**** BEGIN LOGGING AT Sat Feb 26 02:59:57 2011 Feb 26 04:03:55 rpms for facebook status support in netbook 1.1 Feb 26 04:04:05 https://dc414.org/~ngharo/meego/meego-facebook-plugins-1.1.1-1.i386.rpm Feb 26 04:04:12 https://dc414.org/~ngharo/meego/meego-facebook-plugins-bisho-1.1.1-1.i386.rpm Feb 26 05:07:56 êàçàõëàð áàëëàð ìà? Feb 26 05:07:58 ýýé Feb 26 05:08:03 êûçäàðàé Feb 26 05:34:06 Can I do these steps on all distributions where I install the program 'osc' ? http://wiki.meego.com/Build_Infrastructure/Packagers_Developers/CLI_Part_1 Feb 26 07:02:57 http://lists.meego.com/pipermail/meego-commits/2010-November/009569.html :D Feb 26 07:09:01 slonopotamus: i can't help but feel that this guy was ironic Feb 26 07:09:10 http://twitter.com/timmartin2/status/23365017839599616 ;) Feb 26 07:10:04 slonopotamus: part of example to state your package ownership :P Feb 26 07:14:00 ar: mwahaha :) Feb 26 07:53:49 morn Termana Feb 26 07:53:54 morning Feb 26 13:23:57 hello, Stskeeps, may you can answer me this, who is creating the weekly trunk images? Is tis some kind of automatic process triggered by OBS? Feb 26 13:25:06 jbos: semi-automatic Feb 26 13:25:11 it's done with mic2 Feb 26 13:25:55 ok, we would like to start creating a automatic testing system / workflow Feb 26 13:26:28 to improve our meego libs / keep them working y' know. Feb 26 13:26:55 what we wondered, on public obs - is there a way to let it produce images? Feb 26 13:27:29 well, obs just builds packages and makes repositories Feb 26 13:28:52 yea but shouldn't it be able to produce complete distros? Feb 26 13:29:14 so but than - i can see that if many people do this :D Feb 26 13:29:39 it exports repositories and then you use a image creator make the repositories into a image Feb 26 13:29:59 the idea for more business logic is the BOSS system Feb 26 13:30:22 mhm yep I read about this, is it already running? Feb 26 13:30:28 i think in some form Feb 26 13:30:34 i mean, mic2 is cron-able already Feb 26 13:32:42 i think i still miss some links between OBS, and BOSS and how this fit in the 'app/lib' development workflow Feb 26 13:32:52 right Feb 26 13:32:52 :) Feb 26 13:33:06 so, BOSS listens to events from OBS and helps to make business logic on top of this Feb 26 13:33:24 like continous integration, etc, building images, etc.. Feb 26 13:33:32 will BOSS be on 'user account basis?' Feb 26 13:33:52 don't know :) Feb 26 13:34:01 so everyone with a OBS account will be able to configure his own BOSS ( .... :D) Feb 26 13:34:05 http://wiki.meego.com/BOSS Feb 26 13:35:50 k, well btw. finaly we managed to setup a fully crosscompiling (arm and i86) toolchain with realtime rpmizsation, deploy and debug on meego - from within the qt creator Feb 26 13:36:00 cool Feb 26 13:36:08 took us some brain Feb 26 13:36:13 contribute to sdk work? ;) Feb 26 13:36:25 its all possible with the current sdk Feb 26 13:36:33 ah Feb 26 13:36:38 yeah, that was what i wondered :P Feb 26 13:37:11 well, point is - its really just a single line to go... we have a setup of 4 projects Feb 26 13:37:18 2 seperate QML Uis Feb 26 13:37:25 2 libraries Feb 26 13:37:28 :nod: Feb 26 13:37:31 blog about your experiences? Feb 26 13:38:24 Yeap. Started here: http://www.jeremias-bosch.com/?p=79 Feb 26 13:38:44 last bits about deployment just finalizing Feb 26 13:39:12 but.... well your options to go are very very stricted to basically 1. Feb 26 13:39:14 :) Feb 26 13:39:33 use qmake, use own sysroot, use linux Feb 26 13:39:49 (I dont get it working in windows) Feb 26 13:39:51 i think there's a OBS plugin Feb 26 13:40:05 there is ? Feb 26 13:40:12 that would be more than nice Feb 26 13:40:29 http://blog.qt.nokia.com/2011/01/21/qt-creator-build-service-plug-in/ Feb 26 13:40:30 i actually considered to make one Feb 26 13:41:02 this is cool Feb 26 13:41:06 need to take a look Feb 26 13:41:22 one thing, we found very much not funny Feb 26 13:41:43 the gcc version in obs native builds != the gcc in the toolchain Feb 26 13:41:46 yes Feb 26 13:41:53 that one is a screwup in 1.1 Feb 26 13:42:00 in 1.2 the situation is much better afaik Feb 26 13:42:06 we got a bit scarred about this ;) Feb 26 13:42:55 i heard the rumor that 1.2 will provide meegocore - full dev lib - toolchain / sysroots Feb 26 13:43:02 is that correct? Do you know? Feb 26 13:43:16 possibly, but ask #meego-sdk guys during the work week days Feb 26 13:44:02 my worry is scratchboxisms :) Feb 26 13:44:52 like the software depending on things that are in sysroot but not part of meego api, so it doesn't declare dependancies Feb 26 13:45:05 but i guess that the RPM's ld.so based dependancies might take care of it.. Feb 26 13:45:11 (libGLESv2.so.2 Requires:) Feb 26 13:45:31 so people don't declare correct deps in the spec for building Feb 26 13:45:41 does someone an rpm arm package on archlinux host ? Feb 26 13:46:10 mhm Feb 26 13:46:41 You know I see several people trying to work on meego apps Feb 26 13:47:07 and we tried a lot of different approaches ourself Feb 26 13:47:17 is it possible to build