IRC logs for #automotive for Monday, 2015-09-21

*** riazantsev has joined #automotive03:32
*** Luming has joined #automotive03:44
AlisonChaikenpaulsherwood for your amusement:04:56
*** riazantsev has quit IRC05:18
*** KlausUhl has joined #automotive07:09
*** Tarnyko has quit IRC07:20
jeremiahI won't make the TT meeting on IRC today, apologies.07:26
jeremiahCtpollard: Is everything all set with the browser PoC or do you need more from me?07:27
*** jacobo has joined #automotive07:39
*** poxar has joined #automotive08:22
*** toscalix__ has joined #automotive08:26
*** fredcadete has joined #automotive08:32
*** jonathanmaw has joined #automotive08:34
*** Tarnyko has joined #automotive08:42
*** CTtpollard has joined #automotive09:02
*** wschaller has joined #automotive09:12
CTtpollardgood morning09:13
TarnykoI just sent a mail to the ML, demonstrating the type of test suite which we could have on the AGL platform image : (it is "wayland-fits")09:16
Tarnykojeremiah: hi Jeremiah :) ! ^^ Just answered back, please tell me what you think if you have 1 min09:17
Tarnyko(ho, and I need to clarify, I am "Manuel Bachmann" actually)09:19
jeremiahI think we change browser-poc09:21
jeremiahTarnyko: Looks great. :-) How can I reproduce for the Porter board?09:53
jeremiahCTtpollard: Sorry, I think I'm having some pebkac issues.09:53
Tarnykojeremiah: saw your answer :). I am doing a small setup, so you can get it running easily09:57
*** wschaller has quit IRC10:04
Tarnykojeremiah: answered !10:17
*** wschaller has joined #automotive10:20
*** toscalix__ has quit IRC10:31
Tarnykojeremiah: by the way, I had a patch for something other (building the AGL QEMU image with Yocto 1.8), and I tried to push it for review by doing "git push origin HEAD:refs/for/master" (I have an account). It doesn't work, maybe I need additonal rights for that ?10:32
*** fredcadete has quit IRC10:58
*** Tarnyko is now known as Tarnyko_away11:02
*** toscalix__ has joined #automotive11:43
*** Tarnyko_away is now known as Tarnyko11:55
*** wschaller has quit IRC12:11
*** alex_h has quit IRC12:12
*** alex_h has joined #automotive12:19
*** gunnarx has joined #automotive12:26
*** wschaller has joined #automotive12:26
*** jacobo has quit IRC12:33
paulsherwoodGENIVI tools team meeting wil lbe here in just under 30 mins...12:36
gunnarxIf you've heard me rant about bitbake being buggy in its fetch when using a local source mirror...12:40
gunnarxThen it's documented here already:12:40
gunnarxSpecifically the quote "So don't specify SRC_URI like this:" SRC_URI = "git://;protocol=git;tag=v${PV}"12:41
gunnarxwhich is exactly what we do in meta-ivi for 4-5 components and precisely those are failing in my BB_NO_NETWORK setup12:41
gunnarxIf those are converted to SRCREV = <hash> the bug (imho it's a bug anyway) is not triggered.12:42
gunnarxbut using TAG=$PV seems useful I think12:42
paulsherwoodwhich suggests that the folks doign the work maybe did not read all the documentation... which is perhaps unsurprising given it's over 1000 pages :)12:43
gunnarxdo you mean I have not read it? :)12:48
gunnarxIt's true up until recently... :-/12:48
paulsherwoodgunnarx: i confess, i'm not going to read it12:49
gunnarxA few things to notice.  First this is a wiki and not the manual, so not as authoritative.12:49
paulsherwoodi was referring to the 'mega-manual'12:50
gunnarxYes, but it that then means that it's not certain the Wiki's "don't do that" is actually describing real best practice.12:51
gunnarxI've looked and it seems that poky uses almost exclusively SRCREV = <commit hash> everywhere.  So for them that is best practice, apparently...12:51
paulsherwoodi read somewhere that there's no such thing as 'best practice' really...12:51
gunnarxtag=$PV makes sense in theory, it means you just have to rename the recipe to bump the version up - everything else should then be automatic.12:52
paulsherwoodjust better, and worse :)12:52
