Saturday, 2011-10-22

Jucatomorning :)00:20
vgradehey Jucato, morning00:20
*** akira_tsukamoto has joined #mer00:21
Jucatohow's the linuxcon presesntation preparations going? :)00:21
vgradeI've a plasma build done for the N900 but am on a slow link so can't test it00:21
Jucatooh nice00:21
* Jucato drools00:21
vgradeprep going well I think00:22
vgradenot very used to doing presentations00:22
Jucatoheh I know the feeling. my first (and so far only) time was a mess :D00:22
Jucatoand it wasn't even for a big/popular event00:23
*** M4rtinK has quit IRC00:24
vgradeI've done a couple over the years but more comfortable behind the kb. Looking forward to it though00:25
vgradeN900 plasma link if your interested,
Jucato*very* interested. thanks! :)00:27
vgradeit will probably crash and burn00:28
Jucatohehe as long as it doesn't brick :)00:28
*** vetsin_ has joined #mer00:29
vgradei've not managed to brick anything yet apart from a couple of 16gb sdcards, but I think  they were duds anyway00:29
*** KaziKluBey has quit IRC00:30
*** KaziKluBey_N900 has quit IRC00:30
vgradedevice adaptation can be hard00:30
Jucatoas I feared :(00:31
vgradeI think some of the vega kernal guys blew some speakers a while back00:31
*** KaziKluBey has joined #mer00:31
*** KaziKluBey_N900 has joined #mer00:31
Jucatosomething I kinda wanted to learn. I mean, it would be great if we were able to use those not-so-expensive-but-quite-ok Android tablets.00:31
vgradeJucato, yea the prbelm with the cheap andriod tablets is the kernel support is dire00:32
vgradeI don't know if its because once a kernel is out its source is protected by the OEM as the hardware can be cloned easily but not the sw00:34
Jucatovgrade: yeah. it's still a very distant daydream for me.00:37
* Jucato is still debating whether to get a Vega ...00:38
vgradeI think mdfe has a vega clone he wants to get up00:40
*** jamien has joined #mer00:42
*** akira_tsukamoto has left #mer00:44
*** dominikb has quit IRC00:45
*** lynxis has joined #mer00:52
*** arcean_ has quit IRC00:56
Jucatovgrade: iirc you mentioned something about Mer/Plasma Active running on the Asus Transformer?01:00
*** kiviluoto has joined #mer01:02
vgrademer runnig tablet-ux I think but should do plasma01:04
vgradelilstevie has it01:05
Jucatoah. thanks01:06
*** vetsin_ has left #mer01:19
*** stefanopi has quit IRC01:22
*** stefanopi has joined #mer01:23
lilstevievgrade: well I will as soon as I can get this graphics driver solved01:25
lilstevievgrade: it is being silly :p01:25
*** stefanopi has quit IRC01:28
*** stefanopi has joined #mer01:30
Jucatolilstevie: so getting a transformer is too much of a gamble for Mer development right now?01:39
lilstevieJucato: if you don't already have one it is a gamble for anything01:48
lilstevieJucato: however the plasma UX stuff works, and I already have a full ubuntu installation running01:48
lilstevieit is just the tablet-ux/meego stuff that is being a little tricky01:48
Jucatolilstevie: true. it's a matter or which is more expensive :)01:48
lilstevieno, the if you don't already have one gamble is because of SBK01:49
lilsteviesecure boot key01:50
Jucatooh ...01:50
lilstevienewer units have a new SBK01:50
Jucatois there a way to check before purchasing? like a number?01:51
lilsteviewell the key changed somewhere in the B7O range01:51
lilstevieso serial nubers B6O and lower are certainly the old, known one01:53
* Jucato is actually going out to look today ... just look ...01:54
lilsteviefair enough01:54
JucatoI guess since the ExoPC is a no-go, it will be a choice between a less-powerful but more affordable Vega and a more powerful but expensive Transformer. :)01:56
pdanek1there is still a powerfull and yet very affordable Amazon Kindle Fire :)01:58
Jucatosame problem with the ExoPC.local availability :)02:00
pdanek1Amazon will start sending Kindle Fire internationally, with no additional tax02:01
pdanek1means they will cover all additional taxes, if you order from US and you are from Europe, you don't have to pay additional tax, $199 fixed price02:02
wmaronelilstevie: wait wait, ASUS deliberately locked down the boot of the Transformer?02:02
Jucatopdanek1: hackable too? and when will they start doing that (shipping)?02:04
*** berndhs has quit IRC02:04
lilsteviewmarone: yes02:07
lilsteviebut so do a lot of companies02:07
wmaroneI wasn't expecting such dickishness from ASUS though02:08
lilsteviewell they are slightly better than ACER02:09
wmaronewhat'd acer do?02:09
lilsteviesame thing, but each device uses a different key02:09
wmaroneoh that's wonderful02:10
lilstevieso if we both had A500's they would both have different SBKs02:10
ali1234how does that work then?02:10
lilsteviewhere ASUS only changed it once, in response to it being leaked02:10
wmaroneso I take it the secure boot bars the installation of other kernels much like on motorola handsets?02:10
lilstevieali1234: the boot process gets encrypted on upload02:10
lilsteviewmarone: not so much,02:11
lilstevieit makes it harder02:11
lilstevieeverything up to the bootloader is encrypted02:11
ali1234so it's generally impossible for the user to update the bootloader, even if there was a "signed" update teleased?02:11