an rpm package on arch host ? Feb 26 13:47:36 djszapin900: ask in archlinux irc channel? Feb 26 13:48:11 ... from both projects: Firefox Mobile and Peregrine Communicator..... Both with totaly different build requierments... Feb 26 13:48:13 jbos: yeah - we need a straightforward development story Feb 26 13:48:19 jbos: firefox mobile is just crazy to build Feb 26 13:48:20 :P Feb 26 13:48:49 well I work on that little babe since about 2 Years or so... Feb 26 13:48:58 it improved a lot Feb 26 13:48:59 :) Feb 26 13:49:01 stskeeps ask on both ? Feb 26 13:49:07 Stskeeps: jbos: The downside of a "straightforward development story" is usually gooey (GUI)... Feb 26 13:49:11 here and there ?! Feb 26 13:49:29 djszapin900: people here aren't likely to know if you can build rpms for arm on archlinux Feb 26 13:49:30 Stskeeps: jbos: And I like my terminal. :-) Feb 26 13:49:32 djszapin900: so it's offtopic :) Feb 26 13:49:36 gabrbedd: osc build ;p Feb 26 13:49:52 stskeeps meego package making is oo topic Feb 26 13:49:54 Stskeeps: Then I have to be connected to a network to do anythin. :-p Feb 26 13:49:59 gabrbedd: not really Feb 26 13:50:01 * off Feb 26 13:50:03 well yes,... the point is, both options must work, terminal and gooey Feb 26 13:50:10 time to ignore... Feb 26 13:50:21 terminal for i.e. firefox mobile Feb 26 13:50:33 Anyway... I'm just putting in a word for us CLI geeks. :-) Feb 26 13:50:34 djszapin900: how mature Feb 26 13:50:36 you wont be able to build firefox mobile with qt creator... Feb 26 13:50:59 but you would also not like to build firefox mobile within a freaking qemu Feb 26 13:51:04 arm thingy Feb 26 13:51:52 its slow, unstable and very scary... So what options do you have left? If you need to use a terminal Feb 26 13:52:13 ... OBS, and i.e. your native N900 :D Feb 26 13:52:46 comming back to Firefox - it take about 20 minutes to compile on my quad core i7 Feb 26 13:52:51 with HT Feb 26 13:53:00 obs ARM building is really nice Feb 26 13:53:20 guess how long it take on a n900 with meego running *laugh* Feb 26 13:53:27 :o Feb 26 13:54:06 jbos: well, you'd have to be insane to do it on a n900 ;) Feb 26 13:54:27 yeap... so currently we use some very bad hackish scratchbox kind of thing Feb 26 13:54:34 jbos: why, though? Feb 26 13:54:51 i mean, osc build and osc chroot is very capable for this Feb 26 13:55:00 mhm Feb 26 13:55:08 rapid development? Feb 26 13:55:14 still? Feb 26 13:55:27 i mean, usually i just have a chroot laying about with cross compilers already installed Feb 26 13:55:42 ok. I need to consider this Feb 26 13:55:50 ie, run osc build once, re-use chroot Feb 26 13:56:49 i mean, it'd feel like a really fast arm machine :) Feb 26 13:56:54 to build on Feb 26 13:57:13 i think we have basically the same this scratchbox thingy is somekind of chroot ... well duno. most times its enough to build for desktop Feb 26 13:57:40 i mean, that's how we build fennec-qt for arm :) Feb 26 13:58:39 mhm yea. Lokesh is mostly taking care about all meego integration :) Feb 26 13:58:59 so you guys were basyskom or how was it? Feb 26 14:00:01 well I'm employee there, yep. But we are a group of people from ixonos, digia, nokia, basyskom, mozilla Feb 26 14:00:04 :nod: Feb 26 14:00:29 working on getting the best mobile browser on meego Feb 26 14:00:31 :) Feb 26 14:01:24 right now, well, its hard work to shift most responsibilities to mozilla, upstream our changes and patches Feb 26 14:01:54 getting rid of non meego core dependencies Feb 26 14:01:58 like meegotouch Feb 26 14:02:18 (and still keep functionality) Feb 26 14:02:19 :) Feb 26 14:03:05 well and my second project - which i do in my spare time is peregrine communicator Feb 26 14:03:54 We get sponsored by basyskom (server, website, hardware (like weetab, n900, ideapads...)) Feb 26 14:04:13 but no working time :( ;) Feb 26 14:31:13 nate@blue[1014]:~/meego/opt $ ls Feb 26 14:31:14 openmpi--sysconfdir= Feb 26 14:31:25 that kinda explains why it failed on the OBS Feb 26 14:33:38 Is there now a Trunk:HPC group in OBS? :-p Feb 26 14:39:45 gabrbedd: that doesn't surprise me to be honest Feb 26 14:41:29 I'm curious, though... why package openmpi for meego? Feb 26 14:42:10 * CosmoHill shurgs Feb 26 14:42:26 practice run? boredom? running low on brain power at 11pm Feb 26 14:44:11 ...just curious, that's all. :-) Feb 26 14:44:42 gabrbedd: everything is collapsed into Trunk anyway Feb 26 14:59:35 * CosmoHill takes a breath Feb 26 14:59:39 FFFFFUUUU..... Feb 26 15:00:44 there wasn't a space between "--prefix=/foo--sysconfig=/etc" Feb 26 15:05:09 salut DawnFoster Feb 26 15:05:21 hey CosmoHill Feb 26 15:08:56 http://lists.meego.com/pipermail/meego-commits/2011-February/017283.html Feb 26 15:09:01 * Stskeeps looks Feb 26 15:09:15 that's nice :) Feb 26 15:10:10 Emgd for 2.6.37. Feb 26 15:10:29 at least gma500 people are getting some support :) Feb 26 15:12:57 * CosmoHill gasps Feb 26 15:12:59 Will try a joggler adaptation. Feb 26 15:13:01 I think I made a package Feb 26 15:14:35 were are the finished .rpm files? Feb 