IRC logs for #automotive for Wednesday, 2016-08-03

*** tjamison has quit IRC00:06
*** jlrmagnus has quit IRC00:27
*** waltminer has quit IRC01:07
*** AlisonChaiken has quit IRC01:33
*** fury has quit IRC03:25
*** fury has joined #automotive03:28
*** chbae has quit IRC04:03
*** AlisonChaiken has joined #automotive05:16
*** Saint_Is_ has quit IRC06:46
*** rajm has joined #automotive07:06
*** leon has joined #automotive07:10
*** leon is now known as Guest9864307:11
*** Guest98643 is now known as leon-anavi07:14
*** jobol has joined #automotive07:21
*** jobol has quit IRC07:24
*** jobol has joined #automotive07:34
*** jobol has quit IRC07:36
*** toscalix has joined #automotive07:38
*** rajm has quit IRC07:56
*** rajm has joined #automotive07:59
*** jonathanmaw has joined #automotive08:18
*** rajm has quit IRC08:29
*** rajm has joined #automotive08:33
toscalixin 90 min aprox. we release GDP 11 beta for qemu09:28
leon-anavicool :)09:29
dl9pfleon-anavi: ping09:46
*** jjardon has quit IRC09:47
dl9pflibgsystem and ostree_git use AUTOREV. Can you name me git rev's that should be used for blowfish (so we can pin it) ?09:48
*** jjardon has joined #automotive09:50
CTtpollardautorev can be dangerous09:52
leon-anavihi dl9pf09:55
dl9pfhi leon09:56
leon-anaviok, I have to look at that and find the proper revision09:56
leon-anaviIt will take me some time to do proper testing. I will do it right now.09:56
dl9pfright now we're using git head, we might just take git head of today09:56
leon-anaviI will have a look at the git logs and I will pick either the latest version (head of today) or another stable version.09:57
dl9pfok - that is even better09:57
leon-anavidl9pf, should I submit a patch with git rev for both master and blowfish branch?09:57
dl9pfblowfish in any case. master is your choice. A good example is:09:59
dl9pfSRCREV = "ad245bdd60434dd46d6461f585d49db1b3b0d75b"09:59
dl9pf#SRCREV = "${AUTOREV}"09:59
dl9pfmaster could have "${AUTOREV}"09:59
dl9pfthe release gets the fixeed rev09:59
CTtpollardI'd vote against that, even in master10:00
dl9pfup to you how "moving" you want to have master10:00
CTtpollardunless you're 100% certain that said upstream will never break, and break within your build system10:00
dl9pfcase-by-case decision - true10:00
dl9pfwe do it in master if we're pointing to "our own" git repos which makes sense in a way10:01
dl9pffor external upstreams the fixed SRCREV likely the way to got.10:02
CTtpollardagreed, I hope to eventually have that model10:02
leon-anaviCTtpollard, yes, you are right that there is a risk in using AUTOREV10:02
CTtpollardbut update to master of even your own repos can count on other recipes updating in parallel which is hard to achieve especially in CI10:03
dl9pfwe have a few in master - plain convenience/ease-of-use (aka laziness ;) ) for developers10:03
dl9pfCTtpollard: agree that might bite10:03
dl9pfhmm, true, interesting ... got an idea ...10:05
dl9pfSRCREV = "12345"10:06
dl9pf# override-enabled autorev10:06
dl9pfSRCREV_uselatestrev = "${AUTOREV}"10:06
dl9pfgotta try if that works, then we can flip that switch and still hardcode a rev10:06
*** bshah has quit IRC10:16
*** rjek has quit IRC10:16
*** rjek has joined #automotive10:17
leon-anavidl9pf, is there a JIRA issue for libgsystem? Should I create one?10:20
dl9pfno, it falls under  last comment.10:20
dl9pfreference SPEC-20210:20
*** bshah has joined #automotive10:28
CTtpollardhas anyone ever had the problem with bitbake seemingly racing when applying a patch, leading to it rejecting the patch as the source has already been changed?10:43
dl9pffor the same patch ?10:43
dl9pfhuch ?10:43
CTtpollardseems to be10:44
CTtpollardone minute10:44
dl9pfhmm ... did the patch end up twice in the src_uri maybe ?10:44
