2014/10/08/#kubuntu-devel.txt

[03:56] <ScottK> Riddell: kapidox uploaded to Debian.
[06:19] <soee> good morning
[07:02] <lordievader> Good morning.
[07:57] <jussi> morning all. still broken plasma here, seems about 5 or 6 things crash :( Updated again and still didnt sort it, wonder whats goign wrong 
[07:59] <shadeslayer> Riddell: http://paste.kde.org/po5wdn2od
[08:01] <shadeslayer> apachelogger: ping
[08:01] <shadeslayer> oh
[08:01] <shadeslayer> apachelogger: nvm 
[08:01] <apachelogger> -.-
[08:02] <shadeslayer> :D
[08:02] <shadeslayer> apachelogger: good morning!
[10:16] <sgclark> morning
[10:17] <Riddell> hola chica
[10:19] <sgclark> seems I did not get the artistic genes :( what is left on that list kopete skype icons and bug triage?
[10:19] <sgclark> Riddell ^
[10:20] <Riddell> sgclark: and working out what to do about the korean font issue that was posted to the mailing list
[10:20] <sgclark> unless you are happy with stick figures!
[10:20] <Riddell> xkcd style :)
[10:20] <sgclark> lol
[10:22]  * sgclark thinks I have to set kmail aside for the time being :(
[10:23] <Riddell> what's up with it?
[10:24] <sgclark> consentent retrieving folder contents.. restating akonadi sometimes helps
[10:27] <sgclark> it was working great... wonder if there was a new release of akonadi
[10:34] <Riddell> plasma 5 images have lots of problems :(
[10:34] <Riddell> i386 just crashes everything :(
[10:34] <Riddell> jussi: you using i386 or amd64?
[10:46]  * Riddell cries at the number of plasma5 bugs to fix https://bugs.launchpad.net/kubuntu-ppa/+bugs?field.searchtext=&orderby=-importance&search=Search&field.status%3Alist=NEW&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&assignee_option=any&field.assignee=&field.bug_reporter=&field.bug_comment
[10:46]  * Riddell cries more at the length of the url
[10:46] <Riddell> http://goo.gl/B527rj
[10:46] <sgclark> yikes
[10:46] <Riddell> no wallpaper, no clock shown, i386 all crashes :(
[10:48] <sgclark> well still no mail for me, any other good clients out there? 
[10:49] <Riddell> mutt!
[10:49] <Riddell> pine!
[10:49] <Riddell> note: these may not be to everyone's taste
[10:49] <sgclark> I actually started with that years ago lol, installing
[10:49] <Riddell> I use a combination of gmail and mutt
[10:50] <sgclark> as much as I break my X, mutt is a good backup lol
[10:50] <Riddell> oh arm64, why do you confuse kf5 compiles so much?
=== kbroulik is now known as kbroulik-lunch
[11:14] <BluesKaj> Hey folks
[11:14] <soee> hioho BluesKaj
[11:14] <sgclark> hiyas
[11:15] <soee> Riddell: yeah it would be good to have working Plasma5 isos, a lot of people asking on #kubuntu how can they try it
[11:15] <BluesKaj> Hi soee, sgclark
[11:15] <soee> and its always safer to suggest them to use iso than upgrade :)
[11:17] <Riddell> soee: the amd64 one works, just not very well
[11:17] <soee> :)
[11:30] <ScottK> Riddell: https://launchpad.net/ubuntu/utopic/+source/kapidox/5.3.0-1
[11:31] <Riddell> whee
=== kbroulik-lunch is now known as kbroulik
[11:31] <Riddell> ScottK: so you think put the packaging in pkg-kde git?
[11:31] <ScottK> I'm fine with it either way.
[11:31] <ScottK> Probably better to move it to pkg-kde.
[11:32] <ScottK> Not much point before Debian's ready to upload kf5.
[11:33] <apachelogger> (unless it is in pkg-kde it cannot be covered by CI)
[11:36] <ScottK> Can't really be covered by CI anyway AFAICT.
[11:36] <ScottK> It's not really a framework in the same sense as the rest of kf5.
[11:39] <Riddell> vishesh is grumpy again about the scheduler not being changed in 14.04, should we do a SRU or was that considered?
[11:41] <ScottK> What was the result of the discussion with the kernel team on 14.10?
[11:42]  * ScottK lost track.
[11:42] <ScottK> It doesn't seem like SRU material to me though.  Maybe publish something on kubuntu.org about how to change it.
[11:43] <Riddell> he will probably write a grumpy blog post then
[11:44] <Riddell> I don't think writing on kubuntu.org is a great solution
[11:45] <ScottK> Not sure.
[11:45] <ScottK> I don't understand how "that release is done" is so hard to understand.
