[00:08] <apachelogger> ScottK: it woudl appear that I have no usable minions right now...
[00:08]  * apachelogger notes that shadeslayer did surprisingly little since UDS, making him think that maybe it was too son to go there... :(
[00:09] <Riddell> he's had exams
[00:09] <apachelogger> me too :S
[01:11] <apachelogger> http://aplg.kollide.net/images/avatar/snapshot117.png grantlee \o/
[03:12] <nigelb> um, what is sc in respect to kde?
[03:13] <ScottK> nigelb: Software Collection.
[03:13] <ScottK> nigelb: KDE is now the community and not the software.
[03:13] <tsimpson> KDE is like the group, KDE SC is the product
[03:14] <nigelb> ScottK: Ah, I read the article twice trying to work out what's SC :)
[03:14] <ScottK> Sorry.
[03:14] <nigelb> s/article/blog post 
[03:14] <nigelb> ScottK: Nah, its probably because I got used to hearing SC for software center :/
[03:15] <ScottK> Updated
[03:17]  * nigelb hugs ScottK :)
[03:20] <nigelb> Now, I remember again why I like the Kubuntu people a lot :D
[03:21] <apachelogger> hm
[03:21] <apachelogger> [Debug] You fool!!! This HttpDaemon does only supports 'GET'!!!! Go away!!!! You are so stupid for thinking that I am smarter than a tomato!!!
[03:21] <apachelogger> I wonder who wrote that debug message
[03:21] <apachelogger> doesnt even make sense
[03:21] <apachelogger> oh well
[03:21]  * apachelogger invokes the blame
[03:21]  * apachelogger hands nigelb a cookie while he is around
[04:17]  * nigelb hugs apachelogger :)
[09:17] <markey> hmm
[09:17] <markey> why do you refer to ARM as "Armel"?
[09:55] <Tm_T> markey: good question, apparently "Armel" refers to ARM emulator
[09:55] <kubotu> why do you refer to ARM as "Armel"?
[09:55] <Tm_T> kubotu: you repetitive you
[10:02] <agateau> markey: Tm_T: iirc armel is arm eabi little-endian
[10:02] <markey> ah, thanks
[10:02] <markey> the term is a bit confusing :)
[10:02] <markey> cause usually you would say, "We have this build for the ARM architecture."
[10:19] <Tm_T> agateau: ah, different sources gives different meaning apparently
[10:19] <agateau> Tm_T: :)
[10:19] <kubotu> I still don't get too excited :) you should still be able to test, should work fine
[10:20] <Tm_T> markey: ^ the bot owner might like to turn the bot less verbose
[10:20] <agateau> markey: yeah, except it potentially won't work on one half of the arm machines :)
[10:20] <agateau> (assuming big endian are as common as little endian arm, which I doubt actually)
[10:21]  * agateau is still a bit of an arm noob
[10:24] <markey> Tm_T: bot owner?
[10:24] <markey> kubotu: version
[10:24] <kubotu> I'm a v. 0.9.15 (master branch, revision a95675a [rss: watch handle case during rename]) [19 days, 17 hours, 30 minutes and 54 seconds ago] rubybot, (c) Tom Gilbert and the rbot development team - http://ruby-rbot.org
[10:24] <markey> it's not mine...
[10:26] <markey> agateau: hmm, I think ARM rocks. in fact I'd love to get rid of X86, and see ARM becoming the successor. in mobile space this has already happened, but Intel is far too huge for letting them grow much in the server space
[10:26] <markey> and desktop, it would only matter for Linux
[10:26] <markey> but still, ARM is great
[10:26] <markey> very very efficient
[10:26] <markey> and even the ISA is nice, programming it in Assembler is almost fun :)
[10:26] <markey> like in good ole times
[10:27] <agateau> markey: heh
[10:27] <agateau> I kinda agree, at least I am considering replacing my home server with an arm-based machine
[10:28] <markey> agateau: wait for the next-gen ARM, the "Eagle"
[10:28] <markey> it's going to rock
[10:28] <markey> (A 15)
[10:28] <agateau> and I was lucky to get an efika sb during uds
[10:28] <agateau> but that one is not replacing my laptop for now!
