Saturday, 2011-12-17

*** dijenerate has joined #mer00:06
*** KaIRC has quit IRC00:06
*** lynxis_ has quit IRC00:12
[JT]I'm thinking about KDE Plasma Active Two on a Toshiba Thrive (NVidia Tegra 2). Has anyone attempted that yet?00:13
*** zabomber has joined #mer00:18
*** KaIRC has joined #mer00:19
*** ssirkia has left #mer00:27
*** sebsauer__ has joined #mer00:39
*** sebsauer_ has quit IRC00:40
*** trbs has quit IRC00:44
*** raignarok has quit IRC00:54
*** buser has joined #mer00:54
*** purban has quit IRC00:58
*** raignarok has joined #mer01:05
*** buser has quit IRC01:18
*** tilgovi has joined #mer01:23
*** tilgovi has joined #mer01:23
*** jvd_ has quit IRC01:27
*** beford has joined #mer01:31
*** M4rtinK has quit IRC01:31
*** sebsauer has joined #mer01:34
*** phaeron1 has quit IRC01:35
*** sebsauer__ has quit IRC01:35
*** GeneralAntilles1 has joined #mer01:38
*** GeneralAntilles has quit IRC01:40
*** tilgovi has quit IRC01:59
*** raignarok has quit IRC02:05
cxl000[JT] it is not listed on http://wiki.merproject.com/wiki/Community_Workspace. If you want to try an adaptation you will 1st need to be able to boot your own kernel.02:05
*** raignarok has joined #mer02:06
*** raignarok has joined #mer02:09
[JT]cxl000: Okay; I think I can probably boot my own kernel (the device is rooted and I've loaded ClockworkMod). I'll do some research on how to get that done.02:22
*** tilgovi has joined #mer02:24
*** tilgovi has joined #mer02:24
*** sebsauer_ has joined #mer02:26
cxl000[JT] Once you can know how to boot your own kernel you could look at http://wiki.meego.com/ARM/TEGRA2 and http://wiki.meego.com/ARM/TEGRA2/Notes02:27
*** sebsauer has quit IRC02:27
*** HazardousWaster has quit IRC02:28
*** raignarok has quit IRC02:28
*** sebsauer_ has quit IRC02:40
*** kiviluoto has joined #mer02:41
*** kiviluoto_ has quit IRC02:41
*** zabomber has quit IRC02:46
*** lynxis has joined #mer02:51
*** jvd_ has joined #mer03:09
*** dijenerate has quit IRC03:31
*** zabomber has joined #mer03:32
*** tilgovi has quit IRC03:42
*** dijenerate has joined #mer03:43
*** sledges has quit IRC03:52
*** berndhs_950 has joined #mer04:01
*** dl9pf_ has joined #mer04:05
*** dl9pf has quit IRC04:05
*** berndhs has left #mer04:06
*** KaIRC has quit IRC04:51
*** DocScrutinizer has quit IRC05:42
*** DocScrutinizer has joined #mer05:42
*** sebsauer_ has joined #mer05:42
*** disco_stu_droid has joined #mer05:53
*** disco_stu has quit IRC05:53
*** disco_stu_droid is now known as disco_stu05:54
*** beford has quit IRC05:55
*** sebsauer__ has joined #mer05:56
*** sebsauer_ has quit IRC05:57
*** Alison_Chaiken has quit IRC05:57
*** zabomber has quit IRC06:05
*** zabomber has joined #mer06:11
*** sebsauer__ has quit IRC06:22
*** sebsauer__ has joined #mer06:25
*** sebsauer__ has quit IRC06:30
Stskeepszz06:56
*** kiviluoto_ has joined #mer07:30
*** kiviluoto has quit IRC07:31
*** kiviluoto_ is now known as kiviluoto07:31
dm8tbrz07:48
dm8tbrriga airport, weather sux07:48
Stskeepswindy?07:53
Bostikmore likely rains perpendicular to the ground08:00
StskeepsBostik: what was the conflicts of having a dual stack qt4-qt5 btw? besides some things above it having to be built for both?08:06
Stskeepspkgconfig conflicts is one, i guess, what else?08:06
*** rdqfdx has joined #mer08:07
BostikStskeeps: library paths (/usr/lib/libQt*)08:11
Bostikthe plugins, modules and imports are in versioned paths, libs are not08:11
Stskeepsokay, but i would hope they have different so versions08:11
Stskeepslike libQtCore.so.5 ?08:11
Bostikokay, that holds true08:12
Bostikjust not for -devel :)08:12
Stskeepsyes, of course08:12
Stskeepsthat's another side of the story :)08:12
Stskeepsjust wondering if we can make mer a dual stack, so people on top can decide which one to use08:12
Bostikthen there's the issue of qmake, which would probably need to follow debian's route (qmake-qt4, symlink)08:13
Bostikah, the tools require the same08:14
Bostikbinaries, naked .so symlinks, pkgconfig files08:16
*** npm_ has joined #mer08:16
Stskeeps:nod:08:16
Stskeepsbut this is mostly for building not so much runtime08:17
*** npm has quit IRC08:17
Bostikoh for that it already works08:18
Stskeepsmy aim is basically that people can start making qt5 based stuff while rest of stack doesn't break :)08:19
Stskeepsperhaps with a pkgconfig(QtCore) >= 5.0 ish kind of thing08:19
BostikI feel a deluge of bug reports coming in :)08:20
Bostik"pkgconfig files missing/broken/faulty/empty/..."08:22
Stskeepsyeah..08:23
Stskeepsthey should work(TM)08:23
Bostikfor most parts they don't08:24
BostikI put patches in to create them properly but that's it08:24
Stskeeps:nod:08:24
Stskeepsthat's an upstream bug isn't it?08:26
Bostikyep08:26