lilstevieexploits still allow kernel changed02:11
lilstevieali1234: no the bootloader has the ability to encrypt updates02:11
ali1234then... that means the bootloader has the key?02:12
wmaronethat might be why I haven't heard a firestorm of hate towards asus and acer02:12
lilstevieit runs through the AES engine02:12
lilsteviebut the bootloader clears the key02:12
lilstevienot that it matters02:12
*** lynxis has quit IRC02:12
lilstevieinitiating an update from userland works02:12
ali1234it matters to me :)02:12
lilsteviejust if something goes wrong you can get fucked02:12
ali1234is there a wiki about this stuff somewhere or something?02:12
lilstevieum, a little on xda02:13
lilsteviebiggest thing is, the SBK is needed to initiate communication with APX mode02:13
ali1234i don't understand... where it gets the key from...02:14
lilstevieyou don't know the tegra arch then I takeit02:14
lilstevietake it*02:14
lilstevieok, it is in tegra fuses02:14
ali1234does it have like, a hardware keystore?02:14
lilsteviethey are efuses02:14
ali1234such that you can encrypt things but not read out the key?02:15
ali1234and also each one is unique set at the factory?02:15
lilsteviewhich when initialized they are part of the keystore02:15
lilstevieali1234: thats what ACER do02:15
lilstevieASUS no though02:15
ali1234yeah i was talking about the acer thing all along02:15
lilstevieand the keyslot is cleared from the AES engine before you boot02:15
ali1234that's pretty lame02:17
ali1234why is hardware encryption always set up to work against the owner of the device?02:17
*** stefanopi has quit IRC02:17
lilsteviewell that is how tegra is set up :p02:18
*** stefanopi has joined #mer02:18
ali1234i don't have a problem with the architecture02:20
*** kimitake_idle is now known as kimitake02:21
ali1234the problem is who gets custody of the private key. it should be me, and nobody else02:22
*** lynxis has joined #mer02:37
lilstevieits symetric btw02:49
*** OLaee has quit IRC03:08
*** lynxis has quit IRC03:13
*** srikanth_rst has joined #mer03:13
*** lynxis has joined #mer03:14
*** lynxis has quit IRC03:27
*** lynxis has joined #mer03:34
*** lynxis has quit IRC03:39
*** lamikr has joined #mer03:40
*** smoku has left #mer03:43
*** jjmarin has joined #mer03:44
*** KaziKluBey_N900 has quit IRC03:50
*** srikanth_rst has quit IRC03:51
*** lynxis has joined #mer03:54
*** stefanopi has quit IRC04:05
*** stefanopi has joined #mer04:06
*** kimitake has quit IRC04:06
*** keithzg has joined #mer04:09
*** kimitake has joined #mer04:38
*** jjmarin has quit IRC04:40
*** kidproquo has joined #mer04:41
*** stefanopi has quit IRC04:51
*** stefanopi has joined #mer04:51
*** kimitake is now known as kimitake_idle04:53
*** kimitake_idle is now known as kimitake04:55
*** stefanopi has quit IRC04:56
*** stefanopi has joined #mer04:58
*** Alison_Chaiken has quit IRC05:02
*** stefanopi has quit IRC05:31
*** stefanopi has joined #mer05:32
*** disco_stu has quit IRC05:36
*** stefanopi has quit IRC05:37
*** disco_stu has joined #mer05:37
*** stefanopi has joined #mer05:38
*** messerting has joined #mer06:10
lilstevielol Stskeeps06:22
*** srikanth_rst has joined #mer06:25
*** andre103 has joined #mer06:25
*** stefanopi has quit IRC06:37
*** stefanopi has joined #mer06:38
*** stefanopi has quit IRC06:42
*** stefanopi has joined #mer06:42
*** KaziKluBey has quit IRC06:48
Stskeepsmorn lilstevie06:50
*** lamikr has quit IRC06:51
*** stefanopi has quit IRC07:05
*** stefanopi has joined #mer07:05
*** srikanth_rst has quit IRC07:12
*** maour has joined #mer07:12
*** druid23 has joined #mer07:20
*** afiestas has quit IRC07:23
*** afiestas has joined #mer07:24
*** afiestas has quit IRC07:28
*** faenil has joined #mer07:34
*** norayr has joined #mer07:44
*** srikanth_rst has joined #mer07:45
*** rdqfdx has joined #mer07:53
*** afiestas has joined #mer08:16
*** afiestas_ has joined #mer08:24
*** afiestas has quit IRC08:25
*** stefanopi has quit IRC08:27
*** stefanopi has joined #mer08:28
*** raignarok has joined #mer08:34
*** jjmarin has joined #mer08:47
*** faenil has quit IRC08:48
*** bergie has joined #mer08:50
*** leinir has joined #mer08:52
*** stefanopi has quit IRC08:53
*** stefanopi has joined #mer08:53
*** bergie has quit IRC08:54
*** M4rtinK has joined #mer08:55
*** lizardo has quit IRC08:57
*** jjmarin has quit IRC08:59
*** niqt has joined #mer09:01
*** kidproquo has quit IRC09:05
* Stskeeps stretches09:05
*** norayr has quit IRC09:05
niqtwhere is image for this ?09:12
Stskeepsniqt: not public yet as we want to clean it up a bit09:13
Stskeepsmorn lbt09:13
lbtshould be around today - nasty issues with plumbing yesterday :(09:13
lbtheating still not working properly ...09:13
Stskeepsi'd like to discuss process when you have an hour or two09:14
*** maour has quit IRC09:14
lbtOK - I was going to take a look at the CE situation09:15
lbtmay need to start some rebuilds09:15
Stskeepsjust to slowly kickstart use of BOSS in Mer so09:15
Stskeepsand what we specifically need to do09:15
lbtOK - you're playing the "carrot" card :)09:15