26 15:16:42 Report.pub.meego.com Feb 26 15:16:55 Repo. Feb 26 15:17:12 I can't seem to find it Feb 26 15:18:18 ½. Feb 26 15:21:14 oops Feb 26 15:21:20 hena: ½ ? I call Feb 26 15:21:26 hmm, I found a link to the download location and it says 404 Feb 26 15:28:59 * CosmoHill pokes X-Fade Feb 26 15:32:25 CosmoHill: You'll be lucky. bergie & X-Fade seem to have dropped off the face of the planet (except for bergie's regular Twitter updates) Feb 26 15:35:21 X-Fade's in the us Feb 26 15:35:25 afaik Feb 26 15:37:00 I don't seem to have a download repo of my own Feb 26 15:37:06 even tho my package built Feb 26 15:38:25 what's your user name there ? Feb 26 15:38:55 "cosmo" Feb 26 15:39:07 yes, its not there Feb 26 15:40:12 normally it would be in /home:/cosmo Feb 26 15:40:31 http://repo.pub.meego.com//home:/cosmo/meego_1.1_extras_Netbook Feb 26 15:40:38 that's the link it gives me Feb 26 15:40:51 Don't know if my last message was sent as I'm mobile but it was reported earlier I think that repo publishing was us Feb 26 15:41:25 vgrademk: last thing I heard from you was at 15:15 Feb 26 15:42:54 You may need to ping lbt Feb 26 15:43:26 I'd imagine he's doing birthday stuff this weekend Feb 26 15:43:37 or was it last week Feb 26 16:36:14 how come auke has two repos and I don't have any? Feb 26 16:41:15 CosmoHill: auke is more special than you are ? Feb 26 17:18:37 how do distros know if ethernet is connected? dbus? Feb 26 17:24:12 :( someone could have told me the meegon was missing from other pages Feb 26 17:37:00 * timeless_xchat ponders Feb 26 17:37:19 how painful is it to set up obs? Feb 26 17:37:25 is there an rpm for it? :) Feb 26 17:38:09 appliance Feb 26 17:40:53 timeless_xchat: it is still pretty painful - setting up the OBS is only about 10% of the actual work you need to do Feb 26 17:41:28 http://wiki.meego.com/User:Stskeeps/10_easy_steps_to_a_local_OBS Feb 26 17:41:36 (may be a bit out of date) Feb 26 17:41:49 yeah, set 1/10 is "set up the OBS" Feb 26 17:42:00 stskeeps: i'd want it to sandbox pain for mxr Feb 26 17:42:09 timeless_xchat: why do you need obs for mxr.. Feb 26 17:42:13 i'm not nesting a vm into a vm Feb 26 17:42:35 stskeeps : rpmbuild -bp seems to randomly run configure for various rpm's Feb 26 17:42:42 which seems absolutely bogus Feb 26 17:42:48 but there you have it Feb 26 17:42:51 you don't have to run it in a VM, but you basically do have to install opensuse Feb 26 17:43:08 this is a detailed version of step 1: http://wiki.meego.com/User:Ali1234#Notes_on_Setting_Up_Private_OBS_.28WIP.29 Feb 26 17:43:09 timeless_xchat: why you do -bp ..? Feb 26 17:43:13 i never got any further than that Feb 26 17:43:32 stskeeps: i want patched sources Feb 26 17:43:43 otherwise what you have isn't what you're running Feb 26 17:44:00 why not just grab debuginfo then? Feb 26 17:44:01 :P Feb 26 17:44:06 mhm the Meego SDK Linux Setup of Intel Tablet UX does not support opensuse... :( Feb 26 17:44:09 hi Feb 26 17:44:13 ab: hi ! Feb 26 17:44:27 vlj, Feb 26 17:44:29 debuginfo!=sources Feb 26 17:44:33 hey timeless_xchat, how are you. Feb 26 17:44:41 it's just references to line numbers Feb 26 17:44:47 jbos: pretty good Feb 26 17:44:50 timeless_xchat: debuginfo contains sources in meego Feb 26 17:44:57 timeless_xchat: for gdb'ing Feb 26 17:44:59 stskeeps : wtf? Feb 26 17:45:07 timeless_xchat: in opensuse, they have debugsource Feb 26 17:45:44 stskeeps: got a url for a sample rpm? Feb 26 17:46:05 timeless_xchat: repo.meego.com somewhere, too lazy to find it Feb 26 17:46:15 jbos: i hit your "missing time on debian" issue when setting up mxr.meego.com Feb 26 17:46:21 but it's mostly up Feb 26 17:46:32 including a meegon! Feb 26 17:46:33 ab: quillimage does not allow to store image in a 32bits format ? Feb 26 17:46:56 timeless_xchat i was checking latley mxr but it seems to miss certain "symbole click" feature Feb 26 17:47:04 which i like the most about mxr :) Feb 26 17:47:15 jbos : should be working now Feb 26 17:47:23 at least temporarily Feb 26 17:47:40 k, yea i tried to understand how meegotouch is doing rotation of apps... Feb 26 17:47:56 i didn't build the xref initially because i wasn't happy w/ the source extraction story Feb 26 17:47:57 to copy certain missing steps to firefox mobile... Feb 26 17:48:06 ye Feb 26 17:48:08 i see Feb 26 17:48:42 what's the current meego n900 build? Feb 26 17:48:49 vlj, as it is using qimage, it does allow it. Feb 26 17:49:05 just finaly cron'd peregrine mxr today Feb 26 17:49:20 ab: qimage stores image as 8 bits rgb Feb 26 17:49:23 still want to improve the html code / layout Feb 26 17:49:36 :) want to have it more web2.0 ish Feb 26 17:49:39 it is possible to store float rgb value instead ? Feb 26 17:50:17 jbos: there are some temporary patches on mxr.meego.com which might interest you Feb 26 17:50:22 right now it looks a bit to oldschool ( www.jeremias-bosch.com/mxr ) Feb 26 17:50:33 vlj, QImage::Format_RGB32 or ARGB32 Feb 26 17:51:01 vlj, but no float rgb Feb 26 17:51:21 yep, have a buglist? Feb 26 17:51:22 ab: qt can then