[00:01] and off to dinner [00:01] sorry for being scarce, Riddell [00:10] Hi , adblock doesn't work in konqueror === rdieter_laptop is now known as rdieter [07:47] someone broke my login [07:51] * yofel hands apachelogger some duct tape and coffee [07:51] which login? [07:51] * valorie has about 100 [07:51] also: if you statartx from a login tty you end up with kaput polkit handling [07:51] much more important [07:52] valorie: lightdm in general [07:52] oooo [07:52] not nice [07:52] * apachelogger thinks proposed might have a bad linux-firmeware package [07:52] that is the login that allows all those other logins [07:52] indeed ^^ [07:53] hm [07:53] there's nothing quite as exciting as spending all morning debugging ^^ [07:54] other fun result: pulseaudio doesn't actually produce audio [07:55] which may either have the same cause as polkit being broken or lightdm being broken [07:55] for me it managed that by witching the default sound card, which I had to change in some underground level config dialog to get it back [07:56] my point is: it worked yesterday it doesn't today [07:56] so it's either because of startx or because of broken firmware [07:56] I haven't tried sound yet today, but I had to kill pulseaudio since it was consuming 178% of my cores [07:57] house was cold; at least the laptop was trying to warm things up [07:58] hehe === vinay is now known as Guest14183 [08:18] valorie: you're not scare, mearly in an antipodal timezone :) [08:18] valorie: you're not scarce, mearly in an antipodal timezone :) [08:19] :-) [08:19] valorie: was going to ask about this testing website you e-mailed about, I don't really understand what was being suggested [08:20] I summed up what apachelogger and I had discussed [08:20] and also the input of the KDE Quality team [08:20] seems that there isn't much interest though [08:20] I found the google docs we used last time hard to use [08:21] not as bad as the launchpad site, though [08:23] is the notes page unclear? [08:23] I didn't even hear back from apachelogger [08:23] so, can't really carry it forward [08:25] which notes page? [08:26] hmm, which email are you referring to? [08:30] ah I see it https://notes.kde.org/p/Testing_Web_Application [08:30] heh, I found the email at the same minute [08:31] I write too many friggin emails [08:35] ah, so, I concluded polkit and pulseaudio not working through startx likely is legit behavior [08:36] apachelogger: ? [08:37] the thing is, both are somewhat seat driven (if you login in any form or fashion you get a seat associated with the VT, e.g. if you go to the f1 terminal and log in that starts a seat), now pulseaudio for example uses seat activity tracking to automatically mute and unmute output based on whether its seat is active [08:38] but since I used startx my seat would still be on VT1, the X session however runs on VT7, so I am technically on my seat so pulseaudio doesn't want to generate output [08:39] supposedly it works similarly with polkit where it either would prompt authorization from the active seat (which is none because the X I started is not registered as seat and I am not at VT1, so that's not active either), or generally speaking since I am not at my seat it simply cannot get authorization [08:39] proper communication in the error message of polkit would help a bunch [08:58] Riddell: we have very different work flows, and I don't know which one is right now :P [08:58] so half the changelogs are done my way and the other half your way [08:59] Riddell: re .gitattributes, they'll end up in the packaging, whereas they won't in the debian packages ( or rather, I hope they don't ) [08:59] shadeslayer: the order of the changelog is not too important, I doubt it'll matter [08:59] .gitattributes is not important at all === Haudegen_ is now known as Haudegen [09:18] Riddell: that still leaves the issue of how to handle 4.13.2 [09:20] * Riddell blogs http://wire.kubuntu.org/?p=164 [09:21] shadeslayer: IMO just fix what causes problems in proposed by hand for now [09:21] mmm [09:21] or as I said, add magic to ignore released .2 packages [09:22] or be lazy and edit the packages list to not include the ones which are .2 [09:22] shadeslayer: I can take a look if you like [09:23] Riddell: yeah would be awesome [09:23] I am going to SRU 4.13.2 [09:31] yofel: wat http://paste.kde.org/pmu1mvuo9 [09:31] o.O [09:31] you are doing something weird ^^ [09:33] it would be useful to know what package and version