CTtpollardno previous patch has touched the file in question, and the patch applies cleanly to the given ref10:45
CTtpollardI can't reproduce it locally, but have had 2 users report it10:46
CTtpollardand apparently the error can still occur after clean unpack etc10:50
dl9pfhmm ... bitbake -e | grep SRC_URI ...10:53
CTtpollardany ideas dl9pf leon-anavi ?10:53
dl9pfbitbake -e <recipe>  > foo.txt  and paste that one10:54
dl9pfit applies it twice for whatever reason ...10:54
CTtpollardI've grepped through -e before and not seen it claiming two do_patch tasks for it10:54
CTtpollardlet me check again10:54
toscalixGDP 11 beta is out
toscalixups, wrong link10:55
leon-anavihm... CTtpollard if I were you I would use devshell to investigate and see what is really going on.10:55
leon-anaviafter checking bitbake -e :)10:55
dl9pfhmm ... will devshell come up if patching fails ?10:55
leon-anavidl9pf, I think so10:56
leon-anaviI have use it in the past to adjust patches while I am working on the recipes.10:56
toscalixor this shorter link:
CTtpollardthat's the offending recipe btw
dl9pfleon-anavi: probably drops you right at the failing patch ...10:57
leon-anaviyes, I think so but I am not 100% sure10:58
leon-anavianyway in my opinion devshell is really convenient for browsing the code and fixing issues with failing patches10:59
leon-anavimost of the time they are failing due to wrong patch but CTtpollard's case sees different.10:59
dl9pfCTtpollard: what about using the default "quilt" for patchtool ?10:59
CTtpollardyeh, I can't reproduce the failure which makes it harder to debug11:00
dl9pfhmm ... aah ...11:00
CTtpollarddl9pf: I could propose that to a user with the issue to see if it works11:00
dl9pfiirc there is no "git-native" - we rely on the system's git11:01
dl9pfbut there is a "quilt-native" so you're sure it's built with yocto11:01
CTtpollarddiffering git versions has been a problem I've seen11:01
CTtpollardthanks for the help with that one11:02
CTtpollardI presume not setting patchtool will default to quilt?11:08
CTtpollard( it does )11:19
*** Jakubin has joined #automotive11:55
JakubinCTtpollard: Hello, do you happen to know whom should i ask about Navit issues on GDP? I'm trying to run it on Raspi3 using GDP10 but there are errors :(11:57
leon-anavidl9pf, I submitted patches to both branches and I updated the JIRA issue.12:01
CTtpollardJakubin: one of those people is me :) could you possibly log the errors in the GDP JIRA?12:01
leon-anavidl9pf, I have assigned it again to you. Please let me know if anything else needs modifications.12:01
toscalixJakubin: GDP JIRA:
*** waltminer has joined #automotive12:11
*** fury has quit IRC12:19
JakubinCTtpollard: Ok i created the issue
CTtpollardthanks Jakubin12:33
CTtpollardI'll move it to the GDP jira12:34
*** Net147 has quit IRC12:35
*** Net147 has joined #automotive12:37
dl9pfleon-anavi: we should be fine now. tnx.13:31
leon-anavigreat, you are welcome :)13:32
*** nisha has quit IRC13:51
*** Jakubin has quit IRC14:00
*** Jakubin has joined #automotive14:01
*** rajm has quit IRC14:01
*** Jakubin has quit IRC14:11
*** tjamison has joined #automotive14:11
*** rajm has joined #automotive14:22
*** Jakubin has joined #automotive14:38
*** nisha has joined #automotive14:41
*** jlrmagnus has joined #automotive15:17
*** chbae has joined #automotive15:32
*** chbae has quit IRC15:59
*** Nicholas_ has joined #automotive15:59
*** mvick has quit IRC15:59
*** gunnarx has joined #automotive16:00
*** waltminer has quit IRC16:03
*** jlrmagnus has quit IRC16:07
*** leon-anavi has quit IRC16:07
*** rajm has quit IRC16:10
*** AlisonChaiken has quit IRC16:12
gunnarxCTtpollard, you there?16:16
pedroalvarezgunnarx: nope16:19
*** jlrmagnus has joined #automotive16:22
*** praneeth has quit IRC16:22
*** jonathanmaw has quit IRC16:23
gunnarxpedroalvarez, CTtpollard.  I'm finding that a bunch of the GDP pipelines are failing on navigation service.16:25