convert them into 8bit rgb ? Feb 26 17:51:27 so i can take a look Feb 26 17:51:29 see .hg/patches Feb 26 17:51:29 i finally refactored some of it Feb 26 17:51:29 especially the tree switcher Feb 26 17:51:47 vlj, that's one limitation as QImage is mostly about painting something before drawing it to screen. THey are not really about image processing Feb 26 17:51:59 not really, patches tend to have vaguely meaningful names and one line hg commit messages Feb 26 17:52:08 ok Feb 26 17:52:21 i have perhaps a dozen i'm close to pushing to hg proper Feb 26 17:52:29 maybe tomorrow Feb 26 17:52:34 vlj, I'm working on a new qt imaging module but it will take some time before code becomes usable Feb 26 17:52:45 especially with current circumstances Feb 26 17:52:49 ok Feb 26 17:52:50 the tree switcher one will help you Feb 26 17:53:12 it'll automatically hide the tree switcher if you only have one tree Feb 26 17:53:17 ab: I was asking because I'm using float array to store hdr image right now Feb 26 17:53:31 I would like a "better integrated" way of storing image Feb 26 17:53:58 i see Feb 26 17:54:08 ab: btw are image in QImage stored in row major or column major way ? Feb 26 17:54:32 i will take a deeper review of those patches tomorrow or so :) for today my server can sleep ;) Feb 26 17:55:45 * timeless_xchat nods Feb 26 17:56:14 you'll definitely want to use mq - so you can qpop+qdel when they land upstream Feb 26 17:56:34 yep, Feb 26 17:57:55 vlj, they are row major Feb 26 17:57:56 ab: qt imaging should be an "official" part of qt or ? Feb 26 17:57:58 ok Feb 26 17:58:17 vlj, that's unknown yet. Depends on what we'll achieve :) Feb 26 17:58:45 and btw Feb 26 17:59:17 can qt-mobility Camera part use webcam on desktop computer ? Feb 26 18:00:11 it should be possible, barring obvious limitations Feb 26 18:01:01 because I'm lacking samples to test the accuracy of my hdr routine Feb 26 18:01:09 jbos : btw, your readme could use some line breaks ;) Feb 26 18:01:56 :D the readme need some update Feb 26 18:01:58 right now I'm using images from links here http://pages.cs.wisc.edu/~lizhang/courses/cs766-2007f/projects/hdr/index.htm Feb 26 18:02:24 but they are probably not licenceable to run unit test for instance Feb 26 18:02:59 its totaly out of sync after last week jump from cmake to qmake... i just see the guys left cmake files ;) Feb 26 18:02:59 yep Feb 26 18:03:44 hrm, there's something wrong w/ the source header, it has an extra blank area Feb 26 18:03:54 need to figure out why Feb 26 18:03:56 I'm thinking of using my N8 to get some picture using qt -mobility but it would be easier if qt-mobility can take images from webcam Feb 26 18:04:15 (no stabilising issue) Feb 26 18:06:36 ab: hdr construction and compression works acceptably well using jpeg Feb 26 18:06:57 it came to me as a surprise but using the jpeg sample produces correct image Feb 26 18:07:04 there are some minor artifacts however Feb 26 18:07:13 interesting Feb 26 18:07:16 * timeless_xchat breaks the index and cries Feb 26 18:07:27 you may probably take some CC images from Flickr for unit tests Feb 26 18:09:29 good suggestion Feb 26 18:11:00 jbos : oh, you might not have noticed, but look at the bottom of e.g. http://mxr.meego.com/meego.gitorious.org/source/qemu/qlist.c#158 Feb 26 18:11:38 Jaffa: I'm in Vienna so connectivity to the infra is a bit limited, and X-Fade is in the US Feb 26 18:12:37 jaffa: what's up? Feb 26 18:12:51 hey bergie, whatcha doing in vienna? Feb 26 18:13:25 timeless_xchat: http://aloha-editor.org/wiki/index.php/Aloha_Editor_Dev_Con_11 Feb 26 18:13:33 hacking semantic web editing tools Feb 26 18:15:51 * timeless_xchat settles for hopping off a train in search of a movie night Feb 26 18:19:57 ab: thank for information :) Feb 26 18:19:59 brb Feb 26 18:53:37 bergie: Can you poke someone with connectivity then about ticket #1981? Feb 26 19:17:39 MeeGo_current_Core_standard is not a valid repository, use one of: meego_1.1_core Feb 26 19:17:41 huh??! Feb 26 19:23:55 gabrbedd: yesterday you asked about whether i saw waveform display error in mixxx 1.9 on meego 1.2 -- answer: YES Feb 26 19:24:50 gabrbedd: i tried several different things that had fixed the problem in 1.8 and earlier, but it didn't work in 1.9 -- as you reported as well Feb 26 19:28:16 npm: I figured that you were seeing them... I was just curious if you were using OpenGL ES ? Feb 26 19:28:45 npm: Before, you had talked about switching to software rendering to get it to work. Feb 26 19:29:30 i tried using 'raster' 'opengl' and 'x11' systemgraphics per http://labs.qt.nokia.com/2008/10/22/so-long-and-thanks-for-the-blit/ and nothing helped Feb 26 19:30:10 gabrbedd: that was for 1.8 on fedora... 