Friday, 2012-11-02

*** jstaniek has quit IRC00:00
*** arnaud1 has quit IRC00:01
*** arnaud1 has joined #mer00:01
*** imunsie has quit IRC00:08
*** imunsie has joined #mer00:09
*** M4rtinK has joined #mer00:13
*** drussell has quit IRC00:14
*** popey has quit IRC00:17
*** clopez has joined #mer00:20
*** jayrulez has joined #mer00:22
*** npm has joined #mer00:23
*** himamura has joined #mer00:27
*** himamura_ has joined #mer00:28
*** himamura has quit IRC00:32
*** zuh has quit IRC00:41
*** zuh has joined #mer00:41
*** jayrulez has quit IRC00:48
*** sonach has joined #mer00:49
sonachmorning all00:50
*** ccssnet has joined #mer00:56
*** rcg has quit IRC01:00
chriadamgmorning sonach01:00
sonachchriadam: hey :)01:02
*** nibbler has quit IRC01:03
*** nibbler has joined #mer01:03
*** Venemo has quit IRC01:03
*** imunsie has quit IRC01:04
*** imunsie has joined #mer01:05
*** terietor has quit IRC01:10
*** JvD_ has quit IRC01:10
*** JvD_ has joined #mer01:11
*** reels__ has joined #mer01:25
*** panda-z has joined #mer01:27
*** reels_ has quit IRC01:28
*** yunta has quit IRC01:42
*** tilgovi has quit IRC01:42
*** sonach has quit IRC01:46
*** sonach has joined #mer01:48
*** plfiorini has quit IRC01:52
*** dakovaci has quit IRC01:55
*** M4rtinK has quit IRC02:00
*** dakovaci has joined #mer02:30
*** FSCV has joined #mer02:34
*** arcean has quit IRC02:36
*** KaiRo_Mozilla has joined #mer03:13
*** KaIRC has quit IRC03:17
*** KaiRo_Mozilla has quit IRC03:30
*** ka6sox is now known as ka6sox-away03:42
*** furikku has joined #mer03:55
*** DocScrutinizer05 has quit IRC04:03
*** DocScrutinizer06 has joined #mer04:03
chriadamgmorning Bostik04:10
*** dakovaci has quit IRC04:29
*** morphis|away is now known as morphis04:40
*** dakovaci has joined #mer04:41
*** clopez has quit IRC04:48
*** dakovaci has quit IRC04:53
*** npm has quit IRC05:10
*** imunsie has quit IRC05:14
*** panda-z has quit IRC05:24
*** panda-z has joined #mer05:27
*** ka6sox-away is now known as ka6sox05:38
*** rdqfdx has joined #mer05:41
*** kontio has joined #mer05:43
the-gibsonhas anyone tried to put together an image for the nexus 7?05:50
*** merder has joined #mer05:51
*** dakovaci has joined #mer05:54
*** dakovaci has quit IRC06:00
situMorning everyone06:04
*** ka6sox is now known as ka6sox-away06:04
*** plfiorini has joined #mer06:13
*** M13 has joined #mer06:17
* Stskeeps desperately needs coffee06:20
*** pirut has joined #mer06:26
* ljp hands Stskeeps a double espresso06:32
Stskeepsthanks :P06:32
*** VDVsx has joined #mer06:43
kulvethe-gibson: that's a very interesting question :)06:47
Stskeepslbt: ok06:47
the-gibsonkulve: im gunna give it a shot this weekend, well see how it goes06:47
*** Mihanizat0r has joined #mer06:48
kulvethe-gibson: I'm probably doing the same. I already have hw accelerated video/3d/window composition working on a reference board which I think resembles nexus 7 quite closely06:48
the-gibsonone of the cardhu dev boards? Thats what ive been playing with.. havnt done much more than gettin L4T playing nicely06:50
kulveyes. I've pretty much created a hw adaption for tegra in mer and the according kickstart files06:51
*** Mephustopheles has joined #mer06:51
*** M13 has quit IRC06:52
the-gibsonkulve: cool! do you have any notes on that? ive been mostly doing debian things on it for HPC stuff06:52
kulveI haven't got the wireless working though. Any help with that is appreciated06:52
kulveI did copy some bcm firmware binaries and with those I can see the wireless networks but then bad things start to show up in the kernel log06:53
the-gibsoni have that working.. at least in debian/ubuntu06:53
kulvewith the official l4t kernel? Which bcm fw binaries?06:53
the-gibsoni started on the l4t kernel and tweaked it a bit.. and i grabbed the binaries off a transformer prime or something06:54
*** Mihanizat0r has quit IRC06:55
kulveyeah, I tried something similar but they seem to be unstable or a bit incompatible06:55
kulvedo you have URLs for those binaries? I would like to try them (or see if they are the ones I'm already testing)06:55
*** korhojoa has joined #mer06:56
the-gibsonhmm ill have to play with it again, been a few weeks since ive poked at the tablet vs the HPC board06:56
the-gibsonlet me see if i can find them, one sec06:56
the-gibsonkulve: ^06:59
the-gibsoni believe thats the latest i had working07:00
kulveright.. I guess my cardhu has a different chip07:00
the-gibsonhuh, interesting is it one of the nvidia dev tablets?07:01
*** dakovaci has joined #mer07:02
the-gibsonhmm looking around on xda-developers they reference the bcm4330.. the ubuntu image I plopped on earlier seemed to work.. well see what that one is07:04
the-gibsonanyway nap time07:05
*** lamikr has joined #mer07:06
*** rcg has joined #mer07:07
*** DocScrutinizer06 is now known as DocScrutinizer0507:08
*** blam_ has quit IRC07:11
*** calvaris has joined #mer07:15
*** calvaris has quit IRC07:16
*** calvaris has joined #mer07:17
*** veskuh has joined #mer07:18
*** Sfiet_Konstantin has joined #mer07:29
*** Sfiet_Konstantin has quit IRC07:30
*** cxl000 has joined #mer07:40
*** terietor_ has joined #mer07:44
*** vgrade_ has quit IRC07:47
*** morphis is now known as morphis|away07:48
*** CosmoHill has joined #mer07:49
*** FSCV has quit IRC07:52
*** Mephustopheles has quit IRC07:53
*** vgrade_ has joined #mer07:57
*** calvaris has quit IRC08:05
*** gabriel9|work has joined #mer08:08
*** ssirkia has joined #mer08:14
*** melonipoika has joined #mer08:19
CosmoHillI have no work today and wake up at 7am :(08:21
*** ngharo has quit IRC08:28
*** ngharo has joined #mer08:28
*** Cosmo[PB] has joined #mer08:32
*** jpetersen has joined #mer08:32
rcgwell, yeah.. got a day off and also woke up at about 8am :/08:33
rcgor even earlier.. i just tried very hard to sleep longer xD08:34
*** CosmoHill has quit IRC08:34
Cosmo[PB]I'm on assassin's creed 3 atm08:35