*** l32606 has joined #mer08:41
*** M4rtinK has joined #mer08:46
*** marquiz_ has joined #mer09:12
*** vivijim has joined #mer09:13
*** jvd_ has quit IRC09:23
*** NIN101 has joined #mer09:27
*** mingwandroid has joined #mer09:36
*** M4rtinK has quit IRC09:38
*** jvd_ has joined #mer09:41
*** phaeron has joined #mer09:46
*** Venemo_N950 has joined #mer09:55
*** phaeron has quit IRC09:55
*** phaeron has joined #mer10:01
*** Venemo_N950 has quit IRC10:02
*** andre__ has quit IRC10:19
*** mdavey has joined #mer10:30
*** bergie has quit IRC10:42
*** buser has joined #mer11:00
*** chokzi has joined #mer11:04
*** phaeron has quit IRC11:07
mingwandroidStskeeps: Morning/afternoon.11:08
*** chokzi has left #mer11:09
Stskeepsmorn11:09
mingwandroidtrying out ICS on my t2 tab.11:10
mingwandroidthere's been some kernel work done that I need to get working for Mer for the vega.11:11
mingwandroidthis thing is booting slower than mer PA.11:12
*** mdavey has quit IRC11:14
Stskeepsthat's depressing11:24
mingwandroidthink I needed to reflash it or it's broke somehow. plus first boots are the worst boots.11:25
*** lynxis has quit IRC11:27
mingwandroidit booted! cool.11:27
*** npm_ has quit IRC11:29
mingwandroidnot bad, smooth.11:33
mingwandroidso I've got some kernel work to do on mer for vega..11:33
Stskeeps:nod:11:38
Stskeepsi'm fixing up last package atm, poppler-qt (not TC related)11:38
*** NIN101 has quit IRC11:45
*** mdavey has joined #mer11:45
*** NIN101 has joined #mer11:45
*** sebsauer__ has joined #mer11:54
*** Venemo_N950 has joined #mer11:54
*** mdavey has quit IRC11:59
*** arcean has joined #mer12:05
*** mdavey has joined #mer12:06
mingwandroidGetting unresolvables:12:11
mingwandroidlinux-firmware newt-devel u-boot-tools12:11
mingwandroidI don't need u-boot-tools, though getting uboot to work would be a good idea.12:12
*** sebsauer has joined #mer12:18
*** sebsauer__ has quit IRC12:19
Stskeepsu-boot-tools are usable for making uimage12:23
Stskeepshave you seen sage's new kernel packaging?12:23
mingwandroidno, got a link. I think I've not setup some dependencies right.12:26
Stskeepshttps://build.pub.meego.com/package/files?package=kernel-adaptation-pandaboard&project=CE%3AAdaptation%3APandaBoard12:27
*** Eismann has quit IRC12:30
*** Eismann has joined #mer12:31
mingwandroidok great. I'll investigate.12:36
mingwandroidgoogle say they'll release tegra-2 ICS kernel sources very soon.12:36
*** raignarok has joined #mer12:36
mingwandroidofc with lots of android crap in there :-(12:36
Stskeepswhat kernel version are they on these days anywa??12:37
Stskeepsanyway12:37
mingwandroid2.6.39 AFAIK12:37
Stskeepsok12:37
Stskeepsso not that bad anymore12:37
mingwandroidwe're currently 2.6.38-3 or something like that with the vega12:37
mingwandroidthere's a script in ICS AOSP to download nv drivers, wonder if modifying the url a bit would lead to hardfp versions ;-)12:39
Stskeepshehe12:39
StskeepsICS isn't hardfp is it?12:39
mingwandroidnaah. should be, but would break all apps12:40
Stskeepswell, all using native code, yeah12:40
mingwandroidwhen they initially started supporting armv7a they should've used that as a chance to move to hardfp12:40
*** tomeff1 has quit IRC12:40
*** berndhs has joined #mer12:40
mingwandroidand forced devs to supply binaries for both arm6 soft and armv7a hard.12:40
mingwandroidapp devs I mean.12:41
Stskeepsyeah..12:41
Stskeepsone idea that's been brewing around the place is something alike the LLVM based native client stuff12:41
Stskeepsportable NaCl one12:42
mingwandroidyeah, LLVM's gained so much momentum lately.12:42
mingwandroidan NaCl looks great. We should definitely get chromium browser building on Mer.12:42
Stskeepsi'm intentionally not promising binary compatibility in mer on app side because i think this is an area we really need to look into12:42
Stskeepsfor Qt side, i had a thought .. just like shader programs are supported and practically compiled on the spot, embedded in QML12:43
Stskeepsand executed on GPU12:43
Stskeepswhy couldn't you put CPU program there too, built with LLVM on the fly :)12:43
mingwandroidI read some discussions on LLVM dev list that they didn't think using their IR (bitcode is it?) for cross-platform was a great idea...12:44
mingwandroidlemme dig it out.12:44
mingwandroidhttp://www.phoronix.com/scan.php?page=news_item&px=OTk2Nw12:44
Stskeepsi don't think so either, so i was pondering actually to include code, like we include code for shaders12:44
mingwandroidDan argues that the intermediate representation for this Apple-sponsored compiler infrastructure is "a poor system for building a Platform."12:44
Stskeeps:nod:12:45
*** zenvoid has joined #mer12:45
Stskeeps'lo zenvoid :)12:45
mingwandroiddo you mean actual C/C++ source code compiled on the client by LLVM?12:45
Stskeepsyes12:45
zenvoidhello Stskeeps12:45
Stskeepsi mean, it's socially acceptable to build shaders12:45
mingwandroidshaders have very few deps12:45
mingwandroidmakes it sort of easy12:45
Stskeepsyeah, sure12:46
mingwandroidno automatic dynamic loading etc12:46
mingwandroidI had wanted to go down the IR route with a project until I read that stuff.12:46
