IRC logs for #baserock for Sunday, 2015-04-12

*** zoli__ has joined #baserock00:49
*** zoli__ has quit IRC00:53
*** edcragg has quit IRC01:06
*** pacon has joined #baserock02:18
*** pdar_ has joined #baserock05:48
*** pdar has quit IRC05:53
*** zoli__ has joined #baserock05:55
*** zoli__ has quit IRC07:20
straycat21:35  paulsherwood$ is it me or is it harder to try branches from gerrit than it was with the ml?09:24
straycatit's very straight forward actually09:24
straycatgit review -d <change number>09:24
straycatwill download the change and check it out into a branch09:25
paulsherwoodah, thanks!09:25
paulsherwoodthere are so many tricks to learn, straycat :-)09:25
straycatyes, it's all much easier with git-review for sure, submitting a change is just git review -R (-R to prevent automatic rebase against the target branch)09:28
straycatSotK, why not have the get-repo command optionally take a ref as well, so you default to what's in definitions unless the user specifies a ref09:34
straycatlife would also be easier with gerrit if we put .gitreview files into all the repos now hosted on gerrit, this got one +1 and there were no objections, i've only not sent a patch because i don't have time to right now09:39
straycatjjardon, it seems the systemd-timesyncd unit has a ConditionVirtualization=no so it's not getting started in any baserock vms, so there's no time sync happening right now inside vms. it's fixed by which should apparently be in the v220 release, just thought i'd let you know :)09:57
paulsherwoodstraycat: that explains a lot!09:59
straycatyes i think they were expecting the virtualisation to set the hardware clock of the vms or something, but they don't all do that, at least from what i've read
paulsherwoodaha! i've realized why ybd needs max-jobs: 1, and morph doesn't10:17
paulsherwoodi think morph only honours MAKEFLAGS in build-commands, not install-commands10:17
paulsherwoodso for example subversion-tarball, acl, libapr, harfbuzz - install steps all fail with -j6, not with -j110:20
*** brlogger has joined #baserock10:39
SotKstraycat: that makes sense10:56
straycatpaulsherwood, that sounds like a bug11:01
straycatSotK, cool11:02
SotKhm, I was also allowing the destination path to be optionally specified11:21
SotKis there a sane way to do both?11:21
*** zoli__ has joined #baserock11:23
* SotK ponders finishing the yaml lorries patch he was messing with a while ago11:52
paulsherwoodstraycat: thanks for merging the acl patch :)11:57
paulsherwoodSotK: i've successfully built your xfce work, am now checking i can also build genivi from there11:57
SotKpaulsherwood: great, thanks12:11
straycatpaulsherwood, no problem :)12:16
straycatSotK, would be nice to have yaml lorries12:18
*** pacon has quit IRC12:20
* paulsherwood wonders if it's unavoidalbe to have an increasing number of foo.configure files in root of definitions12:32
straycatwe need to tidy it up at some point12:34
straycatthrow all the conf exts into a single directory12:34
straycatpossibly all the install-files too12:35
*** zoli__ has quit IRC12:39
paulsherwoodSotK: genivi-common-api-dbus-runtime build fails for me, investigating12:50
paulsherwoodhmm... broken in master12:52
SotKmaybe genivi should be in our CI?13:02
paulsherwoodi was thinking that :)13:02
pedroalvarezGrr.. Genivi13:07
pedroalvarezThere is a release this Friday, but my week will be plenty of things to do13:16
pedroalvarezOtoh, nothing has changed in genivi13:16
* straycat wonders whether pedroalvarez has a highlight set for genivi13:17
pedroalvarezNo new requirements13:17
pedroalvarezstraycat: nope, but maybe I should :)13:17
paulsherwoodpedroalvarez: don't worry, either i  can do it, or persuade someone else :)13:20
pedroalvarezGreat, thanks13:37
pedroalvarezI've pondering not doing one, given that nothing changed13:38
paulsherwoodmaybe that will be the answer, we'll see13:39
paulsherwood(still needs fixing in master - i'm bisecting)13:39
* SotK also notes that the morph SHA in definitions is from mid-March13:40
paulsherwoodSotK: thanks, but it won't be morph. i get the same result with ybd :)13:40
SotKI know, just mentioning it in case we should be updating it?13:59
paulsherwoodah, in the genvi system?14:08
pedroalvarezHehe, no no, in devel systems. I guess that now with Gerrit we forget to send patches for definitions.git whenever we patch Morph14:10
*** zoli__ has joined #baserock14:24
* jjardon sees the get-repo patches and say thanks to SotK :)14:25
jjardonstraycat: nice, maybe we should cherry pick the patch?14:28
*** zoli__ has quit IRC14:29
pedroalvarezOh great, once it's merged we can use it for our license check script, and stop depending on morph edit14:33
pedroalvarezpaulsherwood: ideally firehose will send patches to update morph version in definitions.git14:34
jjardonbtw, for the curious, after a request, baserock statistics are up-to-date again: Seems February 2015 is the new commit record \o/14:51
paulsherwooddoes that include ybd? :)14:54
paulsherwoodand does it include definitions?14:54
* paulsherwood sees that the answers are no, and yes
jjardonpaulsherwood: feel free to add it14:57
paulsherwoodjjardon: thanks, but i'll wait until it's worth talking about... which hopefully will be soon14:57
paulsherwoodany ideas why gerrit might bounce me with this...
