Friday, 2012-04-27

*** himamura has joined #mer00:16
*** himamura has quit IRC00:17
*** himamura has joined #mer00:17
*** mlfoster has quit IRC00:23
*** himamura has quit IRC00:25
*** Siosm has joined #mer00:27
*** Siosm has quit IRC00:29
*** disco_stu_droid has joined #mer00:31
*** disco_stu has quit IRC00:33
*** disco_stu_droid is now known as disco_stu00:34
*** disco_stu is now known as disco_stu_00:36
*** dionet has quit IRC00:50
lilstevievgrade: you around?01:14
*** disco_stu_ is now known as disco_stu01:14
vgradelilstevie, yea01:16
vgradeI should be in bed though, whats up01:16
lilstevieheh ok01:17
lilstevieI was just wondering if through any of your contacts at nvidia if you have heard any rumblings of armhf r15beta accel drivers01:18
lilsteviefor tegra obviously :)01:18
vgradeno, I noted a new release with update ABI but nothing hardfp01:20
lilsteviehm01:20
lilsteviecause I'm sure you have seen but as of ubuntus 12.04 release that is now hf as well01:21
vgradeyes, I'm sure that will spur nVidia on to release hardfp again.01:21
*** Zeroedout has joined #mer01:22
lilstevieheh01:22
lilstevieit has been such a long time since we have seen a hf driver release :(01:22
lilstevieand they aren't compatible with tegra301:22
vgradenot done anything with T3 yet, still waiting for someone to give me one to play with01:23
lilstevieheh01:23
lilsteviewell we recently ported 3.1 from nv-tegra to the prime01:24
*** andre__ has quit IRC01:26
vgradelilstevie, good to see work is still ongoing01:31
lilstevie:)01:41
*** shadeslayer has quit IRC01:55
*** swer has joined #mer01:57
*** slx has quit IRC02:02
*** shadeslayer has joined #mer02:41
*** AndrewX192 has quit IRC02:56
*** AndrewX192 has joined #mer02:57
*** AndrewX192 has quit IRC02:57
*** AndrewX192 has joined #mer02:57
*** KaiRo_Mozilla has joined #mer02:57
*** KaIRC has quit IRC03:01
*** KaiRo_Mozilla has quit IRC03:08
*** lamikr has quit IRC03:22
*** GeorgeH has joined #mer03:26
*** teve has quit IRC03:28
*** teve has joined #mer03:30
*** pirut has joined #mer03:33
*** wala is now known as amateur03:55
*** amateur is now known as burro04:09
Stskeepsmorn04:09
*** clopez has quit IRC04:10
*** kimitake_idle is now known as kimitake04:13
*** zaneswaf has joined #mer04:19
StskeepsKindle Fire makes up over half of Android tablet market share -- ouch04:29
*** kimitake has quit IRC04:30
*** pirut has quit IRC04:39
*** pirut has joined #mer04:40
*** kimitake has joined #mer04:43
*** jukkaeklund has quit IRC04:44
*** marquiz has quit IRC04:45
*** marquiz has joined #mer04:46
*** Jade- has joined #mer04:59
*** Jade has quit IRC05:02
* Stskeeps merges all the things05:21
E-PMorning05:21
Stskeepsmorn05:22
*** burro has quit IRC05:44
*** lamikr has joined #mer05:45
StskeepsE-P: so, i have an idea..05:52
E-PStskeeps: go ahead05:52
StskeepsE-P: not sure if you know, but there's a couple of interesting features in RPM and OBS.. one is that test cases of packages are run in a seperate %check section of a package spec, the other is that it's possible to customize with RPM macros to run things pre/post that from RPM packaging point of view.. the third thing is that if you put results (rpmlint result, for instance) in RPMBUILD/OTHER, this gets extract-able through OBS05:54
StskeepsE-P: what if we could package up the source/binary tree of the %check time and at a later time run it on actual hardware?05:55
Stskeepsthrough those methods05:56
E-PStskeeps: for unittests that might be doable05:58
Stskeepsyeah, in fact, that's the ones i'm mostly interested in05:58
Stskeepsbecause they don't run reliably on cross compilation05:59
E-Pany idea how to transfer the source/binary tree to somewhere or can we create a own package out of that?05:59
Stskeepsi think we could possibly do a combination of 'build' on device and extracting it through OBS api06:00
Stskeepsbuild just sets up a chroot anyway06:01
E-Pyep, we should try p-o-c about that06:02
*** Alison_Chaiken has joined #mer06:03
vgrademorning06:05
E-PStskeeps: do you have any example where the %check is used?06:07
Stskeepssec06:07
Stskeepshttp://gitweb.merproject.org/gitweb?p=mer-core/rpm.git;a=blob;f=rpm.spec;h=5561194de4ced61b5af9811df8553dc3946c9623;hb=HEAD#l18306:08
*** Alison_Chaiken has quit IRC06:08
*** pohly has joined #mer06:08
E-Pif the check fails, the built is marked fail as well?06:10
Stskeepsyes06:10
Stskeepsthis would make it async instead06:10
Stskeepsso we'd collect observations on a release06:10
Stskeeps== faster builds, too06:10
*** andre900 has joined #mer06:13
*** jukkaeklund has joined #mer06:14
Stskeepsmorn jukkaeklund06:14
*** Aloz1 has joined #mer06:14
Stskeepshello Aloz1  :)06:15
Aloz1Hello :)06:16
dm8tbrmoaning06:16
