it requires ? [15:08] for eg python3-update-manager [15:13] I think the requirement is just to not have crappy dependencies [15:19] 4 files changed, 303 insertions(+), 43 deletions(-) [15:19] down from 6k or so \o/ [15:19] sgclark: kio uploaded! [15:19] sgclark: I decided to go with only 1 -dbg package for simplicity [15:19] Riddell: ok thanks :) [15:19] sgclark: and rename the main package kf5-kio for simplicity [15:20] sgclark: also added build-deps on acl and gssapi [15:21] Riddell: how did you find that? it passed pbuilder [15:21] sgclark: I don't think there's any point in manpages for those binaries, they don't do very much [15:21] sgclark: look at the cmake output [15:21] sgclark: they are optional [15:21] Riddell: ahhh right, good idea. Also agreed on manpages [15:22] sgclark: I think the dbus patch is the important one to get upstream but it should be simplified to just edit the cmakelists files, do you want to do it or shall I? [15:23] Riddell: go for it :) I have a bunch more to get ready for you, I need kparts up today [15:23] oh boy, maybe I won't get through this e-mail backlog after all :) [15:23] lol [15:27] woot I was right [15:28] I never doubted you :) [15:28] ohm ohm, kcm now 90% less broken http://paste.ubuntu.com/6832572/ \o/ [15:30] awooga [15:31] that'll please the corner of belgium which speaks german [15:32] apachelogger: at what point is it supposed to show the release notes [15:32] because for some reason, I just directly jump to "Preparing to upgrade" [15:32] oho [15:33] maybe the release notes are displayed by the invocing thing [15:33] invocing thing? [15:33] update-manager-core or somesuch [15:33] I mean, something needs to download the upgrader tar and start it :P [15:34] that doesn't magically happen :P [15:35] uhu [15:35] *uhuh [15:35] so your instructions are useless ;P [15:36] well I do not know [15:36] investigating stuff usually is a good idea [15:38] oh how I hate nano, I hate it so fing much [15:40] better than pico [15:42] Riddell: s/better/bigger [15:42] shadeslayer: displaying is done by the fetcher [15:43] or so says grep [15:43] right [15:51] well then [15:51] Riddell: libkubuntu and revised kcm should be good to go into archive this week [15:51] aweet [15:51] sweet too [15:51] still need to shuffle a whole bunch around to get into a state where the patch can be upstreamed too [15:52] also need to discuss whether we'd want a binary plugin interface for backing or scripts or whatever [15:52] though given the complexity we require it will probably have to be a plugin [15:52] plus that allows building a nicer experience inside the KCM around it anyway [15:55] apachelogger: fwiw one can't just run the fetcher [15:56] so I think I really need to install 13.10 [15:56] you have a precise vm, no? [15:56] either way you don't need to install any specific version [15:56] you can meddle with your os-release/lsb-release file [15:57] or you can change /etc/update-manager/meta-release to point elsewhere indicating something newer than trusty [15:57] or you can simply hardcode some stuff [15:57] but I reckon you should have a VM :P [15:57] I have a precise VM [15:57] right then, see LTS testing instruction from deadlines board [15:58] and what I did was install python3-distupgrade which has the KDE fetcher [15:58] or you could run kubuntu-devel-upgrade or what its called [15:58] and it installs under the python3 namespace [15:59] so I symlinked the old fetcher to the new one under the python3 namespace [16:00] turns out python-kde4 isn't built with python3 [16:11] shadeslayer: does not compute [16:11] precise has its own fetcher [16:11] apachelogger: this is giving me a headache -.