1.9 seems to have fixed the problem on fedora but now meego is broken Feb 26 19:31:08 IMHO what's going on is this: WaveformViewerFactory :: Sharing existing GL context. Feb 26 19:31:27 after doing WaveformViewerFactory :: Making new GL context. Feb 26 19:31:42 as in it's probably not even using the same graphics context it's drawing into Feb 26 19:31:51 npm: so when you said "meegolem hack" you were pretty much using a Fedora Qt setup (with full Open GL). Feb 26 19:31:53 ? Feb 26 19:31:55 i always end up seeing dropping from other screens in there Feb 26 19:32:24 actually the first time i compiled with fedora qt devels but meego libs (bad idea) Feb 26 19:32:42 the second time i removed all fedora qt and used only meego 4.7.1 qt devel and libs Feb 26 19:32:47 npm: lol! yes, bad idea. :-) Feb 26 19:32:48 and it still happened Feb 26 19:33:04 it actually worked fine.... Feb 26 19:33:33 the only thing that puked was mixing meego and fedora libtag Feb 26 19:34:05 but i decided to clean it up and remove as much fedora stuff as possible and still have all the apps run Feb 26 19:34:17 so, what did you do to make the waveforms /work/ ? I'm still confused. Feb 26 19:34:42 they don't work Feb 26 19:34:47 thats what i was confirming Feb 26 19:34:54 OH! :-) Feb 26 19:34:58 the trick i mentioned worked in fedora previously Feb 26 19:35:11 but i tried it various differenrt ways and it didn't work on meego Feb 26 19:35:25 and yes it has to do with GL graphics contexts being used improperly imho Feb 26 19:36:19 npm: Well, FWIW... I'm starting to think it's a bug in Qt. Feb 26 19:36:24 so feel free to pester the MIXXX people into fixing the bug... i've done enough pestering for the day :-) Feb 26 19:36:38 ...but the problem is that I can't isolate the bug. Feb 26 19:36:57 I've pulled all the waveform code out of Mixxx and into a test app... and the code works fine. Feb 26 19:37:08 it could be... i guess the other experiment i could try is completely overwrite meego's qt with fedora's Feb 26 19:37:26 but i'm trying to make stuff work, not break everything these days :-) Feb 26 19:37:38 No, I don't recommend doing that. Feb 26 19:37:42 :-) Feb 26 19:37:49 you can... it is interesting Feb 26 19:38:20 my least proud moment of meegolemization: lrwxrwxrwx 1 root root 10 2011-02-25 21:35 /usr/lib/python2.7 -> python2.6// Feb 26 19:38:41 but i just installed gpodder and all it's python modules and everything works beautifully Feb 26 19:38:45 Well, the code works for everyone else (with regular OpenGL)... and it's the only thing really different on MeeGo. Feb 26 19:39:32 one thing i've been thinking about is upgrading meego's python to 2.7 instead of using the symlink to make meego and fedora's python modules work together Feb 26 19:39:53 The EGL/GLESv2 code in Mesa and QPaintEngine (AFAIK) haven't had nearly as much testing as the regular OpenGL libs. Feb 26 19:40:17 For example, if you drag and drop a track to load it... Feb 26 19:40:29 Then mcompositor (meegotouch-compositor) crashes. Feb 26 19:40:39 That's the window manager... so EVERYTHING crashes. Feb 26 19:40:52 worksforme... Feb 26 19:40:54 That was an unchecked pointer in Mesa's EGL driver. Feb 26 19:41:14 i was dragging and dropping files in mixxx for several hours last night Feb 26 19:42:00 I believe you... but it still crashed for me and I was able to get the smoking backtrace. :-) Feb 26 19:42:07 the main instability I saw is using ALSA rather than jack... usual timing wonkiness from not running realtime Feb 26 19:43:19 oh, i have the smoking backtraces: -rw------- 1 root root 7417964 2011-02-26 11:26 /var/log/messages Feb 26 19:43:37 7 megabytes of them. Feb 26 19:43:42 No, I meant for the crash. :-) Feb 26 19:43:51 ...the one specific crash. Feb 26 19:45:18 speaking of crashes -- http://nielsmayer.com/meego/netbook-ia32-1.2/meego-networking-fails-syslog.txt Feb 26 19:46:29 npm: are you using clutter? (the netbook ux) Feb 26 19:47:29 yes Feb 26 19:47:47 that's why dragging didn't crash for you. Feb 26 19:48:06 so what are you running? Feb 26 19:48:44 isn't indamixxx based on the netbook ux or does it have its own? Feb 26 19:49:37 Up to January we were netbook. And right now the beta program tablets will come with netbook. Feb 26 19:49:59 r u using intel tablet ux? Feb 26 19:50:00 But netbook is not for tablets... and the devs say it will never be for tablets. Feb 26 19:50:05 And we're selling tablets. Feb 26 19:50:18 Not yet on the Intel Tablet UX. Feb 26 19:50:37 I switched over to handset to do what... well, they did. Feb 26 19:50:54 yes the netbook ux is not optimized for tablet usage... for example, you unpin the icons when you hit them with your finger instead of clicking them Feb 26 19:51:06 I haven't switched over my dev environment yet... since I see that they're using mostly the same stack that I'm using. Feb 26 19:51:22 the whole idea of a button panel that appears when you mouse over it completely fails on a tablet Feb 26 19:51:28 stack: meego touch framework (mcompositor window manager instead of clutter) Feb 26 19:52:25 ali1234: Well, believe it or not we've gotten good at dragging our fingers over to get the pin to appear. :-) Feb 26 19:52:25 so when i'm using 'mixxx' with netbook ux -- it's not using opengl? Feb 26 19:52:33 and in your case it is? Feb 26 19:52:44 gabrbedd: me too but it's not exactly something you