[11:46] <Riddell> he considers it a bugfix
[11:47] <Riddell> I tend to agree espcially since the kernel team said it was ok
[12:02] <Riddell> ScottK: bug 1378789 uploaded by shadeslayer, if you did consider it sru worthy that would stop vishesh biting our heads off
[12:02] <ubottu> bug 1378789 in kubuntu-settings (Ubuntu Trusty) "[SRU] Set the default IO scheduler to CFQ in Kubuntu Trusty" [Undecided,New] https://launchpad.net/bugs/1378789
[12:04] <ScottK> Riddell: Can someone from the kernel team ack that as a reasonable change for post-release?  I certainly don't know what the potential for regression is.
[12:10] <alvin> Isn't Deadline better when you have an SSD disk? So, what if you have Baloo (wants CFQ) AND an SSD disk (wants deadline)....
[12:12] <ScottK> Could be.
[12:12] <Riddell> I'll wait for rohan to come back from lunch
[12:12] <ScottK> I have an SSD and haven't noticed much in the way of issues with Baloo.
[12:13] <sgclark> ditto
[12:13] <alvin> Same here. No complaints about Baloo with an SSD.
[12:13] <ScottK> WAY better than nepomuk in that regard.
[12:14] <sgclark> agree
[12:20] <apachelogger> ScottK: you wouldn't notice with SSD
[12:21] <apachelogger> deadline performs better than cfq on !rationale media
[12:21] <yofel> You won't really notice a difference with CFQ on SSDs either. All it does is waste some CPU power for useless scheduling in that case
[12:21] <ScottK> Ah, so you're going to worsen my battery life?
[12:22] <yofel> hm, I didn't think of that..
[12:22]  * sgclark doesn't have battery life to spare
[12:30] <sgclark> Riddell: there is a merge request pertaining to the korean thread already, I do not have permissions to work with any of branches, so back to you .
[12:30] <sgclark> https://code.launchpad.net/~jincreator/ubuntu-seeds/kubuntu.utopic-korean
[12:30] <debfx> if you look at the udev file, it only changes the scheduler for rotational disks
[12:31] <debfx> (not sure why it's called 60-ssd-scheduler.rules though)
=== rdieter_work is now known as rdieter
[12:47] <shadeslayer> ScottK: alvin udev rule is smart, only changes to CFQ on rotational media
[12:48] <shadeslayer> not on SSD's
[12:48] <ScottK> Systems with one of each?
[12:48] <shadeslayer> same thing
[12:48] <shadeslayer> I have one of those
[12:48] <ScottK> Does what?
[12:48] <ScottK> Does it treat it like SSD or rotational?
[12:48] <shadeslayer> no
[12:49] <shadeslayer> ScottK: http://paste.kde.org/pvkynvnpj
[12:49] <shadeslayer> applies CFQ to HDD, deadline to SSD
[12:49] <ScottK> I see.
[12:49] <shadeslayer> by deafult deadline is used till udev gets loaded -> udev reads rules -> sets CFQ for all rotational media
[12:50]  * ScottK would like to see a kernel team or maybe platform team endorsement on this.
[12:50] <shadeslayer> ScottK: we've already had this discussion with the kernel team for 14.10, do we have to do another one for 14.04?
[12:50] <shadeslayer> though that discussion was actually generic
[12:51] <ScottK> Rules for SRU are different.
[12:51] <shadeslayer> and I didn't mention 14.10/14.04 IIRC
[12:51] <ScottK> Zero regressions.
[12:51] <shadeslayer> but that can't be guranteed
[12:51] <shadeslayer> of any SRU
[12:51] <ScottK> Which is why we go fix them when we find them.
[12:51] <shadeslayer> we can fix the most obvious issues
[12:51] <shadeslayer> but there *might* be a regression in a corner case
[12:52] <ScottK> Where's the discussion documented?
[12:52] <shadeslayer> irc logs, though I don't remember when
[12:53] <shadeslayer> aha
[12:53] <shadeslayer> ScottK: http://irclogs.ubuntu.com/2014/06/10/%23ubuntu-kernel.html#t16:49
[12:54] <shadeslayer> though that's the short one
[12:54] <shadeslayer> hm
[12:56] <ScottK> Which says it's not the default generally because performance was worse.
[12:56] <ScottK> So if we optimize for baloo, are we suboptimizing everything else?
[12:57] <shadeslayer> we're suboptimizing for everything that servers run
[12:57] <shadeslayer> because deadline is optimized for the server usecase
[12:57] <shadeslayer> ( thats why the default was changed, made server performance better, while not costing alot of desktop performance )
[12:58] <Riddell> sgclark: do you know if the korean fonts will explode the size of the images?