*** arfoll has quit IRC08:36
*** stefan_schmidt_w has joined #mer08:39
*** rymek has joined #mer08:42
*** dakovaci has quit IRC08:44
*** M4rtinK has joined #mer08:47
*** hazchemix has joined #mer08:52
*** popey has joined #mer08:52
*** pvuorela has quit IRC08:55
*** notmart has joined #mer09:02
*** notmart has joined #mer09:02
*** yunta has joined #mer09:06
*** reels__ has quit IRC09:09
*** reels has joined #mer09:09
*** M4rtinK has quit IRC09:10
*** dakovaci has joined #mer09:11
*** trbs has joined #mer09:12
*** arcean has joined #mer09:25
*** ortylp has joined #mer09:26
*** sonach has quit IRC09:27
*** auri__ has joined #mer09:28
*** aportale has joined #mer09:33
*** dakovaci has quit IRC09:34
*** melonipoika has quit IRC09:38
*** ced117 has joined #mer09:38
*** ced117 has joined #mer09:38
*** slaine has joined #mer09:38
*** terietor_ is now known as terietor09:39
*** Jucato has joined #mer09:46
*** blitz00 has joined #mer09:46
*** blitz00 has joined #mer09:46
*** phinaliumz has joined #mer09:46
Stskeepslbt: fix binutils and we're good09:47
*** himamura has joined #mer09:47
Stskeepslbt: sure we need to bother adding sb2-* tools ?09:47
*** himamura has quit IRC09:47
lbtno, I'm not sure there09:48
*** himamura_ has quit IRC09:49
lbthmm - so no gold on mipsel?09:49
Stskeepsi think it's a sideeffect of the "%{name}" stuff09:50
lbtI was a bit all over the place yesterday09:50
lbtFYI ICS are sponsoring travel to QtDevDay for Mer09:51
Stskeepsi'm not going, sadly09:52
lbt*nod* - you mentioned that. It's the main reason I offered to do it :)09:54
*** stepiro has joined #mer09:54
lbtright lets see what I screwed up in binutils09:54
Cosmo[PB]I've been unemployed for 3 and a half days and I'm already watching Jeremy Kyle >.<10:03
lbtdo not, ever, watch daytime TV... ever.10:05
*** calvaris has joined #mer10:07
*** ortylp has quit IRC10:10
*** Cosmo[PB] is now known as CosmoHill10:16
*** calvaris has quit IRC10:16
*** plfiorini has quit IRC10:17
lbtmorning auri__ and aportale10:19
lbtany blockers from my side for you guys?10:19
lbtmy build issue yesterday appears to be unrelated to qtc10:19
lbtso I think I borked my sb2 target10:20
auri__hmm.. still I'm wondering about the mkspec10:20
auri__anyways I'm just fine tuning the target addition process10:20
Stskeepsauri__: got time to walk me through that mechanism then? i dealt with sb2 a fair bit and the cross compilers so10:20
aportalelbt: I did not yet try the package creation. So I don't know if that is blocker free10:20
auri__Stskeeps: will u be free in an hour? right now in the middle of something10:21
aportalelbt: Does .yaml to .rpm-package creation already have wrapper scripts?10:21
lbtaportale: (cd rpm; specify)10:21
Stskeepsauri__: i -think- so, else rest of day10:22
auri__ok.. suits me10:22
lbtaportale: erm - that makes the .spec10:22
aportalelbt: So there is a that calls specify remotely?10:22
lbtnot yet10:23
aportalelbt: Then "mb build -t nemo_n9_min rpm/qmltrial.spec" does the rest, right?10:23
lbtwe probably want an 'update spec' button that just runs specify and an update+make rpm that does both that and ^^ command10:24
lbtthe reason is that the user may want to edit the spec file manuall10:24
lbtnb - I'm commenting on the UI without seeing it yet so not sure if that makes sense10:25
lbt(also typing whilst peering around a kitten at the moment :)  )10:25
aportalelbt: Well, I can immagine it makes sense for power users. But thos e power users would perhaps not use the .yaml file anyways, right?10:25
lbtnot so sure actually10:26
Bostikawwwwww crap, the potential fallout from a single %reconfigure macro is ludicrous10:27
*** ortylp has joined #mer10:27
StskeepsBostik: so what was the end result of getting qtwebkit building?10:27
aportalelbt: I suggest we can look at that update spec button, in a later iteration. I think it is possible to do without ugly stunts.10:28
lbthmm it should be run every time you apply a change to the yaml10:29
aportalelbt: it will be "hidden" in the project mode gui, in the deploy/package step.10:29
BostikStskeeps: with all the tricks & voodoo, combined with forced "ld == gold" and removing WebKit2 entirely - that allowed to keep the linking memory (barely?) within the limits10:29
BostikStskeeps:  (had to do minor tweaks afterwards but that's the big piece)10:30
Stskeepsthere's really no way of forcing gold for gcc 4.6-linaro?10:31
Stskeepsthat seems odd10:31
BostikStskeeps: I hunted down some obscure developer forums and it seems to be a gcc build-time option10:31
Bostikat that point I decided I didn't really need to know, when the stock source didn't have the option even available...10:32
lbtStskeeps: that was what I meant yesterday10:32
lbt%define crossextraconfig --disable-libstdcxx-pch --with-arch=i486 --with-gnu-as=/opt/cross/bin/i486-meego-linux-gnu-as --with-gnu-ld=/opt/cross/bin/i486-meego-linux-gnu-ld --with-as=/opt/cross/bin/i486-meego-linux-gnu-as --with-ld=/opt/cross/bin/i486-meego-linux-gnu-ld10:33
Stskeepsyeah, but not related in this particualr case10:33
lbtshould we not be setting an option *like* that for the non-cross setting of ld to be gold, not gnu-ld ?10:33
lbtor do we not want gold as the default linker for some reason ?10:34
Stskeepswell, the fact it may break existing stack10:35
Bostik  <- use of gold depends on "linker plugin support"10:35
Stskeepsi'd be interested to see the impact though10:35
*** icota has joined #mer10:35
Bostik"-fuse-ld=gold" is available only in gcc 4.710:35
*** niqt has joined #mer10:36
*** lamikr has quit IRC10:36
Stskeepswell we can handle it in two ways10:37
Stskeepsone, we try to see what happens if we make gold default10:37
Stskeepstwo, we see if we can make proper settings10:37
* Stskeeps ponders10:38
Bostikthree, we can pull dirty magic in spec to cheat :(10:38
*** mdfe_ has joined #mer10:40
Bostikof course #3 makes it pretty tricky to rebuild qtwebkit locally in SDK, which is probably not what we want10:41
*** lamikr has joined #mer10:42
CosmoHillright, I'm going to the garage to take something apart before i lose my mind10:43