Stskeepswelcome :) So what brings you here to #mer ?06:16
E-PStskeeps: is your idea that the results from the device execution would effect somehow to the build or would it be a separated process?06:16
Aloz1I'm interested in the project06:16
E-PStskeeps: meaning that the %check doesn't effect to OBS build result if the tests are executed in the device06:17
StskeepsE-P: correct - moving %check to post-build phase, ie, when we validate a release06:17
E-PStskeeps: ok06:17
StskeepsAloz1: cool - if you have any questions, feel free to ask, or else feel free to hang out and learn :)06:17
StskeepsE-P: so for example we can do rpm functionality tests in the end06:18
*** JLP has joined #mer06:21
*** JLP has joined #mer06:21
*** niqt has joined #mer06:25
*** jukkaeklund_ has joined #mer06:26
*** jukkaeklund has quit IRC06:29
*** Alison_Chaiken has joined #mer06:30
*** jukkaeklund_ has quit IRC06:30
*** jukkaeklund_ has joined #mer06:30
*** slaine has joined #mer06:31
Stskeepsmorn slaine o/06:39
slaineMorning Stskeeps06:39
*** lamikr has quit IRC06:43
Aloz1Morning? It's 4:40 pm :P06:43
StskeepsAloz1: UGT ;)06:43
StskeepsAloz1: http://www.total-knowledge.com/~ilya/mips/ugt.html06:44
*** arc_mat|tp has joined #mer06:44
Stskeepsarc_mat|tp: happen to know how the GPS is wired up on G9/06:46
Stskeeps_06:46
Stskeeps?06:46
arc_mat|tpStskeeps: UART106:46
Stskeepsok, and it speaks NMEA?06:46
arc_mat|tpnot at all06:46
arc_mat|tpit speaks AI206:46
Stskeepsah.. know of anything in open source that can interpret it?06:47
arc_mat|tpStskeeps: no, there is nothing, I'm afraid. AFAICT, there isn't even a public spec06:47
Stskeepsok06:47
Stskeepsi guess that's what wilink vendor package provides then06:48
arc_mat|tpyes, thoug on the G9 the GPS is not part of wilink, it's a dedicated NL5550 from TI06:48
Stskeepsok06:49
StskeepsAlison_Chaiken: ^06:50
arc_mat|tpwith some snooping and reverse engineering it could be made to work. not an easy task without docs, and you need to know quite a bit about GPS06:51
Alison_ChaikenThanks, Stskeeps, my plan was to plugin a USB GPS (which I already have) for testing, but maybe I won't be putting it away, after all.06:51
Stskeepsyeah, USB GPS seems like a sane option06:52
Stskeepsor bluetooth06:52
arc_mat|tpindeed06:52
Alison_ChaikenYeah, BT is an idea, but I already have a USB one . . . and Scantool is Bluetooth already.06:52
Alison_ChaikenI'm not sure how well two BT devices close together will work?06:53
dm8tbrespecially as the n900 gpsd can not be reused as it talks to CMT not APE06:53
Alison_Chaikenvgrade naturally already built gpsd package for ARMv7l for Mer.06:53
Alison_ChaikenI started testing it yesterday.06:54
Alison_ChaikenRan without error with USB GPS, but didn't provide a fix.06:54
Alison_ChaikenSince usual independent way of testing GPS (gpsd-clients) is strongly tied to GTK+, I'm not sure yet which bit is broken.06:55
Stskeepscan't you connect up to the NMEA stream?06:55
Alison_ChaikenI'm optimistic that approach will work eventually.06:55
Alison_ChaikenLack of GPU support looms larger.06:55
Stskeepsbtw, the location infra on qt side is a bit quirky06:55
Stskeepslack of GPU support? i thought pvr worked on PA images06:55
Alison_ChaikenI'm quirky, so I sympathize.06:56
*** lamikr has joined #mer06:56
Alison_ChaikenNot clear.   Eglinfo runs on mine, but UI is slow and jerky and stuttery.06:56
Alison_ChaikenMy investigation of the situation is still in an early phase.06:57
Stskeepsok, what renderer does it say ?06:57
Alison_ChaikenThanks for detective work, though: good to know where I stand!06:57
Alison_ChaikenYou mean, what is eglinfo?06:57
Alison_ChaikenLet's see if I copied that output to $LOCALHOST.06:58
Alison_ChaikenNope.06:58
Stskeepseither way, just make sure that doesn't say meas07:01
Stskeepsmesa07:01
*** harbaum has joined #mer07:02
*** bergie has joined #mer07:04
Alison_ChaikenI forget, which pastebin are we using here?07:15
Stskeepswhatever works07:15
Stskeepspastie.org was used but it got taken down07:15
Alison_ChaikenI will post damning eglinfo about llvmpipe and mesa.07:15
Alison_ChaikenSensitive people do not look!07:15
*** ZiQiangHuan has joined #mer07:17
Alison_Chaikenhttp://pastebin.com/dBVWjSv607:18
Stskeepsok, so it doesn't use g9 acceleration07:18
Stskeepsthat's interesting07:18
Alison_ChaikenWell, eglinfo says it does, but es1_info says no.07:18
Alison_ChaikenI should compile es2_info and try that too.07:18
Stskeepssounds like bad symlinking07:19
Alison_ChaikenMaybe es2_info shows HW is working.07:19
Stskeepsls -l /usr/lib/libEGL*07:19
Alison_ChaikenI thought DRI2 was x86?07:19
Alison_ChaikenSo that wrong driver is invoked for ARM?07:19
Stskeepswell it seems like some entry points are odd07:20
