IRC logs for #baserock for Friday, 2015-02-06

*** inara has quit IRC04:04
*** inara has joined #baserock04:11
*** inara has quit IRC04:27
*** zoli__ has joined #baserock05:53
*** zoli__ has quit IRC05:54
*** zoli__ has joined #baserock05:58
*** inara has joined #baserock06:03
*** inara has quit IRC06:13
*** inara has joined #baserock06:24
*** zoli__ has quit IRC06:30
*** inara has quit IRC06:37
*** inara has joined #baserock06:38
*** zoli__ has joined #baserock06:47
*** zoli__ has quit IRC06:49
*** zoli__ has joined #baserock06:57
*** zoli__ has quit IRC07:20
*** zoli__ has joined #baserock07:46
*** zoli__ has quit IRC08:06
*** grahamfinney_ has joined #baserock08:16
*** grahamfinney has joined #baserock08:16
*** grahamfinney_ has quit IRC08:18
*** grahamfinney has quit IRC08:18
*** grahamfinney_ has joined #baserock08:19
*** grahamfinney has joined #baserock08:19
*** mariaderidder has joined #baserock08:21
*** CTtpollard has joined #baserock08:30
*** franred has joined #baserock08:35
*** ssam2 has joined #baserock08:52
*** ChanServ sets mode: +v ssam208:52
*** bashrc has joined #baserock09:03
ssam2i'm a bit worried that 2 stories on show up as being created by me, even though they weren't.09:04
mwilliams_ctwhich stories ssam2 ?09:05
ssam2the only ones09:05
ssam2!/story/1 and!/story/209:05
ssam2i'm sure paul created them (since he told me he had created some stories, and those are the only ones)09:05
mwilliams_ctcould he have done it on your laptop? If not, maybe CTtpollard can help, I know he's worked on Storyboard a bit09:06
radiofreeThose links don't work on Firefox mobile09:06
ssam2radiofree: interesting. they don't work for me either unless I enable Javascript09:06
ssam2so maybe it's Javascript-related09:07
radiofreeMight be certificate related!09:07
ssam2the certificate is just some self-signed one I generated, yeah09:07
radiofreeManaged to get it to show the "not trusted" page now09:08
* radiofree doesn't trust storyboard, so will go no further09:08
*** edcragg has joined #baserock09:08
ssam2radiofree: I don't think it means you any harm09:08
CTtpollardI never saw it relate posts to the wrong user09:08
ssam2ok. could be a problem in the OpenID provider then09:09
mwilliams_ctssam2: also, logging in I get "The site has requested verification of your OpenID. I have failed to reach it and thus cannot vouch for its authenticity. Perhaps it is on your local network." -- maybe timeout or something, but it's not an ideal error description09:09
mwilliams_ctalso, those stories now appear to have been created by me!09:10
*** rdale has joined #baserock09:10
ssam2mwilliams_ct: about the warning: see story #2 ;)09:10
ssam2very interesting that they show as being yours though.09:10
mwilliams_cthehe sorry09:10
ssam2they show up as yours for me too09:11
mwilliams_ctBut then since I'm listed as having reported it clearly I knew that ;)09:11
ssam2this is weird.09:11
CTtpollardI presume it's associating the current http remote user as story creator09:11
ssam2I'm wondering if it somehow gets the same OpenID URL for all of us09:12
petefothIS there / will there be any link to the OpenId provider from w.b.o?09:12
ssam2petefoth: might make sense, yeah09:13
petefothssam2: I would add such a link, but I have no idea where the OpenID provider lives :)09:13
ssam2the 'users' table in Storyboard has only one entry which is mwilliams_ct. And it has your OpenID as
ssam2so I guess it's somehow getting the same OpenID for all of us. I must have configured it wrong09:14
mwilliams_ctYeah that sounds like a bug in the openID provider. is there any source code for that I can look at?09:15
* pedroalvarez logs in storyboard to see what happens09:15
pedroalvarezyay! I'm the stories creator now09:15
mwilliams_ctpedroalvarez: congrats!09:15
mwilliams_ctthanks ssam209:15
pedroalvarezcan anybody log in using a different openid?09:16
ssam2no, storyboard ties you to one openid provider09:16
ssam2we could point it to e.g. launchpad09:16
ssam2but then we'd be tied to launchpad09:16
petefothat the meetup it was suggested that we should modify StoryBoard to support any OpneID provider, but I guess that will be some way off09:18
*** jonathanmaw has joined #baserock09:18
ssam2we do have some people at Codethink who are familiar with the Storyboard codebase, I think09:18