*** calvaris has joined #mer10:45
StskeepsBostik: i'll do a test run of mer with gold binutils for your enjoyment10:46
Stskeeps(as default)10:50
Bostikwaiting for the fallout10:52
Stskeepsthat's the fun part10:52
* lbt watches local mips build of binutils10:58
Stskeepslbt: also, depending where you're invoicing for, your signed-off-by may be inaccurate10:59
*** Sfiet_Konstantin has joined #mer11:01
lbtStskeeps: yeah - I should watch that - but 486-cross is not on any bugzilla/task-list :)11:05
*** morphis|away is now known as morphis11:10
*** nibbler has quit IRC11:15
*** panda-z has quit IRC11:18
*** panda-z has joined #mer11:19
lbtStskeeps: quick question on gerrit-ci steps - does it stop at first build failure? ie can I be sure binutils built on all other arches?11:20
Stskeepsit stops at first build failure11:21
Stskeepsand it does it serially11:21
lbtah - so this is nothing to do with mips11:21
StskeepsBostik: Core:i486:gold on shall start to build later11:24
Stskeepsi'll do a full rebuild and see if things work11:24
BostikI'll upload the latest specs+patches for qt5... tomorrow?11:25
Stskeepsmight be a busy weekeend for the builder11:25
Bostikaye, good point11:25
*** ali1234 has quit IRC11:26
Bostikah well, what's couple of days more when it has taken this long to even get this far11:26
lbtFWIW we have 5 build hosts atm11:26
lbtno point letting the cobs ones sit idle11:27
lbtand I thought a cross-gcc rebuild may cause some rebuilds if it got accepted :)11:27
*** ali1234 has joined #mer11:27
Bostik"some" :D11:27
Bostikfor what it's worth, I'm personally convinced that in less than two years a 64-bit build environment is mandatory even in IVI land11:30
StskeepsBostik: i'm thinking a bit ahead in that area but i need some people with perl experience, OBS, and no history of mental illnesses11:31
Stskeepsmostly because they'll have one afterwards11:31
* Bostik glances in his past... yep, gotten past perl phase11:31
lbtBostik: and x32 ?11:32
Bostiklbt: hum, that's an interesting point and one I hadn't considered11:32
*** ortylp has quit IRC11:32
Bostikbut I've peeked at OBS code and I'm pretty sure I wouldn't dare touch that11:33
lbtat one point I thought it may help us but actually it's the inverse - I think it means that linkers need to rethink things11:33
lbtOBS code is easy in small chunks11:33
lbtthe main issue is RPC-via-spooky side effects11:33
* Stskeeps heads out for a little while11:33
lbtbinutils built :)11:34
lbtjust -%find_lang %{?cross}gold11:34
Bostiklbt: the logical steps should not be that different from arm-THUMB code generation, so it shouldn't be a complete dive into unknown waters11:34
*** ali1234 has quit IRC11:36
Stskeepsi do want to make mer smaller than it is at the moment.. the thought of my original idea of toolchain-core does come back a bit11:36
Stskeepsbut i'm not sure yet how i want to do it11:36
Bostika dedicated toolchain-subset would be very handy11:38
Stskeepsi know sb2 can do it, i've built busybox without any on-target toolchain11:40
Stskeepsand acl11:40
Stskeepsand such11:40
Bostik[e]glibc, coreutils, binutils, gcc, ... I think I saw you mention that it takes 7(?) packages to do an absolute minimum set11:47
*** yunta has quit IRC11:47
*** popey has quit IRC11:58
lbtwtf - it is mips specific12:05
*** arcean has quit IRC12:07
lbtaportale: ping - am I right in thinking we'll have 2 qtc plugins? 1 for Mer which does kits/VM/packaging and 1 for Nemo which does components in designer ?12:13
aportalelbt: As far as I can oversee it, having a Mer plugin and a Nemo plugin (which depends on the mer plugin) would be desirable and a clean division.12:15
lbtgood - just checking12:15
aportalelbt: I assume that Nemo is not be the only Mer based platform.12:16
lbtcorrect - I expect that plugin to be the one that our vendors would adapt12:17
lbtmaybe we should get the plasma active guys involved12:18
aportalelbt: How would Qt Creator currently automatically tell a nemo mer-target/kit apart from a plasma-active one?12:21
aportalelbt: That distinction would be necessary to property trigger stuff in the Nemo or Plasma-Active QtC plugin.12:21
lbtNew project?12:22
aportalelbt: Or maybe I am just overcomplicating things, here :) The Nemo plugin just adds an palette of Qt Components and perhaps some IDE branding.12:23
lbtI really don't know where the component part plugs in - partly why I'd like to get an alpha up - in case it impacts the Kit side12:23
*** ali1234 has joined #mer12:24
lbtaportale: you could be right - but somehow we need "use these components"12:24
lbtheck - maybe the presence of a certain file in the sysroot?12:24
lbtdoes Qt have a components.d/ directory :)12:25
lbtyet :)12:25
aportalelbt: Afaik not a standard location (but I did not yet look into that)12:25
lbtno, I was kidding - but there may be a mechanism to register them?12:26
*** KaIRC has joined #mer12:27
*** kallela has joined #mer12:28
aportalelbt: Components are usually installed as QML plugins in a location on the device/system where the QML plugin loader looks at app startup time. Qt Creator has a (not too elegant) way to let a QtC plugin register components in the Designer.12:28
*** ridikulus_rat has joined #mer12:36
*** ridikulus_rat has left #mer12:36
StskeepsBostik: yay, glibc fails12:38
*** VDVsx has quit IRC12:38
Stskeepseglibc 2.15 idiotically relies on ld --version output..12:38
*** ortylp has joined #mer12:42
*** arcean has joined #mer12:48
*** panda-z has quit IRC12:56
*** xmlich02 has joined #mer12:56
Stskeepslbt: can i ask a perl question?12:58
Stskeepslbt: - how would i sort all_files_in_dir ?12:59
Stskeepsie, a deterministic order12:59
lbtfor $file (sort(all_files_in_dir('parts/inc')))13:01
lbtin case you need more than a basic one13:02
lbtso maybe the () are wrong too :)13:02
Stskeepsthere's some cases where i'm getting non-sorted outputs and it pisses me off a bit.. gives obs rebuilds13:04
aportaleWhy does the the instruction say guestwww:guestport = 9292, although the screenshot says 80?