- [16:12] apachelogger: yes, and I'm trying to make it use my fetcher [16:12] which is python3 [16:12] why? [16:12] because my fetcher has the relevant code to show the html release ntoes? [16:12] why not copy the code to the old fetcher? [16:13] then how do I know that it'll work in the new code base :S [16:13] if it works in the old, chances are it will work int he new [16:13] plus to test the new you need a trusty+1 release [16:14] testing the new code on precise gives you absolutely no quality assurance whatsoever [16:14] not necessarily, I think it will work if I use Saucy [16:14] yeah, that's fooling-yourself-testing [16:15] Calligra is broken on Kubuntu. This works on Debian, but not on Kubuntu: http://www.youtube.com/watch?v=7LOHANw_dYc [16:16] how is the former not fooling-yourself-testing [16:16] because it's testing whether twe can throw that change at precise === greyback is now known as greyback|food === greyback|food is now known as greyback [16:57] Riddell: ping, if you have a moment, saw your opengl comment in #Calligra, and wondered what qt/gl support you use on arm. do you build qt with -opengl es2 (or other)? [17:01] apachelogger: hurray, I can make python3 crash with : http://pastebin.kde.org/pd8eiw1bp on line 24 [17:04] aha nvm [17:06] * shadeslayer ponders about icon for upgrade [17:16] rdieter: yeah it's opengles [17:17] Riddell: es2 though? or es1 ? [17:18] extra_configure_opts += -arch armv6 -no-neon -opengl es2 and for arm64 -no-javascript-jit [17:18] rdieter: ↑ [17:19] thanks [17:20] kio just started building, wow that was a long wait [17:20] sgclark: PPAs often take a while alas [17:20] hehe np, just want to get kparts done today, quite a few depend on it [17:21] sgclark: you can just self compile, you don't need to wait for the PPA to compile for you [17:22] Riddell: already did that, just want to run it through pbuilder to be sure it is perfect :) [17:23] it depends on kio [17:34] Riddell: failed to build symbols error. Can you put your source in my ubuntu one so I can fix? [17:55] apachelogger: :D [17:55] apachelogger: http://i.imgur.com/DqWyfDP.png [18:00] Sorry to bother, but a user in #k shared a folder from Dolphin and can access it from a Windows machine, etc. But he can't find where that config for that share is being stored. eg: grep share-name /etc/sama/smb.conf or ~/.kde/share/config/dolphinrc don't show it. Is there an actual place this gets saved or is it all in RAM ? [18:07] genii: moment [18:07] sgclark: you can get it [18:07] from the PPA [18:07] Riddell: yeah amd64 built fine, any idea why i386 died? [18:08] sgclark: probably different symbols [18:08] just needs a symbol update [18:08] that happens sometimes [18:09] ok [18:12] sgclark: I uploaded with the update [18:13] ok thank you [18:13] but you should be able to install amd64 now [18:13] sgclark: are you running i386? [18:13] apachelogger: The user left now and will probably post something to mailing list. If he reappears and you're around in there, the nickname they were under is berndk [18:14] yeah :) kparts going through pbuilder, one of the tests takes forever >.< [18:14] Riddell: no amd64 [18:23] Riddell: kparts ready for you [18:24] ooh [18:24] exciting :) [18:47] there is some ati xorg package, safe to do reboot without breaking things? [18:48] dunno we only do the kde packagse [18:49] Riddell: kdeclarative ready for you [18:51] * Riddell pays himself £50 for AWS usage for the last three months [18:53] sgclark: hmm your kparts depends on a load of libraries which should be brought in by kf5-kio-dev, I'll fix up kf5-kio-dev now [18:54] Riddell: yeah I thought that was odd, thanks [18:54] I will fix kdeclarative, not ready yet :) [18:56] sgclark: where did you get this description "Framework providing elaborate user-interface components." [18:56] I'm not sure it does provide components, rather it allows for other programmes to do so [18:57] Generally I search api doc, readmes etc. sometime techbase [18:57] exactly where I got that one, I am not sure. [18:58] sgclark: I think best to use Multi-Arch: same rather than Multi-Arch: allowed seems "same" is definate while "allowed" is some magic detection that I'm a bit skeptical will cause breakage if something in the package changes [18:58] Riddell: get reay for the lintian screaming at you then. [18:59] I can take on lintian any day :) [18:59] lol [18:59] mv libkf5Parts5.symbols libkf5parts5.symbols package names have lower case [19:06] sgclark: kparts uploaded! [19:08] thanks! [19:34] Riddell: can you give me the output of : qdbus --system org.freedesktop.login1 /org/freedesktop/login1 org.freedesktop.login1.Manager.CanSuspend [19:37] o_o [19:37]