*** M4rtinK has joined #mer12:46
Stskeepszenvoid: how are things? mer's coming along nicely, for armv6 too :)12:46
*** l32606 has left #mer12:47
*** l32606 has joined #mer12:47
*** sebsauer has quit IRC12:47
*** marquiz_ has quit IRC12:48
zenvoidwell... I'm trying to find a phone to buy now, can't use the openmoko anymore12:50
Stskeepsn900's quite hackable12:50
*** phaeron has joined #mer12:50
Stskeepsn9 too, for that matter12:50
*** phaeron has quit IRC12:51
zenvoidI would prefer the n900 actually, even with the smaller screen but...12:51
zenvoidnot easy to buy one new12:51
zenvoidnowadays12:51
*** Venemo_N950 has quit IRC12:51
Stskeepsyeah12:52
zenvoiddo you know if those problems with the usb port are real?12:52
Stskeepsyes they are12:52
zenvoidouch12:52
Stskeepsi've killed one myself12:52
Stskeepsthough it was due to a jackass stepping over my cord12:52
zenvoidhahaha12:53
zenvoidwell, getting one used and without warranty could be a problem then12:53
Stskeepsyeah12:53
Stskeepsn9 situation: you can hack the device fairly easily, flash own kernel, but with a "no warranty" notice.. kernel is 2.6.32 and we don't have redistributable rights for NFC and gps12:54
*** tomeff has joined #mer12:54
zenvoidI'm outdated and lost about the latest developments of maemo/meego/tizen/mer/whatever, trying to read a bit now12:54
Stskeepstizen = nothing public12:55
Stskeepsat least :P12:55
*** tomeff has left #mer12:55
StskeepsNemo mobile (based on Mer), http://www.youtube.com/watch?v=dGrvs7TbVBs&feature=BFa&list=UUt3jrWgutwPvS79NfGs9MQQ&lf=plcp12:56
Stskeepszenvoid: btw, didn't you use to use SB2 back in old Mer days?12:57
zenvoidthe n9 is nice but I like hw keyboard and don't like amoled... I think if nokia just shell the n950 is the one that I'll buy12:58
zenvoidI'm looking at chinese clones and some android phones too, like samsumg's optimous q212:58
zenvoidStskeeps: yes, I used sb2 but it has some bugs, then I coded custom software in C13:00
Stskeeps:nod:13:00
Stskeepswe're moving to using SB2 within OBS now, i've made a very nice prototype with it13:00
Stskeepsas a new cross compilation approach13:00
*** ashok has joined #mer13:01
zenvoidah, a few hints may interest you, learning during my experiments:13:01
ashokcan any1 help to install maemo5 on n80013:02
Stskeepsashok: sorry, not possible13:02
*** l32606 has quit IRC13:02
zenvoidStskeeps: my custom software was working ok, for static binaries I coded a very simple detection test and run them under complete emulation13:02
*** ashok has quit IRC13:03
Stskeepsok13:03
zenvoidbut there were complicated cases that scaped the fake chroot13:03
zenvoidvery difficult to detect13:03
Stskeepsyeah, i've spoken a bit with the author on the pitfalls13:03
*** mdavey has quit IRC13:03
zenvoidok13:04
Termanazenvoid, GTA04 on the cards perhaps?13:04
zenvoidStskeeps: for example I've found that the realpath binary sometimes failed, and it turns out that gcc/glibc optimization code inlines a call to the realpath syscall13:05
zenvoidthen I've found many other cases13:05
Stskeepszenvoid: interesting13:05
zenvoidyeah, it took days of debugging xD13:05
zenvoidif the syscall is inlined, ld preload does not work :(13:06
Stskeepsyeah, i guess in that case you could perhaps add a pragma to glibc not to inline it13:07
zenvoidTermana: yeah, its a posibility, I just would prefer to change the openmoko case and screen too :)13:07
*** l32606 has joined #mer13:08
zenvoidStskeeps: probably yes, compiling with -O0 worked too13:08
Stskeepszenvoid: other things i'm brewing on is an extension of the SB2 work, where we basically split Mer into Toolchain and Core, effectively that we bootstrap glibc, gcc and such from x86 side into target and our target no longer needs to be self-hosting13:09
zenvoidbut I just taken the motherboard of my touchbook and used it for native compilations instead13:09
Stskeeps:nod:13:09
*** l32606 has quit IRC13:13
*** marquiz_ has joined #mer13:13
marquiz_afternoon13:13
Stskeepsafternoon marquiz_13:13
marquiz_on my way to tampere, this time :)13:14
marquiz_in a bus, of course13:14
Stskeepshehe, you certainly get around in finland :)13:14
Stskeepsbut i guess choirs are quite active in december season13:14
marquiz_i thought this would be good time to ver bump mce and dsme13:14
marquiz_yep13:14
* Stskeeps glances at ikea chairs to assemble vs software to assemble..13:16
marquiz_i recommend physical excercise :)13:17
marquiz_although ikea chairs might prove to be lot more complex that software13:18
*** HazardousWaster has joined #mer13:24
dm8tbrand they sometimes miss dependencies13:28
mingwandroidalways13:30
*** tomeff has joined #mer13:30
*** Venemo_N950 has joined #mer13:34
zenvoidStskeeps: your comment about ikea led me to think about "ikea phones", and it turns out that they exist :D13:34
zenvoidhttp://www.engadget.com/2009/05/21/flow-is-like-the-ikea-bookshelf-of-android-phones/13:34
Stskeepshehe13:37