lbtbug probably13:05
aportalelbt: is 80 correct?13:05
lbtI suspect 929213:05
henrik_aportale: 929213:06
aportalelbt, henrik_: Ok, thanks.13:06
lbtthe webapp will run as a user so can't use 80 .... maybe we should use 8080 for both to minimise chance of swapping13:07
*** phinaliumz_ has joined #mer13:08
* lbt stops suggesting silly polishing items just yet :)13:08
*** phinaliumz has quit IRC13:08
*** dod1 has joined #mer13:12
aportaleSmall fix for the instructions: I think it should be "cp ~/.ssh/ ~/.ssh/authorized_keys" in
*** VDVsx has joined #mer13:16
auri__I think it should rather be <Shared_home_folder_as_provided_in_VM_Settings>/.ssh/authorized_keys13:16
*** gabrbedd has joined #mer13:16
*** ortylp has quit IRC13:20
lbtmmm gold is not built for MIPS but it doesn't complain about --enable-gold13:21
Stskeepsugh.. seems like some kind of copyright assignment issue13:22
* lbt slaps yunta13:22
lbtStskeeps: I'll disable gold for mips at the file/install level but leave --enable-gold for all arches13:24
lbtshould make it easier to not forget when we upgrade binutils13:24
Stskeepsmake a comment that gold doesn't function for mips to our knowledge13:25
lbtyep - url?13:25
lbtI found some patches but not a resolution13:25
*** dakovaci has joined #mer13:25
lbtaportale: correct, that's nonsense13:25
lbtand tbh cp is a bad idea13:26
lbtif anything cat >>13:26
lbtauri__: we assume a shared $HOME13:26
lbtbut if not, yes13:26
lbthenrik_: do we have a better/later image yet?13:28
auri__is there on where one need not do the untar, retar with --hard-dereference and untar again?13:31
henrik_lbt: latest vm image:
lbtauri__: just do the untar on the host, not in the VM13:31
*** Jucato has quit IRC13:32
auri__aportale: ^^13:32
aportalehenrik_: that latest one has the non-hardlink fix, right?13:32
lbtit won't have13:32
henrik_aportale: yes. if you use sdk-manage to unpack13:32
lbthenrik_: oh, clever - what did you do ?13:32
henrik_lbt: yunta did magic :)13:33
lbtunpack to vdi/repack ?13:33
henrik_untar in vm and then rsync to target13:33
henrik_slow, but seems to work13:33
lbtyeah - be careful around that - it's very alpha and quite complex13:34
lbtI have a simpler solution coming out soon13:34
auri__\me thinks wikis ned to be updated!13:34
* auri__ thinks wikis ned to be updated!13:35
*** Jucato has joined #mer13:35
*** yunta has joined #mer13:35
lbtoh, it's still valid (afaik) - just a WIP13:35
* lbt points and laughs at yunta13:35
*** Hoolxi has joined #mer13:38
yuntalbt: ^13:41
Stskeepsno pointing and laughing here13:41
lbtaww - even at yunta?13:42
Stskeepsbe nice :P13:42
lbtoverwriting ssh authorized_keys13:42
lbtbut its *yunta*13:42
* lbt gestures and grins instead ;)13:42
Stskeepslbt: think you have to explain in depth what exactly you're talking about..13:42
auri__Stskeeps: ping13:43
Stskeepsauri__: pong13:43
auri__free for ~15 min?13:43
lbtnah, it's yunta - he's clever enough to know :)13:43
yuntacan I laugh at myself?13:43
auri__so coming back to mkspecs13:43
yuntalbt: why should I? :)13:43
auri__normally the mkspecs from /usr/share/qt4/mkspecs/ is picked up13:44
auri__the one that is symlinked to default13:44
auri__to get the qmake.conf and qplatformdefs.h13:44
*** Hoolxi has quit IRC13:44
yuntalol, lbt, that actually made me laugh when I wrote it :)13:45
*** Hoolxi has joined #mer13:45
auri__I am not sure which one is apt for armv7hl-meego-linux-gnueabi-gcc13:45
yuntalbt: I just followed installation, and didn't read the part explaining that my home will be mounted on /home/mersdk13:45
yuntalbt: so I totally scp-ed my keys there :D13:46
Stskeepsauri__: so, to understand it clearly, when you launch a build with the vm, this happens with sb2 running qmake and make?13:46
auri__I can see linux-arm-gnueabi-g++ and linux-g++-maemo13:46
lbtStskeeps: yes, it does13:46
lbtit issues a sanitised "ssh <vm> sb2 qmake"13:46
lbtand "ssh <vm> sb2 make"13:47
Stskeepsauri__: let me quickly take a look at mkspecs13:47
*** phinaliumz_ has quit IRC13:47
* lbt notes that this worked just fine very recently13:47
auri__qtc tries to select the mkspecs and passes that as argument when calling qmake13:47
Stskeepsauri__: is there any problem in simply selecting 'default' within the sysroot?13:49
* auri__ wonders how it wokrd so smooth for mer!13:50
aportaleauri__, henrik_: (I am first trying the re-taring cure) Do you remember the tar command options for tar-ing it without the links? /me bad at tar13:50
Stskeepsauri__: because mer's actually linux-g++ even on arm target13:51
auri__Stskeeps: depends! if you want a build for linux-g++ which I m pretty sure provides x86 spec.. then it should work13:51
Stskeepswell, sb2 gcc -> the arm toolchain13:51
auri__why havent you configure Qt for arm ?13:51
auri__but you use Qt built for x86 in an arm environment?13:52
Stskeepsauri__: should i walk you through mer's qt building? :P13:52
auri__yes plz13:53
henrik_aportale: well you can untar in host as well. or use the latest vm image with sdk-manage13:53
henrik_aportale: or if you want, untar in host and re-tar with --hard-dereference13:53
aportalehenrik_: Ah, it was -hard-dereference whaich I was missing. Thanks. In case that fails, I will try the latest image :) Thanks.13:54