[10:29] <markey> I must say, I recently bought an HTC Desire phone (Android) with 1Ghz Snapdragon SoC
[10:29] <markey> it's insanely fast
[10:29] <markey> (for a phone)
[10:29] <markey> kinda amazing
[10:29] <agateau> what worries me a bit is that now that intel is working with nokia on meego, they may try to sneak atom there
[10:29] <markey> agateau: of course
[10:29] <markey> Intel is well on the way of shrinking Atom to 22nm
[10:29] <markey> and they will try to get it into phones
[10:30] <markey> but ARM might just be too good for them to tackle
[10:30] <agateau> unfortunately technical superiority is not always enough to succeed
[10:30] <markey> right
[10:30] <markey> Intel has better fabs than anyone else
[10:31] <markey> they can take crap ISA, through billions at it, make it work anyway
[10:36] <Tm_T> apachelogger: howdy
[11:05] <apachelogger> oggy oggy oggy
[11:07] <apachelogger> markey: snapdragon would be way faster if it didnt have to cope with crappy software
[11:07] <apachelogger> such as *cough* java
[11:08] <markey> apachelogger: Darlek is quite fast
[11:08] <kubotu> dont work with other install (although i don't know who tomwilso is :/
[11:08] <markey> google ain't daft
[11:08] <markey> but yeah, C++ with Qt could be even cooler
[11:09] <kubotu> plugs into the matrix and Trinity downloads it in main could be done for now. Thanks.
[11:09] <apachelogger> nah, only if done right
[11:09] <apachelogger> and since only 0.09% of Qt developers know how to do efficient Qt...
[11:09] <markey> nah
[11:09] <markey> nonsense
[11:10] <markey> modern GUI apps like e.g. Amarok, they do *no* number crunching at all
[11:10] <markey> we don't have any complex algorithms 
[11:10] <markey> all delegated to libs
[11:10] <markey> so that makes it very easy to write efficient code
[11:10] <apachelogger> yes
[11:10] <apachelogger> but
[11:10] <kubotu> we need to revert to native as the hashing functions I wrote yesterday, which consistenly only of a, b, i, r, x and k
[11:10] <apachelogger> if you do not know how to use the libs the sophisitcated algorithms within the libs will not help you
[11:10] <markey> apachelogger: ok, you can make some stupid mistakes, e.g. with blocking the GUI
[11:11] <markey> that is probably the #1 issue
[11:11] <apachelogger> like say people use qstandarditemmodel for every sort of model since it is a *standard* model
[11:11] <markey> with Qt apps
[11:11] <markey> but still, learning that is not rocket science
[11:11] <apachelogger> it is not just blocking really
[11:12] <markey> e.g. our very new contributor, Ralf Engels from Nokia, he has no issues at all with coding Qt stuff
[11:12] <markey> it's almost like he does it all day
[11:12] <apachelogger> if you haven't read krake's blog post about async app magic stuff, I can very much recommend this
[11:12] <markey> :p
[11:13] <apachelogger> well, actually, let me just expand the argument to the lot of software development in general ;)
[11:13] <apachelogger> like applications that initialize *cough* all scriptengine stuff in blocking manner, or *cough* initialize all their widgets in blocking manner...
[11:14] <apachelogger> running plasma-mobile on the n900 I only realized how much time the latter really takes
[11:14] <apachelogger> supposedly plasma could start in the blink of an eye if the widget startup was lazy initialized
[11:59] <ScottK> apachelogger: It's this kind of thing getting fixed could cause mobile work to help desktop performance too.
[12:05] <ScottK> New skrooge release someone should package up.
[12:36] <ScottK> Quintasan: I'm looking at sip4-qt3 again.  Did you fix the stuff POX told you to fix (sorry, forgot)?  Apologies also for taking to long to look at it.
[12:39] <maco> thats funny. afternoon utc, and no sign of the scot that isnt scott yet
[12:45] <ScottK> maco: Weren't you going to fix some armel FTBFS for me the other day?
[12:45] <maco> yeah...how abou tomorrow after my bio exam, since then i'll be on campus and thus have bandwidth for downloading/uploading packages?
[12:46] <maco> (and also because i'll be done with this week's finals)
[13:02] <ScottK> Great.