*** arc_mat|tp has joined #mer13:40
*** Venemo_N950 has quit IRC13:42
*** KaIRC has joined #mer13:44
*** Venemo_N950 has joined #mer13:51
Stskeepsah, ikea chairs gone, now for some python..13:56
marquiz_sounds like good activity for saturday evening :)#13:57
Stskeepsyeah.. going to some birthday party in 2 hours or so, so trying to squeeze in some code before that13:57
marquiz_rebasing mce patches13:58
marquiz_argh, not a single patch rebases cleanly13:58
Stskeepsi'm afraid to ever touch pulseaudio again :P13:59
marquiz_ah, i have birthday today!!!13:59
Stskeepsoh congratualtions :)14:00
marquiz_thx!14:00
Stskeepser, congratulations :)14:00
*** mdavey has joined #mer14:01
*** Venemo_N950 has quit IRC14:01
*** tomeff has quit IRC14:03
marquiz_i guess i gotta grab a beer later today :)14:03
*** mdavey has quit IRC14:06
*** kiviluoto has quit IRC14:15
*** kiviluoto has joined #mer14:16
Stskeepslbt_away: ping14:28
*** alien_ has quit IRC14:40
*** alien_ has joined #mer14:45
*** andre__ has joined #mer14:45
*** javier has left #mer14:47
* marquiz_ is arriving in tampere14:56
marquiz_mce patches rebased, nice timing :)14:57
Stskeeps:)14:57
marquiz_doesn't build, though :(15:02
marquiz_have to fix that tomorrow or so15:02
*** lbt_away is now known as lbt15:07
lbtStskeeps: pong15:07
marquiz_bye15:15
*** marquiz_ has left #mer15:15
Stskeepslbt: so, i'm a vendor and i am maintaining a mer fork for product purposes.. how do i rebase from one core release branch to another?15:16
Stskeepsbased on our idea of branching off one release branch per weekly release15:17
Stskeepsi mean, there's commits in the stabilization branch that likely doesn't exist in the linear progression from the branch point15:18
lbtyep - that's the kind of "day in the life of" item I want to do15:18
Stskeepsi'm coding a bit on MDS and it doesn't clean to have a seperate file indicating what branch is the current prerelease branch15:19
Stskeeps+seem15:19
Stskeepsso i'm pondering how to encode that information the best15:19
* lbt presses reset button in brain15:20
lbtOK - so commits in stabilisation ... not our problem15:21
lbtclarify15:21
lbtcommits in Mer release stabilisation15:21
lbtwe do another release from linear + some stable patches in a weeks time15:22
lbtvendor needs to rebase15:22
lbtthat may not be clean15:22
*** tomeff has joined #mer15:22
lbtand they only rebase *their* patches15:22
lbtagree?15:23
Stskeepsright, linearly we push commits (read: core changes) into 'master', once a week we branch off, apply changes to the branch until it's releasable according to a number of criteria, do same next week15:24
lbtyes15:24
Stskeeps1) how do we indicate best what's the current prerelease branch15:25
Stskeeps(sorry for that tangent)15:25
Stskeepsok, forget i ever asked that question :)15:25
lbtyeah15:25
Stskeepsisn't it too rough on the vendor to do having to do a rebase? is there any way we can make it easier to jump from release to release?15:26
lbtlets think of the cases15:27
lbt1. vendor has own package ...15:27
lbt2. vendor has a patch not accepted (yet)15:27
Stskeepsthe thing is that the vendor is indeed not seeing the changes from week A to week A+1 in a linear manner15:27
lbt3. vendor has a patch that won't get in15:28
lbtyeah, but why are they even modifying Mer?15:28
Stskeepsright15:28
Stskeepswhy? because they have to for various technical reasons :)15:28
lbt1,2,3 above15:28
Stskeepsbut yes, those are the three cases15:29
lbtso, only 3 has patches they have to carry and re-apply15:29
lbt2 should vanish15:29
Stskeepsthe idea was to provide vendors a linear stream of core updates, wasn't it?15:29
lbtI think master is a linear stream ... and if they have no patches then yes15:30
lbtif they have patches then that's not really possible15:30
lbtthey have to rebase15:30
Stskeepsok, if we imagine they apply their patches on master, then they can easily rebase on a release branch15:30
lbtconsider 2 vendors applying the same semi-trivial patch15:30
lbtyes15:31
Stskeepsthat makes more sense, i guess15:31
lbtand we hope that release branches will have minimal deltas from master15:31
lbttypically I expect any fixes will go straight to master too15:32
lbtunless someone jumps in with an API change minutes post-fork15:32
Stskeepsthat makes sense15:33
* Stskeeps adds a postit for vendor advice15:33
lbtso the normal case is linear15:33
lbt"day in the life"15:33
Stskeeps"apply and follow on master but rebase on release branches"15:33
lbtyes15:33
Stskeepsok15:34
lbtas we get to more major releases we expect longer branches15:34
Stskeepsyes15:34
lbtI also think this becomes a whole lot easier when we use pristine tar15:34
lbtand have a 'proper'-ish git tree15:35
Stskeepsi'm pondering to drop tarballs and simply have a distfiles dir instead, preferred to in the git15:35
arc_mat|tpits pretty inconvenient to rebase your own work on upstream updates, if you don't intend to contribute them back15:35
lbtarc_mat|tp: what's the alternative?15:35
arc_mat|tpdifficult to say...15:35