gunnarxword games :)  I'm serious about that local mirror should work - see no reason for the behavior of going to the network simply because a lightweight tag is specified in the recipe instead of a hash.12:53
paulsherwoodgunnarx: ok, so is your recommendation that genivi should create its mirror by running bitbake, and publish the results from there?12:54
gunnarxnot making a recommendation at this point, just trying options12:54
gunnarxBy the way if you publish a mirror over http, there is no bug.12:55
gunnarxwell, it depends.  You can't run with BB_NO_NETWORK flag anyway since your mirror is on a network.  The recently added BB_ALLOWED_NETWORKS (sp?) could limit to your local network only, and I presume the bug is triggered also then, but not sure.12:57
paulsherwood== GENIVI Tools Team Meeting Starts ==13:00
paulsherwoodhi folks...13:00
paulsherwoodproposed agenda is here
paulsherwoodanyone have additions?13:00
*** klausbirken has joined #automotive13:00
*** Bjoern_ has joined #automotive13:00
paulsherwoodi guess we could discuss whether folks preferred the return to telephone last week :)13:01
gunnarxsource mirroring?  :)13:01
paulsherwoodcan add that, ok13:01
paulsherwoodanything else?13:01
*** waltminer has joined #automotive13:01
gunnarxMore generally, is CIAT topic no longer a topic?13:01
paulsherwoodi woud do it under automated testing13:02
*** steve_l has joined #automotive13:02
paulsherwoodthe agenda items are taken from the charter13:02
paulsherwood(well most of them)13:02
gunnarxok, let's do that13:02
paulsherwood= Confluence Progress =13:03
*** pavelk has joined #automotive13:03
* paulsherwood notices joel is not here13:03
paulsherwoodany other progress to report on this?13:03
paulsherwoodi see some pages being updated13:03
paulsherwoodin general i'll take silence as a justification to move onto next topic? :)13:04
paulsherwoodi wonder if it's only you and me today, gunnarx13:05
* steve_l waves13:05
pavelkany updates on by when a stable url will be available?13:05
klausbirkenI am here, too. Just silent.13:05
* waltminer is here13:05
paulsherwoodwelcome all :-)13:05
* Bjoern_ is, too.13:05
gunnarxmaybe a roll call is a good idea13:05
paulsherwoodpavelk: last i heard is there would be a further report at tomorrow's pmo13:05
* paulsherwood takes note to do rollcall in future13:06
paulsherwoodany other comments on confluence?13:06
steve_lits getting nearer to the top of my todo list13:06
gunnarxno, same as before.  Everyone can fill in content, but not rely on the URL.  For me those are independent things but others think differently.13:06
pavelkgunnarx: "/who #channel"13:07
paulsherwoodwell, i'm using it, won't be returning to mediawiki13:07
paulsherwoodpavelk: that doesn't necessarily mean much :)13:07
gunnarxpavelk, I'm not sure that's what we mean by attendees :)13:07
pavelkpaulsherwood: meaning of data is a tricky business13:07
paulsherwoodfor example jeremiah expressly noted he can't attend13:08
gunnarxpavelk, maybe it was a joke?13:08
paulsherwoodanyways... shall we move on?13:08
pavelkgunnarx: no, I am dead serious :-)13:08
gunnarxHe uses a bouncer I bet13:08
gunnarx+1 move on13:08
paulsherwood= Build Tools =13:08
paulsherwoodsteve_l: was there some movement on the generated code topic in bit last week?13:09
*** toscalix__ has quit IRC13:09
steve_lno substantive input on topic from baseline maintainers13:09
*** Manfred_ has joined #automotive13:10
steve_lso seems on hold until their is external input13:10
paulsherwoodany other comments wrt build tools?13:10
* paulsherwood notes, as an aside, that we decided to use jira , but doesn't think we've actually started to do so yet13:11
gunnarxpaulsherwood, is the plan to transfer the trello cards then?13:11
paulsherwoodwe could do that. i guess that's an action on me :)13:12
paulsherwoodACTION: PS transfer trello cards to Jira13:12
paulsherwoodanything on build tools? or move on?13:12
paulsherwood= UML Modeling =13:13
* paulsherwood hands over to UML experts for this....13:14