would put in a product you expect people to buy Feb 26 19:53:07 ali1234: Right! Feb 26 19:53:23 without doubt the tablet UX works better on a netbook than the netbook UX works on a tablet Feb 26 19:53:37 npm: Yes, and I think Trunk is even using OpenGL ESv2. Feb 26 19:53:47 npm: It's just not using mcompositor for the window manager. Feb 26 19:53:51 how would i check? Feb 26 19:54:04 check for mcompositor? Feb 26 19:54:26 * npm reads xdpyinfo Feb 26 19:54:31 yes Feb 26 19:54:44 ps aux | grep mcompositor Feb 26 19:55:15 ah thans Feb 26 19:55:17 thanx Feb 26 19:56:20 btw, your wifi crashes look like they're happening in the kernel. Feb 26 19:56:24 Have you filed a bug? Feb 26 19:56:26 yep Feb 26 19:56:43 i attached it to previous bugs i'd filed Feb 26 19:56:50 ath9k looks problematic Feb 26 19:57:15 but also having problems w/ wired Feb 26 19:57:58 and tracker seems to be involved somehow as well (migbht be a secondary issue) http://nielsmayer.com/meego/netbook-ia32-1.2/tracker-banshee-kernel-fail.txt Feb 26 19:58:11 https://bugs.meego.com/show_bug.cgi?id=13778 Feb 26 19:58:14 Bug 13778 nor, Undecided, 1.1.90.4, martin.xu, ASSI, Wireless connection loses "router" setting for Static-IP Feb 26 19:58:27 https://bugs.meego.com/show_bug.cgi?id=13780 Feb 26 19:58:29 Bug 13780 nor, Undecided, 1.1.90.4, martin.xu, ASSI, Wired connection stops working when transferring lots of data Feb 26 19:59:08 i just turned off tracker so i'll see if that has something to do with it... Feb 26 19:59:44 ok. Feb 26 20:04:21 gabrbedd: by the way what's your fav qt skin for apps... i'm finding the meego one makes existing interfaces and dialogs not fit but 'plastique' does and isn't bad to use with the finger Feb 26 20:05:28 http://nielsmayer.com/meego/qtractor-dialogs-meego.png vs http://nielsmayer.com/meego/qtractor-dialogs-plastique-style-meego.png Feb 26 20:05:45 npm: I've always like plastique. What I think is interesting, though... is how it looks "old" when you use it in MeeGo. Feb 26 20:06:05 npm: Indeed, QJackCtl's settings won't fit on the screen unless you switch to something like plastique. Feb 26 20:06:48 and also the meego style looks all "greyed out" and the buttons are hard to see if clicked Feb 26 20:06:55 plastique the clicked button is bright red Feb 26 20:07:06 npm: since it doesn't look like a Qt style, it must be Gtk+ Feb 26 20:07:28 which? the meego style? Feb 26 20:07:46 yes Feb 26 20:07:55 the default must be Gtk+, which loads the Gtk style of choice Feb 26 20:08:45 on my fedora desktop, i seem to be using a style between gtk and qt that looks similar Feb 26 20:08:57 (plastique actually) Feb 26 20:09:12 (i use kde) Feb 26 20:09:49 is there a way to set that for meego, so that the gtk and qt all look like my fedora desktop? Feb 26 20:10:36 and when is kde going to save the day and release a nice tablet ux for meego :-) Feb 26 20:11:27 my favourite Qt style is QGtkStyle Feb 26 20:12:39 and my standard challenge for anyone who claim that KDE is better than gnome because KDE is more configurable, is to make KDE look exactly like gnome... Feb 26 20:13:09 so does QGtkStyle do that? Feb 26 20:13:12 no Feb 26 20:13:23 it makes Qt apps look like gtk apps Feb 26 20:13:30 but it doesn't make KDE look like gnome Feb 26 20:15:42 npm: One of the KDE devs said they had a small team working on a Tablet environment -- though I don't know if they're targeting it being a "MeeGo UX" Feb 26 20:16:18 npm: Check the ML archives on meego-community for one of the Tablet UX threads Feb 26 20:16:50 npm: ...or on meego-dev where a KDE dev introduced their intention to collaborate with MeeGo. Feb 26 20:17:20 KDE would make a Plasma-Tablet UX Feb 26 20:17:28 it's up to MeeGo to adopt it or not Feb 26 20:17:34 http://www.youtube.com/watch?v=OjO5X1ADUrE Feb 26 20:17:48 in any case, there's a MeeGo tablet UX being done by Intel (shown at MWC last week) Feb 26 20:19:01 ali1234: thanks for suggesting QGtkStyle -- looks good Feb 26 20:19:32 npm: it's used by default if you use a gnome desktop... if you try to use it on KDE, it doesn't really work Feb 26 20:19:51 it doesn't change how plasmoids or the panel or the window decorations look Feb 26 20:20:08 and all the KDE versions of those just look even worse if you use a gnome style for widgets Feb 26 20:20:46 KDE is fixated on ugly gradients everywhere for some reason Feb 26 20:21:04 doesn't work too well with the usual flat gtk styles Feb 26 20:21:07 since there's no open-source tablet uX for meego, and people are in agreement that the netbook ux doesn't cut it for touchscreens... seems like plasma-based QML stuff could be the open source tablet solution we need Feb 26 20:21:34 I'm pretty sure Intel will open-source it Feb 26 20:22:06 it certainly has the apache license there. Feb 26 20:22:24 if you look in the code, it's really simple Feb 26 20:22:38 simple, but hard to put together Feb 26 20:22:47 that's the problem w/ UI code. Feb 26 20:22:57 designing it maybe Feb 26 20:23:17 reimplementing it from scratch wouldn't be that hard Feb 