IRC logs for #automotive for Tuesday, 2015-08-18

*** jlrmagnus has left #automotive00:12
*** waltminer has quit IRC02:03
*** Luming has quit IRC03:57
*** brlogger has joined #automotive05:08
*** FaH_auto has joined #automotive05:34
*** Luming has joined #automotive06:31
*** marcomentor has joined #automotive06:48
marcomentor /me waves hello06:49
marcomentorping marcomentor06:50
marcomentorwhois marcomentor06:51
*** marcomentor has quit IRC06:52
*** Egy has joined #automotive07:02
*** txenoo has joined #automotive07:03
*** FaH_auto has quit IRC07:06
*** KlausUhl has joined #automotive07:06
*** txenoo has quit IRC07:09
*** FaH_auto has joined #automotive07:11
*** fredcadete has joined #automotive07:21
*** txenoo has joined #automotive07:40
*** jonathanmaw has joined #automotive07:55
*** mdunford has joined #automotive07:59
*** jacobo has joined #automotive08:25
*** FaH_auto has quit IRC08:32
*** txenoo has joined #automotive08:34
*** Egy has quit IRC08:35
*** Egy has joined #automotive08:38
*** txenoo has quit IRC08:40
*** FaH_auto has joined #automotive08:40
jeremiahrjek: You're warming a rodent?08:43
jeremiahAh, now I read the backlog and it is perfectly logical.08:44
jeremiahmarcomentor: Is that you Gianpaolo?08:44
rjekjeremiah: :)08:46
*** jeremiah has quit IRC08:49
*** jeremiah has joined #automotive08:50
jeremiah/c/c08:55
jeremiahwee!08:55
* rjek looks at jeremiah.08:56
paulsherwoodwhoever it was, he/she is no longer here?08:58
rjekinetnum:        62.154.245.64 - 62.154.245.7909:00
rjeknetname:        XS-EMBEDDED-VILLININGEN-SCHWENNINGEN-NET209:00
rjekdescr:          XS Embedded GmbH09:00
CTtpollardughh someone posted re the common api error in meta-ivi for gdp on the mailing list09:01
jeremiahOh! I bet that was Marco Residori then. :-)09:01
jeremiahI.e. the Villiningen IP09:01
jeremiahCTtpollard: Why ughh?09:02
CTtpollardjeremiah: because we still don't have a solution09:02
jeremiahAh!09:03
*** txenoo has joined #automotive09:03
jeremiahThis is a bit of a problem with meta-ivi, the lack of maintenance.09:03
jeremiahThere are some patches that are rotting on the mailing list as well.09:03
jeremiahI plan to bring this up at today's PMO meeting.09:03
jeremiahCTtpollard: Do we have a bug number in bugzilla? I want to add it to my email.09:08
* jeremiah goes to query the 'zilla.09:08
CTtpollardjeremiah: my post on the mailing list has everything you should need09:08
jeremiahCTtpollard: Cool, thanks.09:09
CTtpollardjeremiah: http://lists.genivi.org/pipermail/genivi-projects/2015-August/000688.html09:09
jeremiahThanks, I just read that.09:11
jeremiahWhat about us creating a mirror/branch of the Yocto meta-ivi repo in GENIVI's git repos and patch it there?09:12
jeremiahIs that a lot of work then to build meta-ivi?09:12
jeremiahThis way we don't have to wait for the meta-ivi maintainer to come back.09:12
paulsherwoodjeremiah: i've proposed that approach at AGL... problem is it opens up the no-forking debate09:12
CTtpollardmirroring is an option, but there's been a lot of talk on how we don't want to fork upstream09:12
jeremiahBut creating local branch is not forking09:13
paulsherwoodjeremiah: i agree09:14
paulsherwoodexcept that in practice, it leads to it09:14
paulsherwood(given how folks use bb in practice... iiuc)09:14
paulsherwoodcase in point...09:14
paulsherwoodstephen does work on a local branch at https://github.com/slawr/meta-renesas09:15
paulsherwoodthen in practice, folks start building on that09:15
paulsherwoodresult he is an upstream09:15
jeremiahOkay.09:16
paulsherwooddistinct from the two other upstreams of meta-renesas09:16
paulsherwoodnobody intends to fork, but reality conspires against us09:16
jeremiahBut the fork occurs not when the git repo moves around, since git is designed to distribute09:16
jeremiahThe fork occurs when you move support somewhere else.09:16
jeremiahIf we intend to send patches up to master, then we're not forking09:17
paulsherwoodack09:17
jeremiahBut I agree with your careful reading09:18
jeremiahof the situation09:18
paulsherwoodin a way the flexibility of bb leads to implicit forking... and we only realize later09:18
jeremiahI'll bring it up anyway and let's see if we can get some movement from the meta-ivi maintainers because it shouldn't turn into a blocker.09:18
paulsherwoodjeremiah: to be clear i'm in favour of creating a meta-ivi mirror and doing work there09:20
