IRC logs for #baserock for Wednesday, 2016-03-02

*** gtristan has quit IRC01:40
*** gtristan has joined #baserock02:16
*** edcragg has quit IRC02:58
jjardonpaulsherwood: Hi, how fresh is the ybd cache? Its trying to build openssl-new here06:51
jjardonalso, I got this error, sounds familiar?:07:00
jjardon(with the latest ybd tag)07:00
jjardonFYI, 16.08 seems to work fine07:03
paulsherwoodjjardon: no, that's not familiar :/07:58
paulsherwoodlooks like i'll have to revert the pyfilesystem stuff :/08:00
*** fay has joined #baserock08:07
*** fay is now known as faybrocklebank08:08
paulsherwoodjjardon: i haven't refreshed it since openssl changed08:09
jjardonAh ok, nothing wrong here then :)08:15
paulsherwoodwell, apart from that bug08:15
*** toscalix has joined #baserock08:27
*** ctbruce has joined #baserock08:27
*** CTtpollard has quit IRC08:28
*** CTtpollard has joined #baserock08:41
* gtristan is going to start pushing tristan/wip/aboriginal branches to g.b.o if that's alright08:47
gtristanbut I'm also going to need:
* gtristan doesnt think at this point any other lorries are needed08:48
paulsherwoodgtristan: please go ahead08:50
paulsherwoodgtristan: how far have you got?08:50
gtristanI am building... there is a lot of unsorted stuff but I'm close to a stage where people can try building (but not deploying)08:51
gtristanprobably today I can make a status report with instructions on how to reproduce this on an arbitrary machine, but I'll need that tarball lorried08:52
*** bashrc_ has joined #baserock08:59
gtristanpaulsherwood, here's a preview of that report:
gtristanwith odd syntax highlighting on a text file ;-)09:05
*** edcragg has joined #baserock09:07
*** jonathanmaw has joined #baserock09:13
paulsherwoodgtristan: cool :)09:34
* paulsherwood notices sed-tarball has not appeared in lc so far09:36
rjekOh, lorry-controller, not locallycompact09:36
*** NaN has joined #baserock09:38
*** NaN is now known as Guest5170809:39
*** Guest51708 is now known as N-a-N09:42
*** franred has joined #baserock10:10
*** ssam2 has joined #baserock10:10
*** ChanServ sets mode: +v ssam210:10
*** locallycompact has joined #baserock10:11
*** edcragg has quit IRC10:16
* gtristan bumps into waves at David Woodhouse10:26
gtristanlooks like no locale data for aboriginal builds for now10:26
pedroalvarezright, all the components of Baserock infra have been upgraded now :)10:31
*** edcragg has joined #baserock10:34
locallycompactpaulsherwood, What *do* the numbers mean in ybd?10:36
locallycompact[257/284/284] [weston-system-x86_64-generic] Running command:10:36
pedroalvareznumber of component you are building right now/ components you need to build / total components of the system you are building10:38
locallycompactWhy did it finish at 257 then10:39
pedroalvarezhehe, no idea then10:41
paulsherwoodlocallycompact: it's a bug. i've not investigated10:44
locallycompactk nice10:45
paulsherwoodlocallycompact: i think it's to do with the Counter() implementation10:45
locallycompactcan't reproduce this failure10:45
paulsherwoodi'm thinking it needs to be re-done in any case...10:45
locallycompactjjardon, what were you doing here?