*** srikanth_rst has quit IRC09:16
Stskeepsnah, just practicalities09:16
lbthehe - it's allowed09:16
Stskeeps looks interesting09:16
Stskeepsanother gerrit issue i noticed, a slightly annoying one: i change the hostname and because of the way openid/google accounts work, it sees logins as completely new identities09:17
lbtyes - they're there - but iirc they re-invented git hooks. So if you commit to gerrit hooks you can't run a non-gerrit git server09:17
*** Sazius has quit IRC09:18
Stskeeps-ish, those aren't stored in the actual git repos09:18
*** Sazius has joined #mer09:18
Stskeepsso it's more like hook into gerrit process instead of individual gits09:18
*** LoCusF_ has quit IRC09:18
lbtyes - exactly09:18
lbtbut again, I've not looked in detail - this was part of someones explanation09:19
Stskeepsi think we can abstract ourselves a bit so we don't specifically rely on a specific review board09:19
Stskeeps is my initial thoughts on what we probably need to do as a start09:20
Stskeepsintentionally abstracted as to not rely on gerrit vs reviewboard vs whatever09:23
*** panta has joined #mer09:34
*** tarantism has joined #mer09:34
Stskeepslbt, Sage_ , vgrade, w00t, Mordae: please re-login to , go to "settings" -> Profile, give me "account id" and don't do anything else09:39
Stskeepshmm, ok, so it's a google account only problem09:41
Stskeepsas your identity is static09:41
lbtOK, good09:41
*** stefanopi has quit IRC09:42
w00ttyped by hand as am mobile so may be off by a 009:44
*** stefanopi has joined #mer09:44
Stskeepsw00t: ok, that was after logging?09:44
Stskeepslog in , i mean09:45
Stskeepsfor the logs: this was the problem:
lbtI'm not liking that you have to sign in to google to view archives of mailing lists09:47
Stskeepslbt: that's kind awkward, true09:47
lbthow ironic09:48
w00tStskeeps: yes09:49
Stskeepsw00t: ok09:49
Stskeepsw00t: and it seems to work okay, it doesn't see it as a new account?09:49
Stskeepsie "anonymous coward"09:50
w00tdunno hard to check atm09:52
*** panta has quit IRC09:56
*** slx is now known as swerden09:59
*** raignarok_ has joined #mer10:00
*** niqt has quit IRC10:01
*** raignarok has quit IRC10:02
*** raignarok_ is now known as raignarok10:02
*** lynxis has quit IRC10:11
*** pdanek has joined #mer10:22
*** pdanek1 has quit IRC10:25
*** lynxis has joined #mer10:28
* Stskeeps sets up gitweb10:30
*** norayr has joined #mer10:33
lilsteviemorning vgrade10:33
*** norayr has quit IRC10:34
Jucatogood evening10:34
Jucatolilstevie: it seems most of the Transformer units around here are B60*10:34
lilstevieJucato: B6 is ok10:36
lilstevieB7 is when the transition happened10:36
Jucatooh crap. forgot I have to test vgrade's n900 image. and that I forgot to buy a card reader for the desktop -_-10:37
lilstevieStskeeps: they changed the secure boot key on newer units10:37
Stskeepslilstevie: ah10:37
lilsteviebut it is very random10:37
lilsteviethere is no specific serial that is the change over10:38
vgradeStskeeps, 100000410:38
*** lizardo has joined #mer10:39
*** stefanopi has quit IRC10:39
*** stefanopi has joined #mer10:40
Stskeepsvgrade: ok, looks like you're good10:40
vgradeJucato, are you planning to test that N900 image?10:40
vgradeStskeeps, ok10:40
Jucatovgrade: kinda. just dd it directly to a microSD card right?10:41
Jucatoas long as it won't brick the n900, I'm good :P10:41
lilstevievgrade: I need a bit of a hand with the accel drivers still10:42
vgradeJucato yea if you have the dual boot stuff on there.10:42
lilstevievgrade: I am getting some weird error10:42
Jucatovgrade: yeah. u-boot and stuff. I used to try out meego ce releases10:42
vgradelilstevie, about matching formats10:43
vgradeJucato, ok your good to go10:43
Jucatothere's a fire extinguisher nearby anyway. just in case10:44
lilstevievgrade: about version matching modules10:44
vgradelilstevie, what kernel version are you on10:45
lilsteviecapturing the exact wording is a bit of a pain, cause it flashes on screen, but it doesn't show up in logs10:46
*** lynxis has quit IRC10:47
*** gprade has joined #mer10:47
*** lynxis has joined #mer10:48
*** kthomas_vh_ has quit IRC10:48
vgradelilstevie, which ux and image is this with10:49
*** kthomas_vh_ has joined #mer10:49
lilstevieNvRmModuleGetCapabilities: MOD[85]10:50
lilsteviethats all I have captured so far of the error10:50
lilsteviethat I generated10:50
vgradelilstevie, I see that on my images as well10:51
lilsteviecannot find matching version of module10:51
lilstevieok, well my issue is that it flashes10:51
lilstevieso the image shows10:52
lilsteviefor like 2-3 seconds, then the screen goes black10:52
vgradebut only on tablet, plasma is ok10:52
vgradecan you try 16 bit depth on xorg config10:53
*** gprade has quit IRC10:53
lilstevietrying now10:58
*** arcean has joined #mer11:00
*** lynxis has quit IRC11:00
lilstevievgrade: awesome worked11:01
vgradelilstevie, np11:05
lilstevienow I just need to figure out buttons :p11:06
vgradeil, did you do some pics/video?11:06
lilstevievgrade: I sent you a link with plasma11:06
StskeepsMordae: ping11:07
lilsteviehmm how do you quit an app?11:07