vgradeI'm not sure G9 has pvr acceleration yet07:21
*** InformatiQ has joined #mer07:23
Stskeepsmorn InformatiQ07:23
*** enc0de has joined #mer07:25
*** mdfe_ has joined #mer07:28
arc_mat|tpno it doesn't07:28
Stskeepswhat's blocking, if you don't mind me asking?07:28
arc_mat|tpPVR/omapdrm07:29
Stskeepsok07:29
dm8tbrSage__ has the same problem with his Samsung phone IIRC07:30
arc_mat|tpretrofitting omapdrm and the pvr xorg stuff into our kernel proves to be difficult, again07:30
Sage__:nod:07:31
Sage__I guess it might be better if I could disable the omapfb but they have patch that made dss depending omapfb07:32
arc_mat|tpoh07:32
arc_mat|tpbad07:32
Sage__arc_mat|tp: https://github.com/supercurio/android-omap-tuna/commit/b442e52cba6dcffdf5880ca1c2f2625bfc5ee7f207:32
Sage__my kernel has that same patch07:33
Sage__+ bunch of patches on top of that of course07:33
Alison_ChaikenIt's worth pointing out that just because es2gears and es1_info are looking for DRI2, that doesn't mean KDE is.07:33
Alison_ChaikenI will check that tomorrow.07:33
Stskeepsexperiences of people porting Tizen (handset) to snowball, https://www.tizen.org/irclogs/%23tizen.2012-04-27.log.html#t2012-04-27T06:53:4607:36
*** the-gibson has quit IRC07:38
*** the-gibson has joined #mer07:39
*** the-gibson has joined #mer07:39
*** furikku has joined #mer07:40
*** vilpan has joined #mer07:44
*** eocanha has joined #mer07:44
StskeepsSage__: what exactly is wrong with that patch?07:47
arc_mat|tpStskeeps: it creates a hard dependency between hdmi and omapfb07:48
arc_mat|tpStskeeps: but omapfb is likely to interfere with omapdrm07:48
Stskeepsok, but if you wanted just LCD..07:49
arc_mat|tpStskeeps: maybe it works if you disable HDMI support in kernel, but not sure07:50
Sage__hmmp...07:50
*** DocScrutinizer has quit IRC07:56
*** DocScrutinizer has joined #mer07:57
*** jonnor has joined #mer07:57
dm8tbrStskeeps: ouch, worse than I thought...07:58
Stskeepsdm8tbr: i'm intentionally marking 'handset', as IVI is much much saner07:59
dm8tbr*nod*07:59
*** himamura has joined #mer08:00
Sage__CONFIG_OMAP_HSI <- what that is used for?08:00
Stskeepsmodem, maybe08:05
Sage__o_0 it compiled without fbdev if I dropped hdmi08:05
Stskeepsgood?08:05
Stskeeps:P08:05
Stskeepsor not08:05
Sage__err... omapfb that is08:05
Sage__which is how it is supposed to go08:05
* Stskeeps heads out08:06
Aloz1Goof night08:07
Aloz1Good*08:07
*** Aloz1 has quit IRC08:07
Sage__*headdesk* battery run out of the nexus08:09
arc_mat|tpSage__: well, well ;)08:15
lbtmorning all08:16
arc_mat|tpSage__: HSI is probably modem, yes08:17
arc_mat|tpSage__: but it's a generic interface, maybe there is another USB port behind it or something08:17
*** araujo has quit IRC08:20
*** smoku has joined #mer08:20
*** araujo has joined #mer08:20
*** arcean has joined #mer08:21
*** byako has joined #mer08:21
*** bergie has quit IRC08:29
*** bergie has joined #mer08:30
*** kavurt has quit IRC08:41
Termanamorning08:42
*** kavurt has joined #mer08:43
*** jukkaeklund_ has quit IRC08:47
*** kavurt has quit IRC08:49
* Stskeeps ponderx if we could use testrunner for production testing, ie, at factories08:54
lbttesting what?08:55
Stskeepswell, the usual test lcd, hardware, components, etc08:56
Stskeepsto state if a given just off solder board/casing is ok08:57
lbtit would be fine for random samples08:57
lbtI don't know if every device is tested then reflashed with PR08:58
lbtprobably a "low" prio just yet though :D08:58
Stskeepsthink that's the usual way08:58
Stskeepssure, but has to kept in mind too08:58
Stskeepssee raspberry pi production line tests08:59
Stskeepsthey used a self made test image08:59
byakoDo we discuss drivers source code here or should I visit arm irc channel? :)09:05
byakoIt's about mali400 driver09:05
Stskeepsmali400 is a traditional arm gfx driver.. open kernel module, closed userland09:15
*** tommis has joined #mer09:15
Stskeepswe discuss hw adaptation at least09:17
arc_mat|tpStskeeps: what is testrunner?09:29
vgradebyako, whas the mali question09:31
Stskeepsarc_mat|tp: http://releases.merproject.org/~carsten/hostbased.png is ui for it.. sshs into device and runs automated,semiautomatic or manual tests09:33
*** clopez has joined #mer09:33
arc_mat|tpStskeeps: a bit too much for production line testing09:33
arc_mat|tpbut of course a stripped down test count could help09:34
arc_mat|tpStskeeps: how does it control the DUT?09:34
Stskeepsarc_mat|tp: well, test on ui is just an example -- ssh through our test enablers09:35
arc_mat|tpwell for certain prod tests you need interaction with the tester, or how to check if the LCD is connected correctly and the buttons are working etc...09:36
Stskeepssure09:36
Stskeepsthere are manual test types09:36