paulsherwoodlocallycompact: if you fancy,
*** paulw has joined #baserock10:46
locallycompactI do not10:47
*** paulw has joined #baserock10:48
*** ssam2_ has joined #baserock10:56
*** ssam2 has quit IRC10:56
pedroalvarez plans for upgrading libxml2 for python3 too?10:58
pedroalvarezjjardon: ^^10:59
pedroalvarezoh sorry, that was done before11:02
*** CTtpollard has quit IRC11:23
*** jonathanmaw has quit IRC11:26
*** jonathanmaw_ has joined #baserock11:26
*** jonathanmaw has joined #baserock11:26
*** CTtpollard has joined #baserock11:27
*** jonathanmaw has quit IRC11:29
*** jonathanmaw has joined #baserock11:29
*** ctbruce has quit IRC11:29
*** jonathanmaw has quit IRC11:33
*** tiagogomes has quit IRC11:34
*** ctbruce has joined #baserock11:43
*** tiagogomes has joined #baserock11:45
*** jonathanmaw has joined #baserock11:45
*** tiagogomes has quit IRC11:50
*** tiagogomes has joined #baserock11:50
*** jonathanmaw_ has joined #baserock11:53
*** jonathanmaw has quit IRC11:55
jjardonlocallycompact: build master of definitions11:56
locallycompactDon't see the error unfortunately11:57
locallycompactRevert the thing if it's a problem though11:59
*** tom__ has joined #baserock12:08
*** CTtpollard has quit IRC12:11
*** ssam2_ has quit IRC12:12
* locallycompact cries out for subsystems12:33
*** ssam2 has joined #baserock12:33
*** ChanServ sets mode: +v ssam212:33
locallycompactI'm redoing the masony system and I realised I wanted it to inherit strata additions to devel on an ongoing basis.12:46
pedroalvarezlocallycompact: is that needed for your usecase?12:59
locallycompactWell, kind of. I mean the mason thing is going to be anything that devel can potentially build, plus the ci tools, forever and ever.13:00
*** gtristan has quit IRC13:01
locallycompactIt's not that much of a pain right now, I can just type the buttons13:01
*** paulw has joined #baserock13:05
*** bashrc_ has quit IRC13:33
*** bashrc_ has joined #baserock13:35
jjardonlocallycompact: I agree subsystems would be very helpful13:52
* persia doesn't really understand the difference between "subsystem" and "stratum"13:53
paulsherwoodstratum = collection of chunks13:53
paulsherwoodsubsystem = collection of strata/chunks13:53
paulsherwoodis my guess13:53
persiaThat would make sense, in which case I'd consider that strata can't include strata a bug :)13:54
paulsherwoodsystem = special kind of collection of strata13:54
persiaHow is a "system" special?13:54
paulsherwoodthis has been discussed before, i agree with you :)13:54
paulsherwoodthey have, for example, system-integration-commands13:54
persiaAh, cool.  I've always wanted a good semantic distinction between "system" and "stratum" and that a "system" includes the results of running system-integration is a good one.13:56
paulsherwoodi would like to tidy this up, and *can* for ybd i think, but i think morph is heavily fixed on the current schema/design13:56
paulsherwoodi may be wrong...13:56
ssam2feel free to propose radical changes13:58
ssam2i don't think morph is unfixable, but if it is that's not a reason to brain-damage the data model :-)13:58
rdalei was hoping the new spec repo would allow to fix our terminology and data model, without it needing to depend on a build tool14:01
persiaI think we're getting closer to that.  We just have to make sure we don't conflate discussions on the data model with discussions about the tooling.14:02
ssam2rdale: I wouldn't want to merge stuff to master of spec.git that had no support in any tools14:04
ssam2rdale: making branches of spec.git would be useful for exploring ways to change things, before then implementing those changes in 1 or more tools14:05
rdaleyes,  we can experiment with a combination of a spec branch, and 'adapter code' in the tool - for instance replacing Definitions in ybd with some that takes a new definitions format and creates the same data structures as before so that the rest of ybd wouldn't need changing14:06
* paulsherwood was hoping someone would offer that functionality as a library :)14:07
rdalei think the internal structures that ybd uses are too specific to factor them out into a library14:08
paulsherwoodincidentally, could someone shed any light on whether the pros and cons of representing Definitions as objects vs dicts in python?14:08
paulsherwoodmorph does objects, and i found it hard to follow. dicts seemed obviously easier to me when i started14:09
ssam2morph kinda does both14:10
ssam2its data model doesn't actually mirror the definitions one exactly14:10
ssam2and you can access the actual .morph file content as a dict, which some code does14:10
ssam2i would suggest using the raw data as a dict, if possible, and switching entirely to representing them as objects if that seems better for some reason14:11
ssam2or use javascript where objects are dicts :-)14:11
ssam2actually, objects can be dicts in Python too -- you could have a subclass of 'dict' which could expose the actual data as a dict, but could also contain accessor functions that return extra info that needs to be calculated from the data14:12
locallycompactI thought everything was a dict in virtue of the fact that everything implicitly has the potential to have __getitem__, defined or not14:14