* paulsherwood notices that silences are less likely on the phonecalls13:15
klausbirkenNothing new on UML from my side...13:15
klausbirkenMaybe Bjoern?13:15
Manfred_sorry guys - first day after holiday13:15
gunnarxNot sure if we have any identified open topic in this area?13:15
gunnarxYou said you had just taken the agenda from the tools charter (which is a list of _potential_ topics I presume?)13:15
klausbirkenHowever, some news on Franca: Trademarks issues are resolved now - Franca could move on to
paulsherwoodklausbirken: w00t! :)13:15
Manfred_have some news about about Franca validation - the OSS approval for my patches are under way13:16
klausbirkenI am currently reading a lot of documents… to prepare the migration.13:16
pavelkwhich means that all links will have to move ... again?13:16
* gunnarx notices silence is broken13:16
Manfred_great new Klaus13:16
paulsherwoodgunnarx: correct. however if the charter items were not generating any discussion in general, i would wonder if we consider modifying the charter13:16
klausbirkenpavelk: Probably not all links, I try to keep the repos on github.13:16
gunnarxpaulsherwood, the charter is like a constitution.  It can be modified, but only after a long mismatch.  :)13:17
paulsherwoodgunnarx: ack13:17
klausbirkenBut we have to expect that many links will change…esp. package names etc. will have to be changed org.franca.* => org.eclipse.franca.* :-(13:17
pavelkpaulsherwood: charter is about the scope; active projects come and go13:17
*** kooltux_ has quit IRC13:17
paulsherwoodpavelk: ack, also :)13:17
gunnarxManfred_, what's in the pipe, can you share?13:17
*** kooltux_ has joined #automotive13:17
pavelkklausbirken: or dear...13:17
Manfred_Klaus: any time plan for the migration13:18
pavelkcorrection: "s/or/oh/"13:18
*** wschaller has quit IRC13:18
gunnarxManfred_: what are your patches related to?13:19
klausbirkenpavelk: But this should be doable with a full-text-search-and-replace tool…13:19
gunnarxsed :)13:19
paulsherwoodklausbirken: i hope you have a smoother process than the confluence migration :)13:19
pavelkklausbirken: well, I don't have much to find and replace yet, ;-), but I am will have soon13:19
paulsherwoodok, any more on UML, or shall we move forward?13:20
klausbirkenNothing more from my side.13:20
paulsherwood= Debugging and Analysis =13:20
Manfred_gunnar: I´ve done some corrections to the original FrancaValidator Jscripts for Enterprise Architect i can share with that approval. Those patches lead me to a completely new approach which covers new validation use cases. The new approach can go OSS after the approval got signed too.13:21
gunnarxOK, thanks13:21
paulsherwoodManfred_: cool! :)13:21
Manfred_klaus: any time plan for the migration to eclipse? this year or next year?13:21
paulsherwoodfwiw i can't remember if we've ever had any discussion about debugging and analysis so far?13:22
klausbirkenManfred_: Definitely this year. But no detailed plans yet.13:22
paulsherwoodthis seems like the lost topic of TT13:22
paulsherwoodi do note in passing that there's been some work to integrate wayland debugging into GDP13:23
paulsherwoodCTtpollard: ^?13:23
pavelkpaulsherwood: we had touched upoin this last week with Andreas Warnke mentioning the possibility for Elektrobit to share their tool13:23
gunnarxDebug & Analysis, it's basically part of the expectations on a shared dev environment / SDK13:23
Bjoern_sorry guys, I guess irc just doesn't work for me very well. But I don't have any open topics about UML atm.13:23
paulsherwoodBjoern_: fair enough. i think you are not alone. maybe in general the participants here prefer phone13:24
CTtpollardwe have been working with wayland debug and gdp yes13:24
gunnarxCTtpollard, quick summary?13:24
Manfred_well UML seems a quite graphical topic to me13:24
paulsherwoodi'd be happy to do phone and log on irc (as last week) but if i'm actually discussing in a phone call i find it difficult to type much13:24