Stskeepsor semiautomatic09:37
Sage__arc_mat|tp, Stskeeps: After disabling HDMI stuff the screen goes blank quite soon and device doesn't boot far. http://pastebin.com/xENfU1ba09:37
arc_mat|tpwe have a simple menu based production test software that runs on the device itself09:37
piruthas anyone experience with btrfs, I would like to bind mount a directory that is in a snapshot without first having to mount the snapshot ?09:38
Stskeepsnod, my point was there is a lot of different homemade frameworks09:38
Stskeepsfor this purpose09:38
arc_mat|tpSage__: that is all? hard hang after disabling VCXIO?09:39
Sage__arc_mat|tp: yes09:40
Sage__after that serial goes silent so I suspect hard hang there09:40
arc_mat|tpweird09:40
byakovgrade: mali_fbdev.c:mali_drm_open_master sends xf86DrvMsg about DRM OPEN with X_ERROR flag at the time when message is informative, shouldn't it be X_INFO ?09:40
arc_mat|tpand with hdmi enabled, how does a boot look like?09:41
Sage__arc_mat|tp: and with hdmi enabled it doesn't compile http://pastebin.com/Wrf893Tq if I don't enable the omapfb as well09:42
arc_mat|tpSage__: yes, of course09:42
vgradebyako, thats one for the mali dev mailing list I think.09:43
arc_mat|tpSage__: if you suspect the disabling of VCXIO to be a problem, there is a call to "regulator_constraints_complete()" or something in the board start code that you could comment for testing09:43
vgradebyako what device are you working on? If you don't mind me asking09:43
byakovgrade: I'm not working on any device :) I'm still trying to bring up MeeGo HS UI09:45
byakovgrade: using snowball A1209:45
Sage__arc_mat|tp: ok, will try that. Have to go now though but I'll report back later.09:46
Sage__arc_mat|tp: meant regulator_has_full_constraints();09:47
Sage__?09:47
arc_mat|tpyes, that one09:47
arc_mat|tpmaybe VCXIO is needed, but not requested when the hdmi driver doesn't initialize09:48
*** arfoll_ is now known as arfoll09:49
Stskeepsbyako: found out what caused hdmi trouble?09:50
Sage__arc_mat|tp: yes, after that it boots further. Got serial.09:51
*** teve has quit IRC09:52
byakoStskeeps: in progress, I've found huge log file with lots of same errors, but some of them quite critical as for me: mthemedaemon: can't send pixmap sharing message; and : MRmi::invoke(): "QLocalSocket::connectToServer: Connection refused" "/home/nemo/.mabstractdecorator"; I guess some QML daemon didn't get up and apps just can't connect to it09:52
byakoStskeeps: you were right about log in home folder it was just named differently09:53
Stskeepsbyako: ah.. try get mer-gfx-tests running first, not entire ui09:53
byakoStskeeps: I will, thanks09:53
* Stskeeps really needs to write some 10 commandments of mer hw adaptation09:54
Stskeeps1. Thou Shalt Start From A Core Bootup to Console09:54
*** notmart has joined #mer09:55
*** notmart has joined #mer09:55
* Sage__ has to go now09:55
Stskeepscya09:55
*** cxl000 has joined #mer10:04
*** araujo has joined #mer10:18
*** araujo has joined #mer10:18
*** teve has joined #mer10:20
*** lardman has joined #mer10:21
*** tommis has quit IRC10:25
*** Attie has quit IRC10:25
*** Attie has joined #mer10:26
*** tommis has joined #mer10:28
*** lizardo has joined #mer10:35
*** decibyte has quit IRC10:43
*** steff_ is now known as steff10:54
*** pdz- has joined #mer10:59
*** teve has quit IRC10:59
*** pdz has quit IRC10:59
*** steff has quit IRC11:03
*** dionet has joined #mer11:03
*** trbs has joined #mer11:04
*** himamura_ has joined #mer11:07
*** teve has joined #mer11:08
*** himamura_ has quit IRC11:09
*** himamura has quit IRC11:11
*** jluisn has joined #mer11:11
*** Behold has quit IRC11:12
*** BeholdMyGlory has joined #mer11:13
*** dionet has quit IRC11:17
*** BeholdMyGlory has quit IRC11:17
*** BeholdMyGlory has joined #mer11:21
*** InformatiQ has quit IRC11:21
phaeronso ruby-1.8 timeout is not reliable11:21
phaerondoesn't timeout if the enclosed code crosses to C land11:22
phaeronlbt: ping11:22
lbt:/11:22
lbtwhilst making coffee this morning I considered dropping packaging for our ruby stuff11:23
phaeronagain ?11:23
lbtWe have the parable of King Canute11:24
phaeronlbt: there's a gem that implements reliable timeout for ruby 1.811:24
phaeronbut this is ridiculous :D11:24
lbtyep11:24
lbtthe nice thing about ruby is that there are so many stable versions to pick from11:25
lbtall with their own little quirks and foibles11:25
phaerononly 1.8 in suse11:25
*** sirdancealot has quit IRC11:26
lbtI'm just watching my life fade away whilst trying to make obs schedule some jobs with X-Fade11:28
phaeronwhat life11:29
lbtindeed11:29
X-FadeYeah, easy to debug :)11:30
*** steff has joined #mer11:37
*** KaIRC has joined #mer11:39
phaeronlbt: there are two requests waiting on OBS , for tools , who's supposed to check them11:42