[12:59] <shadeslayer> ScottK: mind joining #ubuntu-kernel
[12:59] <ScottK> Done
[12:59] <sgclark> Riddell: no idea, and not really sure how to find out
[13:01]  * sgclark is fighting her own crashtastic upgrade
[13:03] <ScottK> Comment just noticed in code: /* I am VERY, VERY sorry about the gotos. */
[13:16] <ScottK> OK.  How about this for a plan:
[13:17] <ScottK> Since it's changed in 14.10 and that's about to get large scale testing, we wait until ~ a month after 14.10 release to see if we get negative feedback and if not go ahead.
[13:17] <ScottK> In the scheme of 5 years of LTS, that time is a detail.
[13:20] <ScottK> Riddell: ^^^ does that seem like a reasonable approach?
[13:23] <Riddell> ScottK: that seems excessively conservative, we know upstream linux recommends it, we know baloo recommends it, ubuntu kernel team don't seem too sure why they switched, we don't have any reason to think it'll cause problems
[13:24] <Riddell> Pali: did you see the feedback on your 4.14.1 packages
[13:24] <ScottK> OK.  See if another SRU team is willing to move faster or go to the TB then.  I'll add this to the bug.
[13:32] <Riddell> meh, poor vishesh has left
[13:37]  * jmux following #ubuntu-kernel just noticed that 10.04 trusty HWE also has deadline => guess I'll open a RfC to change it to cfq
[13:39] <Riddell> jmux: RfC? wanting the change in ubuntu or in limux?
[13:40] <Riddell> jmux: I've no idea if nepomuk is affected as much as baloo, and maybe you don't even use nepomuk, but I guess the same issues apply
[13:44] <jmux> Riddell: Just LiMux
[13:44] <jmux> Well - we'll be using Kolab and were told they need baloo
[13:46] <Riddell> jmux: ah typo above, s/10.04/14.04/ so you can just grab the same change we are arguing about on bug 1378789
[13:46] <ubottu> bug 1378789 in kubuntu-settings (Ubuntu Trusty) "[SRU] Set the default IO scheduler to CFQ in Kubuntu Trusty" [Undecided,New] https://launchpad.net/bugs/1378789
[13:47] <Riddell> sgclark: I suggested something to try on that khelpcentre issue
[13:47] <sgclark> ok
[13:51] <ScottK> Riddell: re the kubuntu-meta in unapproved, shouldn't the "active" metapackage be removed enitrely?
[13:52] <ScottK> I'm accepting what's there, but that seems the logical next step.
[13:52] <Riddell> yeah I think so
[13:54] <sgclark> Riddell: fyi kio is still having the file conflict, didn't you fix that?
[13:54] <Riddell> sgclark: umm, I think I did, let me check
[13:55] <sgclark> kcm_trash.so
[14:09] <Riddell> sgclark: it's fixed in kio_5.3.0-0ubuntu1 but that's still in utopic-proposed
[14:09] <Riddell> because arm64 is being super slow :(
[14:10] <Riddell> but maybe it'll work this time, keep your eye on https://launchpad.net/ubuntu/+source/kio/5.3.0-0ubuntu1
[14:11] <Riddell> sgclark: korean fonts merged, may well be time for you to go through kubuntu-dev and show the world your eliteness
[14:12] <sgclark> go through kubuntu-dev? apply to be a dev? not sure I am elite enough yet :( sure am trying though
[14:14] <ScottK> sgclark: Generally one of the best signs you're ready to apply is when an existing dev says you are.
[14:15] <ScottK> Getting upload rights doesn't mean you HAVE to do it, just the you can.
[14:15] <sgclark> aww okies, I believe you then, I will adventure into applying then Riddell
[14:32] <genii> Holy libkf5* updates today Batman!
[14:33] <Riddell> meh, still only half done
[14:33] <Riddell> grumble, arm64 grumble
[14:33] <genii> Ah, ARM
[14:33] <Riddell> armhf is fine
[14:33] <Riddell> you'd think 64 would be faster but nooo
[14:34] <shadeslayer> genii: nothing new for me, I use CP :p
[14:38]  * Riddell syncs up kapidox pkg-kde git to debian package
[14:38] <Riddell> ScottK: should I change the Vcs-Browser: lines to pkg-kde ?
[14:38] <ScottK> When is Debian qt-kde going to upload?
[14:39] <Riddell> ScottK: I've not heard any plans
[14:39] <ScottK> Then no.
[14:39] <Riddell> okay dokay
[14:39] <ScottK> There's no reason to change it until they are ready to do the rest of KF5.
[14:42] <shadeslayer> I thought I already changed the vcs headers, or was that for Plasma 5 only
[14:44] <Riddell> shadeslayer: kapidox has quite different packaging now
[14:44] <shadeslayer> :(
[14:49] <ScottK> It's a python app used for kf5 stuff, not a Qt based framework.