Stskeepsauri__: okay, i'll do the ARM build. assume for a second that you're on an actual ARM machine, what do you do, you extract qt, you configure, linux-g++ is the configuration that matches standard linux, and you build, with using gcc/g++ etc13:55
Stskeepsauri__: (continued in a moment)13:55
auri__can you copy paste the configure comamnd with the parameters that you pass to it?13:57
*** Jucato has quit IRC13:57
Stskeeps , search for ./configure14:00
Stskeeps+ ./configure -runtimegraphicssystem raster -graphicssystem runtime -confirm-license -platform linux-g++ -prefix /usr -bindir /usr/bin -libdir /usr/lib -docdir /usr/share/doc/qt4 -headerdir /usr/include/qt4 -datadir /usr/share/qt4 -plugindir /usr/lib/qt4/plugins -importdir /usr/lib/qt4/imports -translationdir /usr/share/qt4/translations -sysconfdir /etc/xdg -demosdir /usr/lib/qt4/demos -examplesdir /usr/lib/qt4/examples -opensource ...14:00
Stskeeps... -no-sql-ibase -no-sql-mysql -no-sql-odbc -no-sql-psql -plugin-sql-sqlite -no-sql-sqlite2 -no-sql-tds -system-sqlite -no-qt3support -xmlpatterns -no-multimedia -audio-backend -no-phonon -no-phonon-backend -svg -no-webkit -script -scripttools -system-zlib -system-libtiff -system-libpng -system-libjpeg -no-rpath -optimized-qmake -dbus-linked -no-separate-debug-info -verbose -no-nas-sound -opengl es2 -arch armv6 -no-openvg -lfontconfig ...14:01
Stskeeps... -I/usr/include/freetype214:01
Stskeepsauri__: can you tell where the mkspecs are used then specifically?14:03
Stskeepswhen doing sb2 qmake, pass linux-g++14:03
auri__-platform linux-g+14:03
auri__this should be -xplatform linux-arm-gnueabi-g++14:03
auri__if you want a platform build14:03
auri__or if you have a specific board then14:04
*** kjokinie has left #mer14:04
w00tno, it shouldn't, because there's no cross compilation required - the whole system is already doing the cross compilation14:04
Stskeepsnop, think of it as that we're building on a native arm machine14:04
Stskeepsgcc in sb2 actually invokes the cross compiler itself14:04
lbtand each sb2 target is ~ a discrete native machine FWIW14:05
Stskeepssb2 enables a practically "fast" arm machine, backed up by x86 binaries (cross compilers, make, bash, etc)14:05
Stskeepsso when you do sb2 gcc -o foo foo.c, it'll actually transparently call armv7hl-meego-linux-gnueabi-gcc14:05
Stskeeps(as that's what we set up in the target)14:05
auri__ok.. I am absolutely a layman when it comes to sb214:05
w00t( is a very useful read)14:05
Stskeepsauri__: no worries, just explaining14:06
*** plfiorini has joined #mer14:06
Stskeepsso all you have to do is select default from the sysroot as qmake parameter, it'll work14:06
auri__ok.. if sb2 does the magic then its great14:06
auri__unfortunately you have set the default as linux-g++ when you passed -platform linux-g++ with ./configure14:07
auri__so it takes that as the "correct" mkspec14:07
Stskeepsit is the correct mkspec :)14:07
Stskeepsfrom our perspective14:07
Stskeepswhat does that cause in qt creator?14:07
auri__well it call sb2 qmake -spec linux-g++14:08
Stskeepsthat's fine14:08
auri__will that be a problem?14:08
Stskeepswill work fine14:08
auri__but for the code model it might be ( i am not sure tho')14:08
auri__since the platform definitions are picked up from that location14:09
Stskeepscode model?14:09
auri__in QtC14:09
auri__for the #ifdefs14:09
lbtStskeeps: what's the magic to say :     %ifnarch %{mips}  in a spec?14:10
Stskeepslbt: isn't there a %{mips}?14:10
Stskeepsif not, thats a bug14:10
lbtI was worried you'd say that14:10
Stskeepselse just say mips mipsel and go on with your life14:10
auri__would it be a prob if you call -platform with the arm spec? will sb2 get confused? or will it still call the correct cross compiler?14:11
lbtnot familiar with that idiom ?14:11
*** merder has quit IRC14:11
Stskeepslbt: get on with, ie, %ifnarc mips mipsel14:11
*** arcean has quit IRC14:11
lbtI see14:11
*** Jucato has joined #mer14:12
Stskeepsauri__: the correct platform is linux-g++ - i'm just wondering where we're seeing an actual problem at the moment :)14:12
Stskeepslike, what reads the name of the target and assumes it's not arm, as an example14:13
*** xmlich02 has quit IRC14:13
* lbt notes again that his report is likely to be a local misconfiguration since it used to work14:13
lbt(but it *should* work - I did double check it)14:14
lbtftr I do :   sudo mic create fs /mer/mer/devel/mer-sdk/nemo_n9_min.ks --pkgmgr=yum --arch armv7hl -o /srv/mer/targets/14:14
auri__lbt: reported a problem and I was checking the configuration and found this.. which I assumed was inconsistent14:14
auri__I am not sure if there is an actual problem ..yet14:14
lbtI'm not sure there is aproblem either - but it's good to go through this level of detail14:15
auri__less work for me then :)14:16
*** Venemo_N9 has joined #mer14:16
auri__Stskeeps: thnx for the explanation.. didnt know sb2 was clever.. :)14:17
Stskeepsauri__: it can even accelerate autoconf and other evil scripts14:17
Stskeepsbut that's off topic ;)14:17
* auri__ maybe now likes sb2 a bit 14:18
Stskeepsit makes us able to deal with embedded systems as if they were full systems, so14:18
Venemo_N9sb2 is one of the best things since sliced bread14:20
Stskeepsyes, thank you nokia for paying me to get it into mer ;)14:23
*** blitz00 has quit IRC14:23
lbt3rd time lucky on binutils ?14:24