*** stefanopi has quit IRC11:07
*** stefanopi1 has joined #mer11:07
*** lizardo has quit IRC11:12
vgradelilstevie, with the optical sensor on the wetab11:22
vgradelilstevie, windows key on a usb kb if you have usb host11:23
vgradei really needs a gesture or hardware button11:23
lbtOK - we have a queue of build trials for n950 now11:24
lbtand I've logged a bug against islam for the silly osc session timeout that boss is having11:25
lilstevievgrade: I have a home button on the keyboard dock which seemed to do it11:25
lilstevie(same id as windows)11:25
vgradelilstevie, yes probably. I missed the link to the pics/video can you post here11:27
lbtStskeeps: there was a rejection for you : Package policy-settings-basic-n950 not built successfully in project home:stskeeps:mer:hw:n950 repository CE_Mer_Core_armv7hl for architecture armv8el11:27
Jucatovgrade: seeing the plasma active splash screen! :D11:28
Stskeepslbt: that's not going to be fixed right away since we don't have policy framework atm11:28
Jucato(and got a notice about power management)11:28
vgradeJucato, ok good11:28
*** lynxis has joined #mer11:29
vgradeJucato, first boot takes a while11:30
Stskeepslbt: same error should exist for n90011:30
lilstevieno videos yet11:30
*** lizardo has joined #mer11:30
lbtno, just the rescue-initrd11:31
lbtsimply not present11:31
Jucatovgrade: oh subsequent boots will be faster? I mean if I boot back into Maemo 5 and book later again into Active, it will not "reset" right?11:31
Stskeepslbt: correct11:31
vgradeJucato, right11:32
Jucatocool :) *puts it away for a while*11:32
*** trbs has joined #mer11:32
Jucatovgrade: to see even just the splash screen was already nerdgasmic for me :D11:32
Stskeepslbt: so, i should help review?11:33
Jucatovgrade: this is based on mer already?11:33
lbtStskeeps: I scripted a review accept as this is just a priming run11:34
Stskeepslbt: ok11:34
vgradeJucato, yes on the kickstart Stskeeps posted yesterday for Nemo11:34
lbtfailing on armv7el - I've not even looked why11:34
vgradeJucato, can you take a snap, how far did it boot?11:34
Jucatoseems to have frozen up11:35
Stskeepslbt: -bin will likely fail on armv7el, don't bother building for it11:35
lbtI'm about to run the generic x8611:35
lbt works though ?11:35
Stskeepslbt: that's armv8el11:35
lbtoh, nm11:35
Stskeepsn900/n950 is hardfp only11:35
Jucatovgrade: up to the Introduction activity, but it failed to load the Welcome widget, which takes up more tha the screen (so the box goes beyond the screen, but the activity controls are placed properly)11:35
lbtyanking the bad build targets from CE:Adaptation:N950-N911:36
vgradeJucato, sec for key combination, can you do a picture11:37
lbtI just saw 2 columns and thought they'd been setup as duplicates :)11:37
*** khetzal has joined #mer11:37
lbtboss is churning through - here's the trial build for 2504
Jucatovgrade: trying to get one. lighting at home is terrible at night :)11:38
lbtso if it goes wrong it (theoretically) shouldn't delete it and we can diagnose11:38
vgradeALT+D+R to remove welcome screen11:39
Jucatovgrade: what's alt on the n900? :) the "Fn" key? any it seems frozen. might have to restart it11:40
Stskeepslbt: Project home:sage:Mer:adaptation:x86 does not contain a repositorythat builds only against project CE:Adaptation:x86-generic repository Mer_Core_armv7hl for architectures armv8el11:42
Stskeeps.. x86 against ARM?11:42
lbtyeah - that's the single priming run - I thought the config was done yesterday11:43
lbtwe don't have much of a dashboard - thinks are too scattered :(11:44
lbtI intend to have a 'tab' in OBS showing the config and process selected for a project11:44
lbtI'v proof-of-concepted it and it's not hard11:44
lbt(cf the 'Project Config / Users / Subproject ' tabs)11:45
Jucatovgrade: sorry, best I could do on such short notice:
lbtthat's better11:48
Jucato(usually don't do shots of my phone :)11:48
*** jjmarin has joined #mer11:53
Jucatovgrade: ah well, looks like it really hangs up there. (already tried restarting)11:54
lbtStskeeps: so the x86-generic packages really are 'bad' - Sage_ was fixing them up12:03
lbtIf anyone is bored we could do with changes files  in home:sage:Mer:adaptation:x86  mtdev slang newt and linux-firmware being fixed up12:04
*** lynxis has quit IRC12:04
*** arcean has quit IRC12:05
*** arcean_ has joined #mer12:05
*** druid23 has quit IRC12:07
*** druid23 has joined #mer12:07
*** raignarok has quit IRC12:12
*** druid23 has quit IRC12:12
*** lynxis has joined #mer12:18
vgradeJucato, tried the alt+d+r?12:19
Jucatovgrade: I only know Ctrl and Fn and Shift keys on the n900. I don't know how to do an "Alt"12:20
vgradeah just seen your message sry12:20
*** Jucato has quit IRC12:21
*** Jucato has joined #mer12:21
Jucatosorry, crashed.12:22
Jucatovgrade: also, if you could see in that shot, the panel above is a bit broken. specifically the icons to the left show up for a few second and then get replaced by that X button. the rest of the system is unresponsive *except* when the screen is locked. I seem to have been able to unlock it. going to try to reproduce that now12:25