paulsherwoodif you agree, and we do it nowish, it would help us make some progress i think09:20
jeremiahI agree, I think we need to bring it up in PMO today for transparency's sake, but I also agree it has to be done nowish.09:22
jeremiahCan it wait until later today or tomorrow?09:23
jeremiahI can create the repo so there will be no wait on that, but I don't want to NMU (if you no what I mean) because that might feel like hijacking.09:24
*** wschaller_ has joined #automotive09:26
jeremiahThis only affects the x86_64/QEMU version of GDP, right?09:36
CTtpollardnope all of them09:36
CTtpollardthey all use meta-ivi09:36
jeremiahEew.09:37
jeremiahOkay09:37
jeremiahSent the mail, hopefully we get to it in today's PMO meeting.09:53
CTtpollardthanks jeremiah09:55
*** FaH_auto has quit IRC09:56
*** CTtpollard has quit IRC10:01
*** CTtpollard has joined #automotive10:04
*** FaH_auto has joined #automotive10:04
paulsherwood+110:05
*** txenoo has quit IRC10:16
*** KlausUhl has quit IRC10:23
*** KlausUhl has joined #automotive10:33
*** FaH_auto has quit IRC10:33
*** txenoo has joined #automotive10:39
*** txenoo has quit IRC10:44
*** txenoo has joined #automotive10:45
*** txenoo has quit IRC10:47
*** txenoo has joined #automotive10:51
*** wschaller_ has quit IRC11:00
*** txenoo has quit IRC11:03
*** txenoo has joined #automotive11:38
*** txenoo has left #automotive11:41
jeremiahIts on the agenda for today.12:11
*** FaH_auto has joined #automotive12:13
jeremiahI've also explicitly mentioned the need for fast action on the patches and the request to create a branch for patching which will then get pushed back upstream to master when we have a maintainer.12:17
paulsherwoodcool!12:22
*** wschaller_ has joined #automotive12:24
jeremiahEww. meta-ivi skipping certificate checks when downloading? Why? They were put there for a reason.12:27
jeremiah wget -t 2 -T 30 -nv --passive-ftp --no-check-certificate <-- from a yocto log12:27
paulsherwoodjeremiah: i do wonder why this stuff is in 'recipes' at all...12:28
fredcadetewhich package is that?12:29
paulsherwoodif i want to build something reliably, my build instructions should not include '... go to the internet and get a load of stuff'12:29
fredcadetepaulsherwood: what's your opinion on maven?12:31
paulsherwoodfredcadete: i confess i have not used it, can barely spell it12:31
fredcadetepaulsherwood: you are in luck12:32
paulsherwoodheh12:38
paulsherwoodis it java?12:38
fredcadeteyes12:44
paulsherwood:)12:44
KlausUhlfredcadete: do you have problems with maven?12:45
fredcadetenot specifically12:46
fredcadeteit's just that in general maven buildsystems download the whole internet to compile one tool12:47
KlausUhl:-)12:47
fredcadetelike yocto on steroids12:47
KlausUhlThat's a nice description of what Maven does.12:47
KlausUhlBut it caches what it has downloaded locally.12:48
KlausUhlAnd you can force it to work offline.12:48
KlausUhlIf you want to go crazy (or are just fed up with a slow internet connection) you can also install a local mirror to do the caching for you.12:49
fredcadeteyeah, I guess it does what it does well enough12:50
fredcadeteand it's probably very handy as a developer to have all those libraries12:51
KlausUhlIndeed. But it is a hell to configure.12:51
fredcadetebut as a user when I have to compile such a thing I always shed a few tears12:51
fredcadeteI guess it's the same feeling people have with yocto the first time they run a build12:52
KlausUhlProbably. But on the other hand you do not need to reiterate trying to compile something just to realize that some (devel) package is not installed on your system.12:52
fredcadeteas with autoconf, you mean?12:53
fredcadeteyeah12:53
KlausUhlyes12:53
KlausUhlor any other C/C++12:53
KlausUhlbuild system12:53
fredcadeteyes12:54
*** Egy has quit IRC12:54
fredcadetebut with the C/C++ ones I have stockholm syndrome12:54
fredcadetewhereas with maven it's too foreign12:54
KlausUhlOh, there is a simple cure for that:12:55
KlausUhljust develop some medium sized Java project for a few months :-)12:56
KlausUhlif you also throw in some Eclipse plug-in development you will think that hell is a comfortable place to stay12:57
*** Egy has joined #automotive13:09
*** waltminer has joined #automotive13:14
*** Egy has quit IRC13:16
waltminergood morning13:19
waltminerJust read through the meta-ivi convo and I +1 the meta-ivi mirror idea13:19