[14:49] <ScottK> So it's appropriately different.
[14:50] <shadeslayer> is it useful outside KDE?
[14:50] <Riddell> it makes sense to have python style packaging
[14:50] <apachelogger> ScottK: where would we get feedback on performance in 14.10?
[14:50] <Riddell> and it'll move to pkg-kde once pkg-kde people start to upload it
[14:51] <ScottK> I'll accept lack of complaining.
[15:20] <debfx> Riddell: deadlock scheduler? :)
[15:22] <Riddell> something like that
[15:22] <Riddell> deadlock is what this problem feels like
[15:24] <ScottK> Conversation on #ubuntu-kernel moved my opinion from "no" to "wait and see what happens after 14.10", so progress from your PoV.
=== vinay is now known as Guest87101
[16:50] <soee> ding gong
[16:50] <soee> i have a plan
[16:50] <sgclark> hiyas
[16:51] <soee> i will create fancy 1 page website that will explain what kubuntu is and promote it, there will be only basic and most importand infomration
[16:51] <soee> what do you think
[16:52] <soee> if new user ask what is Kubuntu what it can offer he can jupm there than and see in 'fancy' way
[16:52] <sgclark> with links to the userbase (somewhere) for help sounds good to me
[16:52] <soee> for more detailed info he will have to visit kubuntu.org
[16:52] <soee> this will be liek promotional poster
[16:52] <sgclark> gotcha
[16:53] <soee> multilanguage so all can easily read it
=== rdieter is now known as rdieter_work
[19:15] <_Groo_> hi/2 all
[19:15] <_Groo_> question
[19:16] <_Groo_> i backported plasma-network-manager and libnetworkmanagerqt to trusty
[19:16] <soee> nice :)
[19:16] <_Groo_> can i send the diffs to you guys?
[19:16] <_Groo_> to ninja ppa or whatevr?
[19:16] <_Groo_> for review?
[19:27] <soee> Riddell, sgclark, shadeslayer ^
[19:32] <shadeslayer> Tomorrow, busy getting drunk
[19:32] <sgclark> mail them to kubuntu-devel list?
[19:33]  * sgclark is also done for the day
[19:35] <soee> _Groo_: catch one of them tomorrow, they will know what to do :)
[19:53] <_Groo_> will do
[22:27] <shadeslayer> someone shrunk the tardis!
[22:27] <shadeslayer> https://lh5.googleusercontent.com/-YMIkYSJDD-E/VDW041va4FI/AAAAAAABBx4/6zYWINpkRuw/w886-h590/Flatline%2BSmall%2BTARDIS.jpg
[22:27] <shadeslayer> it's now smaller on the outside
[22:27] <shadeslayer> wait ..
[22:28] <sgclark> LOL
[23:11] <valorie> that's hilarious
[23:23] <valorie> hmmm, today's update/dist-upgrade on my 14.04 box not too good: http://paste.kde.org/pkbrnxcsu
[23:23] <valorie> I'm thinking of purging the neon5 ppa
[23:23] <valorie> :(
[23:24] <sgclark> neon... but just do dpkg -i --force-overwrite /var/cache/apt/archives/project-neon5-kio_0.0+git20141008.0321+neon4~faa423d~14.04_amd64.deb
[23:24] <sgclark> should work
[23:28] <valorie> http://paste.kde.org/pua7sog95
[23:30] <valorie> sgclark: ^^^
[23:31] <sgclark> looks successful or was it not?
[23:32] <valorie> I can't tell
[23:33] <valorie> so many warnings
[23:33] <sgclark> right, because it is overwriting.
[23:33] <sgclark> line 13 shows it setting up
[23:33] <sgclark> upstream released plasma 5 and then made major changes to frameworks and released, caused problems all over the place
[23:34] <sgclark> dunno who works on neon, but maybe notify them?
[23:36] <valorie> right now, no one is working on it
[23:36] <sgclark> eek
[23:36] <valorie> shadeslayer says it is unsupported
[23:37] <valorie> which is why I think I'll just purge it
[23:37] <sgclark> yeah if no one is working on it, that would be wise
[23:37] <valorie> let us get settled before the 14.10 release
[23:38] <valorie> and then take up testing again
[23:38] <valorie> seems like we're just jammed now
[23:40] <valorie> !ppa-purge
[23:40] <ubottu> To disable a PPA from your sources and revert your packages back to default Ubuntu packages, install ppa-purge and use the command: « sudo ppa-purge ppa:<repository-name>/<subdirectory> » – For more information, see http://www.webupd8.org/2009/12/remove-ppa-repositories-via-command.html
[23:43] <valorie> he, that won't work very well either
[23:44] <valorie> only wants to remove libepoxy and the newly added linux headers
[23:44] <valorie> pfff