Stskeepsarc_mat|tp: well, best thing we can do is to make it easier15:36
lbtyeah - that point is actually the cost of OSS+proprietary15:36
lbt"patch weight"15:36
lbtit's what Samsung are about to learn :D15:36
arc_mat|tpfrom own experience, with e.g. linux kernel, we have a lot of patches that will never make it upstream because they are hacks15:36
lbtNokia learnt it - hence MeeGo (I think)15:36
lbtarc_mat|tp: exactly ... so the cost is rebase vs cleanup15:37
lbtand that must be calculated over the product lifetime15:37
Stskeepsarc_mat|tp: yeah, so we have to make it easy to follow and merge our local changes15:37
arc_mat|tpindeed15:37
Stskeepsyour local changes, that is15:37
Stskeepslbt: ok, so, from a git POV, how do i best indicate what's the current prerelease branch? or even latest release15:38
arc_mat|tpStskeeps: what we usually do is either git merge upstream and then rebase on our internal branch, or cherry pick a range of patches from upstream, fixing conflicts as we go along15:38
Stskeepsarc_mat|tp: yeah15:38
Stskeepsarc_mat|tp: that's what we want to enable with mer, both on package and core level15:38
lbtarc_mat|tp: we're proposing the inverse15:38
Stskeepswe are?15:39
lbtI think so15:39
arc_mat|tprebasing our internal branches for upstream "maintenance" updates is impossible15:39
lbtwe suggest vendors cherry-pick/re-apply their patches to new upstream15:39
lbtfixing conflicts as they go15:39
arc_mat|tpwe'd have to lock the whole repo for hours to avoid loss of work15:40
arc_mat|tplbt: that's maximum inconvenience15:40
lbtwhich package?15:40
arc_mat|tplbt: I'm not talking about a specific package, or even mer15:41
lbtbut this is typically HA space issue?15:41
arc_mat|tpwhat is HA space?15:41
lbthardware adaptation15:41
Stskeepsi think what arc_mat|tp is trying to say is that in his experience with linux kernel, even rebasing internal branches for upstream maintaining updates is difficult?15:42
arc_mat|tpnot really15:42
Stskeepsmaintenance, that is15:42
*** raignarok has quit IRC15:42
lbtStskeeps: kernel is different - and not in Mer15:42
arc_mat|tpmy experience is mainly with android15:42
Stskeepslbt: right, i know, but in principle a similar issue15:42
arc_mat|tpStskeeps: +115:42
lbtagreed - but kernel really is different15:42
lbtwhat arc_mat|tp is talking about is taking over internal maintenance of a package15:43
lbtisn't it?15:43
arc_mat|tpyes15:43
lbtso to a large degree that's what Mer is supposed to avoid15:44
lbtso it's not a primary use-case15:44
lbthowever15:44
lbtmer doesn't preclude it15:44
arc_mat|tplbt: I think its mainly a matter of bandwidth15:44
lbtwhose? Mer or vendor?15:45
arc_mat|tplbt: what we would do is to pick a stable release, and productize it.15:45
arc_mat|tplbt: vendor15:45
lbtOK - and that's fine15:45
lbtbut we're back to the "co-operative" nature of Mer15:45
Stskeepslbt: we should really set up a list of vendor use cases..15:45
lbtproductisation happens in the open15:45
lbtfor the non-differentiation packages15:46
arc_mat|tplbt: it depends on what a "stable" release is, and how many there are within one of our product development cycles15:46
lbtand we need to minimise overhead15:46
lbt*nod*15:46
lbtcurrently we're talking about weekly iteration15:46
lbtStskeeps: agreed15:46
arc_mat|tplbt: and if course if there is a market driven necessity to closely follow "stable" releases for marketing reasons15:46
arc_mat|tplbt: like for android15:46
arc_mat|tplbt: it's close to impossible to sell a device with an "old" version of android15:47
lbtyep - and Mer is about being able to influence what's in stable15:47
lbtyou don't have to contribute15:47
lbtbut then your patch weight goes up15:47
lbtbut the cost of contribution is not zero15:47
lbtbalance15:48
lbtI think the initial focus of this chat was how to minimise cost of contribution15:48
lbtI agree we'd like to support carrying patches forward15:48
lbtI think the split to HA and MWare is good15:48
lbtthe MW is less likely to have closed patches15:49
lbtand HA is down to the vendor with 'stable' apis15:49
lbtI have toyed with describing Mer as "from the kernel .config to Qt"15:49
arc_mat|tpHA comes to play at the start of each new product generation, when new hardware is designed. It's a critical step, because you want a full system up as soon as possible for hardware verification15:49
lbtyep - and we should be clear about what we need there - particularly things like DRI/udev etc15:50
arc_mat|tpbut it's mostly manual labor. not really easy to carry on patches that you made for the previous product generation15:51
arc_mat|tpfor example, our kernel is usually "thrown away" each year.15:51
lbtyep15:51
arc_mat|tpand we start to productize from whatever version is feasible at a certain point in time15:52
lbtat the kernel level I'm not surprised15:52
arc_mat|tpby that point our last product kernel and upstream have diverged so much that we cannot reasonably rebase our patches15:52