26 20:23:18 Hi, I was wondering why contact feature is not included in the core part of meego. It is not planned for all the adaptations (handset, netbook, etc)? Feb 26 20:23:35 yeah... if the code looks simple that's a testament to QML being the right tool for the job Feb 26 20:23:50 heh, well, i wouldn't say it's exactly a testament to Qt Feb 26 20:24:11 Qt is just a tool Feb 26 20:24:11 it would be half the size if it wasn't full of comment like "// ugly workaround for bug xxx" Feb 26 20:24:40 and shortcomings are expected in the first version Feb 26 20:24:57 intel's been developing with QML way before it was released at 1.0 Feb 26 20:25:11 thiago_home: agree -- QML is awesome Feb 26 20:25:37 like i said before, QML reminds me a lot of early days of Html, where everything had to be tweaked by hand and hacks were used like fixed width container tables etc Feb 26 20:25:50 i should quit screwing around w/ the netbook and get back to writing some :-) Feb 26 20:25:56 at least, that's the impression i get from the tablet UX code Feb 26 20:25:58 * thiago_home doesn't like the fixed-width part, but the designers of QML seem to Feb 26 20:26:45 ali1234: the tablet ux code is the result of a deadline i'm sure... Feb 26 20:27:04 if they open-source it we could easily rip out the fixed scaling and make it adaptive to display sizes Feb 26 20:27:14 sure. when i said it was simple, i meant it was implemented in the quickest and dirtiest way possible Feb 26 20:27:28 it's also that QML is new and we don't yet understand the "design patterns" it necessitates Feb 26 20:27:38 npm: i actually attempted to do that, and had some success Feb 26 20:27:41 plus there' no obious "toolkit" of compoentns to use Feb 26 20:28:22 in order to change the size of the panels on the main page i had to change the relative width of about 5 different sub components Feb 26 20:28:40 there is no top level "scale everything by 50%" Feb 26 20:28:43 that's one of the reasons why i've been wanting the netbook ux alongside n900 for apps.. . so as to figure out how to make apps work on difft screen res/sizes Feb 26 20:28:50 actually there is, but it doesn't work the way anyone would expect it to Feb 26 20:29:31 if you scale the panel by 50%, the distance between panels remains the same (because it is in a fixed width container) Feb 26 20:29:47 but if you scale that container by 50%, then the panels get scaled to 25% Feb 26 20:29:59 because they inherit both the parent container size and also the scaling of the parent Feb 26 20:30:11 so you have to put in loads of really nasty hacks Feb 26 20:30:41 * gabrbedd was actually asking about the fixed size problem on #qt Feb 26 20:31:51 gabrbedd: there's a task about other units Feb 26 20:31:57 like ems and cm Feb 26 20:32:33 thiago it is a shame there is no simple 'fill window' mode Feb 26 20:32:41 fill with what? Feb 26 20:32:56 content Feb 26 20:32:59 lcukn900: there is, it fills the window by stretching everything in the UI Feb 26 20:33:19 lcukn900: Do you mean like `cat /dev/urandom > content.dat` ?? Feb 26 20:33:22 but it doesn't quite work right because some bits are fixed size Feb 26 20:33:24 yeah so why does that not work? Feb 26 20:33:30 it more or less does Feb 26 20:33:39 ali then it is not dynastretching Feb 26 20:34:14 lcukn900: instead of using layouts, they're using alignment anchors. Feb 26 20:34:14 what is dynastretching? is that like callanetics? Feb 26 20:34:33 lcukn900: ...and it more or less does the same thing. Feb 26 20:34:34 several decades ago, we had this thing called a "table widget" which would do all that quite nicely: http://nielsmayer.com/winterp/doc/Table.txt Feb 26 20:34:39 :-) Feb 26 20:34:44 lol Feb 26 20:35:26 ali1234 component might request certain sizes but should be displayed as large as possible within the actual allocated space Feb 26 20:35:33 lcukn900: even though you can only see two panels on the ideapad instead of the three you would see on the wetab, nothing is actually getting drawn outside the screen, and the UI is completely functional Feb 26 20:35:47 i know Feb 26 20:36:03 and if you tell it to scale the entire UI by 50%, that works Feb 26 20:36:19 you still see a little bit less than two panels, but in a window that fills 1/4 of the screen Feb 26 20:36:38 this isn't a failing of QML, it's a failing of how the UI has been written Feb 26 20:36:45 because it's a hacky demo, not a serious product Feb 26 20:36:48 IMO :) Feb 26 20:37:24 thiago_home: how do I get the top-level QML element to "do the right thing" without me setting width/height manually? Feb 26 20:37:25 and because it's a hacky demo... well, it wouldn't take someone long to make an equally hacky open source version Feb 26 20:37:51 and that's just assuming they don't release the source, which they probably will Feb 26 20:38:15 gabrbedd: I don't know Feb 26 20:38:42 thiago_home: ok, thanks. Feb 26 20:39:12 thiago_home: I asked on #qt and somebody started a QML troll session. :-( Feb 26 20:39:33 * thiago_home pays attention there Feb 26 20:39:50 