*** lynxis has quit IRC12:28
*** stefanopi1 has quit IRC12:30
vgradeStskeeps, is there any key combo in N900 adaptation to generate and Alt keypress12:30
*** stefanopi has joined #mer12:31
Stskeepsdon't think so12:31
Stskeepscheck the x11 keymap12:31
Sage_Stskeeps: 100000212:31
StskeepsSage_: ok, so it was just my login that was messed up12:31
Sage_I'll have a fix round tomorrow at some point12:33
Sage_Today I don't have time to do much.12:33
*** pdanek1 has joined #mer12:41
Jucatovgrade: strangely enough, I can drag the unlock icon (the visual is slow, but the action/response is fast)12:42
*** pdanek has quit IRC12:44
*** raignarok has joined #mer12:44
*** kiviluoto has quit IRC12:45
*** kiviluoto has joined #mer12:48
*** lynxis has joined #mer12:51
vgradeI think we need to get rid of that welcome dialog, we had the same issue on vega but were able to remove the widget using the alt-r-d combo12:51
Jucatoit seems I could actually "scroll down" *very slowly*. but I did something that kinda stopped that ... heh12:53
*** berndhs has joined #mer13:03
*** norayr has joined #mer13:03
*** norayr has joined #mer13:03
*** druid23 has joined #mer13:10
* Stskeeps upgrades his n95013:15
*** khetzal has quit IRC13:25
*** stefanopi has quit IRC13:37
*** stefanopi has joined #mer13:38
*** srikanth_rst has joined #mer13:39
*** jamien has left #mer13:40
*** faenil has joined #mer13:41
*** Mordae has quit IRC13:43
*** srikanth_rst has quit IRC13:43
*** Mordae has joined #mer13:43
*** stefanopi has quit IRC13:45
*** stefanopi has joined #mer13:45
*** stefanopi has quit IRC13:48
*** stefanopi has joined #mer13:49
*** norayr has quit IRC14:10
*** druid23 has quit IRC14:11
*** druid23 has joined #mer14:18
Stskeepslbt: /me tries to understand BOSS14:20
*** slx has joined #mer14:22
*** swerden has quit IRC14:25
*** berndhs has quit IRC14:26
*** norayr has joined #mer14:26
*** berndhs has joined #mer14:27
*** faenil has quit IRC14:28
*** druid23 has left #mer14:29
*** stefanopi has quit IRC14:32
*** stefanopi has joined #mer14:32
*** maour has joined #mer14:38
*** norayr has quit IRC14:46
*** faenil has joined #mer14:48
*** trbs has quit IRC15:03
*** norayr has joined #mer15:04
*** jjmarin has quit IRC15:06
*** stefanopi has quit IRC15:09
*** stefanopi has joined #mer15:10
*** trbs has joined #mer15:12
*** raignarok_ has joined #mer15:16
*** raignarok has quit IRC15:20
*** raignarok_ is now known as raignarok15:21
*** maour has quit IRC15:22
*** maour_ has joined #mer15:22
*** kimitake has quit IRC15:29
lbtStskeeps: back now - what do you want to know :)15:29
*** maour__ has joined #mer15:33
*** maour_ has quit IRC15:33
Stskeepsi'll ask when done watching the big bang theory :P15:40
*** NIN101 has joined #mer15:41
lbthehe - we recorded all the reruns and are on series 2 now15:44
*** kimitake has joined #mer15:44
Stskeepsso, long story short: workflow and AMQP is what's used to do general workflow handling and message queueing in BOSS, right?15:46
Stskeepsruote and AMQP i mean15:47
*** NIN101 has quit IRC15:48
*** kimitake has quit IRC15:48
lbtruote is the workflow engine15:49
lbtamqp simply does reliable message queues15:49
lbtclients (python) do the work and listen to amqp queues15:50
lbtclients can be any lang really15:50
Stskeepswhat has been the biggest issues in the development up to now, ie, setbacks, etc?15:52
Stskeepsjust to keep track15:52
*** andre103 has quit IRC15:52
lbtof BOSS ?15:52
lbtin my opinion we've taken it to a production style development process too early.15:53
lbtit has also been hurt by not having lib-osc15:53
Stskeepsi can't find any good explanation of what SkyNET is, can you elaborate on that?15:54
lbtmmm ... it's like a framework for running participants15:54
lbteach participant is a simple plugin that only does logic15:55
lbtskynet is all the daemon, logging, start, stop, register15:55
lbtso "skynet register check_changelog"15:55
lbt"skynet start check_changelog"15:55
lbt"skynet log check_changelog"15:55
Stskeepsso it's kinda like svchost from windows, it hosts one or more 'services', ie, participants15:55
*** poohmaan has joined #mer15:56
lbtyeah - sounds right15:56
lbtit's only a shell script15:56
lbtso that's a small participant - no need to worry about daemons, handling signals, system shutdown, logging15:57
lbtalso no need to think about amqp15:57
lbtjust 3 basic interfaces15:57
Stskeepsso, process defines to take worksheet/json structure first to get_changelog, then to another check, then to another one, etc?15:57
lbtthat process has to start sometime - on an event15:58
lbtso anything that does 'events' has to have a launcher15:58
lbtto convert event to a process :)15:58
lbtrobogrator handles OBS events.... does a filesystem lookup15:59
lbtand then launches the process it finds15:59
lbtoop  that was a bad url to show you15:59
Stskeepsconvert event to a process sounds weird in my ears - i would think rather that an event comes in as a worksheet and then gets passed on to the right process - i assume processes can start other processes?16:00
*** tarantism has quit IRC16:00
lbtan event happens inside OBS16:01
lbtor in bugzilla16:01
lbtor in the real world ...16:01