[13:19] <drdanz> Hi, I tried to create a package for KParts Plugin (http://kde-apps.org/content/show.php?content=125066) because I couldn't find it anywhere. Is it really missing or maybe it is packaged with some other name? If it is not packaged, can anyone review my package and tell me if I'm doing it right or if I am missing something?
[13:19] <kubotu> i was doing the dummy transition package and see ARM becoming the successor. in mobile space this has already happened, but Intel is well on my netbook
[13:20] <droidslayer> Riddell: coming to http://www.ubuntu.com/ubuntudeveloperday?
[13:20] <droidslayer> Riddell: coming to http://www.ubuntu.com/ubuntudeveloperday
[13:20] <droidslayer> Bah....
[13:22] <Riddell> hmm, might be a bit far
[13:22] <droidslayer> :P
[13:55] <ScottK> Quintasan: How do you determine the API version for sip4-qt3?
[14:14] <apachelogger> aloha
[14:14] <apachelogger> why do I never get invited to ubuntu developer days in india
[14:14] <apachelogger> it is a shame
[14:16] <Riddell> I believe it's an open invitation
[14:19] <JontheEchidna> hmm, I'm not sure I like the device notifier notifying me of me worker's windows shares...
[14:35] <apachelogger> JontheEchidna: same here
[14:35] <apachelogger> it is horrible on campus
[14:41] <ScottK> Can we default that to off in k-d-s?
[14:41] <kubotu> later tonight/tomorrow Can we talk about the right packages
[14:42] <Riddell> since 4.6 isn't out yet best thing would be to find the relevant upstream and moan at them
[14:42] <Riddell> apachelogger: is kubotu making random comments again?  is this done just to annoy me or for some greater purpose?
[14:47] <ScottK> Riddell: Isn't that enough?
[14:49] <Riddell> ScottK: isn't what?
[14:49] <ScottK> Being an annoyance enough of a reason?
[14:49] <Riddell> depends how evil apachelogger wants to be
[14:52] <Riddell> ScottK: "I'm going to create the first tarball for KDevelop 4.2 later today"
[14:52] <ScottK> Cool.
[14:52] <ScottK> Moar progress.
[15:26] <ari-tczew> kde-zeconf update failed. subprocess installed post-installation script returned error exit status 1
[15:27] <Riddell> ari-tczew: lucky you, a new bug to investigate :)
[15:27] <Riddell> agateau: git question
[15:27] <Riddell> if I have a revision adec5bbc4893b290395f8ced35bf597d0fdefdba how do I find the diff to the previous revision?
[15:27] <ari-tczew> Riddell: go ahead :P
[15:27] <agateau> git diff adec...^
[15:28] <Riddell> agateau: that shows me diff against current version
[15:28] <agateau> Riddell: ah right
[15:28] <agateau> git diff adec5bbc^..adec5bbc
[15:28] <agateau> Riddell: or simpler: git show adce5bbc
[15:28] <agateau> *adec5bbc
[15:29] <Riddell> perfect, thanks
[15:29] <agateau> yw
[15:40] <CIA-116> [muon] jmthomas * 1206713 * trunk/extragear/sysadmin/muon/installer/ (3 files in 2 dirs) (log message trimmed)
[15:40] <CIA-116> Fix a bug where a transaction cancelled due to an authorization error/cancelment
[15:40] <CIA-116> wasn't removed from the transaction stack, as well as a bug where the
[15:59] <CIA-116> [muon] jmthomas * 1206723 * trunk/extragear/sysadmin/muon/installer/ApplicationDetailsView/ (ApplicationDetailsWidget.cpp ApplicationDetailsWidget.h) Add a KRatingWidget next to the application icon/name/description in the header to display the popcon rating.
[16:01] <apachelogger> Riddell: very evil I want to be
[16:01]  * apachelogger wants to become a darth vader eventually
[16:01] <apachelogger> muahahahah
[16:01] <apachelogger> anyhow
[16:01] <apachelogger> matrix fun
[16:01] <apachelogger> o/
[16:01] <JontheEchidna> http://i.imgur.com/nQNyI.png
[16:07] <Riddell> JontheEchidna: what's new?