gunnarxIsn't it the old race condition that was solved by reducing the parallelism in the builds?16:26
gunnarx(which we agreed to try to remove, but maybe it was a bad idea)16:26
pedroalvarezoh, I remember CTtpollard wanting to change that to "solve" an issue, but I don't remember what component16:27
paulsherwoodhow much ram do these runners have?16:30
paulsherwoodgunnarx: maybe paste a log?16:30
gunnarxThe logs are on the go server16:35
*** AlisonChaiken has joined #automotive16:39
pedroalvarezthis would be it
gunnarxyes that's it16:39
pedroalvarezrace condition when building that plugin16:43
pedroalvarezit seems16:43
pedroalvarezreal solution would be to fix that race condition  :)16:43
paulsherwoodpedroalvarez: how are you deducing race condition from that info?16:44
gunnarxwizard magic16:44
paulsherwoodwe've seen intermittent fails due to races, but also due to running out of memory. hence my question earlier.16:45
paulsherwoodreducing parallelism ironically may fix both16:46
gunnarxI'll try to put back the parallelism limit and see if it goes away16:49
*** praneeth has joined #automotive16:51
*** toscalix has quit IRC16:51
gunnarxsorry, no battery left and power adapter has gone AWOL16:52
pedroalvarezpaulsherwood: when running out of memory, gcc throws an "Internal compiler error: Killed program..."16:56
paulsherwoodah, ok16:57
pedroalvarezthis error is more like "hey things I need to build are not there"16:57
paulsherwoodack, so race condition16:57
paulsherwoodor broken recipe :)16:57
pedroalvarezyep, given that sometimes works16:57
*** gunnarx has quit IRC16:57
* pedroalvarez nods16:58
pedroalvarezmy guess is a bug in the navigation.git CMake instructions17:01
pedroalvarezAuthor: Philippe Colliot17:02
pedroalvarezmaybe he can help with this if we don't find people with cmake foo17:02
*** jlrmagnus has quit IRC17:13
pedroalvarezI don't understand how navigation-service sometimes builds17:19
paulsherwoodto be clear, this is not navit?17:22
*** jlrmagnus has joined #automotive17:26
*** mvick has joined #automotive17:34
*** nuohan has quit IRC18:28
*** nuohan has joined #automotive18:29
*** robin__ has quit IRC18:39
*** fury has joined #automotive19:05
*** jlrmagnus has quit IRC19:05
*** Ford_Prefect has quit IRC19:20
*** Ford_Prefect has joined #automotive19:22
*** nuohan has quit IRC19:22
*** nuohan has joined #automotive19:23
*** jlrmagnus has joined #automotive19:41
*** jlrmagnus has quit IRC20:06
*** jlrmagnus has joined #automotive20:20
*** waltminer has joined #automotive20:29
*** jlrmagnus has quit IRC20:36
pedroalvarezpaulsherwood: just looked at that. No, is something that uses navit21:28
pedroalvarezI'm just very very confused by the fact that the project carries patches inside, and a script to clone navit and apply the patches21:31
pedroalvarezgah, better to wait for someone with more knowledge about this. maybe CTtpollard :)21:32
*** tjamison has quit IRC22:06
*** mvick_ has joined #automotive22:06
*** tjamison has joined #automotive22:06
*** mvick has quit IRC22:08
*** tjamison has quit IRC22:14
*** mvick_ has quit IRC22:14
*** nisha has quit IRC22:14
*** tjamison has joined #automotive22:16
*** mvick has joined #automotive22:19
*** fury has quit IRC22:25
*** jlrmagnus has joined #automotive22:26
*** Jakubin has quit IRC22:26
*** nisha has joined #automotive22:27
*** fury has joined #automotive22:29
*** tjamison has left #automotive23:01
*** jlrmagnus has quit IRC23:04
*** jlrmagnus has joined #automotive23:18
*** waltminer has quit IRC23:27
*** Nicholas_ has quit IRC23:37
*** jlrmagnus has quit IRC23:47
*** jeremiah has joined #automotive23:59

Generated by 2.14.0 by Marius Gedminas - find it at!