petefothWe can investigate how we go about asking them to have a look at this issue09:20
*** tiagogomes_ has joined #baserock09:20
paulsherwoodmwilliams_ct: didn't you look at that earlier?09:20
mwilliams_ctpaulsherwood: petefoth: Yes, we looked briefly into having multiple openID providers in previous work I did. From memory, it was concluded that it was a lot of effort for not much reward on our part (though I think upstream would be grateful for it, though they'd be unlikely to use it themselves)09:21
petefoththe 'reward' would be for Baserock, which would then have a defect and story tracker which can be used by anyone without having to register with
radiofreeWhat is baserock using storyboard for?09:26
mwilliams_ctpetefoth: sure, but when I said reward I meant relative to what we were doing at the time, not to Baserock - apologies for being unclear09:27
ssam2radiofree: we want an issue tracker and task tracker09:27
persiaradiofree: To report problems with morph, lorry-controller, lorry-minions, system-version-manager, definitions, infrastructure definitions, etc.09:27
paulsherwoodradiofree: issue logging, for example, since yesterday i think09:27
petefothmwilliams_ct: I did understand :)09:27
radiofreeDoes storyboard email you now?09:28
CTtpollardradiofree: no09:28
radiofreeSo you're just expected to constantly refresh?09:29
radiofreeWhat about attachments?09:29
CTtpollardthey're working on email, but for the time being yes09:29
ssam2radiofree: out of interest, how do you feel about Trac ?09:29
radiofreessam2: it's better than storyboard09:30
CTtpollardThese are the current patches in jenkins,n,z09:30
CTtpollard*gerrit sorry09:31
petefothMention of Trac in the meetup summary session caused robtaylor to cry out in anguiush - 'Nooooooo!'' was what he said iirc :)09:31
CTtpollardThere's email patches up for review09:31
petefothCan we have a defect tracker war now? ;)09:31
ssam2eventually we'll get to using Bugzilla like everyone else. But we can spend more time setting up new instances of different things for a while yet.09:32
Kinnisonpetefoth: Trac often invokes horror in people, but it's actually a fairly competent project organisation system -- it kinda wants to take over your world though which is likely why robtaylor didn't like the idea09:32
radiofreeI didn't say it was good, I said it was better than storyboard09:32
paulsherwoodpetefoth: if you want. in the meantime, i've logged a couple of things in storyboard09:32
* paulsherwood has better things to do than participate in wars09:32
Kinnisonssam2: It is indeed a sorry indication of the state of the defect/issue tracking world that people still misuse a bug tracker for it in preference to the dedicated solutions09:32
Kinnisonssam2: But I fear you may be right09:32
petefothpaulsherwood: I really *don't* want! Yesterday's 'decision' was storyboard09:33
* Kinnison makes some peas for paulsherwood 09:33
paulsherwoodwhirled peas? please? can we have whirled peas?09:33
* Kinnison spins 'em around09:34
* petefoth is now the author of the two storyboard stories :) How long will that last?09:36
ssam2til someone else logs in, sadly09:36
ssam2i'm looking at that problem09:36
petefothssam2: do you think we should link to storyboard from w.b.o now, or should we wait for a few wrinkles to be ironed out?09:39
persiaI think we should iron before hanging09:39
ssam2we should wait09:39
CTtpollardis any of the bootstrap theme going to be changed to match w.b.o?09:41
ssam2CTtpollard: what bootstrap theme?09:41
tiagogomes_why do we have kexec in tools? Are we making us of it?09:42
CTtpollardit uses bower bootstrap for the theme, you can over ride it though09:44
* pedroalvarez likes storyboard webpage being responsive09:44
ssam2cttpollard: storyboard does? ok. I don't have a clue about theming, personally, I'd appreciate help09:45
ssam2same goes for -- I'll go crazy if I try to write a proper .css for it09:45
CTtpollardI've only changed stuff such as the navbar colour, storyboard icon & favicon though09:46
CTtpollardnothing major09:46
persiaCTtpollard: Even changing that stuff would be lovely :)09:46
ssam2cttpollard: how do you change those things?09:46
* persia wonders if that sort of change is just a patch vs. defintions, or whether we end up forking storyboard, and then patching definitions to use the fork09:46