[16:07] <JontheEchidna> Rating widget
[16:07] <JontheEchidna> I'm fairly happy with the layout of the app pages now
[16:07] <kubotu> when bulldog98 is happy with it, and it'll go to go for
[16:07] <Riddell> ah hah, where does that go?
[16:08] <kubotu> Riddell: I'm offended!
[16:08]  * JontheEchidna giggles
[16:08] <JontheEchidna> The ratings come from the popcon scores, nothing too special.
[16:09] <Riddell> right
[16:10] <JontheEchidna> A standard 1-5 star rating system that users of the Ubuntu Software Center or Muon could submit their ratings to via the application would be better, though
[16:10] <JontheEchidna> dunno if the USC guys are planning on a non-popcon rating system
[16:11] <Riddell> yes I thought that was what they had/were doing
[16:12] <JontheEchidna> iirc they had some sort of rating system in right before lucid, but they pulled it due to code immaturity or something
[16:12] <JontheEchidna> doesn't seem to do ratings atm, at any rate: http://i.imgur.com/spoWF.png
[16:13] <Riddell> JontheEchidna: are you looking at the extras.ubuntu.com updates stuff?
[16:14] <JontheEchidna> I noticed a new URI in the sources.list for that in maverick, but it always 404'd for me
[16:15] <KRF> apachelogger: not sure who's responsible for the german announcements, but: http://www.kubuntu-de.org/nachrichten/software/programmierung/2045-kdevelop-4-1-1-mit-kubuntu-paketen-verfuegbar references the wrong ppa
[16:15] <Riddell> JontheEchidna: it should be there now, I believe there's some XML description of the contents
[16:16] <KRF> or Nightrose ^
[16:17] <Riddell> KRF: probably best to ask in #kubuntu-de
[16:20] <JontheEchidna> Riddell: so visiting the repo in chromium I noticed a "daily-journal" package. Is that what you were talking about?
[16:21] <JontheEchidna> oh, it it just like another partner repo?
[16:31] <ScottK> JontheEchidna: No.
[16:31] <ScottK> It's much scarier than that.
[16:31] <Riddell> JontheEchidna: it's the new partner repo but now has a process for free software without paying canonical
[16:31]  * ScottK looks for references.
[16:31] <Riddell> oh and ScottK disapproves :)
[16:32] <ScottK> JontheEchidna: https://wiki.ubuntu.com/PostReleaseApps
[16:35] <ScottK> Tm_T: http://cdimage.ubuntu.com/kubuntu/daily-live/20101215.1/natty-desktop-powerpc.iso
[16:44] <ScottK> Any reason (beyond no one did it yet) that kdepim-runtime is updated to 4.4.8 and kdepim isn't?
[16:46] <Riddell> mm, that should be in
[16:47] <ScottK> I'll have a look.
[16:52] <ScottK> Looks like it's in bzr, just didn't get uploaded.
[16:52] <Riddell> yes I guess so
[16:53] <Riddell> want to upload it?
[16:53] <ScottK> Sure.
[16:53] <Riddell> thanks
[16:53] <ScottK> I'll merge from Debian first.
[16:59] <JontheEchidna> Riddell: since extras.ubuntu.com behaves as a regular repository, muon will show it as one, so there shouldn't be anything preventing it from working
[17:00] <Riddell> JontheEchidna: right but there is (I believe) an XML file somewhere to say what's new so software centre can put it in its what's new section
[17:00] <ScottK> JontheEchidna: It would possibly make sense to put some extra warnings around it (as presumably you do for third party repositories)
[17:01] <JontheEchidna> Riddell: ah, I know the one you're talking about
[17:02] <JontheEchidna> bleh, lousy connection
[17:02] <JontheEchidna> Riddell: ah, I know the one you're talking about. It's /usr/share/app-install/menu.d/new.menu which is shipped in the software-center package
[17:03] <Riddell> that could be it
[17:03] <Riddell> so it has similar data as the .desktop files for other packages
[17:03] <Riddell> other applications I should say
[17:03] <JontheEchidna> actually, it's the featured.menu in the same folder
[17:04] <JontheEchidna> USC uses some sort of zeitgeist magic to determine new packages
[17:04] <JontheEchidna> so new.menu is basically just a shell that USC uses for making the "new" section GUI
[17:05] <JontheEchidna> http://pastebin.com/7dpXrrZ8 It's similar to the categories.xml files that USC, KPK and MSC use, but I don't think that we'd really want to use USC's since it's basically all gnomey apps
[17:06] <JontheEchidna> grr, connection dropped again. What was my last message?