Stskeepsand based on that event, a process is started with an initial worksheet?16:01
lbtso that needs a beskpoke event-2-process mechanism16:01
lbtfor OBS we have a plugin16:02
Stskeepscan that process then launch other process'es with a worksheet?16:02
lbtand it looks here
lbtit can launch fresh ones or it can do concurrence16:02
*** kimitake_idle has joined #mer16:03
lbtthe workflow can handle many types of pattern16:04
Stskeepsis there a good document describing ruote process definitions?16:04
Stskeepslike, how to write them16:04
*** lynxis has quit IRC16:04
lbtbut somewhat technical16:04
lbtfirst see
*** maour__ is now known as maour16:05
Stskeepsright, i worked with coloured petri nets / was in that research group so this isn't a stranger to me16:05
lbtthat is from an academic analysis of wflow16:05
lbtOK so
lbtis about how various patterns in that analysis are realised in ruote16:06
lbt(incomplete but getting there)16:06
lbtanyhow - the issues for BOSS are not in the ruote or amqp space16:07
Stskeepsok, so i can see ruote has workflow and defines participants too - where does AMQP fit in?16:08
lbtjust above TCP16:08
lbtand should be almost ignored16:08
Stskeepslooking at 'the engine and some participants' on
lbtamqp gives reliable json message queues over a network - sure there is fun to be had there but it's just a tool16:09
lbtwe will eventually be able to do master-master replication etc16:09
lbtwhich matters when scaling up16:09
lbtwe could have chosen zeroMQ and probably done just as well16:10
Stskeepsok, so, what handles the ruote workflow? skynet?16:10
lbtnot really16:11
Stskeepsspecifically what hosts it16:11
lbtskynet just launches the participants16:11
lbtruote has a worker class16:11
lbtwe make an instance16:11
lbtand run it under daemontools/runnit16:11
lbtso it reloads if it dies16:11
lbtit's 'robust'16:12
lbtso if it crashes the DB should always be consistent16:12
lbtboss is simply an instance of a ruote worker16:12
*** messerting has quit IRC16:12
Stskeepsok, so, let's say we have - check_submitter_maintainer will talk over AMQP to skynet participants check_submitter_maintainer?16:14
lbtthink of it as like RPC16:15
*** norayr has quit IRC16:15
Stskeepsthis is RuoteAMQP::ParticipantProxy ?16:15
lbtsee the msg I sent16:15
Stskeepsso theoretically i can start prototyping a Mer process definition16:15
lbtnow I *deliberately* wrote boss to allow personal hacking16:16
lbtso you can run boss participants on shell.in16:16
Stskeepshow much footprint does BOSS take?16:16
Stskeepsinstallation, memory, cpu16:16
lbtmany kb16:16
Stskeepsis it feasible to run an instance in a 1gb vm?16:17
Stskeepsie, memory16:17
Stskeepsthe packaging checks you've done for CE.. are they tied to OBS?16:18
lbtsome are - since they typically pull the files from the OBS project16:19
Stskeepsjust trying to assess how much we realistically can reuse16:19
Stskeepscan you give me a simple example of something that launches a process with an initial worksheet in JSON?16:20
lbtnow ... those are probably a touch outdated - but they'll only get simpler16:22
lbtthe client won't work now as the register() method went away16:22
Stskeepsok, i have to eat so brb16:23
Stskeepslooks good so far16:23
lbtl8r ...16:23
*** lynxis has joined #mer16:27
Stskeepsso, - fields is the initial worksheet?16:33
lbtsec - rebooting16:34
lbtneed to fix usb for n950 to flash Nemo16:35
*** faenil has quit IRC16:35
Stskeepsyeah, err, if you did upgrade over the air, don't16:35
*** lbt has quit IRC16:36
*** kimitake_idle is now known as kimitake16:40
*** lynxis has quit IRC16:52
*** lynxis has joined #mer16:57
BostikStskeeps: qt5 isn't quite ready for tablets, but it should be getting there; case in point:
Bostikonce that gets fixed, we're all better set17:04
Bostiktook ages to debug that closely17:06
*** dariornelas has joined #mer17:16
*** dariornelas has left #mer17:18
*** joejoe_ has quit IRC17:24
*** joejoe_ has joined #mer17:25
*** Venemo has joined #mer17:27
pdanek1Stskeeps: you own N950?17:46
*** toggles has quit IRC17:55
*** toggles has joined #mer17:57
*** maour has quit IRC18:17
*** maour has joined #mer18:18
*** mr_jrt has joined #mer18:19
*** maour has quit IRC18:23
*** messerting has joined #mer18:24
*** jarkko^ has quit IRC18:29
*** harbaum has joined #mer18:33
*** lbt has joined #mer18:33
*** stefanopi has quit IRC18:35
*** stefanopi has joined #mer18:36
*** slx has quit IRC18:36
*** poohmaan has quit IRC18:38
*** slx has joined #mer18:39
*** stefanopi has quit IRC18:44
*** stefanopi has joined #mer18:45
*** andre103 has joined #mer18:50
*** Jonno has joined #mer18:52
*** maour has joined #mer19:01
*** lizardo has quit IRC19:06
*** Zaran_ has joined #mer19:07
Stskeepspdanek1: i don't own it19:15
Stskeepslbt: don't try to flash nemo when you have beta2 + over the air, there's no public flasher image for the PR1.1 thing19:15
lbtStskeeps: I avoided the OTA upgrade19:30
*** Venemo has quit IRC19:30
*** Zaran_ has quit IRC19:32
*** stefanopi has quit IRC19:40
*** swer has joined #mer19:40
*** stefanopi has joined #mer19:40
*** slx has quit IRC19:42
Stskeepslbt: so something like this could be a beginning of a process?