ssam2anything in Python can have __getitem__, though14:17
locallycompactyes exactly14:17
locallycompacteverything is everything14:18
* locallycompact wonders where's a suitable strata to add jq?15:17
ssam2its own stratum15:18
locallycompactWhat's wrong with the same place sed is?15:18
mwilliams_ctsed is in coreutils-common right? doesnt seem like jq would fit there and for most systems/users I would imagine it is extraneous15:20
mwilliams_ctthough ssam2 might have better reasoning than that :)15:20
ssam2same place as sed would be ok, I guess15:21
ssam2the Baserock openid provider should accept registrations again15:34
ssam2i couldn't find any way to get debug logs from django.registration or django.contrib.auth ...15:34
ssam2without actually editing the code and adding some of course15:35
persiassam2: Thanks for that.  Should password reset also work now?15:45
*** tom__ has quit IRC15:53
ssam2persia: i didn't find anything wrong.. but didn't manage to enable any diagnostics15:55
ssam2so it might not work for you15:55
*** CTtpollard has joined #baserock15:56
ssam2i can add a 'print' statement to the django.contrib.auth module somewhere as a last ditch resort15:56
persiaI've submitted the form.  If I don't get an email in a while, I'll try registering again, which might provide more interesting errors.15:57
pedroalvarezyay, fail2ban seems to work in baserock :)16:01
*** jonathanmaw_ has quit IRC16:02
paulsherwoodw00t :)16:03
ssam2persia: i've discovered that if you put an email into 'reset password' link that doesn't have any users associated, it silently succeeds16:06
persiaAha!  Then maybe I'm not registered, or registered with an email address different to what I think I was using.16:06
ssam2looking at the DB, I think you aren't registered in the first place16:07
persiaAs a side note, on the form, I would suggest "Given Name" rather than "First Name", as not everyone orders their names that way.16:08
persiaAnd, with fixed registration, registration worked, and all is good.  The only outstanding mystery is why I had a username and password recorded for the site in my password safe, but that's not related to site issues.16:14
*** will2 has joined #baserock16:15
*** wdutch has quit IRC16:15
*** franred has quit IRC16:17
persiaSotK: I'm sorry to report that my "[object URL]" seems to be preventing login as well.  If I go to "", I end up at "[object URL]#!/page/about, and using the log in feature results in "The requested URL /[object URL] was not found on this server.".  I can replicate this with multiple different browsers.  I do not experience this for storyboard.openstack.org16:19
SotKpersia: that is very weird, and I have no idea why it is happening :/16:24
persiaAs long as it only happens for me, I don't mind that much, but if I can provide any information to help figure out *why*, I'd be happy to do so.16:24
persiaThe part that baffles me the most is that it seems to be only Baserock, rather than Storyboards in general.16:25
SotKmaybe something weird in the apache config on storyboard.b.o for some reason?16:26
persiaThat's my guess.16:26
SotKpedroalvarez: ^16:27
pedroalvarezpersia: what browser are you using?16:31
pedroalvarezi can't reproduce the failure :/16:34
persiapedroalvarez: I tried rekonq and uzbl : both are just wrappers around webkit, but I have an old webkit16:35
* paulsherwood wonders whether SotK should really consider every possible browser combination16:37
persiaI don't seek to have everything supported, which is why I don't mind if it is just me, but I do believe there should be at least one known solution that both works and is free.16:37
*** jonathanmaw has joined #baserock16:38
persia(free in the sense of free software, so no trademark encumbrance, no auto-update-and-mothership-tracking agreement, etc.)16:38
jjardonpaulsherwood: trying to deploy with ybd Im getting this:16:43
pedroalvarezso... fail2ban needs python-systemd16:43
jjardonam i doing something wrong?16:43
pedroalvarezwhere could it go, and.. python2? python3?16:44
pedroalvarezjjardon: can I see the deployment morphology?16:44
pedroalvarezanybody fancies a python2vs3 discussion? :)16:45
jjardonpedroalvarez: its clusters/weston-system-x86_64-generic-deploy.morph in current definitions16:45
pedroalvarezjjardon: my guess is that ybd is not deploying the subsystem16:46
pedroalvarezbut I don't know why16:46
jjardonpedroalvarez: I guess depends on what fail2ban needs16:47
pedroalvarezI seriously don't know what to do16:48
pedroalvarezIf i install it as python2, but python2 is not present, then it won't work16:49
pedroalvarezif I install it as python3, but python2 is present overriding the symlink, it won't work16:49
persiapedroalvarez: Hrm?  What it is that depends on the symlink but has two dialects, rather than using six or nine?16:50
pedroalvarezright, that's something I didn't know16:51
* SotK reproduces the storyboard oddness with uzbl, and sees it on too16:52
persiaI'm having trouble imagining why it fails, but I'm guessing either the fail2ban install script does funny things to a wrapper script depending on the version, or there's some library hookery that makes assumptions somewhere.16:52