ssam2the current storyboard isn't cleanly deployed from the infrastructure repo09:47
CTtpollardhow are you running the instance ssam2, via vagrant or tox+grunt?09:48
persiaIn that case, I suppose I'm wondering what that sort of change *should* be, rather than *would* be :)09:48
*** Krin has joined #baserock09:48
ssam2persia: we can get there in stages. committing /etc/storyboard/storyboard.conf to git would be the first step, I guess09:48
ssam2CTtpollard: I don't think it's using either of those things09:49
ssam2I ran the puppet scripts, and then fixed up some bits that had failed to work manually09:49
ssam2apache is runnig the storyboard workers09:49
persiassam2: That's a sensible first step, I expect.  Configuration management of declarative systems is something that probably needs more thought though, so I'm undecided if that ends up as part of a final solution.09:50
* persia has not managed to reconcile "declarative system definition" with "ability to adjust configuration to reflect operational requirements" yet09:50
ssam2the model I've been trying in the infrastructure.git repo is to have 2 ansible scripts, one for 'image-config' and one for 'instance-config'09:51
ssam2the 'instance-config' one is run *after* instantion, and can be edited and rerun whenever you want09:51
ssam2it's far from perfect, but I think it's the right model09:52
CTtpollardI've found my instance where I changed the favicon and brand colour09:52
*** Krin has quit IRC09:53
pedroalvarezssam2: just to add info to your debugging - storyboard thinks I'm you09:53
*** petefotheringham has quit IRC09:53
*** simonh_ has quit IRC09:53
*** mwilliams_ct has quit IRC09:53
persiaThat does sound like a sensible model.09:53
*** mwilliams_ct has joined #baserock09:54
*** petefotheringham has joined #baserock09:54
ssam2pedroalvarez: it thinks we're all the same user, identified by
ssam2Storyboard requests that OpenID from, which is a special OpenID that means 'let the user choose what OpenID they want to use'09:55
ssam2so I guess Storyboard expects the OpenID provider to return but instead it returns the identifier_select URL again09:55
pedroalvarezgood debugging09:56
ssam2seems like a bug in django_openid_provider so far09:56
petefothplease can we have a projcet in storyboard for w.b.o? The UI / Documentation session at the meetup decided that we need some 'What I do with baserockm and how do I do it' stories. I would likie to be able to track the progess of getting those written09:57
ssam2ok. what should it be called?09:57
pedroalvarezpetefoth: "please don't spend hours filing new issues in there yet"09:58
petefothssam2: 'BAserock wikiw'?09:58
ssam2unconventional, but I like it09:58
petefothpedroalvarez: I won't spend hopurs, but we need to use it a bit09:58
petefothssam2: or you could spll ittt writttte09:59
ssam2I now realise the bug in django_openid_provider was introduced by me. :(09:59
ssam2in in case you are curious10:01
pedroalvarezlines 94 95, right?10:02
*** inara has quit IRC10:02
ssam2yeah. the reason for that was that if the user requested 'foo' and the provider returned 'foo/' it'd trigger an assertion10:04
ssam2so i changed it to return whatever was requested10:04
ssam2but that breaks the 'identifier_select' case.10:04
petefothssam2: looking again , I guess it should be 'baserock/wiki', or we could use the baserock/infrastructure project. Personally I would prefer it to have its own project10:05
ssam2projects are cheap, baserock/wiki sounds good to me10:05
ssam2or would baserock/documentation be better?10:06
ssam2i guess part of the point of storyboard is you don't need 'catch-all' projects like 'documentation' though10:06
ssam2since stories can span multiple projects10:06
*** Krin has joined #baserock10:06
*** simonh_ has joined #baserock10:07
petefothssam2: where would I then put stories to do wwith updating the documentation of morph?10:07
Zarapedroalvarez: yeah, I was worried about the maintenance side of things. One option I see is to make the tutorial very small (eg: walks through the wiki instructions from launching the baserock vm up to the end of the 'upgrade your vm to a devel vm' stage).10:07
ssam2petefoth: good question.10:07