it wasn't me :) Feb 26 20:46:25 IMHO, the problem is there needs to be a toplevel component that abstracts away screen size in pixels and has everything underneath work in "floating point" percentages Feb 26 20:46:50 you don't want to do that Feb 26 20:46:54 scaling is not the answer Feb 26 20:46:59 it's also CPU intensive Feb 26 20:47:43 npm: it already does do that Feb 26 20:47:44 there needs to be an abstraction beyond pixels, esp if you're working "declaratively" Feb 26 20:48:17 cm and other units are useful Feb 26 20:48:24 if you set scale: 0.5 on the top level container it will draw the whole UI in 1/4 of the screen, therefore it is already resolution independent Feb 26 20:48:31 expanding elements are also useful Feb 26 20:50:37 part of the issue is also the need for actual "declarative programming" (aka backtracking in prolog) Feb 26 20:51:07 in that constraints can propagate down from the top (e.g. screen size) and up from components (such as fixed images etc) Feb 26 20:52:09 but it seems like most of the layout in QML is top-down, when it actually needs to be solving a constraint problem simultaneously from bottom up and top-down Feb 26 20:53:10 the thing is, this stuff is handled for you automatically if you use QWidgets. i would argue that by trying to simplify it too much, QML actually makes it harder to get this stuff right Feb 26 21:32:50 qwidget's layout system is incredibly complex Feb 26 21:33:47 perhaps, but reimplementing it in QML is even worse Feb 26 21:35:04 it makes creating a QVerticalLayout and adding some widgets into it seem trivial Feb 26 21:37:01 * javispedro finds gtk's "layouts"="widgets" concept easier to explain Feb 26 21:39:35 * Stskeeps watches quim's talk at scale9x Feb 26 21:41:34 Stskeeps: live? or is it streaming? Feb 26 21:41:45 gabrbedd: streaming, i'm not in the US right now :) Feb 26 21:41:58 Stskeeps: url? Feb 26 21:42:32 http://www.socallinuxexpo.org/scale9x/presentations/foundations-meego-project Feb 26 21:44:31 missile lock.... Feb 26 21:44:34 ON. Feb 26 21:44:37 oi Feb 26 21:44:54 that's not friendly :P Feb 26 21:45:58 ha! it may sound bad... but in the US that's actually friendly. It means "I'm dialed in and you have my undevided attention." Feb 26 21:46:49 to europeans it sounds like vietnam and iraq.. Feb 26 21:46:51 :P Feb 26 21:46:56 but yeah, cultures Feb 26 21:50:24 * gabrbedd makes mental note not to use military idioms around europeans. :-) Feb 26 21:50:35 s/not/note/ Feb 26 21:51:01 erm... stet. Feb 26 21:51:04 bah! Feb 26 21:53:04 poor qgil... I knew that would be the first question. Feb 26 21:53:09 :P Feb 26 21:53:10 yeah. Feb 26 21:53:27 * Jaffa opens stream now. What was it? "Why does Nokia suck?" Feb 26 21:53:46 Jaffa: "will nokia's focus change in meego" or something Feb 26 21:56:38 really the worst kind of question to get, especially since there's not too much info yet Feb 26 21:57:01 and there will never be... Feb 26 21:57:52 they just said they're working on a offical statement ;) Feb 26 22:00:19 my qml/qt button are so square Feb 26 22:01:03 niala: Tried Qt Components or one of the other toolkits? Feb 26 22:23:52 If I ask the same question in a whiney voice... will I get a different answer? Feb 26 22:24:15 * gabrbedd is still listening to qgil's talk Feb 26 22:24:23 same Feb 26 22:26:24 ...but what if I ask it /again/? Feb 26 22:27:26 there's a lot of misinformation to be beaten down Feb 26 22:27:27 :P Feb 26 22:28:43 Ok, let me phrase it differently.... Feb 26 22:29:04 Are these KDE guys asking this? OMG! Feb 26 22:32:31 * Stskeeps plugs out headphones Feb 27 00:20:12 well that's clamav updated Feb 27 00:20:19 ~seen lbt Feb 27 00:20:28 lbt <~david@pool-173-50-232-79.ptldor.fios.verizon.net> was last seen on IRC in channel #meego, 2d 5m 13s ago, saying: 'I am always optimistic and can still hope that Nokia will find a way out of the darkness...'. Feb 27 00:21:03 man that's profound Feb 27 01:05:14 hello Feb 27 02:06:21 why does select all have to be one key below quit program? Feb 27 02:18:24 I think the OBS might have broken on the 22th of Feb Feb 27 02:18:46 hey CosmoHill you're still standing. not sleeping Feb 27 02:18:53 readoing fc ? Feb 27 02:19:01 that's the last community package to be created in the public repo Feb 27 02:19:31 niala: 3 all against Crystal Palace and I'm actually lying down Feb 27 02:20:37 bonne nuit all Feb 27 02:20:37 again a draw !! incredible :) Feb 27 02:20:58 bonne nuit. i don't see Feb 27 02:21:04 your package in obs? Feb 27 02:21:28 niala: that's the problem Feb 27 02:21:28 your's packages CosmoHill Feb 27 02:21:45 it compiled successfully but it hasn't appeared on the repo Feb 27 02:21:51 http://repo.pub.meego.com/home:/ ? Feb 27 02:21:56 correct Feb 27 02:22:11 I need to speak to lbt about it on monday but for now I'm going to sleep, cyas Feb 27 02:22:16 there was some email from others on the dev ML, same symptoms Feb 27 02:22:20 have created your home ? Feb 27 02:22:27 ok good night nice dream **** ENDING LOGGING AT Sun Feb 27 02:59:57 2011