are... [09:33] Downloading https://launchpad.net/~kubuntu-ppa/+archive/ppa/+files/audiocd-kio_4.13.1.orig.tar.xz [09:33] why would it do that [09:33] hm [09:33] hm.... [09:34] you might have to delete the .1 packages from the ppa ^^ [09:34] e.g. anything that's .1 but not .2 [09:34] heh [09:35] or maybe just wrap that chdir in try/except so it ignores wrong packages [09:37] lets do that [09:45] learning to use the pinky finger on the violin has had the awesome effect of being able to reach the page up/down button without having to move my hand alot [09:55] Riddell: btw did you setup the seed for kubuntu next? [09:57] shadeslayer: still on the todo, but go ahead and setup your scripts and just install plasma-desktop as a temporary standin [09:57] Riddell: right, about the scripts, will we able to release the ISO on cdimages? [09:58] or how do we want to handle distribution [09:58] should we ask Canonical for an exception to the PPA rule? modify scripts et all [09:58] ( for cdimages.ubuntu.com ) [09:58] or do we go on our own [09:58] setup has to be done accordingly [10:01] hmm [10:01] dunno [10:01] I'm starting to think it would be easiest just to use the archive [10:06] 3 hours of my life down the drain [10:06] Riddell: apachelogger ScottK yofel btw if there were a proposal to change the scheduler to CFQ in utopic/trusty to improve baloo performance, how would that go down :P [10:06] linux scheduler? [10:06] IO scheduler [10:06] I don't know what CFQ is [10:07] shadeslayer: that would go down by proposing it to the foundations team [10:07] it's a IO scheduler which supports IO niceness ( current one is deadline ) [10:07] apachelogger: foundations team is fine if we do it for Kubuntu [10:07] why would we do it for kubuntu only? [10:07] because it's easier [10:08] shadeslayer: I think changing the IO scheduler to support one application is insane. [10:08] it also makes IO scheduling on kubuntu fundamentally different from all other buntus [10:08] FWIW, I don't see any issues with baloo performance here. [10:08] ScottK: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1310402 [10:08] That too. [10:08] Ubuntu bug 1310402 in linux (Ubuntu) "Userland depends on ionice idle but default scheduler is "deadline". " [Wishlist,Triaged] [10:09] apachelogger: yes, however changing it in all of ubuntu is going to be a long winded discussion, and I'd rather not get into that [10:09] why would it not be a long winded dicussion for kubuntu? [10:09] I reckon changing the scheduler warrants a lot of wind TBH :P [10:10] Seems like something people who want to change ought to change. [10:11] the thing is [10:11] deadline is not the default scheduler of kernel upstream [10:11] so IMO this ought ot be taken up with foundations because their divergence from upstream default is causing problems for people so it should probably be evaluated whether deadline by default makes sense [10:12] Things that would awesome if you could find out - [10:12] 1. Real reasons why Ubuntu changed it from the default (which is CFQ). I haven't managed to find proper reasons. I can see that it was done when the server + desktop kernel were merged. [10:12] 2. Evaluate if IO niceness is actually a thing you guys want in Kubuntu. Do you want one app doing crazy amounts of reads to cause problems with others? [10:12] apachelogger: won't be a long winded discussion in kubuntu because we have very clear reasons to switch, and it would have a smaller impact, and CFQ is what baloo upstream recommends [10:12] apachelogger: plus we can switch back and forth a bit more easily [10:12] shadeslayer: we also have a very clear reason not to switch because we do not understand why foundations diverges from upstream to begin with [10:13] we have no technical expertise to make any assertion on scheduler defaults really [10:14] so learn about the different schedulers and evalualte them? [10:14] you guys understand code, and there is ample documentation about these schedulers available. [10:14] or simply make the foundations team figure it out [10:15] apachelogger: kernel team actually. [10:16] theres't two ways I see this going into a default a) daedline is considered best default for minimal setups and servers so desktops need a udev rule to switch to CFQ b) cfw is considered best default for desktops and servers/minimal setups need a rule to swtich to deadline [10:16] http://irclogs.ubuntu.com/2014/06/10/%23ubuntu-kernel.html#t16:49 [10:16] might want to read that ^^ [10:16] considering deadline clashes with the concept of indexing all the world on a desktop setup in an IO nice way I don't see one scheduler being able to accomodate both a server and a desktop setup [10:18] In any case, no way for trusty. [10:18] because? [10:18] Because it's already released. [10:18] We don't make invasive changes like that post release. [10:19] again, because? [10:22] ScottK: as seen on the IRC logs, the foundations team says that they can't think of any issues [10:22] so I don't think it's an invasive change as such [10:22] it's basically a drop in replacement [10:22] they also say that the reason we switched to deadline was because CFQ had performance problems [10:23] which was ... how many years ago exactly? [10:23] also, could someone find the data they had to support this? [10:23] so that statement doesn't hold true unless we can't prove otherwise [10:23] yeah, so make the kernel team switch to CFQ again [10:24] apachelogger: ok, for utopic we can bring that discussion up with foundations, I'm talking about trusty [10:24] vHanda, shadeslayer: you should carry this out via a paper trail really... send to their mailing list [10:24] shadeslayer: we cannot change trusty [10:24] ubuntu-sru will say nonononon [10:24] apachelogger: why? [10:24] possibly also nononononononononono [10:24] this is a kubunut specific change, no? [10:25] how does the ubuntu sru come into this? [10:25] kubuntu is part of ubuntu, so if the ubuntu-sru team says nonononononno that's the law :) [10:25] vHanda: that is ubuntu as in the project not as in the product mind you ;) [10:26] this makes no sense to me. I don't understand most of the jargon. [10:26] I get that you're a part of ubuntu, but if you are making a "Kubuntu" specific change, what is the problem? [10:26] you have control of your project, don't you? [10:27] vHanda: ubuntu-sru is a team that is the authority saying what is in line with the update policy and what is not [10:28] so as for possible reasons why they would say nononnonono: like you expected CFQ to be used, someone else may expect deadline to be used (because we released with deadline active)... so changing to CFQ post release would then break someone elses scheduler expectation [10:29] we could provide a new optional package carrying out the switch but that would be opt-in, or make the change via kubuntu-ppa only (which would earliest kick in with 4.14 then I guess) [10:30] actually forcing existing 14.04 installations from deadline to CFQ seems like a very not legit update change [10:30] so outside the two presented cases I don't see that happening [10:31] shadeslayer may of course feel free to try xD [10:31] there are no "deadline" specific features. There is "IO niceness" which certain schedulers support and certain do not. [10:32] anyway, I can only argue so much. Kubuntu gets a worse name (unfortunately that also drags in Baloo) [10:32] vHanda: there is deadline specific performance though, server setups are known to perform much better with deadline than with CFQ, so imagine a server that has whatever random package we would add the udev rule to, that server would now have decreased performance [10:32] apachelogger: is Kubuntu targetting servers? [10:32] yes or no [10:32] it is not, ubuntu is, so you can have a kubuntu package on a server [10:33] vHanda: we are all in the same archive, so we can do nothing to prevent someone from installing kubuntu on a server [10:33] so for those few (say 5%) you're willing to screw the rest. [10:33] same rules for everyone [10:33] urgh. Guys, take a call. [10:33] there are always going to be pros and cons. [10:33] don't just say - oh, there are possible cons, so lets not do anything [10:33] that's how the ubuntu SRU policy works... [10:34] also, please don't use vague terms like "performance decrease", get exact numbers of how much it would be. [10:34] they include a regression evaluation [10:34] is it not possible to change it at runtime? (i.e. it doesn't have to be set during compile?) [10:34] it is [10:34] then we could just put it in kubuntu-settings [10:34] Riddell: for trusty [10:34] oh, unlikely, that was released already [10:34] so, change it and push it as an update? [10:34] Riddell: that was my idea [10:34] to put it in kubuntu-settings [10:35] as a udev rule [10:35] vHanda: FWIW IMO you could/should put a udev rule in baloo upstream forcing CFQ [10:35] and you guys would be fine, with that? [10:36] I dunno, I am saying