ssam2I think ideally Morph's docs would live in morph.git and would be tracked in the Morph project10:07
ssam2so I think tasks for morphs docs should go in baserock/morph ...10:07
franredthat sounds sensible to me10:08
petefothssam2: agreed. I think wiki should be a separate project. Some dcoumentation stories mu=ight involve tasks in both wiki and morph10:08
pedroalvarezZara: my idea (when i had it) was to start the tutorial within a "fresh baserock vm"10:08
petefoththe wiki is a separate entity and deserves its own project I think10:09
ssam2petefoth: fair enough. i'll create baserock/wiki once I've fixed this openid provider bug10:09
pedroalvarezZara: so you go through basic morph commands10:09
ssam2actually, I may as well do it now10:09
petefothssam2: thank you :)10:10
CTtpollardI think i've tracked down the location of the icon as well10:11
ssam2cool, where is it?10:11
Zarapedroalvarez: Ah, I see, so the tutorial would be within baserock itself, and thus easier to update? Sounds good. I'd like some way to make it simpler to set up the vm in the first place, though, because my goal here is to make it really easy for someone to quickly play around with baserock (even a skeletal version of baserock) and feel familiar with it.10:12
CTtpollardfor the favicon and brand colour in the webclient go to /src/theme/custom and take a look at the readme10:13
CTtpollardfor storyboard icon it's to do with /src/fonts/custom_icons10:14
ssam2ok, ta10:14
CTtpollardwhich I think points to src/fonts/src/storyboard.svg10:15
pedroalvarezZara: this if from the creator of that tutorial:
CTtpollard@brand-primary in theme.less will let you change the top bar colour, and placing a favicon in that directory will override the default one10:16
pedroalvarezZara: he says that the best option would be to create a container somewhere for every user accessing the tutorial, but he didn't do that because docker doesn't work inside of docker10:17
ssam2cool. anyone have a favourite colour for the top bar?10:17
ssam2i'd like to replace it with a picture of some bikes or something10:17
pedroalvarez(wich is the background colour of
Zarapedroalvarez: does this mean that we'd need the user to get baserock running on docker, first? because (going from the wiki), that seems like it could be complicated itself.10:21
ssam2baserock in Docker is simple, actually10:21
*** lachlanmackenzie has joined #baserock10:21
ssam2cat baserock-15.02-chroot-x86_64.tar.gz | docker import - baserock/15.0210:21
pedroalvarezZara: no, the idea is that the user opens a webpage, and in that webpage he has a shellinabox (shell emulator) running accessing to a baserock instance, which behind the scenes is baserock in docker10:22
pedroalvarezcrazy uh/10:22
Kinnisonssam2: ARGH, I KILL YOU10:22
Kinnisonssam2: Pointless use of 'cat'10:22
ssam2it's not pointless. it's sending the file to docker!10:23
Kinnisondocker import - baserock/15.02 < baserock-15.02-chroot-x86_64.tar.gz10:23
* Kinnison also whines that docker says "URL" and then only claims to only accept URLs whose scheme == "http"10:24
* richard_maw is looking forward to the next systemd release, as he would be able to nspawn all this stuff10:25
*** bashrc has quit IRC10:32
*** bashrc has joined #baserock10:33
simonh_How does morph behave if there is no `build-commands` e.g. in does it do `make` and `make install` knowing it's autotools?10:38
*** simonh_ is now known as mauricemoss_10:39
Kinnisonmauricemoss_: Basically it depends on the autodetected (or specified) build system10:40
Kinnisonmauricemoss_: it can detect (by presence of configure.{in,ac}) autotools10:40
nowster"build-system: autotools"10:41
nowsterif there's only a configure-commands: stanza, the rest is defaulted?10:41
pedroalvareznowster: yes, and the configure-commands would be overriding the default one10:42
pedroalvarezdefaults are here:
mauricemoss_Kinnison, pedroalvarez thanks!10:43
pedroalvarezmauricemoss_: btw, have you seen the "Native baserock for CB5-311 guide" mail in baserock-dev?10:44
mauricemoss_pedroalvarez, I've seen it, figured out the issue and will reply by the end of the day.10:44
pedroalvarezmauricemoss_: thanks :)10:46
Zarapedroalvarez: I see, it sounded complicated to me, at first, but might not be complicated in practice. Where would we put the baserock-in-docker instance that the user accesses?10:46