paulsherwood(as demonstrated by last week's minutes)13:25
paulsherwoodmaybe CTtpollard can comment later13:26
paulsherwoodany more on debugging and analysis?13:26
CTtpollardgunnarx: I'm trying to provide a fix for initial app freezing issues between hmi-controller & ivi-extension13:26
CTtpollardsorry I got pulled aside then13:27
gunnarxCTtpollard. Understood, and in that process you are improving debugability / tools also?13:27
gunnarxor did I misunderstand...13:27
paulsherwoodCTtpollard: iiuc you've added recipe for wayland debug?13:28
CTtpollardgunnarx: I've managed to get wayland_debug integrated, which can be helpful13:28
paulsherwoodmaybe that would be useful for other gdp users13:28
gunnarxok, got it13:28
CTtpollardpaulsherwood: yes I have a buildable gdp system that has wayland debug tool added to it, it's not from upstream but could be useful13:29
paulsherwoodwell, you're the maintainer... so you decide :-)13:30
paulsherwoodi propose to move on13:30
CTtpollardI would add it to poky, if anything13:30
CTtpollardbut yes13:30
paulsherwood= Automated Testing =13:30
paulsherwoodgunnarx: source mirroring? iirc the previous upshot was you and/or jeremiah would look into yocto upstream approach on this?13:31
paulsherwoodin other news, i notice that AGL has its jenkins live now... https://build.automotivelinux.org13:33
*** wschaller has joined #automotive13:33
paulsherwoodand there has been some discussion about automated testing on AGL lists13:34
* steve_l wonders how they are dealing with the gfx. Maybe its private to the CI.13:34
paulsherwoodsteve_l: nothing for that yet, iiuc13:35
paulsherwoodwaltminer: will all CI be public for AGL?13:35
steve_lthat would be a solution :)13:35
gunnarxpaulsherwood, sorry got distracted13:35
*** ToxicGumbo-work has joined #automotive13:36
gunnarxYes iirc jeremiah had an action to check with upstream.  In the meantime I have made investigations on my own.  I think I'm ready to post a message to ask about the fetch bug mentioned before the meeting started.13:36
gunnarxs/upstream/The Yocto Project/13:37
Bjoern_soryr guys, have to leave now.13:37
*** Bjoern_ has quit IRC13:37
paulsherwoodBjoern_: bye for now :)13:37
paulsherwoodgunnarx: ok, so no recommendations yet?13:37
gunnarxAny experience from AGL setup so far?13:38
gunnarxIs it Jenkins only or also Gerrit?13:38
paulsherwoodJenkins + Gerrit13:39
gunnarxI'm curious primarily on Gerrit experiences.  Some are/were negative to Gerrit...13:39
gunnarxin terms of usability.13:39
steve_lgunnar: I've learnt to be somewhat careful about the wiki. Some it is crusty, so I think an email would be the right approach13:39
paulsherwoodit's some friction to get started (ie as a new user, need to get permissions etc, and ui is complex)13:40
gunnarxsteve_l, don't get the context of your response.  Gerrit?  An email for what, patches?13:40
paulsherwoodbut once the initial pain is over it works13:40
gunnarxwhat wiki?13:40
paulsherwoodgunnarx: i think he means yocto13:40
* gunnarx is completely confused now13:41
*** waltminer has quit IRC13:41
paulsherwoodeg jenkins has built and given a +113:41
gunnarxah, ok you were referring far back.  I get it now.13:41
paulsherwoodgunnarx: he's suggesting how to approach yocto upstream13:41
* steve_l withdraws until current sub-topic (Gerrit) is finished13:42
gunnarxGerrit UI really is not that nice...13:42
gunnarxBut it sure beats having nothing13:42
paulsherwoodgunnarx: there are commmand line tools which allow folks to ignore the ui apparently13:43
gunnarxpaulsherwood, have you continued to promote combined AGL/GENIVI infra in this area?13:43
paulsherwoodfor better or worse i think gerrit is the only game in town for code review in this context13:43
gunnarxno that's not right13:43
paulsherwoodok... how else would you propose to integrate code review into ci, gunnarx ? (foss...)/13:44
gunnarxis gerrit superior to all alternatives, or you think there are no alternatives?13:44