*** persia_ is now known as persia13:19
CTtpollardinstead of using patching bb.append in gdp?13:21
paulsherwoodwaltminer: would agl folks be interested in re-using the same mirror (if patches are required for agl), to avoid further proliferation?13:24
jeremiahfredcadete: I think it was in CommonAPI13:24
jeremiahlol, Maven?13:25
jeremiahA cesspool13:25
paulsherwoodjeremiah: let's keep it polite, here :)13:25
jeremiahOh, my bad.13:25
jeremiahI though we were discussing plumbing . . .13:25
paulsherwoodlol13:25
paulsherwoodsoftware is hard, and most of it is somebody's baby... never pays to call someone's baby ugly :)13:26
jeremiahSo true.13:26
* paulsherwood has learned this the hard way :)13:26
jeremiahThen again, Maven has been known to be somewhat liberal with licenses.13:26
jeremiahPerhaps too liberal in some cases.13:26
waltminerI don’t see why not since meta-ivi is GENIVI’s. The patches AGL would do we be to propose moving things out of meta-ivi and into meta-ivi-common. This would seem to be the “polite” place to make such a proposal13:27
paulsherwood+113:27
paulsherwoodwith the clear understanding that the aim is to work with upstream13:28
jeremiahAs long as there is a maintainer or two I don't see any issue.13:28
waltmineralways13:28
paulsherwoodwaltminer: what is happening about further meetings?13:31
*** wschaller has joined #automotive13:31
paulsherwood(agl meetings... are they to be at a different time now?)13:32
*** wschaller_ has quit IRC13:33
fredcadeteregarding the meta-ivi woes, please don't rush a fork. It may be necessary in the end but if it's just for integration issues that's what bbappends are for. I say this as a user: forks are confusing13:36
paulsherwoodfredcadete: i don't think anyone wants to fork... we're just suggesting a mirror to allow progress preparing patches for upstream (iiuc)13:41
fredcadetea mirror has the exact same contents as the source13:42
fredcadetethe moment you put a patch on top it's called a fork :)13:42
fredcadetebut yes, I get where the need is coming from13:42
fredcadetebut let's remember that last week people were surprised that gunnar's gdp-submodules were using a fork of meta-something (meta-ivi or meta-renesas, I don't remember)13:43
paulsherwoodfredcadete: we discussed this somewhat already... and i fear that bbappends achieve exactly the same effect, but hide it :)13:44
CTtpollardivi13:44
fredcadetejust the same way downstream users will be surprised that one working group of Genivi is forking another working group's repository13:44
paulsherwoodbut this is a complex topic13:44
fredcadete+1 on the complexity13:45
paulsherwoodnot on the proposed mirror, i guess :)13:45
fredcadeteno :) but that's just my opinion. And my opinion is worth much less than actual work done13:46
*** yoshi_ito has joined #automotive14:02
paulsherwoodPMO seems to be deciding not to make a mirror for the moment, so CTtpollard should go ahead and figure a bbappends stopgap14:02
*** FaH_auto has quit IRC14:07
CTtpollardpaulsherwood: ack14:09
waltminerwhen does a patch become a “fork”?  Let’s assume that some package will need to be patched by GENIVI or AGL. They need to keep that patch somewhere until it is accepted upstream. Since that process can take time where would the patch or patches live if there is no common place to put them ?14:09
CTtpollardpaulsherwood: a commit on master, or a branch with stopgap fixes on?14:11
waltminerYou need to balance that with the pain of keeping the patches in sync with upstream. I don’t think it is a fork until you recognize that you cannot get your patch into the mainline for the project and it is too difficult keep the patch current with the mainline.14:11
waltminerand you absolutely cannot live without the functionality of the patch14:11
paulsherwoodwaltminer: that's my point exactly. hence i'm strongly in favour of having a mirror with our proposed patches14:15
* paulsherwood seems unable to convince others of this, though (and the disagreement appears to be from both engineers and non-engineers), so may still be wrong14:16
waltminerI am ok with a non-mirror solution if someone could put it on paper, but I would like to understand the thought process and how to account for all of the use cases.14:18
paulsherwoodCTtpollard: which would you prefer to do? both seem less than ideal to me :)14:18