phaeronlbt: system timer gem works ..12:10
*** ieatlint_ has joined #mer12:15
*** Zeroedout_ has joined #mer12:16
*** DocScrutinizer has quit IRC12:16
*** DocScrutinizer has joined #mer12:16
*** Behold has joined #mer12:17
*** Stskeepz has joined #mer12:18
*** ighea_ has joined #mer12:18
*** rrix_ has joined #mer12:19
*** BeholdMyGlory has quit IRC12:23
*** Zeroedout has quit IRC12:23
*** Stskeeps has quit IRC12:23
*** ighea has quit IRC12:23
*** alh has quit IRC12:23
*** rrix has quit IRC12:23
*** ieatlint has quit IRC12:23
*** ZiQiangHuan has quit IRC12:23
*** ieatlint_ is now known as ieatlint12:24
*** FACEFOX has quit IRC12:27
*** drussell has quit IRC12:29
*** zeq has quit IRC12:36
*** drussell has joined #mer12:37
*** FACEFOX has joined #mer12:44
*** zeq has joined #mer12:53
phaeronlbt: what do you think of this so far https://github.com/MeeGoIntegration/open-build-service/commit/1e1699a14cf1e8c0a6b587d4a190901d07f620c213:07
*** harbaum has quit IRC13:14
*** himamura has joined #mer13:14
*** JLP has quit IRC13:19
*** toscalix has joined #mer13:22
*** gabrbedd has joined #mer13:24
phaeronlbt: force pushed https://github.com/MeeGoIntegration/open-build-service/commit/3b3a3e4be77478e5f4679f0213e14b6feedac5c213:38
*** vivijim has joined #mer13:53
*** vivijim has joined #mer13:53
*** djszapi has joined #mer13:53
*** djszapi has left #mer13:53
* Stskeepz yawns13:54
* w00t kicks Stskeepz to rename13:54
*** Stskeepz has quit IRC13:54
*** Stskeeps has joined #mer13:55
*** byako has left #mer13:55
*** jukkaeklund_ has joined #mer13:56
alteregolol13:58
*** bfederau has quit IRC14:01
*** Jade has joined #mer14:05
*** Jade- has quit IRC14:08
*** niqt has quit IRC14:20
*** alh has joined #mer14:20
*** toscalix has quit IRC14:29
*** rrix_ has quit IRC14:29
*** rrix_ has joined #mer14:29
*** rrix_ is now known as rrix14:29
Stskeeps26 C out, i conclude this is a beer day14:42
slaine7C, wet and windy, I conclude this is a beer day14:43
fralsno idea about the weather, i conclude this is a beer and pizza day14:44
frals\o/14:44
* timoph declares this a lay in the sofa and do nothing that requires too much thinking day (having a bones season 2 marathon)14:47
* jukkaeklund_ <3 bones14:47
w00tStskeeps: whatthehell14:50
w00twho are you and what did you do with all the nice weather14:50
* w00t is informed that it's 10c out14:50
fralsjust went and checked, 18 here14:51
* frals is QUITE pleased with being back in Stockholm14:51
Stskeepsgood day for living in warsaw14:52
w00tyeah, but on the downside, sweden14:52
w00t:P14:52
Paimen4 days little vacation <314:53
Paimenwell I need to do few hours work on monday, but it does not count :P14:53
*** raignarok has joined #mer15:08
*** slaine has quit IRC15:22
*** smoku has left #mer15:32
*** mdfe_ has quit IRC15:34
*** jukkaeklund_ has quit IRC15:40
matrixxyeah, beer day today \o/15:46
*** raignarok has quit IRC15:50
*** gimli has joined #mer15:51
*** gimli has joined #mer15:51
Stskeepslbt: http://www.hetzner.de/en/hosting/produkte_rootserver/ex1015:51
lbtyummy15:54
lbt2x ex4 may be better value at the moment15:56
lbtphaeron: pong ... I missed your msgs15:56
lbtphaeron: looks sane15:58
lbt5seconds is quite generous for an LDAP timeout isn't it?15:59
phaeronlbt: yes15:59
phaeronbut it is better than 60 seconds default timeout15:59
lbt*g*15:59
lbtI wonder if ruby has a shuffle method15:59
phaeronso I can uncomment the remove line ?16:00
lbtI hate doing server = ldap_servers[rand(ldap_servers.length)]16:00
phaeronyeah. well I also wanted to remove every server as it is tried , but that makes the "retries" setting have a different meaning or be not needed at all16:01
lbtagreed16:01
phaeronlbt: http://www.ruby-doc.org/core-1.8.7/Array.html#method-i-shuffle16:01
phaeronit does have shuffle16:01
lbtit gets awfully fiddly16:01
lbtthat may be better than the rand crap16:01
phaeronok , not also that the way the error message in the rescue was generated was not working at all16:02
phaeronyou can't extract the error after unbinding16:02
phaeronbecause the conn object is reset16:02
phaeronor so I found16:03
lbttesting failure modes is hard16:03
phaeronI copied the code to a standalone script and tried various failures16:03
lbtI think some of this was "well, if it breaks lets try and do something sane"16:03
phaeronnot particularly proud of the if RUBY_VERSION.split(".")[1] == "8" .. but couldn't find another way to do it16:04
lbtupstream are ruby experts16:04
lbtthey can comment in review16:05
phaeronwhich upstream , SR or Adrian ?16:05
lbtadrian16:05
phaeronok16:06
*** FACEFOX has quit IRC16:08
phaeronusing shuffle instead of rand will still change the meaning of max attempts16:08