paulsherwoodi think there are no credible alternatives, but i can always be wrong13:44
*** riazantsev has joined #automotive13:44
paulsherwood(point being, gerrit handles creation of candidates to push through ci, so you can do pre-merge testing)13:46
gunnarxphabricator, gitlab, atlassian-whatever-it's-called,13:46
gunnarxok atlassian maybe not foss13:46
gunnarxI thought we have had this discussion already.  Let's structure this.  Are we trying to reach TT consensus on what are good alternatives?13:46
gunnarxpaulsherwood, yes, that workflow support is good from gerrit13:46
paulsherwoodiiuc the other alternatives are post-merge13:46
gunnarxok, I have not checked that13:46
KlausUhlWhat would be the alternatives to Gerrit?13:47
KlausUhlSorry, I have missed the line above.13:47
paulsherwoodgunnarx: 'Are we trying to reach TT consensus on what are good alternatives?' - i think so, but i don't know how to achieve it13:48
gunnarxEither timebox the discussion and we have it now, here on this meeting13:48
gunnarxor assign action to analyze and come back with a report13:48
paulsherwoodwhich is preferred?13:48
gunnarxyou're leading the meeting13:49
paulsherwoodyes, i suppose i am13:49
paulsherwoodi'll write to the list with my preliminary thoughts, seek feedback on that13:50
gunnarxgood start13:50
paulsherwoodACTION: PS write on code review options for CI13:50
paulsherwood14:43 < gunnarx> paulsherwood, have you continued to promote combined AGL/GENIVI infra in this area?13:50
paulsherwoodwrt the above, yes but i don't think i'm succeeding (in either forum)13:51
gunnarxyes :) that sounds like an important consideration13:51
gunnarxI thought I saw Walt online?13:52
gunnarxdon't see him now...13:52
paulsherwoodyes he was13:52
paulsherwoodanyway time is passing...13:52
paulsherwood= AOB =13:52
klausbirkenNo AOB from here…13:53
paulsherwoodfrom my pov, ELCE is approaching... are there any engineer attendees who would be willing to join the automotive collaboration panel?13:53
gunnarxI'm likely not attending I'm afraid... :(13:53
*** kooltux_ has quit IRC13:54
gunnarxThe LF fees are too high.  There I said it.13:54
*** kooltux_ has joined #automotive13:54
pavelkpaulsherwood: I plan to go to elce.  What is the subject for the panel?13:54
*** jlrmagnus has joined #automotive13:54
paulsherwoodview from the trenches on software collaboration in automotive13:54
paulsherwoodgunnarx: i think genivi fees are just as high? :)13:55
paulsherwoodACTION: PS request panel members on list13:55
pavelkpaulsherwood: doom and gloom; does not work that well if you ask me :-(13:55
*** Luming has quit IRC13:55
paulsherwoodpavelk: i'm reaching the same conclusion lately, tbh. are you interested to take that discussion at the conference?13:56
gunnarxmaybe, I haven't checked the GENIVI fees.  Me+colleagues attendance is included in membership fee.13:56
*** Luming has joined #automotive13:56
paulsherwoodgunnarx: ack13:56
* paulsherwood has a plane to catch... ao aob?13:56
pavelkpaulsherwood: don't know yet, let's discuss offline13:56
gunnarxnot from me13:56
paulsherwoodpavelk: ok, will do13:56
pavelknothing from me13:56
paulsherwoodok, thanks all13:57
paulsherwood== GENIVI Tools Team Meeting Ends ==13:57
klausbirkenThanks, Paul. Bye, all.13:58
steve_lgunnar: I'm not sure the yocto wiki is referring to a fetch bug. They are just flagging behaviour..14:00
*** waltminer has joined #automotive14:01
gunnarxsteve_l, I agree, they are flagging behavior and saying "don't do this".  As I said above, that's not really an authoritative instruction.14:01
gunnarxwaltminer, we missed you during the meeting :)14:02
gunnarxsteve_l:  Then, I claim that it's a bug.14:02
steve_lwell dont do this if you want to be untethered I guess :)14:02
*** klausbirken has left #automotive14:03
gunnarxsteve_l: And I have made some speculation about what should be considered best practice.  One interesting piece of evidence is that poky devs are using SRCREV= on everything14:03