that with an upstream hat on [10:36] I don't see why we would have a problem with that anyway, you say CFQ must be used, so CFQ must be used if we want to ship baloo [10:37] ^^ I don't see a problem fwiw [10:37] on rotational media, yeah [10:37] I can see a lot of distros fighting with me if I push this, but lets see. [10:38] though, you guys are again, refusing to take a call, and passing the buck. [10:38] just an observation [10:38] vHanda: well, if they disagree they can simply drop the udev rule but then they are willingly not supporting baloo [10:38] it's like the distros that rip vlc apart and break it beyond repair [10:39] alright. I'll figure it out. [10:39] not much you can do about people willingly breaking your things, but you shoudl at least try to prevent them from doing it unwillingly... like we do not intentionally use deadline out of spite :) [10:40] I just find it sad that you guys have no control of your project [10:40] but oh well, that's something you people need to figure out [10:41] yofel: http://paste.kde.org/p2qcse08o [10:41] there you pissed vHanda off [10:41] @_@ [10:42] me? don't put this on me, I actually think shipping a SRU with kubuntu-settings makes sense [10:42] sense != acceptable for ubuntu-sru [10:42] actually I am not sure about the sense either [10:42] why exactly are we talking about trusty before fixing this in utopic? [10:42] Riddell: yes, for the recommends :P [10:42] because we have no data to back up CFQ performance better than what it used to [10:43] well, yes [10:43] Riddell: and I'm considering removing that warning indeed as people are reporting bugs thinking it's a bug [10:43] shadeslayer: I have no idea how you've managed to get a unicode error while logging into launchpad [10:43] dantti: makes me wonder why debian dropped it [10:44] yofel: btw : lp = Launchpad.login_with("pull-ppa-source", "production") < line that crashes? [10:44] yeah [10:44] wtf? [10:44] yofel: possibly bad response from launchpad? [10:44] possibly, but no idea really [10:44] 'Morning folks [10:44] Riddell: hmmm good question does they depend on scp-gnome (if it still exists?) [10:45] dantti: nope https://packages.debian.org/sid/system-config-printer-udev [10:45] no, I meant print-manager [10:45] but it doesn't [10:46] * Riddell asks in debian channel [10:50] feed branches had 25 updates, showing the latest 6 [10:51] I pinged maxy but no reply so far :P [10:51] Riddell: bug 1330445 [10:51] bug 1330445 in Kubuntu PPA "KDE 4.13.2 for precise" [Undecided,Invalid] https://launchpad.net/bugs/1330445 [10:51] I hope you remember what I said when you did 4.13.0 [10:52] I'm not going to backport .2 to precise, if pali wants to help that's all good [10:53] goes against policy fwiw [10:54] if he wants to do it we shouldn't use policy to stop him that's nuts [10:55] feel free to, I am not going to deal with people who upgrade precise to 4.13.2 :P [10:56] because we have no data to back up CFQ performance better than what it used to [10:56] apachelogger: so how do we fix that [10:56] yofel: btw unicode error was just a temporary thing [10:57] heh [10:57] 3rd times a charm, uploading 4.13.2 [10:57] shadeslayer: uploading to where? [10:58] Riddell: trusty-proposed [10:58] where else [10:58] just making sure it wasn't to utopic (which I'm doing) [10:58] right, ofcourse :) [10:59] oh fun [10:59] oh hi Pali [10:59] hi [10:59] just found this in my rc.local: "echo cfq > /sys/block/sda/queue/scheduler" :D [10:59] no idea when I added that... [11:00] heh [11:00] Pali: we have some scripts in lp:~kubuntu-packagers/+junk/kubuntu-automation/ we use to package KDE SC releases [11:00] ok, will look at it [11:01] Pali: I think kubuntu-batch-backport is the one you want to make 12.04 packages [11:02] Pali: note that this is somewhat controvertial because we don't want spend lots of time on things we can't manage like backporting to two LTS releases, so you may get some people making grumpy remarks about policy :) [11:03] might be a bit tricky btw. as .1 wasn't backported [11:03] maybe run the script once on .1, then again on .2, that should catch all packages [11:03] ^^ [11:03] Riddell: http://irclogs.ubuntu.com/2014/04/25/%23kubuntu-devel.html#t10:12 [11:04] * ScottK has about given up on the idea that anyone in KDE even understands what a stable user experience is. Let alone that it's a reasonable goal. [11:04] "=== Skipping konsole because version 4.13.2 is already in the archive" sweet, my lovely