lbtserver = ldap_servers[count % ldap_servers.length ]16:12
lbtassuming % is mod16:12
*** NIN101 has joined #mer16:12
*** FACEFOX has joined #mer16:13
phaeronmmm . wut ?16:14
lbtthat will retry max_attempts balanced evenly over all servers16:15
lbtalthough ... you know... this is really not that worthwhile16:15
*** NIN102 has joined #mer16:16
lbtif we do care then the sane thing to do is to have a list of 'active' LDAP servers and if one times out, restore it after X mins16:17
phaeronok , I wonder how to get upstream to review ? pull request from MerProject ?16:18
*** gimli has quit IRC16:18
lbtyes16:18
phaeronlbt: that would need some kind of persistent marker. in the db.16:18
*** NIN101 has quit IRC16:19
lbtnah, in memory would be fine16:19
*** afiestas has quit IRC16:19
phaeronnot sure if passenger threads are persistent across requests16:19
*** gimli has joined #mer16:19
*** gimli has joined #mer16:19
lbt!16:19
lbtthey must be16:19
lbtsurely16:19
phaeronwhy16:20
lbtDB connection pooling typically16:20
*** afiestas has joined #mer16:20
*** FACEFOX has quit IRC16:21
*** sebas has quit IRC16:23
*** enc0de has quit IRC16:24
*** jonnor has quit IRC16:24
*** FACEFOX has joined #mer16:25
*** pohly has quit IRC16:25
phaeronnot sure16:26
phaeronI believe that's memcache is used for16:27
phaeronin memory persistent objects that have a lifetime16:27
phaeronhttp://stackoverflow.com/questions/2791093/what-data-if-any-persists-across-web-requests-in-ruby-on-rails16:28
lbt"a class variable on a controller and this will be carried from one request to the next as you might expect"16:28
lbtso yeah - the point is that dweebs would naivley expect that a variable is set from one req to the next16:29
phaeron"The catch is that this only applies to the singular instance of that controller"16:30
lbtyes, that's right16:30
*** zeq has quit IRC16:30
lbtwe don't really care16:30
lbtit's an ldap connection pool16:30
phaeronyeah but the correct way is to use memcache16:30
lbtno, it's not16:30
lbtmemcache is for storing persistent session data or truly persistent data16:31
lbtor typically to allow us to be lazy when doing queries :)16:31
lbtin this instance we're implementing connection pooling (which should set off warning bells)16:31
phaeronit's not truly persistent , it has a lifetime (can be very long) and will disappear if memcache is restarted16:32
lbtyes. and memcache is a cache, not a store16:32
lbtso step back ... what are we actually doing?16:33
phaeronwasting time :D16:33
lbtA: storing hints about which LDAP servers are down16:33
lbtif we lose the hint list then ... we assume they're all up16:33
lbtand it costs us a try to recreate the hint list16:33
*** FACEFOX has quit IRC16:34
lbtnot worth memcaching in order to store across all threads16:34
lbtjust make it a class var and we're fine16:34
phaeronI also wanted to put the require block in the class but wasn't sure if it will work16:35
*** FACEFOX has joined #mer16:36
lbtyeah ...  that should be at a higher level - but it's only important if LDAP is :on16:37
*** jukkaeklund_ has joined #mer16:42
*** eocanha has quit IRC16:44
*** lamikr has quit IRC16:49
phaeronwhat if there's only one server and it gets marked as down :)16:50
*** andre900 has quit IRC16:52
phaeronand using modulus would mean that in the good case the same server will get hammered while others will be idle16:55
phaeronanyway I just need to test this before submitting upstream16:56
*** Attie has quit IRC16:59
*** Siosm has joined #mer17:02
lbtno you use shuffle! before each loop17:03
*** sledges has quit IRC17:03
lbtand if there's no servers then you restore them all to the "up" list and try them in order17:04
*** Siosm has quit IRC17:04
*** CosmoHill has joined #mer17:09
*** lamikr has joined #mer17:10
*** zeq1 has joined #mer17:12
*** sirdancealot2 has joined #mer17:25
*** vilpan has quit IRC17:25
*** afiestas has quit IRC17:35
*** afiestas has joined #mer17:37
*** gabrbedd has quit IRC17:41
E-Phmm, good evening to update ubuntu17:50
CosmoHillare the moons in alignment?17:51
* lbt finishes swapping card amongst various machines and sends in a bug report ... waste of half a day :/17:51
CosmoHilllbt: aww17:51
lbthey, at least I  know the cards work17:52
E-Pit is clear sky, so I should have a good change17:52
CosmoHillI got my first paycheck today :)17:52
IanWizard-CloudCosmoHill: I got my last yesterday :P17:52
IanWizard-Cloudhow ironic.17:52
lbtCosmoHill: first pension payment tomorrow ?17:52
CosmoHilllbt: on a saturday?17:52
lbtcan't start too soon17:52
lbt:D17:52
CosmoHillI have no idea about pension17:53
*** harbaum has joined #mer17:53
* lbt -> Tango .... l8r all o/17:53
*** mlfoster has joined #mer17:53
E-Pcya17:53