steve_lbut yes, just providing a tag along does lead to a behaviour that can be painful in certain corporate environments14:03
steve_lagree SHA is preferred14:03
gunnarxsteve_l, what are those problems?  Are you saying..14:03
gunnarxoh ok, you're saying SHA is preferred14:03
steve_la tag *alone*14:03
steve_lwith just a tag bitbake has to go out to  the  git rep to check that it has not moved. That's the behaviour. A feature of git.14:04
gunnarxto me a tag=$PV looks like a useful practice.  It simplifies upgrades by just renaming the recipe.14:04
gunnarxsteve_l, yes, I have speculated that that is the intention, to check that it hasn't moved14:05
*** wschaller has quit IRC14:05
steve_lAgreed on tag. It's much more readable, but has the associated side effect14:05
gunnarxis that reasonable?   The local mirror also has an opinion about what the tag points to, is that a less correct answer?14:05
steve_lwhat is  reasonable? tag=$PV?14:06
*** riazantsev has quit IRC14:06
gunnarxis it reasonable to expect tags to move I mean14:06
steve_lIt's bad practise in an upstream but can occur, hence for correct dependency checking bitbake must check it no?14:07
gunnarxwhat's to say that you prefer to follow upstream if it changes :)14:07
waltminergunnarx I missed you too, but I can’t follow this even back scrolling14:08
steve_lthen defo right the receipe for that and accept that it will go out to check14:08
gunnarxthat makes things quite unpredictable...  will bitbake warn about it moving or just go on?14:08
waltminerthen I wander off to get coffee14:08
steve_lI believe go on (should be straight forward to check) else your approach for following upstram would quickly become painful, but the rebuild would show in the log and task list14:10
gunnarxwaltminer, no problem, we were discussing tooling and then I wondered if shared infrastructure was still on the agenda, paulsherwood was not sure what the organizations want14:10
gunnarxwaltminer, gerrit and jenkins up and running at AGL - a good start14:10
gunnarxsteve_l, anyhow I think asking yocto devs opinions is the right thing now - it will clarify what to expect in this area14:10
waltminerShared infrastructure in terms of using the same tools would be great14:11
steve_lagreed they are nice helpful peeps. I would ask them generally in terms of "we are an oss who wants to do a mirror" advice anyway14:11
gunnarxsteve_l, will do14:12
*** gunnarx has quit IRC14:19
*** riazantsev has joined #automotive14:19
*** wschaller has joined #automotive14:20
*** brlogger has joined #automotive14:37
*** fulup__ has joined #automotive14:38
*** steve_l has quit IRC14:43
*** fredcadete has joined #automotive14:57
*** KlausUhl has quit IRC15:10
*** jimmy1 has joined #automotive15:11
CTtpollardare you about jeremiah?15:24
*** waltminer has quit IRC15:47
*** ulf` has joined #automotive15:49
*** jlrmagnus has quit IRC15:57
*** waltminer has joined #automotive16:18
ulf`Gah just missed him16:24
*** jlrmagnus has joined #automotive16:35
*** waltminer has left #automotive16:37
*** waltminer has joined #automotive16:39
kooltux_ulf`, hey what's up ??? nice to see you here :)16:41
*** Manfred_ has quit IRC16:44
*** Tarnyko has quit IRC17:03
*** wschaller has quit IRC17:24
*** brlogger has joined #automotive17:48 [freenode-info] channel flooding and no channel staff around to help? Please check with freenode support:
*** jonathanmaw has quit IRC18:23
ulf`kooltux_: jet lagged :)18:23
ulf`kooltux_: How's it going?18:23
*** waltminer has joined #automotive19:28
*** erbo has quit IRC20:19
*** erbo has joined #automotive20:19
*** jlrmagnus has quit IRC20:23
*** jlrmagnus has joined #automotive20:36
*** radiofree has quit IRC20:41
*** e8johan has quit IRC20:42
*** e8johan has joined #automotive20:43
*** radiofree has joined #automotive20:43
*** ToxicGumbo-work has quit IRC21:35
*** fulup__ has quit IRC22:38
*** waltminer has quit IRC23:18

Generated by 2.14.0 by Marius Gedminas - find it at!