aportalewhich sb2-* command can I use to remove a target?14:24
*** M4rtinK has joined #mer14:24
*** spiiroin has quit IRC14:24
Stskeepsaportale: i'm not sure there's one14:25
* aportale tried sb2-config -d ..lol14:25
Stskeepsyou remove it's underlying sysroot and you can overwrite the target at any time14:25
Stskeepsbut yeah, there should be one for consistency,..14:25
lbtremove the ~/.scratchbox2/<target>14:25
*** rcg has quit IRC14:27
*** kontio has quit IRC14:28
lbtStskeeps: do we use cross-\\(.*\\)-binutils-accel ?14:28
Stskeepsno, we can remove it in the future14:29
lbtbecause if so I think the spec could be buggy14:29
Stskeepsit is, it's stuff that needs removing14:29
*** dakovaci has quit IRC14:29
*** merder has joined #mer14:30
*** yunta has quit IRC14:30
lbtauri__: aportale I'm likely to be in a meeting for the rest of the day - it feels like we're looking good for me to show off this work on tuesday?14:30
lbtand we'll have a little time on monday am before I leave in case I hit issues over the weekend14:31
*** morphis is now known as morphis|away14:31
*** Jucato has quit IRC14:32
*** Jucato has joined #mer14:33
*** kallela has quit IRC14:33
sledgeslbt, going through, get /srv/mer/sdks/sdk/mer-sdk-chroot: line 128: /srv/mer/targets/mer-core-armv7hl-xorg-basic-qmlviewer/usr/bin/readlink: cannot execute binary file14:34
*** Hoolxi has quit IRC14:36
*** FSCV has joined #mer14:36
*** NIN101 has joined #mer14:36
lbtsledges: mmm sounds like a bad target ? using an arm binary w/o sb2?14:42
*** yunta has joined #mer14:42
* lbt is prepping for a meeting so a bit afk14:42
sledgesI pasted you the output from qtcreators "Compile Output"14:42
sledgesit is using qmake (the fake one from wiki)14:42
sledgesmaybe I should launch qtcreator from within sdk (with DISPLAY=:0 )14:43
auri__lbt: I think its ok for a demo ..14:48
lbtauri__: cool14:48
aportalelbt: This morning you pasted (cd rpm; specify); mb build -t nemo_n9_min rpm/qmltrial.spec14:50
aportalelbt: mb is not found. Did you mean sb2?14:50
lbtno, mb14:50
lbtit should be in the VM14:51
aportalelbt: Not on mine, apparently. I#ll try zypper14:51
lbtas root in the vm try: zypper in sdk-utils14:51
sledgeswhois tswindell ?14:53
Stskeeps-> alterego14:53
sledgeswas just talking to him on #nemomobile14:54
aportalelbt: Cool. mb is now installed, and it does something. But I did not yet find my rpm package.14:54
*** stepiro has quit IRC14:54
sledgesalterego, could you look at my qt5-mer@qtcreator bit ?14:54
lbtthe mb -t arg is the same as the sb2 one14:54
lbtit wasn't developed with this in mind so we will want to tweak it14:55
Stskeeps(why are we using mb, just ooi?)14:55
lbtit works14:56
lbtit does a non-osc spec->rpm via rpmbuild14:56
aportalelbt: I adjusted it to my values, and it gave this output:
Stskeepslbt: yeah, ok14:57
lbtI'm sure there are better ways - but it sets up the SOURCES dir etc14:57
aportalelbt: ...but no .rpm package in sight.. (anyways, we can discuss that on Monday, since now You are busy)14:57
alteregosledges: not sure what's causing that. What are you doing when you get that message? Running qmake or building?14:57
Stskeepslbt: nah, you're right14:58
sledgeswhich triggers qmake14:58
sledges"/srv/mer/scripts/qmake" '/' -r -spec linux-g++ CONFIG+=release14:59
*** PMG has quit IRC14:59
lbtaportale: hmm I suspect your sb2 target is missing a couple of pkgs14:59
*** niqt has quit IRC15:00
lbtrpm-build and meego-rpm-config15:00
lbtI've been putting them into my minimal setups now15:00
lbtaportale: is your code pushed? even not working?15:01
aportalelbt: Only the code thath generates the .yaml file is pushed.15:02
lbtI need to fix the nemo minimal root issue and then I need to push a tarball15:02
lbtpush it all :)15:02
aportalelbt: the rest is in the mer-packaging branch (and pushed)15:02
lbtthen I can look at it over the w/end - if it assumes those command then I'm sure it'll work :)15:02
aportalelbt: But I need some hacking to get the packaging into a deploystep15:02
lbtyeah, I can understand that15:03
sledgesalterego, ^15:04
aportalelbt: with the two additional packages I get a bit further, but then: "error: Bad owner/group: /home/mersdk/rpmbuild/SOURCES/nemoapp.tar.bz2"15:04
* aportale did not create /home/mersdk/rpmbuild/ manually15:04
*** rymek has quit IRC15:04
sledgestrying to break it down by launching the command manually15:04
lbtah :(15:05
lbtaportale: all this stuff is fixed here but I had to do the 486 cross- stuff15:05
aportalelbt: OK, no worry. I am not blocked by this, since I need to do the buildstepp stuff first.15:06
lbtessentially you need to fix etc/passwd and etc/group in the sysroot15:06
aportalelbt: I just wanted to test if the commands work when invoked manually15:06
lbtin the vm grep sdkbuild from /etc/passwd and >> etc/passwd in sysroot15:07
lbtsame for etc/group15:07
sledgesalterego, I have so many --binds and symlinks around my system, that mersdk clearly stops liking things15:07
Stskeepssledges: i'm pretty sure mersdk just adds on top of that15:07
sledgestrue, but it doesn't get lost only amongst the stuff it added itself ;)15:08
*** veskuh has quit IRC15:08
*** PMG has joined #mer15:12
*** Sfiet_Konstantin has quit IRC15:14
*** jpetrell_ has quit IRC15:14
*** yunta has quit IRC15:14
*** Moo-- is now known as Moo-_^15:15
sledgeshalf awake? :)15:15