jjardonpedroalvarez: fail2ban is compatible with python3, so use python316:53
* SotK moves talking about this to #storyboard16:53
persiaSotK: Interesting.  Apologies for the confusion: my old uzbl was behaving differently.16:53
pedroalvarezI guess the only thing I can do is test that I was wrong talking about symlinks :)16:53
pedroalvarezjjardon: foundation for python-systemd?16:54
jjardonpedroalvarez: seems a good place to put it16:54
pedroalvarezfail2ban can go to network-security I guess16:55
jjardonpedroalvarez: nope, those stratum are for libraries that provide SSL/TLS implementations16:57
jjardon(yeah, the name of the stratum is quite bad)16:58
pedroalvarezit kind of run-depends on iptables16:58
pedroalvarezbut i'm not sure if it could go to connectivity16:58
jjardonif only run depends, putting strata/connectivity.morph in your system definition would be enough, not need to put it there16:59
persiaI like the idea of putting it in connectivity: most connected systems would benefit from fail2ban16:59
jjardonpedroalvarez: Id suggest you create a new stratum for fail2ban; doesnt seems something generic16:59
pedroalvarezI'm happy with that17:00
pedroalvareznote that fail2ban can be configured to work without iptables17:00
*** jonathanmaw has quit IRC17:00
*** jonathanmaw has joined #baserock17:00
pedroalvarezit can be configured to do many many many things...17:00
pedroalvareztook me a while to test it  :)17:00
jjardonpersia: conectivity is a stratum that contains generic libraries related with connectivity; Id not want fail2ban in my GNOME or weston system for example17:01
locallycompactWhat's this one about: I build a full devel system successfully, then I add jq to coreutils-common, I rebuild and ybd tries to rebuild bash-completion, and fails with execv: No such file or directory17:02
persiajjardon: Why not?17:02
persiajjardon: Do you not have any network services at all?17:02
jjardonpersia: no17:03
persiajjardon: `netstat -ntl` returns no items?17:03
*** jonathanmaw has quit IRC17:03
persiaActually, `netstat -nul` is just as interesting17:03
persiaI consider myself careful about expsosing network services, but still have a few running on my desktops.17:04
persiaAnd for any application that isn't a desktop/handheld/etc., one needs to run *more* interactive services, as the user isn't expected to do as much themselves.17:05
jjardonpersia: seems only cups:17:05
persiajjardon: Nice job.  That's tighter than my laptop.  Still, fail2ban can protect against CUPS attacks :)17:06
N-a-NHello everyone! I'm new here and I'm trying to learn about the project. Can anybody recommend me best reference system or an example to begin with to make a demo of an embedded x86 application?17:09
edcragghi N-a-N! there are pre-built images for x86 to get you going, typically people develop for baserock in a VM or chroot. what platform are you thinking of using?17:14
edcraggmost of the release images are at
edcraggand there are a fair number of guides on the wiki17:17
*** ctbruce has quit IRC17:20
pedroalvarez !!17:28
paulsherwoodpython-systemd.... what can possibly go wrong? :)17:30
N-a-Nthanks edcragg! I'll take a look at the wiki and the pre-built images. I just want to make a demo of an embedded x86 application, if that's what you're asking17:35
*** ssam2 has quit IRC17:36
pedroalvarezpaulsherwood: hah, thanks :)17:42
*** faybrocklebank has quit IRC17:44
paulsherwoodyw :)17:46
*** gtristan has joined #baserock17:55
*** bashrc_ has quit IRC18:02
*** lc_ has joined #baserock18:03
pedroalvarezwoop woop!18:15
pedroalvarez foundation: Add python-syste18:15
pedroalvarez Add fail2ban-common stratum18:15
pedroalvarezI think I'll make gary_perkins happy with these :)18:15
gary_perkinsYay \o/!18:16
gary_perkinsthank you pedroalvarez :)18:16
pedroalvarezit was fun18:16
pedroalvarezand now basrock systems can be more secure :)18:16
gary_perkins\o/ :)18:17
pedroalvarezonly tested it with ssh, though18:18
pedroalvarezbut that's at least a start18:18
pedroalvarezI guess most of the things will work though18:19
*** toscalix has quit IRC18:22
paulsherwoodcannot be merged?18:43
* paulsherwood dislikes gerrit, sometimes18:43
*** locallycompact has quit IRC18:46
*** lc_ has quit IRC18:46
persiapedroalvarez: Does the Baserock gerrit comprehend Depends-On: ?18:46
*** radiofree has quit IRC18:47
persiaOn, nevermind, that's a zuul thing.18:48
* paulsherwood rebased, and merged18:49
persiaIt's an interesting problem in git automation: given a change with an annotation indicating that a dependency was rebased, is it safe to rebase against the dependency?18:52
*** radiofree has joined #baserock18:53
*** edcragg has quit IRC18:53
paulsherwoodpedroalvarez: i took the view that it's security related, therefore +219:05
pedroalvarezI'm planning to add it by default in trove. If there is a system that makes sense to include it, is trove19:16
*** lc_ has joined #baserock20:27
*** locallycompact has joined #baserock20:28
*** edcragg has joined #baserock20:33
*** locallycompact has quit IRC21:17
*** lc_ has quit IRC21:17
*** edcragg has quit IRC21:59
*** edcragg has joined #baserock22:25
*** edcragg has quit IRC23:02

Generated by 2.14.0 by Marius Gedminas - find it at!