[17:08] <Riddell> ...it's basically all gnomey apps
[17:11] <Riddell> hi kronos 
[17:11] <kronos> Riddell: hey. 
[17:11] <kronos> Riddell: any task for me ?
[17:19] <Riddell> kronos: not yet but kdevelop tars are due out later today
[17:25] <Riddell> kronos: don't have an ICQ account do you?
[17:25] <Riddell> we need some way of testing bug 676663
[17:25] <Riddell> apachelogger: any ideas how to test that?
[17:25] <kronos> Riddell: no ..
[17:27] <ScottK> Riddell: Maybe kronos could work on the MIR for libasyncns since shadeslayer is slacking.  No one else did it, did they?
[17:29] <Riddell> ScottK: I don't believe anyone has
[17:29] <Riddell> kronos: fancy doing that?
[17:29] <kronos> Riddell: u will have to guide me though ..
[17:31] <Riddell> kronos: review the package libasyncns for these requirements https://wiki.ubuntu.com/UbuntuMainInclusionRequirements
[18:11] <kronos> Riddell: i guess an MIR has been filed https://bugs.launchpad.net/ubuntu/+source/libasyncns/+bug/689766
[18:13] <ScottK> kronos: It would be good if you could review the MIR requirements and add the missing details.
[18:14] <kronos> ok
[18:23] <Riddell> afiestas: do you know the answer to this? 
[18:23] <Riddell> 18:21 < shane4ubuntu> Ok, I have a wireless mouse (trackball actually) in gnome it would alert me when the battery was low, is there any thing to do that in KDE?
[18:25] <maco> afiestas: nevermind. its not a bluetooth wireless mouse
[18:26] <afiestas> maco:  Riddell anyway if the usar can provide some dbus output I can help
[18:26] <Riddell> afiestas: he's in #kubuntu now
[18:26] <maco> afiestas: he's in #kubuntu if you want to ask about dbus
[18:36] <kronos> ScottK: a couple of problems i found. Standards-version is 3.7.2 instead of 3.8.4 . Maintainer is not @ubuntu.com and the debian/compat is 5 while debhelper recommends 7.
[18:37] <Riddell> I don't think those should be problems for main
[18:37] <ScottK> Agreed
[18:37] <nixternal> i can say this, my wireless mouse doesn't show battery status in gnome. i want that now :)
[18:37] <ScottK> Riddell: I'd take doko's suggestion to port to another library already in Main if he'll agree to maintain the port.
[18:38] <ScottK> nixternal: http://cdimage.ubuntu.com/kubuntu/daily-live/20101215.1/natty-desktop-powerpc.iso today thanks to your powerpc box.
[18:38] <nixternal> rock on with your badself!
[18:39] <ScottK> apachelogger did the patch for the only tricky one.
[18:58] <Riddell> kronos: bug 683439 would be interesting if you're wanting to do some package changes
[19:00] <kronos> Riddell: i'll try that one.
[19:02] <kronos> Riddell: the MIR request filed has a spelling mistake - at 2 places it says loudmoyth instead of loudmouth . does this matter ?
[19:03] <Riddell> kronos: shouldn't matter unless it's confusing to the reader
[19:04] <Riddell> kronos: kalgebra should be split into three packages, kalgebra-common, kalgebra and kalgebra-mobile
[19:13] <apachelogger> Riddell: get an icq account and try to login
[19:13] <apachelogger> unpatched kopete will fail to login, patched one will succeed
[19:13]  * apachelogger finds it a bit odd that no one has bothered to try that
[19:13] <apachelogger> maybe kopete got no icq users?
[19:14] <apachelogger> or no users at all?
[19:14] <apachelogger> Tm_T: ^?