lbtupstream = android ?15:53
Stskeepsthink that's covers for normal linux products too15:53
lbtso you *can* apply that model to Mer ... and there's not a lot we can do to help15:53
arc_mat|tpupstream for the kernel is rarely "android", as in AOSP, for example, but whatever the system (chip) vendor chose to deliver15:53
lbtOK15:54
arc_mat|tpbut for all other packages, yes, in my case it's android15:54
lbt(at least I can't trivially see anything we can do)15:54
arc_mat|tpyeah15:54
arc_mat|tpit is of course most convenient for us to have a linear history from release to release15:54
lbtthe alternative is to spread the ongoing deltas (especially in non-kernel areas)15:55
lbtthat makes each change *relatively* easy to make - but if you're not using that then why would you?15:55
lbtarc_mat|tp: linear?15:55
lbthmm15:56
lbtthe problem there is that it can't be generalised15:56
arc_mat|tpwithout having to dive into branches15:56
lbtlets say we release 1.1 and you use it15:56
lbtwe *could* rebase master of Mer to 1.1 and then your upgrade would be linear15:56
lbtnow we release 1.1.115:57
lbtsorry15:57
lbtwe release 1.215:57
lbtthen we release 1.1.115:58
lbtwhich do we rebase master onto ... either way, vendors using 1.2 or 1.1.1 can't have a linear upgrade path15:58
arc_mat|tpthat depends on if there is a necessity to move to 1.2, for whatever reason15:58
lbtexactly15:59
lbtso if you don't then we'd have to rebase 1.2 onto 1.1.115:59
lbtwhich doesn't make sense15:59
arc_mat|tpyes15:59
arc_mat|tpits mainly a question of roadmap, and policy15:59
lbt*nod*16:00
arc_mat|tpit depends on how fast you want/need to follow your own upstreams16:01
Stskeepsour situation might also be different in terms of that we're not dealing with one huge source tree to patch for a product, but instead a core will have patches against it (not for hardware) and a bunch of packages outside it by vendor compiled into it16:01
lbtyes - that non-hardware part is crucial16:01
Stskeepswhich means it's just the "Core" programme in the company is not terribly intensive16:01
arc_mat|tpI think maintenance is not really a critical issue (unless things are really screwed up), what matter is how easily you can move from one "major" release to another16:03
arc_mat|tpthat is what gives us most headaches, for example when moving from honeycomb to ics on android16:03
*** raignarok has joined #mer16:03
lbtyeah - and we're hoping that most of that will be in the Qt space16:04
Stskeepsif you don't mind me asking, are you OHA participants or not?16:04
Stskeepsie, able to see the tree before release16:05
arc_mat|tpnot even OHA participants are able to see that16:05
Stskeepsok, so android's more closed than i thought16:05
arc_mat|tpthere is a circle of a few chosen vendors that are doing the "GED", the Google Experience Devices, and only one of them is chosen each year16:06
arc_mat|tpor, for each new version16:06
arc_mat|tpStskeeps: indeed16:06
arc_mat|tpStskeeps: there is AOSP of course, but the updates you get to see are only "maintenance", not upcoming new major releases16:07
arc_mat|tpStskeeps: and even as OHA members you're not involved in the development.16:07
arc_mat|tpStskeeps: contribution is limited to bugfixes, if they are interesting for a broader audience16:08
arc_mat|tpbecause there a are quite a number of areas that google chooses to "neglect", i.e. they barely work ;)16:08
arc_mat|tpmtp for example16:09
arc_mat|tpbut that's fine, really. you would not contribute your differentiators anyway16:10
lbtarc_mat|tp: I think we must remember that *that* is what Mer is about16:10
*** vakkov has joined #mer16:10
lbtWe're expecting differentiatores to sit above Mer16:10
arc_mat|tpwhich is a totally legit assumption.16:10
*** meng has joined #mer16:11
menghello16:11
lbthey meng16:11
menghey , mer project?16:12
lbtarc_mat|tp: we do think that some things may move from differentiated to shared16:12
lbtmeng: we are indeed :)16:12
mengLooking forward to it16:12
lbtmeng: what's your interest?16:12
mengdo you guys know about XMS had release for N9?16:12
mengmy interest? programming, but I don't think i could help in develop Mer project16:12
arc_mat|tpI think the decision about which version to productize on would be driven from stuff above Mer anyway. But not entirely. sometimes you need core stuff to get all aspects of your hardware running...16:12
mengI'm still learning vb.net code..16:12
lbtmeng: hehe16:13
Stskeepsmeng: everybody started somewhere :)16:13
mengthanks16:13
mengbtw,16:13
mengthis is the first time i use IRC16:13
Stskeepsyou seem better at it than most people we see16:13
lbtmeng: try Qt/QML then - you should be able to handle both16:13
lbtarc_mat|tp: yes - and I think Qt and HTML5 apis would be in the shared areas16:14
menghey sts, thanks.16:14
mengQT and QML?16:14
Stskeepsmeng: are you able to see youtube videos?16:14
mengyes.16:14