*** andre103 has quit IRC19:46
lbtthe hard thing about this stuff is deciding where to break the logic down19:47
lbtalso what is passed around on the worksheet19:47
lbthave you read the CE one ?19:47
lbtwhat about the Apps one?19:48
Stskeepsnot that one19:48
lbtjust to give you a flavour19:48
lbtit does some concurrence and voting19:48
lbtit also has a periodic reminder that a supertester needs to look at an unloved package19:49
Stskeepshow well does this stuff deal with "reach out and touch a human"?19:49
lbtthere are 2 answers :)19:50
lbtat the conceptual / design level .... really well!19:50
lbtat the "have we implemented it yet" level - not so much19:50
lbtbut if you consider the app store19:50
lbtthat reaches out to humans for votes19:51
lbtand it slots in really well19:51
lbtessentially your web app (or whatever) simply listens on the queue for tasks19:51
lbtthen the human interacts with the web app19:52
lbtand the app sends the task back19:52
lbtnb ... running CE on my N950 now ... next step is to run Nero19:52
lbtI keep doing that :)19:53
lbtwith your process... what event is it reacting to ?19:53
Stskeepsgerrit patchset crated19:54
Stskeeps(or patch received, or whatever..)19:54
Stskeepswe can probably abstract it a bit19:54
Stskeepsso we can hook into both RB and gerrit19:55
Stskeepsa patchset received is pretty much the entry point for mer contribution, so19:56
Stskeeps is where it fits in19:57
lbtI read that earlier19:57
Stskeepsyeah, i updated it a bit during today19:57
lbtah, so you did19:58
lbtso I think the idea is that boss does the drudge work19:58
Stskeepsyeah, it handles automation19:58
Stskeepsand gets things done/checked/etc19:58
lbtand there are 2 approaches - it controls an end-2-end process with hand-off to people (as per CE right now)19:59
lbtor it handles snippets and people hand off to BOSS19:59
Stskeepswell, or my third approach: it handles patchsets and it acts like an automated reviewer and gives feedback on the change, leaving maintainers/humans to do the final review20:00
Stskeeps / approval20:00
lbtwho applies the patch?20:00
Stskeepsmaintainer clicks submit when patchset is marked as verified and reviewed as +2 (, approved) and gerrit merges to git20:01
lbtso CE does 'your approach' today20:01
*** fw190 has joined #mer20:01
lbtthe difference with gerrit is that boss would then pick up a 'maintainer has made a commit' event20:02
lbtand possibly do a new image or something20:02
Stskeepsit could do that, but probably have done that even earlier in process20:03
Stskeepsdepending on what makes sense with resources available20:03
lbtor mark "change made, do new QA image tonight"20:03
lbtso it only does QA images on nights when changes occur20:03
lbtall kinds of little rules20:03
Stskeepsso, my line of thought is along lines of: before change is even allowed into the master of the package's git, it will have been tested in different ways, localdep builds, image test, submitting to vendors to help verify it etc. when it's inside the master of package's git and staged for updated into "Core"'s git (read: updating the package's git commit that Core points to), it'll do test with the potential combined tests20:06
Stskeepsso we're fairly sure that once Core moves ahead with the admitted changes, it's releasable20:06
Stskeepsand sanity checked20:07
*** harbaum has quit IRC20:07
Stskeeps.. not sure if that made sense20:07
lbtit does - and it raises questions of course20:08
lbtessentially acceptance criteria20:08
Stskeepsand how we collect data on how a change affects vendors20:08
lbtveto ?20:08
*** stefanopi has quit IRC20:08
Stskeepssanity check by maintainer based on reactions to change20:09
Stskeepsi think20:09
*** stefanopi has joined #mer20:09
lbtI would like our open reference implementations to be notified of changes too20:10
Stskeepsof course20:10
lbtmmm this gets back to that drawing of patches from Core impacting the CI process for the CE/Vendor20:10
Stskeepsthe idea is actually that all vendors get notified of QA checks20:10
Stskeepsie, "can you help us out with this and this change and check it"20:11
lbtdo we classify changes20:11
lbtbugfix vs feature20:11
Stskeepseven the smallest can break stuff really :P20:11
Stskeepsin practice they're all changes20:12
Stskeepsperhaps prioritization20:12
*** Jonno is now known as Jonno_20:13
lbtthinking what changes are likely to occur20:14
lbtdo you have any data on that?20:14
Stskeepskyb3r's statistics could be interesting for that20:14
Stskeepswe basically dumped entire meego trunk + history20:14
lbtnew upstream, code patch, packaging,20:14
Stskeepsthe important thing is that we provide a useful core<->vendor interface and allow vendors to help us out keeping core quality and feed back, closed or open20:17
lbtthe vendor notification could be an invitation to participate in CI20:19
lbtI like that20:19
Stskeepsit -is- an invitation for all vendors and that one would be where we hook up CE, hardware adaptations, etc..20:20
lbtso that's a lot like CE  process with the addition of ABI and removal of OTS20:21
Stskeepsthat's why i'd like to share a lot of things20:21
lbtI think it makes sense to have a reference implementation BTW ...20:21
lbtmaybe kvm20:21
Stskeepsno preferential vendors, if we are required to have a proper interface for reference impl, we should offer it to everyone20:22