[19:16] <apachelogger> neversfelde: <KRF> apachelogger: not sure who's responsible for the german announcements, but: http://www.kubuntu-de.org/nachrichten/software/programmierung/2045-kdevelop-4-1-1-mit-kubuntu-paketen-verfuegbar references the wrong ppa
[19:21] <neversfelde> apachelogger: I know
[19:21] <neversfelde> my quasselcore died and everyone who could change this is offline now :)
[19:36] <Riddell> NCommander: can you make out what we should do about this neon stuff in bug 664431 ?
[20:09] <kronos> Riddell: what should kalgebra-common and kalgebra-mobile contain ?
[20:15] <Riddell> kronos: kalgebra should contain /usr/bin/kalgebra and the associated .desktop file, kalgebra-mobile should contain /usr/bin/kalgebra-mobile and kalgebra-mobile.desktop
[20:15] <Riddell> kalgebra-common should contain everything else
[20:15] <Riddell> kalgebra-common should conflict and replace old kalgebra (as should kalgebra-mobile)
[20:15] <rbelem> Riddell, http://identi.ca/attachment/35886187
[20:15] <rbelem> apachelogger, ^
[20:16] <Riddell> oh hi rbelem 
[20:16] <rbelem> ScottK, ^
[20:16] <rbelem> :-D
[20:16] <Riddell> gosh, a little rbelem!
[20:16] <rbelem> :-D
[20:18] <rbelem> :-D
[20:21] <ScottK> rbelem: Congratulations.  Speaking as a parent of multiple teenagers I can honestly say you've no idea what you've signed yourself up for.
[20:22] <rbelem> ehehe :-)
[20:22] <rbelem> thx ScottK 
[20:22] <ScottK> You think I'm joking.  Check in with me in ~16 years.
[20:24] <rbelem> I remember ten years ago when i was teen. I gave lots os headaches to my mother :-)
[20:30] <highvoltage> ScottK: how old are you btw?
[20:31] <DarkwingDuck> 103
[20:31] <DarkwingDuck> ScottK: is 103
[20:34] <ScottK> highvoltage: 47.
[20:34] <ScottK> So DarkwingDuck was ~ right.
[20:34] <DarkwingDuck> hehehe
[20:34] <DarkwingDuck> love you too ScottK :P:P
[20:38] <kronos> Riddell: http://pastebin.com/s9bwuqRE
[20:38] <kronos> Riddell: still need to change synopsis .
[20:56]  * apachelogger dances around with rbelem \\o/
[20:56] <apachelogger> *hugs* cookies !!! unicorns!!!
[20:58] <apachelogger> kronos: how is liblikeback coming along?
[20:58] <kronos> apachelogger: couldnt do much after that .. had been out all day ..
[21:12] <kronos> apachelogger: http://pastebin.ca/2020734 . where should these files go ? liblikeback0 or the -dev package or both ?
[21:12] <apachelogger> certainly not both :D
[21:12] <apachelogger> then there would be a conflict ^^
[21:12] <apachelogger> the cmake files goes into -de
[21:12] <apachelogger> v
[21:12] <apachelogger> the png actually, ideally, would get their own package
[21:12] <apachelogger> -data for example
[21:13] <apachelogger> (with arch: all, so that this package only gets generated on one architecture, since it is architecture independent anyway)
[21:13] <ScottK> Or -common
[21:13] <apachelogger> or -share ;)
[21:31] <kronos> apachelogger: http://pastebin.com/JW1J6Jgc updated control file
[21:33] <apachelogger> kronos: description would not describe it to someone who does not know what likeback is about :)
[21:34] <apachelogger> also you can drop the section lines in liblikeback0 and -data (they will use the one defined for the source if none is defined, and the source is libs, so...)
[21:34] <kronos> any suggestions for description ??
[21:35] <apachelogger> not really, I am knee deep in interpolation stuff right now ^^
[21:35] <apachelogger> oh
[21:35] <apachelogger> kronos: liblikeback0 needs to depend on -data
[21:35] <apachelogger> -data neesd a depends line with misc:Depends and -dev also needs the misc stuff
[21:35] <apachelogger> (without misc debhelper grows every so grumpy with you and starts whining and whatnot ^^)
[21:36] <kronos> hmmm
[21:46] <kronos> apachelogger: hows this "Likeback is a user feedback tool that allows users of an application to communicate their opinions/liking of the application to its developers. It makes the developers understand and satisfy better the needs of the users."