Stskeepsmeng: http://www.youtube.com/watch?v=Fr5FuGhTqm8&feature=results_main&playnext=1&list=PL084A46CE1D502DB3 - it's basically a language to create exciting user interfaces16:15
Stskeepsmeng: quite easy to write, my wife can do it :)16:15
Stskeepsyou can even prototype in photoshop and export to QML16:15
mengalright, thanks for your information.16:16
mengjust found out something special in google16:16
mengsearch let it snow at google.16:16
Stskeepshehe :)16:16
mengStskeeps, owner of n900?16:17
Stskeepsmeng: yeah, i have two16:17
mengWow, that's great. I've only one.16:17
Stskeepslbt: anyway, since i'm implementing mds.. how do you propose we figure out what's most recent prerelease branch? date wise maybe?16:17
Stskeepsmeng: nokia 770, n800, n810, n900, n950 but no n9 yet ;)16:17
Stskeepslbt: prelease_week40 kind of thing16:17
mengO_O16:18
mengawesome.16:18
mengwell, i lost the oppurtunity.16:18
mengi started to learn programming in 2011,16:18
mengi can't develop for n900, can't get N950.16:18
Stskeepsthere'll be more devices, hopefully16:19
Stskeepsi develop for mer in a virtualbox VM16:19
Stskeepsas well16:19
Stskeepscan run mer on old pc hardware as well :)16:20
lbtStskeeps: "most recent" ?16:20
mengneed an intel proc isn't it ?16:20
Stskeepslbt: yes, latest16:20
Stskeepsmeng: nop, we have i486 port which should work on non-Atom too16:20
mengeven AMD?16:21
Stskeepsyes16:21
mengthat's great.16:21
lbtsymbolic branch name ?16:21
Stskeepswhat on earth is that?16:21
lbta branch that we randomly make point to the most recent16:21
lbtforced push16:22
Stskeepsok, that's a git feature?16:22
lbtabuse?16:22
lbtbut have a branch called "latest-pr"16:22
lbtit may share HEAD with week-2316:22
lbtand then next week it shares HEAD with week-2416:22
lbtbut always latest16:23
Stskeepsyeah, okay16:23
Stskeepsthat could work16:23
Stskeepsthere's git symbolic-ref too?16:24
menghey I've a question but not related to mer project :S16:26
Stskeepsjust ask, don't ask to ask :)16:26
mengis it hard to port a QT apps from N9 to N900?16:27
Stskeepsit depends on what APIs yu use16:27
Stskeepswe have Nemo Mobile which is for N900 which is closer to N9 architecture16:27
lbtStskeeps: that may be even better16:27
vakkovhttp://www.youtube.com/watch?v=6HtzwswaN-k16:27
vakkov:)16:28
mengvakkov, thanks for the link, i saw that just now.16:28
mengNemo use deb extension too?16:28
*** tomeff has quit IRC16:28
Stskeepsno, we use rpm16:28
mengsame as meego CE?16:28
Stskeepsyeah, Nemo is meego CE continued on top of Mer16:29
*** tomeff has joined #mer16:29
mengthen how do we install deb file on Nemo? since n9 is deb, nemo is rpm16:30
mengand what can i do to contribute this project?16:31
lbtbbiab16:32
mengI'm not that good as you guys but I'm willing to spend my free time to learn.16:32
*** tomeff has quit IRC16:34
*** tomeff has joined #mer16:34
Stskeepsmeng: it's just packaging, so the binaries are inside16:34
mengokay, thanks for your information .16:35
Stskeepsand extractable with simple commands :)16:35
mengalright. :D16:35
mengso when is the next release?16:36
Stskeepsmer releases each week16:36
Stskeepswe try to incrementally improve it16:36
mengDid nemo become smoother? compare to 1.2.9 CE16:36
Stskeepsyes it did, we already gained quite a lot with the rebase to meego 1.316:37
Stskeepsand mer is much lower memory footprint16:37
wmaroneStskeeps: have you used the %pre section in a kickstart to manipulate files before?16:37
Stskeepswmarone: i haven't used %pre, i usually just use %post16:37
wmaroneok16:37
Stskeepsi didn't even know there was a %pre :)16:37
wmaroneyeah, I need to use %pre to generate the image, unless there's a flag for "part" that allows me to specify the starting offset of a partition16:38
*** mdavey has joined #mer16:38
*** lynxis has joined #mer16:39
Stskeepsi wonder, since we would have to need same thing as pandaboard i would assume16:39
Stskeeps / beagleboard16:39
wmaroneyeah16:39
Stskeepssage would know but i think he's having a proper weekend for once :)16:39
wmaronealways good16:39
wmaronethe only open question is how the filename for the image is handled within the .ks16:40
Stskeeps:nod:16:46
Stskeepsbbl16:46
mengone more question !! :D hows the battery consumption when n900 connect to 3g or wifi?16:46
*** mdavey has quit IRC16:46
mengwill it drain battery like nitdroid online?16:46
*** trumee_ has joined #mer16:48
*** trumee has quit IRC16:50
*** trumee_ is now known as trumee16:50
menghm..16:52
mengaynone there?16:52
mengi still remember that, 1.2.9 CE has a bug in conversation. unable to scroll down to read the latest message when you have too many message from same user..16:52
mengis the problem still exist?16:53
*** trumee_ has joined #mer16:55
*** trumee has quit IRC16:56
*** trumee_ is now known as trumee16:56
*** mdavey has joined #mer16:59
mengtesting one two three..17:02
*** Ans5i has quit IRC17:06
mengHello17:12