straycatjjardon, is there any reason we build ntp without --enable-linuxcaps ?15:12
SotKjjardon, straycat: Thanks for the reviews!15:13
SotKjjardon: you'll have to wait for tlsa to answer your queries about the xorg stratum15:14
straycatSotK, no problem, quite looking forward to having a morph get-repo command15:14
straycatalthough i'd also need to add a ntp user15:16
jjardonstraycat: no idea, but we stop using ntpd in the generic systems when we upgrade systemd and start suse timesyncd15:16
jjardonstraycat: you should probably need to ask jmacs15:17
straycatjjardon, aye but for swift we need ntp, so there'll be a conf ext to disable timesyncd and use ntp for anyone deploying a swift node15:17
jjardonstraycat: ah, ok didnt know ntpd is a hardcoded dependency of swift15:18
straycatyeah i think internal time synchronisation is required to keep the object store consistent15:20
*** zoli__ has joined #baserock15:56
*** zoli__ has quit IRC16:00
straycatjjardon, (cherry-pick) idk maybe, or we could wait for the release16:24
straycatpaulsherwood, re what happens if you run git commit --amend then try to resubmit?16:28
straycatalso, unless you're pushing a new patchset to the same change, try deleting the Change-Id in the commit msg16:29
paulsherwoodwill try (when this build finishes)16:30
straycatwe could also put the git-hook in the repo and have morph checkout/branch copy it into the right place16:31
straycatthat's probably not a good idea16:32
*** zoli__ has joined #baserock16:47
straycathmm, I get Operation not permitted when i chown a directory in this chroot, running chown as a system integration command16:55
straycatand i have created the user and group16:56
straycatand now i'm in the chroot and i can chown the directory just fine16:59
paulsherwoodstraycat: can you use mknod and chmod instead?17:26
* paulsherwood had a similar problem some months ago17:26
paulsherwoodnot sure if what i'm suggesting is relevant though17:27
straycati'll try it anyway18:03
straycatpaulsherwood, chown failed too18:29
straycatroot@salo:/home/richardipsum# ping
straycatping: unknown host
straycatroot@salo:/home/richardipsum# ping
straycatPING ( 56(84) bytes of data.18:33
straycat64 bytes from icmp_req=1 ttl=53 time=48.0 ms18:33
straycatsorry, wrong window >.>18:33
paulsherwoodjjardon: you updated dbus - genivi for some reason has a 'custom marshalling' patch which the authors have not upstreamed18:37
paulsherwoodi'll send a patch to apply it18:37
paulsherwoodjjardon: is there something special about dbus-1.8.16, or should we go even further forward?18:39
straycatrichard_maw, before i start debugging this, is there some obvious thing i'm missing that means chown() chmod() etc can't be called from within a linux-user-chroot ?18:56
straycatpaulsherwood, sorry i meant chmod19:03
*** sambishop has joined #baserock19:10
paulsherwoodstraycat: i'm guessing it's something to do with permissions (ie your operation is somehow denied by the chroot config you've set up)?19:11
paulsherwoodi found this, for example?
straycatit looks like linux-user-chroot is to blame19:14
straycati took the linux-user-chroot- commandline morph is using and stripped it down to a bare minimum, it still failed with operation not permitted, then when i substituted linux-user-chroot for chroot the integration script succeeded19:15
paulsherwoodah :)19:33
paulsherwoodprogress, at least19:33
*** sambishop has quit IRC19:35
*** zoli__ has quit IRC22:03
*** zoli__ has joined #baserock22:18
*** zoli__ has quit IRC22:23
*** zoli__ has joined #baserock23:19
*** zoli__ has quit IRC23:23
*** zoli__ has joined #baserock23:42
*** zoli__ has quit IRC23:47

Generated by 2.15.3 by Marius Gedminas - find it at!