[21:47] <ScottK> kronos: I'd say helps instead of makes.
[21:47]  * kronos is making notes.
[22:13] <NCommander> Riddell: *grumble* that should have been fixed
[22:22] <apachelogger> http://www.codinghorror.com/blog/2007/09/rainbow-hash-cracking.html
[22:22] <apachelogger> epic wallpaper he's got there
[22:23]  * rbelem hugs back apachelogger
[22:23] <rbelem> :-D
[22:26] <ScottK> NCommander: Read the bug.  It's a bit more complicated than we thought.
[22:26] <NCommander> ugh
[22:28] <ScottK> On a more positive note we fixed all the qreal/double stuff without anyone from the arm team needing to help.
[22:28] <ScottK> (for KDE 4.6 beta 2)
[22:29] <NCommander> ScottK: thanks (on the later), great work
[22:29] <NCommander> Riddell: ScottK: I'm not seeing the bug
[22:30] <NCommander> We're not going to provide a NEON compiled version of Qt.
[22:30] <NCommander> It breaks compatibility with a ton of boards
[22:30]  * ScottK doesn't recall which one it is, but thought Riddell linked you to it.
[22:30] <ScottK> NCommander: We can do most of it at runtime.
[22:31] <ScottK> NCommander: What supported boards does it break (I'm aware of one)?
[22:31] <ScottK> The proper fix for Natty is runtime detection, not disabling it.
[22:31] <NCommander> ScottK: Qt doesn't support dynamically loading a NEON version, does it?
[22:32] <ScottK> Sort of.
[22:32] <NCommander> ScottK: anyway, it breaks everything expect OMAP
[22:32] <ScottK> It's the sort of that's the tricky part.
[22:32] <apachelogger> thiago says it does
[22:32] <NCommander> (in theory, imx51 would work, but about half the hardware in the wild have buggy NEON implimations)
[22:32] <apachelogger> with patchy or something like that
[22:32] <NCommander> apachelogger: patchy?
[22:33] <apachelogger> there is runtime detection in some Qt branch (master or 4.7 not sure)
[22:33] <NCommander> Still doesn't help for cases with boards with buggy NEON
[22:34] <NCommander> The only way to do a NEON version at the moment is to use hwcaps, and provide a qt4-neon as a separate package
[22:34] <apachelogger> maybe introduce a blacklist?
[22:34] <NCommander> apachelogger: no sane way to autodetect
[22:34] <apachelogger> <3 hardware
[22:34] <NCommander> There are difficult logistics of providing NEON packages. 
[22:34] <NCommander> Right
[22:34] <NCommander> Read through the bug
[22:35] <NCommander> If runtime detection exists, and is known to work properly, no issues from me
[22:35] <NCommander> But a non-NEON confirmation MUST work, or its a high/critical bug.
[22:36] <NCommander> ^- ScottK apachelogger Riddell 
[22:37] <ScottK> NCommander: We didn't ask for your help to have you tell us the same thing ogra's already bitched about.  We'd like some actual help figuring it out.
[22:37] <NCommander> oh, you mean making autodetection out?
[22:38] <NCommander> s/out/work/g
[22:38] <NCommander> I thought ogra wanted to keep it on
[22:38]  * NCommander might be slightly exhausted
[22:38] <ScottK> NCommander: Since N900 is one of our targets this cycle, lack of Neon support is a real problem.
[22:38] <ScottK> He just wants Neon out.
[22:38] <NCommander> N900 has NEON
[22:38] <NCommander> oh
[22:38] <NCommander> d'oh
[22:38] <ScottK> Exactly.
[22:38]  * NCommander goes to drink coffee
[22:38] <NCommander> ScottK: TBH, I recommend doing a hwcaps version of NEON
[22:39] <NCommander> Even with the hand-optimized NEON code, a full compile with NEON will get you a much better performance increase
[22:39] <ScottK> NCommander: Find Thiago's discussion of it in the bug and help us sort it out.
[22:39] <NCommander> No, I see it
[22:39] <NCommander> Thigao's runtime detection would make specific hand-optimized functions be NEONed
[22:40] <ScottK> Gotta run.