CosmoHillcyas17:53
Stskeepsi've started to put some diagrams on the architecture description page, feedback welcome: http://wiki.merproject.org/wiki/Architecture17:59
*** tommis has quit IRC18:01
*** dionet has joined #mer18:03
*** jluisn has quit IRC18:06
E-PStskeeps: looks nice18:06
*** jluisn has joined #mer18:07
*** jluisn has joined #mer18:09
* Stskeeps discovers AgileBrowser can do batch mode18:15
CosmoHillStskeeps: go on18:16
CosmoHillwait, that's not a web browser is it?18:16
Stskeepsie, automated architecture documentation18:16
Stskeepsand no, it's not18:16
*** jluisn has quit IRC18:17
*** jluisn has joined #mer18:21
*** jluisn has quit IRC18:21
*** jluisn has joined #mer18:22
*** jluisn has joined #mer18:27
*** arcean_ has joined #mer18:35
*** arcean has quit IRC18:38
*** gabrbedd has joined #mer18:38
*** Cosmo[PB] has joined #mer18:40
*** CosmoHill has quit IRC18:40
*** Cosmo[PB] is now known as Cosmohill18:40
*** Cosmohill is now known as CosmoHill18:40
*** CosmoHill has quit IRC18:44
*** CosmoHill has joined #mer18:47
Stskeepsvgrade: http://wiki.merproject.org/wiki/Architecture18:49
Stskeepsis this helpful to understand what's going on?18:50
vgradeIndeed it is , oh its changed18:50
Stskeepsyeah, uploaded a bunch of graphs18:51
vgradecan you drill down into the groups18:51
Stskeepsyeah18:52
Stskeepsthat'll come later, but with external links18:52
CosmoHillmy laptop visited 1970 and now some apps are jibbing out18:52
Stskeepsmoment18:52
Stskeepsvgrade: http://releases.merproject.org/~carsten/startupservices2.png18:53
vgradethats the sort of thing18:54
Stskeepsi'm intentionally making it high level in wiki page, and then giving people guide how to use the tool18:54
vgradethese are much more infromative than the more trad layer diagrams which don't show the interactions18:54
Stskeepsvgrade: warning, massive png: http://releases.merproject.org/~carsten/everything.png18:59
Stskeepsthese are times i wish i had a plotter at home19:00
*** raignarok has joined #mer19:07
*** Behold has quit IRC19:07
ali1234that's funny cos i was looking at the first one thinking "that's massive"19:07
*** BeholdMyGlory has joined #mer19:07
Stskeepspackage dependencies are loads of fun19:08
ali1234Stskeeps: that stuff about "integrated linux core" yesterday sounded a lot like mer to me... any thoughts on that?19:08
*** arcean_ has quit IRC19:08
Stskeepsit does indeed show a direction that we're moving in, yes19:08
Stskeepslike, the linux distro mindset19:09
Stskeepsie, it's no longer about massive repositories of packages19:09
Stskeepsit's about shit that actually works19:09
Stskeeps:P19:09
Stskeepsali1234: did you see my FOSDEM presentation?19:10
*** zeq1 has quit IRC19:10
ali1234no... is there a link?19:10
Stskeepssec19:10
Stskeepshttps://archive.fosdem.org/2012/schedule/event/391/79_beyond.pdf19:10
ali1234i think that teletype picture is a bit unfair :)19:11
Stskeepsit was meant to be :)19:12
Stskeepsin my notes i had it marked down as "unneccesarily provocative" :)19:12
ali1234most of the arguments yesterday were about how the new stuff specifically doesn't work in that configuration19:12
Stskeepswhich specific discussion are you talking about, btw?19:12
Stskeepsthere's been some integrated core discussions around the place19:12
ali1234well i only saw the comments posted on gregkh' google+19:13
ali1234but i figure it was mostly the same stuff getting repeated19:13
Stskeepsah, yeah19:13
ali1234the stuff about integrating udev and systemd more tightly with the kernel to create a standardized base19:13
*** smoku has joined #mer19:14
Stskeepsthey just have to be careful.. world doesn't move as fast as kernel19:14
vgradeStskeeps, work of art19:14
Stskeepsand most importantly, not everything is x8619:15
Stskeeps:P19:15
vgradeinteresting to see the mass of interactions in base essentials19:15
Stskeepsyeah.. base essentials is a mess19:15
Stskeepssome patterns will eventually emerge from there19:15
vgradeis that why its difficult to bootstrap things19:16
Stskeepspart of it, but that's build-time architecture19:16
Stskeepsi have some ideas towards that but it's 2-3 months ahead19:17
smokuStskeeps: what happened to Mer:fake:Core:i486 on COBS?19:17
*** jluisn has quit IRC19:17
Stskeepssmoku: shouldn't have happeend anything19:17
Stskeepssmoku: what problems do you have?19:17
smokuhttps://build.pub.meego.com/project/monitor?project=Project%3ACordia%3AHildon19:17
smokusuddenly everything for i486 is unresolvable :/19:18
Stskeepssmoku: gah.. earlier it was reverse!19:18
Stskeepssmoku: add Prefer: nss    in Advanced -> Project config19:18
smokufixed.  thanks19:19
Stskeepssmoku: http://wiki.merproject.org/wiki/Architecture19:20