alteregoCan't wait for SDK VM integration proper with qt creator ;)15:18
alteregosledges: maybe you should try using the VM :P15:18
*** rymek has joined #mer15:19
kulvethe-gibson: could you also provide your kernel config for reference?15:20
*** pirut has quit IRC15:21
sledgesalterego, I sorted it out, now getting good output from command line, but qtcreator still complains like before (same qmake invocation)15:22
*** kostaja has quit IRC15:26
*** Venemo_N9 has quit IRC15:27
*** mikhas has joined #mer15:30
*** veskuh has joined #mer15:33
Stskeepswhy is the perl build system as crazy as the language?15:34
*** jpetrell has joined #mer15:34
lbtit's old15:35
lbtit runs on windows too15:35
lbtit has tests15:35
CosmoHillwith people it's the older the wiser15:42
CosmoHillwith code it's the older the more hacky and held together with spit and sticks15:43
CosmoHillalso it's perl15:43
sledgesand bugs covered15:43
*** veskuh has quit IRC15:44
lbtto be fair perl has very few bugs ... other than the syntax15:46
*** ravirdv has joined #mer15:49
*** kallecarl_ has joined #mer15:50
*** rcg has joined #mer15:50
*** aportale has quit IRC15:51
* sledges has just successfully ported qt4 to qt5 and compiled it for armv7hl under mer target (outside qtcreator, but qtcreator was still prividing intellisense for qt5-arm) !!!!!!15:52
*** rymek has quit IRC15:53
lbtStskeeps: looks like binutils worked at last15:53
*** icota has quit IRC15:53
lbtsledges: neat - that's been the design goal15:53
sledgesyou guys rock!!!15:54
lbtgood group effort :)15:54
sledgesI did not find any other qt5 tutorials out there, apart from with Mer ;)15:55
sledgestime-to-market springs to mind ;)15:55
*** aportale has joined #mer15:55
*** himamura has joined #mer15:56
*** himamura has joined #mer15:57
sledgeshm, on launching it on Nemo, app says Aborted15:58
sledgesI should start with hello world :{15:58
sledgesor find way to debug it15:58
CosmoHillI wrote a hello world program, it said "Goodbye Cosmo"15:58
CosmoHillit was creepy15:59
*** reels has quit IRC16:00
*** ortylp has joined #mer16:00
*** mikhas has quit IRC16:03
*** calvaris has quit IRC16:07
sledgesguys i'm getting this:
sledges(gdb says the same)16:08
*** stefan_schmidt_w has quit IRC16:12
*** mikhas has joined #mer16:17
*** Buikhj has joined #mer16:17
*** popey has joined #mer16:18
*** Buikhj has quit IRC16:24
sledgesand a simplest qt5 qml2 hello world segfaults16:25
sledgesi do not have nemo target on my PC, I am building against mer-core target - should not be a problem ?16:26
sledges(launching on latest nemo on N9 though, with qt5-* stuff installed)16:26
*** ravirdv has quit IRC16:29
*** popey has quit IRC16:31
*** slaine has quit IRC16:32
*** hazchemix has quit IRC16:34
jonnisledges: that looks like that you just havent compiled/installed xcb platform plugin, then your options is to compile xcb, or start the demoes by specifying with commandline which platform to use.16:39
*** arcean has joined #mer16:40
*** Milhouse has joined #mer16:43
sledgesqt5-plugin-platform-xcb is available on nemo to be installed16:44
sledgesand also on local PC mersdk mer-core could have it installed, what do you mean by compiling ?16:44
sledgesdoes my source code have to bind against it and thus get recompiled ?16:44
*** FSCV has quit IRC16:46
*** ortylp has quit IRC16:49
Stskeepssledges: install qt5-plugin-platform-xcb on device17:00
Stskeepssimply as that17:00
*** CosmoHill has quit IRC17:04
sledgesStskeeps, done that, didn't work17:04
sledgestrying from 017:04
*** FSCV has joined #mer17:04
sledgesI can't launch a qt5 app on my desktop on the first place - same error: segfaults17:04
sledgesif I show qmessagebox or anything else graphics17:05
sledgeson my desktop - archlinux x64, packages qt5-qt{core,quick1(as declarative)}-git are installed from AUR17:05
sledgesqtcreator configured17:05
*** CosmoHill has joined #mer17:05
*** gabriel9|work has quit IRC17:06
sledgeslibxcb in place17:06
sledgesStskeeps, ^17:06
Stskeepssledges: sorry, i don't support AUR :/17:07
Stskeepsif it was in mer sdk it'd be easier17:07
sledgesbut it segfaults the same way as on nemo-arm17:07
Stskeepssure, but .. :P17:07
sledgesd: :)17:08
Stskeepsinstall the right plugin for your distro then :P17:09
*** Sfiet_Konstantin has joined #mer17:12
*** JvD_ has quit IRC17:18
*** morphis|away is now known as morphis17:22
*** vgrade_ has quit IRC17:25
*** Hoolxi has joined #mer17:26
*** sirdancealot has quit IRC17:27
*** NIN101 has quit IRC17:33
*** panda-z has joined #mer17:34
*** rcg has quit IRC17:34
*** sirdancealot has joined #mer17:35
StskeepsBostik: 211 done, 142 to go, 2 failed (openssl, glibc)17:36
Stskeepsnot bad so far17:36
*** ssirkia has left #mer17:37
sledgesIT ALL WORKS!!!17:44
sledgesi was missing qt5-declarative-qtquick-plugin17:44
*** ortylp has joined #mer17:45
sledgeswhat a great start of weekend17:58
sledgesbut it's a mess in qt5: qt-declaratives provides qtquick, and qtquick provides qt-declarative :(17:58
*** gabrbedd has quit IRC18:04
*** tilgovi has joined #mer18:05
*** gabrbedd has joined #mer18:05
*** Attie has quit IRC18:10
*** udsslayer is now known as shadeslayer18:12
Bostiksledges: if my packaging has such bugs, please note them down so I can fix the thing18:26