waltminerthen again. I am not GENIVI14:18
waltminer:)14:18
paulsherwoodwaltminer: you are not. but we have similar situations in AGL i think14:18
CTtpollardpaulsherwood: master for less friction14:18
paulsherwoodCTtpollard: go for it. annotate the commit clearly to explain why. and preferably post it to the ML so folks can +1 it14:19
paulsherwood(or -1, if they are so inclined :-)14:19
waltminerwe do. AFAIK we will use a mirrors.14:20
paulsherwoodeven to the extent of mirroring genivi software if that were required to expedite progress? (sorry to be provocative, but if the answer is yes, i may notify the PMO that this is the case)14:23
waltminerand some smoke and a warmed up mouse14:23
waltmineryes14:24
*** Dthiriez has quit IRC14:26
*** jlrmagnus has joined #automotive14:55
*** ctgriffiths_ has joined #automotive14:55
jlrmagnusMorning14:55
paulsherwoodhowdie14:56
*** KlausUhl has quit IRC14:58
*** ctgriffiths has quit IRC14:59
CTtpollardfredcadete: I'm trying to create the append file for a patch but I'm struggling to find good documentation15:46
paulsherwoodCTtpollard: i assume http://www.yoctoproject.org/docs/1.6/dev-manual/dev-manual.html#using-bbappend-files is not enough?15:48
fredcadetethe yocto manual is fairly good15:48
fredcadeteany specific problem CTtpollard ?15:49
CTtpollardI can't find how to use the append to remove variables15:50
fredcadetethe language itself is documented in http://www.yoctoproject.org/docs/1.6/bitbake-user-manual/bitbake-user-manual.html15:50
CTtpollardto remove the PR variable as the patch does http://git.yoctoproject.org/cgit/cgit.cgi/meta-ivi/commit/?id=e56917139ed15a8cba2a78ba7e31eeffd439c75c15:50
fredcadetethis is the closest thing I found: http://www.yoctoproject.org/docs/1.6/bitbake-user-manual/bitbake-user-manual.html#removing-override-style-syntax but I think it would leave you with an empty variable. For most uses it should be the same thing15:52
*** waltminer has quit IRC15:52
fredcadeteyou could also do PR="", it'll lead to the same thing15:55
*** waltminer has joined #automotive15:55
fredcadeteI'll have to go now15:57
fredcadetesee you tomorrow15:57
*** fredcadete has quit IRC15:58
CTtpollardahh he's gone, will say thanks tomorrow16:04
jonathanmawI've learned an important lesson while trying to crowbar AudioManager into meta-agl, DISTRO_FEATURES specifies whether you use sysvinit or systemd, so a meta-agl system which uses systemd ought to be using the distro poky-ivi-systemd (from meta-ivi)16:10
*** yoshi_ito has left #automotive16:27
paulsherwoodCTtpollard: I would add +1 on a .bbappend inside GDP as a quick stopgap.  It can16:27
paulsherwood(will) disappear when GDP is upgraded to a later baseline version anyhow.16:27
paulsherwoodwould it make sense to do that upgrade and be done?16:28
paulsherwood(sorry, unclear... i'm quoting gunnar's email)16:28
CTtpollardupgrade to a 8.0 baseline?16:28
CTtpollardI'm currently building with an append in place, I will have the results from it tomorrow16:29
paulsherwoodCTtpollard: yes, i was asking about upgrading to current baseling. seems pretty poor form not to16:30
*** mdunford has quit IRC16:30
CTtpollardupgrading to 8.0 bring in the fix, it is just then making sure it builds and works with the other upgrades16:30
paulsherwoodperhaps. unless there is other mess16:30
paulsherwoodbut i would try it, at least.16:31
CTtpollard6.1 - 6.2 audiomanager for example16:31
*** waltminer has quit IRC16:34
CTtpollardhow much of that is just a drop in place upgrade it unknown to me at this point16:36
*** Dthiriez has joined #automotive16:38
*** bharmon has joined #automotive16:41
*** jonathanmaw has quit IRC16:44
*** mdunford has joined #automotive16:45
*** waltminer has joined #automotive16:46
*** mdunford has quit IRC17:09
*** RzR is now known as rZr17:18
*** waltminer has quit IRC17:24
*** wschaller has quit IRC17:30
*** waltminer has joined #automotive18:34
*** jacobo has quit IRC18:46
*** steve__ has joined #automotive18:49
*** steve__ has quit IRC18:51
*** fredcadete has joined #automotive18:53
*** fredcadete has quit IRC18:59
*** rZr is now known as RzR19:37
*** waltminer has quit IRC20:05
*** waltminer has joined #automotive20:09
*** waltminer has quit IRC20:28
*** waltminer has joined #automotive21:26
*** jlrmagnus has quit IRC23:54
*** bharmon has quit IRC23:55

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!