[22:40] <NCommander> What you want for max performance is a full build of Qt in NEON, you have to do multiple packages of Qt to make that work
[22:40] <NCommander> ScottK: cya. glad to discuss more when you return
[22:40] <ScottK> apachelogger: Please discuss with NCommander.
[22:40] <Riddell> bug 664431
[22:40]  * apachelogger is doing homework!
[22:40] <ScottK> Riddell: kdepim done in Natty and the updates PPA.
[22:40] <apachelogger> NCommander: that is 48 hours of building right there
[22:40] <ScottK> BBL
[22:41] <Riddell> ScottK: lovely
[22:41] <NCommander> apachelogger: I didn't say it would be fun ;-)
[22:41] <ScottK> apachelogger: Don't forget we want the qws version too.
[22:41] <apachelogger> ^^
[22:41] <NCommander> ScottK: QWS?
[22:41] <apachelogger> ScottK: uhhh
[22:41] <apachelogger> NCommander: non-X11 Qt
[22:41] <NCommander> AHHHH
[22:41]  * NCommander jumps out of the nearest emergency window
[22:42] <apachelogger> 72 hours of Qt building
[22:42] <apachelogger> oh this is going to be fun :D
[22:42] <apachelogger> people will hate us
[22:42] <apachelogger> :D :D :D :D
[22:42] <apachelogger> love it
[22:42] <Riddell> what's this about QWS?  I haven't heard any suggestion that's going (back) in the archive
[22:43] <apachelogger> Riddell: qws is part of Qt
[22:43] <apachelogger> nowadays anyway
[22:44] <Riddell> yes I know, we had packages of it in maverick but it didn't get past new because people didn't want a second compile of Qt in the archive
[22:45] <apachelogger> well, I gather it is something important for mobile movement
[22:46] <Riddell> apachelogger: moderately, but the advise from upstream is to wait for Lighthouse. Our people only the binary incompatible bits rebuilt for QWS and that's fiddly to do so it remains in PPA only
[22:47] <apachelogger> hm
[22:48] <apachelogger> too bad, only 48 hours of building then :(
[22:49] <Riddell> NCommander: so for middling performance we can have the runtime neon detection?
[22:51] <NCommander> Riddell: it won't make as much as an improvement as your hoping I suspect
[22:51] <NCommander> apachelogger: if we can get the buildds to go TILT, it will be fun :-)
[22:51] <NCommander> (in all honestly, we probably need to make a new qt4-source binary package, then have stub packages which build the source instead of MEGAPACKAGE OF BUILDD DEATH)
[23:01] <apachelogger> well, that is coming with modularization anyway
[23:09] <Riddell> NCommander: but it's still an improvement, so we should do it?
[23:11] <NCommander> Riddell: TBH, I'd have to have someone look at benchmarks, but it would be a pain to do multiple builds
[23:11] <NCommander> and there's the usual maintence overhead, etc.
[23:12] <Riddell> NCommander: I'm not suggesting multiple builds, I'm suggesting the runtime stuff
[23:13] <NCommander> Riddell: oh, that
[23:13] <NCommander> sorry, my brain is not quite working
[23:14] <NCommander> I'm in favor of the runtime detection if it can be backported without breaking the world
[23:15] <Riddell> I just care about natty in the first instance
[23:38] <apachelogger> http://www.youtube.com/watch?v=sor9GzivGbk
[23:41] <Riddell> hmm, random youtube link
[23:41] <Riddell> apachelogger: what is it?
[23:41] <apachelogger> money
[23:43] <Riddell> you haven't convinced me to click it yet
[23:43] <Riddell> any more clues?
[23:43] <apachelogger> monty python's money song
[23:44] <Riddell> ooh, that's more tempting
[23:45] <apachelogger> ^^
[23:45] <Riddell> "KDevelop 4.2 Beta 1 ready to be packaged"
[23:45] <Riddell> where's Kronos when you need him?
[23:58] <rbelem> apachelogger, do you know the status of natty qt on arm?
[23:59] <rbelem> apachelogger, my micro sd class 10 arrived today :-)
[23:59] <apachelogger> rbelem: didnt work with meego kernel when I tried it last week