code works [11:06] Riddell: so that script just pull packages from ppa, change version/series, pack it and push to ppa for other series? [11:07] that + runs a few fixing scripts in backport-hooks/ [11:07] by default it using kubuntu-ninjas/ppa [11:07] packages are in kubuntu-ppa/ppa too which is public [11:07] but I do not have probably permisison to see it [11:08] ok [11:08] you'll have permission to see it but not upload [11:08] shadeslayer: metrics gathering::: option a) SRU completely new opt-in tech to gather metrics (or find a way to make them not opt-in but not compromise quality) b) throw opt-in tech at utopic and gather data until a reasonable set size is reached (alas, no clue what reasonable would be... probably want a metric for how many installations have used utopic vs. how many installations use trusty) [11:09] eh, b) would be opt-out to ensure maximum gathering [11:10] also get kernel team to give data on why deadline was chosen initially [11:11] random thing I just thought about ... wouldn't the deadline locking only affect installations that have huge amounts of data pouring into $home at roughly the same time? [11:11] with incrementally growing homes you'd have no noticable impact as the indexing would run through it quickly [11:12] if so then the scheduling problem would primarily impact people who do a new installation and then copy heaps of old data into the new $home [11:12] gives it a much smaller impact I guess [11:17] Riddell: where I can find pull-ppa-source program? [11:19] Pali: lp:kubuntu-dev-tools [11:22] ok [11:24] Riddell: for the record I filled a bug https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=751966 [11:24] Debian bug 751966 in print-manager "print-manager missing recommends" [Normal,Open] [11:26] dantti: nice [11:26] Riddell: should I create new PPA, or where can I tell that script to upload packages? [11:26] Pali: I think best if you make your own PPA for it, upload there then when it's all working I can copy them over [11:26] which is pretty much what we do normally anyway [11:27] ok [11:34] shadeslayer: I recall sluggishness just after upgrading, but none since. I think optimizing for the just after upgrade/install case would be a mistake. [11:34] Riddell: I do not see 4.13.2 packages yet... should I wait for them or try to backport 4.13.1? [11:35] Pali: I don't think you need to wait, at least shadeslayer did 4.13.2 in trusty without waiting [11:36] I'm working on utopic now [11:36] in which PPA/archive is 4.13.2 version? [11:37] shadeslayer? ↑ [11:38] kubuntu-ppa/ppa [11:38] https://launchpad.net/~kubuntu-ppa/+archive/ppa/+packages?field.name_filter=&field.status_filter=published&field.series_filter=trusty [11:38] ok, thanks [11:39] Pali: also, please backport from trusty [11:39] no utopic [11:39] *not [11:43] I started "./kubuntu-batch-backport -v 4.13.2 -f trusty -t precise -V 0.1 -s kubuntu-ppa/ppa" and script generating version "4:4.13.2-0ubuntu0.1~ubuntu0.1~ppa1" [11:44] -V 0.1 should be 12.04 [11:46] ok, now version string is "4:4.13.2-0ubuntu0.1~ubuntu12.04~ppa1" [11:46] is anyone looking at bugs btw? [11:46] Riddell, shadeslayer ^ [11:46] bugs are overrated ^^ [11:48] even more akonadi mysql fun, weee [11:49] if someone knows a good way to switch to sqlite without screwing up existing setups I'm for it [11:49] UPSTREAM! [11:49] they don't even know what DB actually exists on the system before they use it -.- [11:50] hmm, actually how do I prepare for upload to the archive if it's not in a PPA? [11:50] feed branches had 25 updates, showing the latest 6 [11:52] Riddell: hm, the archive upload script does use bzr to upload, so maybe just adding an option to disable the failure prevention checks would work [11:53] although I doubt things will work fine if you just go and upload now [11:53] .1 and .2 had some changes in them [11:53] but it uses pull-ppa-source to get the tar [11:53] I've run kubuntu-initial-upload to check all the patches apply etc [11:53] right, then it does a diff against bzr to makes sure the PPA and bzr actually match [11:54] or rather that the ppa isn't ahead of bzr [11:56] this code does badly need some comments [11:56] it's magic ^^ [12:06] yofel: in kubuntu-ppa/ppa still missing some 4.13.2 packages [12:06] Pali: our scripts check if the package has changed since the last version and skipt it if not. That's why you don't see many .2 packages [12:07] you should also run the script on 4.13.1 to catch those [12:07] ok, thanks for info [12:28] good morning Riddell: did you have time to review baloo-widgets5? [12:31] how can I upload packages generated by ./kubuntu-batch-backport to my PPA? [12:31] dput does not working due to missing gpg signatures [12:32] run debsign *.changes [12:43] thanks [12:50] feed branches had 25 updates, showing the latest 6 [12:52] now packages are compiling in my ppa: https://launchpad.net/~pali/+archive/kubuntu-backports/+packages [12:54] out of curiosity, why are you doing this? [12:55] sgclark: oh sorry not yet, will do shortly [12:55] np [13:33] mhmh [13:33] seriously, who is looking at bugs? [13:33] me bug inbox has many things :'< [13:39] sgclark: libbaloowidgets5 says Conflicts: libbaloowidgets [13:40] but there's no such package [13:40] the kdelibs4 package is libbaloowidgets4 but that shouldn't conflict, different soversion [13:41] Riddell; ahh ok [13:41] otherwise packaging looks lovely [13:41] ok thank you, fixing and uploading [13:44] good afternoon [13:45] jussi: got those polo shirts (for sending to akademy) [13:45] s/got/I bought/ [13:45] Riddell meant: "jussi: I bought those polo shirts (for sending to akademy)" [13:45] I have been digging a bit into "kde 5" runtime problems, I think I will be fixing some in the next days [13:46] "plasma 5"! [13:46] plasma next :P [13:47] I don't really like the next term, cos in a year's time it'll mean something else and you'll get confused when looking at old irc logs [13:48] probably, didn't they decided the name yet? [13:50] plasma 5 [13:50] I decided, and nobody has argued with me so far [13:50] hahaha [13:50] plasma 5 then [13:50] seems like the obvious choice [13:52] the question now is how to refer to the old version [13:52] plasma 4? [13:52] wasn't that plasma 1? ^^ [13:53] it was [13:53] except it was always released with KDE SC so that term was never used [13:53] "plasma from SC" [13:54] plasma 1 from kde workspaces 4 [13:54] to make PR people happy [13:55] PR people dropped use of "workspace" as being not understandable [13:56] uhuh [13:56] why am I not surprised that I never heard about that [14:13] Riddell: wire post has a pointless 'a' in the last sentence [14:13] You’d pay as much for a just the Windows licence in most other shops. [14:13] thanks, fixed [14:18] Riddell: wrt the epoch thing in pkg-kde-talk, what about doing foo5 packages now, switch later to epoch if needed? [14:18] ug, worst of both worlds [14:19] why? [14:21] renaming packages is very untidy [14:21] shouldn't be foo5 transitional packages enough or am I missing something? [14:22] best avoided if possible [14:23] ok [14:27] apachelogger: did you spot that on the twitter feed or the facebook feed? [14:29] Riddell: the planetfeed [14:29] boring [14:30] * apachelogger isn't very social [14:30] g+ is where the world is [14:31] but it has no api! [14:31] * yofel should read the last 2 weeks of backlog at some point [14:31] least social social media platform ever [14:31] there was some api at some point [14:31] I distinctly remember having made a tiny qml app to browse plus [14:32] shadeslayer: now I'm getting a million jenkins failures from there 4.13.2 packages, how do you see what's failed? [14:32] give me a link [14:33] ah, the api is readonly [14:33] why that is weird [14:33] apachelogger: it's a feature [14:33] prevents you from posting drunk posts [14:33] or drunk pictures [14:33] or both [14:34] right, google needs to api to find those [14:34] *no [14:35] shadeslayer: https://jenkins.qa.ubuntu.com/job/utopic-adt-umbrello/2 [14:36] annoying how jenkins doesn't set the mimetype of log to something sensible, you have to download it [14:36] Riddell: https://jenkins.qa.ubuntu.com/job/utopic-adt-umbrello/2/ARCH=amd64,label=adt/console [14:36] I don't have to download anything [14:37] that's the console, the log file has no mime type [14:37] which is annoying [14:37] https://jenkins.qa.ubuntu.com/job/utopic-adt-umbrello/2/ARCH=amd64,label=adt/consoleText [14:37] seems to work for me? :S [14:37] wait [14:37] shadeslayer: https://jenkins.qa.ubuntu.com/job/utopic-adt-umbrello/2/ARCH=amd64,label=adt/artifact/results/log [14:37] ah [14:38] who even looks at that :P [14:38] I just go to the console output [14:38] shadeslayer: how do you get to your link? [14:38] Riddell: well, on the link you gave me, there are 2 links at the bottom for each arch [14:38] I click on the arch , then on console output [14:39] in your case kde-runtime can't be installed [14:56]