Stskeepswe can set up a few easy targets, though :)20:22
lbtI'm thinking of an early gate20:22
*** pdanek has joined #mer20:23
lbtie we run a kvm image and if it fails there then we don't notify vendors - change (probably) fails20:23
kyb3RStskeeps: stats coming, I have been pushing the other researcher to give me some graphs20:23
Stskeepswell, we can always do early-warning vendors, as in, these gets asked first in order to save everyone else effort20:24
lbtwas just about to say that20:24
Stskeepswith ability for everyone to sign up for early warning20:24
lbtpermit early hook or late hook20:24
kyb3R(I would do those stats my self, but engaged to too many things already )20:24
*** pdanek1 has quit IRC20:25
*** mr_jrt has left #mer20:27
*** smoku has joined #mer20:28
Stskeepslbt: "the review process will last at least 48 hours, for early notification vendors we notify right away, if there has not been any bad reviews within 12 hours, we notify late vendors", ish20:28
Stskeepssounds like a good approach20:29
Stskeepsso that's what i would like to steer towards in next week20:32
Stskeepsyeah, i know20:32
Stskeepsmonday, saturday, sunday ;)20:33
lbtpiece of cake20:33
Stskeepsi'm a little unsure how effectively to implement the vendor notification stuff20:33
lbtI would be happy to do this using BOSS - it's a good match20:34
Stskeepsyeah, that's the plan20:34
lbtthe network transport.... not sure20:34
Stskeepsthe git side of things is a bit of a paradigm switch to how it looks like it's currently done, with OBS20:34
Stskeepsbut i see very good things with the git approach that makes our life a lot simpler and process a lot easier to write20:35
lbtit is20:35
lbtI have a massive preference for the git approach though20:35
Stskeepsi've hooked up gitweb for gerrit now too, a bit more useufl20:36
lbtit gives an authoritative definition of the source20:36
*** arnet has joined #mer20:37
Stskeepsi'd also like to model some typical processes, like, roadmapping, requirements processes, code freeze policies, etc20:37
Stskeepssee how many roles we can replace with proper process20:38
Stskeeps(you have been replaced with a two-liner shell script..)20:38
*** leinir has quit IRC20:38
*** fw190 has quit IRC20:39
lbtbut that's the point - what's left should *need* a human - hence "boss does the drudge work"20:39
Stskeepsone thing though.. how does all this deal with processes that change?20:39
Stskeepslike, i have a process that's ongoing, but definition changes mid-way20:40
lbtBT spent millions *trying* to solve this problem :)20:40
Stskeepsand they didn't succeed?20:40
lbtit's a very very hard problem20:40
lbtif you use a state machine approach you're stuffed20:41
Stskeepsthe rule is then that you go back to square one ideally i guess, ie, as a design principle20:41
lbtbecause we use worksheets...20:41
lbtwe can cope20:41
lbtso for us ... just change the definition20:41
lbtnext launch will use the new process, in flight uses the old one20:41
lbtthere are issues - eg if your participant api changes20:42
lbtso there are 2 options - versioned interface20:42
lbtor run get_changelog_v220:42
lbtversioned interface is slightly nicer - and can be done as syntactic sugar20:43
lbtbut this needs something like the Nokia MINT team behind it20:43
lbtotherwise we lose focus on Mer Core20:44
lbthowever... from my perspective ... the design of BOSS and ruote (and the use of AMQP) makes that very doable.... the foundations are strong20:44
Stskeepswell, i'd like to get process running for mer, utilizing BOSS, so that's my focus as that's one of the powers i see in mer :P20:45
Stskeepsand probably a good way to cut costs..20:46
lbtyes, I see it as a key to high efficiency20:47
lbtand high quality20:47
Stskeepsit's possible to do amqp-less ruote, i pressume20:48
Stskeepsi'm thinking in vendors20:48
Stskeepsfor simple process20:48
lbtanything is possible20:48
lbtbut it's simpler to use amqp20:48
Stskeepsright, time for sleep20:49
Stskeepscya tomorrow20:49
lbtit's easier to install and run than mysql20:49
lbtOK - I'll write up a bit20:49
lbtb4 you go20:49
lbtare we aiming to have Nemo on the N950 for linuxcon?20:50
lbtfrom the BOSS managed process20:50
Stskeepsn950 hw adaptation isn't impressive, n900 is better20:50
lbtOK -20:50
lbtN900, then ExoPC ?20:51
lbtthe N950 will have sexy gadget appeal :)20:51
ShadowJKany unobtainable gadget becomes appealing ;)20:56
lbtShadowJK: true20:56
*** joejoe_ has quit IRC21:07
*** rdqfdx has quit IRC21:11
*** NIN101 has joined #mer21:16
*** npm_ has joined #mer21:26
*** npm has quit IRC21:26
*** arnet has quit IRC21:52
*** xxiao has quit IRC22:49
*** andre103 has joined #mer23:04
*** messerting has quit IRC23:05
*** pdanek1 has joined #mer23:05
*** pdanek has quit IRC23:07
*** peavey has quit IRC23:17
*** peavey has joined #mer23:21
*** NIN101 has left #mer23:21
*** wmarone_ has joined #mer23:31
*** wmarone has quit IRC23:31
*** wmarone__ has joined #mer23:34
*** wmarone_ has quit IRC23:34
*** wmarone__ is now known as wmarone23:51
*** gprade has joined #mer23:53
*** stefanopi has quit IRC23:53
*** stefanopi has joined #mer23:53
*** gprade has quit IRC23:58
*** smoku has left #mer23:58

Generated by 2.11.0 by Marius Gedminas - find it at!