*** meng has quit IRC17:14
*** SharkCrew has joined #mer17:14
*** tomeff has quit IRC17:14
*** lynxis has quit IRC17:35
*** gabrbedd has joined #mer17:47
*** raignarok has quit IRC17:52
*** slx is now known as swerden17:54
*** toscalix has joined #mer18:07
*** kiviluoto has quit IRC18:07
*** kiviluoto has joined #mer18:07
*** mingwandroid has quit IRC18:10
*** mingwandroid has joined #mer18:10
*** Anssi138 has joined #mer18:17
SharkCrewHello18:26
*** raignarok has joined #mer18:35
*** SharkCrew has quit IRC18:42
*** berndhs has left #mer18:42
*** ALoGeNo has joined #mer18:49
*** ALoGeNo has joined #mer18:49
*** phaeron has joined #mer18:51
*** berndhs_950 has quit IRC18:53
*** kiviluoto has quit IRC19:00
*** berndhs_950 has joined #mer19:04
*** kiviluoto has joined #mer19:10
*** berndhs has joined #mer19:10
*** araujo has quit IRC19:10
*** vakko has joined #mer19:13
*** vakkov has quit IRC19:13
*** vakko is now known as vakkov19:13
*** araujo has joined #mer19:24
*** araujo has quit IRC19:26
*** raignarok has quit IRC19:28
*** raignarok has joined #mer19:28
*** araujo has joined #mer19:39
*** alexxy has joined #mer19:46
*** berndhs_950 has quit IRC19:47
*** ALoGeNo has quit IRC19:48
*** berndhs_950 has joined #mer19:48
*** berndhs has quit IRC19:49
*** Sage has quit IRC19:50
*** vivijim has joined #mer20:00
*** toscalix has quit IRC20:00
*** Sage has joined #mer20:01
*** ALoGeNo has joined #mer20:07
*** ALoGeNo has joined #mer20:07
*** gabrbedd has quit IRC20:07
*** lynxis has joined #mer20:09
Anssi138IOError: [Errno 2] No such file or directory: '/usr/local/bin/mic'20:10
*** berndhs_950 has quit IRC20:12
*** Wong has joined #mer20:13
WongHello20:13
WongHello20:13
Wonganyone there?20:13
Anssi138maybe20:14
Wongjust now20:15
Wongi downloaded the image of mer project20:16
Wongi use GnuWin32 to decompress it20:16
Wongand now the image gone20:16
Wongi tried twice.20:16
Wongimage missing in action and hard disk space gone20:16
Wongany idea?20:17
Wonghmm...20:19
Wonganyone there?20:19
*** ALoGeNo has quit IRC20:20
*** Wong has quit IRC20:21
*** Nicholas has joined #mer20:25
NicholasHelllo20:25
Nicholasanyone there?20:25
*** Nicholas is now known as Guest7201620:25
_av500_</crickets>20:25
Guest72016hi.20:25
arc_mat|tpdamn, a deja vu. they're changing the matrix! out!!20:26
Guest72016hi20:28
Anssi138ho20:29
Anssi138i am trying to build mer with http://wiki.merproject.org/wiki/Trying_it_out. the mic-creator fails due missing /usr/local/bin/mic, i wonder if the bootstrap is ok20:31
*** Guest72016 has quit IRC20:33
*** berndhs has joined #mer20:43
StskeepsAnssi138: missing /usr/local/bin/mic? mic2 or mic?20:50
*** slx has joined #mer20:58
Anssi138$ which mic20:59
Anssi138/usr/bin/mic20:59
Stskeepshmm21:01
Stskeepsshow me full log?21:01
*** swerden has quit IRC21:02
*** niqt has joined #mer21:12
Anssi138http://pastebin.com/zsGNTgK021:16
Stskeepsyeah, that's a missing bootstrap.. hang on21:17
StskeepsAnssi138: mic-create-bootstrap -n trunk -k /var/cache/mic -r http://repo.meego.com/MeeGo/builds/trunk/latest/repos/oss/ia32/packages/ -o /var/cache/meego-bootstrap21:18
Stskeepsand get rid of  the --mainrepo stuff21:18
Anssi138http://wiki.merproject.org/wiki/Trying_it_out  < should that  be updated?21:19
Stskeepsyes, please do21:19
Stskeepswe've started writing proper guides for this21:19
*** sandroandrade has joined #mer21:20
*** raignarok has quit IRC21:24
*** raignarok has joined #mer21:33
Anssi138doesn't work for me yet.21:40
*** niqt_ has joined #mer21:42
*** shanem has quit IRC21:43
*** niqt has quit IRC21:44
*** gabrbedd_ has joined #mer21:55
*** niqt has joined #mer21:57
*** shanem has joined #mer21:57
Anssi138*fixed21:59
*** niqt_ has quit IRC22:00
*** niqt has quit IRC22:07
*** dijenerate has quit IRC22:09
*** dijenerate has joined #mer22:14
Anssi138Stskeeps: there was wrong bin_path in /usr/lib/pymodules/python2.7/mic/utils/bootstrap.py on line 686.changed /usr/local/bin to /usr/bin22:17
*** dijenerate has quit IRC22:24
*** dijenerate has joined #mer22:25
*** ALoGeNo has joined #mer22:28
*** ALoGeNo has joined #mer22:28
*** dijenerate has quit IRC22:32
Anssi138actually it was line on 69922:35
*** dijenerate has joined #mer22:35
*** shanem has quit IRC22:37
*** shanem has joined #mer22:39
Anssi138Stskeeps: http://wiki.merproject.org/wiki/Talk:Trying_it_out22:46
*** vivijim has quit IRC22:48
*** jonnor_work has joined #mer22:55
*** lynxis has quit IRC22:57
*** Venemo has joined #mer23:00
*** Venemo has quit IRC23:00
*** Venemo has joined #mer23:00
*** lynxis has joined #mer23:05
*** ali1234 has quit IRC23:09
*** ali1234 has joined #mer23:10
*** ali1234 has quit IRC23:16
*** zenvoid has quit IRC23:19
*** ali1234 has joined #mer23:28
*** raignarok_ has joined #mer23:35
*** raignarok has quit IRC23:36
*** raignarok_ is now known as raignarok23:36
*** rantom has quit IRC23:47
*** rantom has joined #mer23:47
*** NIN101 has quit IRC23:52
*** berndhs_950 has joined #mer23:56
*** raignarok has quit IRC23:58

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!