smokuStskeeps: anything in particular?19:21
Stskeepsno, just showing :)19:22
smokuoh.  looks good :)19:22
Stskeepsthere'll be a lot of Group: changes to base packages (not practical implication) and hopefully we'll be able to encourage people to do sane architecture for their UIs and hardware adaptations on top as well19:23
Stskeepswith their own architectual groups, etc19:23
Stskeepssmoku: and massive png: http://releases.merproject.org/~carsten/everything.png19:23
*** phdeswer has joined #mer19:24
*** kthomas_vh_ has quit IRC19:25
*** furikku has left #mer19:31
smokuwhoa!19:31
smokua bit complicated19:33
*** zeq has joined #mer19:33
*** mlfoster has quit IRC19:34
Stskeepsyeah, dependencies are always a mess19:35
Stskeepsmany of those are simply 'libc.so.6' though19:35
*** mlfoster has joined #mer19:39
*** phdeswer has quit IRC19:48
*** pharon has joined #mer19:51
*** harbaum has quit IRC20:00
*** tilgovi has joined #mer20:03
*** Behold has joined #mer20:06
*** BeholdMyGlory has quit IRC20:07
*** waa has joined #mer20:14
*** bergie has quit IRC20:16
*** smoku has quit IRC20:24
*** Jade has quit IRC20:28
*** dijenerate has quit IRC20:28
*** Jade- has joined #mer20:28
*** raignarok has quit IRC20:33
*** raignarok has joined #mer20:36
*** dionet has quit IRC20:40
*** notmart has quit IRC20:42
*** Behold has quit IRC20:43
*** BeholdMyGlory has joined #mer20:49
*** kthomas_vh_ has joined #mer20:51
*** HazardousWaster has quit IRC20:58
*** HazardousWaster has joined #mer21:01
*** lizardo has quit IRC21:02
*** zeq has quit IRC21:07
*** raignarok has quit IRC21:24
*** raignarok has joined #mer21:28
*** FACEFOX has quit IRC21:29
*** FACEFOX has joined #mer21:29
*** FACEFOX has joined #mer21:30
*** BeholdMyGlory has quit IRC21:31
*** BeholdMyGlory has joined #mer21:31
*** FACEFOX has joined #mer21:34
*** kavurt has joined #mer21:38
*** FACEFOX has quit IRC21:40
*** FACEFOX has joined #mer21:41
*** niqt has joined #mer21:42
*** BeholdMyGlory has quit IRC21:46
*** BeholdMyGlory has joined #mer21:47
*** phdeswer has joined #mer21:49
*** kthomas_vh_ has quit IRC21:49
*** NIN102 has quit IRC21:51
*** kthomas_vh_ has joined #mer22:00
*** gimli has quit IRC22:01
* CosmoHill returns22:08
*** kavurt has quit IRC22:11
*** pharon has left #mer22:16
*** gabrbedd has quit IRC22:24
*** Behold has joined #mer22:31
*** BeholdMyGlory has quit IRC22:32
*** mlfoster has quit IRC22:34
*** bef0rd has joined #mer22:35
*** mlfoster has joined #mer22:35
*** raignarok has quit IRC22:36
*** clopez has quit IRC22:55
*** CosmoHill has quit IRC23:05
*** phdeswer has quit IRC23:07
*** bef0rd has quit IRC23:12
*** afiestas has quit IRC23:22
*** afiestas has joined #mer23:23
*** niqt has quit IRC23:25
IanWizard-CloudOk, I'm back.23:40
IanWizard-CloudBeen away, busy :(23:41
IanWizard-CloudI have a question, and I'm just not sure..  I've always worked with Android, and built things, like kernels, using CodeSourcery's  cross compiler / toolchain.  The time has come that I need to get off my ass, and setup my own toolchain.  My friend is telling me to use the Android NDK, and bionic, but I'm not certain that I trust him.  Suggestions / guidance would be greatly appreciated.23:44
vgradeIanWizard-Cloud, what are you planning to build? Mer used eglibc so not compatible with bionic23:48
vgradeuses23:48
vgradeMer SDK allows you to setup cross toolchains in a scratchbox2 setup23:49
vgradefor i486/i586/arm(hard/soft float)/mips23:49
IanWizard-Cloudvgrade: atm I'm building a small kernel / system, separate from both android, and mer, then I'll start working on porting mer.23:49
IanWizard-Cloudvgrade: I've been told (though I admitedly haven't ever tried), that I can use both bionic, and libc (or elibc) on the same system.23:50
IanWizard-CloudThis is one of those areas that I know enough, to know that I don't know enough :P23:51
vgradeIanWizard-Cloud I don't think thats a good design23:52
IanWizard-CloudI tend to agree23:52
vgradeIf you are not targetting andriod I'd stay clear of Bionic23:53
IanWizard-Cloudvgrade: supposedly bionic has optomizations for the qualcomm chipset.23:54
IanWizard-Cloudhere's my big question.23:54
IanWizard-CloudIf say.. I build mer, with elibc, all of the repos that I use will need to be built with elibc.23:55
IanWizard-CloudCorrect?23:55
vgradeyes23:56
IanWizard-CloudSo then wouldn't we want to use libc for everything, and be able to use all of the (using debian as an example), debian repos for arm23:56
IanWizard-Cloudyes, that's my real confusion I guess. isn't it smarter to use libc for everything?23:58
vgradeyou can build what you need from debian23:58
vgradewe are targetting smaller embedded mobile devices23:58

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!