Bostikthe chains have gone through *literally* half a dozen renames and/or module shuffles18:27
*** arcean has quit IRC18:27
*** panda-z has quit IRC18:28
*** Hoolxi has quit IRC18:31
*** JvD_ has joined #mer18:31
*** ljp has quit IRC18:36
*** ljp has joined #mer18:37
*** plfiorini has quit IRC18:40
*** mikhas has quit IRC18:43
*** lbt is now known as lbt_away18:52
*** aportale has left #mer18:53
*** furikku has quit IRC19:07
*** tilgovi has quit IRC19:07
*** pvuorela has joined #mer19:09
*** ali1234 has quit IRC19:10
*** npm has joined #mer19:14
*** ka6sox-away is now known as ka6sox19:19
*** ali1234 has joined #mer19:21
*** mikhas has joined #mer19:21
*** khetzal has joined #mer19:23
*** tilgovi has joined #mer19:33
*** sp3000 has quit IRC19:35
*** M4rtinK has quit IRC19:36
*** sp3000 has joined #mer19:36
*** nsuffys has joined #mer19:47
*** Jucato has quit IRC19:52
*** M4rtinK has joined #mer19:54
*** kallecarl_ has quit IRC19:54
*** Jucato has joined #mer19:55
*** kallecarl_ has joined #mer19:58
*** kallecarl_ has joined #mer19:58
*** ali1234 has quit IRC20:00
*** NIN101 has joined #mer20:03
*** ali1234 has joined #mer20:04
*** FSCV has quit IRC20:06
*** arcean has joined #mer20:16
*** plfiorini has joined #mer20:17
*** arcean_ has joined #mer20:17
*** arcean has quit IRC20:17
*** arcean_ is now known as arcean20:17
vgradeweekend at last20:18
specialfinally, time for all that work you didn't have time for during the week20:22
*** dod1 has quit IRC20:23
CosmoHillvgrade: :)20:24
*** morphis is now known as morphis|away20:26
sledgesBostik, not you, it"s ArchLinux AUR qt5 packager :) but will do for your side, well done Bostik !20:28
sledgesspecial, :D20:28
CosmoHillbah, so hard to find T9 phones :(20:29
CosmoHilli mean decent ones, not the £9.99 you give to a young child one20:29
CosmoHill"you searched for 'standard phone', did you mean 'shite'?"20:34
ali1234T9 is software20:35
ali1234isn't it?20:35
CosmoHilli mean a phone with a physical keypad20:35
ali1234a T9 phone is one with a number pad20:35
ali1234"Text on 9 keys"20:35
ali1234how about a nokia C3 touch & type20:36
ali1234one of the last phones they made before elop ruined everything. it's quite nice20:36
ali1234and also very cheap20:36
ali1234ie free on any contract20:37
ali1234C2-01 also20:38
ali1234not as nice looking in my opinion20:38
CosmoHilli did have a sony w995 but gone back to my sony k800i20:39
CosmoHilli like sonys20:40
CosmoHillor sonies :/20:40
ali1234k800i is horrible plastic rubbish that breaks in about 2 days20:40
ali1234but to each their own20:40
CosmoHillreally? I've thrown mine at walls, people and dogs and it still works20:40
ali1234it still works but i bet the case is all smashed up20:41
ali1234every one of those phones i have ever seen has button missing and cracked cover20:41
CosmoHilllooks fine20:41
CosmoHillI've had two nokias in my life, both died from a cracked screen20:43
CosmoHillone was a fat friend sitting on it20:43
CosmoHillanother was just random20:43
ali1234what was that matrix phone20:43
ali1234with the sliding cover20:43
ali1234nokia 711020:43
ali1234that was a bad phone20:43
ali1234peope only bought it cos it was in the matrix20:43
CosmoHillI want the phone the woman has the first animatrix video20:44
*** lamikr has quit IRC20:44
ali1234i want the phone zoolander has in zoolander20:45
*** merder has quit IRC20:46
CosmoHillstupid email20:47
CosmoHilljust went bing but the email was sent 7 mins ago20:47
ali1234email isn't supposed to be instant messaging20:47
ali1234it probably got greylisted or something20:47
*** mikhas has quit IRC20:51
Bostiksledges: thanks, nice to know20:55
*** pvuorela has quit IRC20:55
CosmoHillmy friend told me the only phones he'd get20:56
CosmoHillpretty much the most expensive ones and total overkill for what i want20:56
*** mdfe_ has quit IRC20:57
*** notmart has quit IRC21:06
*** JvD_ has quit IRC21:16
*** fcorrea has joined #mer21:19
*** faenil has joined #mer21:24
*** ortylp has quit IRC21:28
*** Jucato has quit IRC21:32
*** kallecarl_ has quit IRC21:32
*** jpetersen has quit IRC21:35
*** Sfiet_Konstantin has quit IRC21:37
*** JvD_ has joined #mer21:41
*** nsuffys has quit IRC21:43
*** morphis|away is now known as morphis21:53
*** fcorrea has quit IRC21:56
*** fcorrea has joined #mer21:57
*** ShadowJK has quit IRC21:58
*** ShadowJK has joined #mer21:58
*** Guest40482 has quit IRC21:58
*** Termana has joined #mer21:59
*** Termana is now known as Guest9457721:59
*** khetzal has quit IRC22:04
*** morphis is now known as morphis|away22:08
*** totte has quit IRC22:09
*** JvD_ has quit IRC22:10
*** totte has joined #mer22:11
*** vasily_pupkin has joined #mer22:17
*** gabrbedd has quit IRC22:17
*** rcg has joined #mer22:20
*** vasily_pupkin has quit IRC22:22
*** himamura has quit IRC22:40
*** rdqfdx has quit IRC22:40
*** NIN101 has quit IRC22:41
*** vasily_pupkin has joined #mer22:45
*** sp3000 has quit IRC22:57
*** faenil has quit IRC23:01
*** sp3000 has joined #mer23:02
*** CosmoHill has quit IRC23:08
*** sp3000 has quit IRC23:23
*** vasily_pupkin has quit IRC23:36
*** vasily_pupkin has joined #mer23:36
*** rcg has quit IRC23:41
*** Hoolxi has joined #mer23:45
*** mer_ge has joined #mer23:47
*** M4rtinK has quit IRC23:53
*** blauzahl has quit IRC23:59

Generated by 2.11.0 by Marius Gedminas - find it at!