pedroalvarezZara: sounds really complicated to me :)10:47
pedroalvarezI'd be interested on ideas about how to implement this10:47
ZaraHeh, if I were implementing it, I'd probably just go the docker route and tweak their JS emulator, without a real baserock instance being involved anywhere in the process, but that's mainly because I don't have the know-how to do something more complicated. But I think I'll look into Docker more, today.10:50
pedroalvarezI'm having problems with iptables + openstack + baserock and I wonder where would be the best place to ask11:12
ssam2I've fixed the OpenID provider. Storyboard should now be able to tell people apart11:18
ssam2the old user account that we wrongly shared is now called 'Storyboard Monkey'11:19
jjardontiagogomes_: hey! did you, by any chance, have time to test the automake 1.15 patch?11:20
jjardonI can not connect to, known issue?11:25
ssam2use https://11:25
ssam2http:// doesn't forward there, need to fix the apache config11:25
persiaIs there a potential to autoredirect that?11:25
persiaAh, cool.11:25
jjardonssam2: still not connecting11:27
tiagogomes_jjardon in which archs have you tested so far?11:28
jjardontiagogomes_: x86_6411:28
petefothjjardon: you have an extra 'o' in yiur url11:29
jjardonpetefoth: oh, good catch! thanks! (sorry ssam2)11:29
franredwhich documentation I can read to know what is in /src/tmp/{deployments,chuncks,staging,failed} directories?11:29
jjardonmmm, now I can not login in http://openid.baserock.org11:30
jjardon"This account is inactive"11:31
pedroalvarezjjardon: create an account first?11:31
jjardonpedroalvarez: I did that11:31
pedroalvarezthat sounds like: "the account is being activated"11:32
*** gary_perkins has joined #baserock11:37
*** lachlanmackenzie has quit IRC11:38
jjardonpedroalvarez: any idea how much time until it gets activated?11:43
ssam2jjardon: you need to click the link sent to you in an email11:46
pedroalvarezssam2: oh really? is that wokring?11:49
ssam2should be11:49
pedroalvarezthat's amazing :)11:49
ssam2django.registration :)11:49
CTtpollardall hail django11:50
*** lachlanmackenzie has joined #baserock11:55
*** persia_ has quit IRC12:00
*** rdale_ has joined #baserock12:08
*** persia_ has joined #baserock12:12
*** persia_ has joined #baserock12:12
*** rdale__ has joined #baserock12:13
*** rdale_ has quit IRC12:16
*** rdale has quit IRC12:19
*** dabukalam has quit IRC12:19
*** perryl has quit IRC12:19
*** dabukalam has joined #baserock12:20
*** perryl has joined #baserock12:20
*** mdunford has quit IRC12:22
*** mdunford has joined #baserock12:23
*** lachlanmackenzie has quit IRC12:27
jjardonI can not assign tasks in storyboard, is this a known issue?12:32
* jjardon thinks mayb its better to report this in the storyboard instance instead ;)12:32
ssam2jjardon: hmm, i'm not sure why12:34
ssam2maybe there are permissions that need to be set up12:35
ssam2I'll have a look later12:35
ssam2I've written descriptions for all the baserock/ repos at <>: could everyone check that they're sane, and that the repos I've marked as obsolete are indeed obsolete ?12:35
*** lachlanmackenzie has joined #baserock12:42
Zarathey seem fine to me12:46
jjardonssam2: I can assign now: it works if you press enter instead click in the pop up12:48
jjardonssam2: about the repo description, Id change "Baserock Robotic Integration Engineer" for "Baserock CI/CD system"12:51
Kinnisonfirehose isn't CI?CD12:52
Kinnisonit's a robotic integration engineer12:52
jjardonmy misunderstanding, nevermind then12:53
petefothjjardon ssam2 persia - I've created the "What *I* do with Baserock" stories on the wiki" story in storyboard, with tasks fro thoise who volunteered in yesterday's UI & Docs session!/story/612:53
petefothI have also copied the text for the story in case it disappears :)12:54
jjardonpetefoth: assigned myself already ;)12:54
ssam2I find it really bizarre how when I loaded that page in Storyboard it said 'No title' for a while, then the content appeared12:54
radiofreeare you Storyboard Monkey petefoth12:54
ssam2hmm, petefoth maybe you need to log out and back in12:55
petefothradiofree: looks that way, but there'll be another ID along in a moment12:55
ssam2Could you log out and log back in? You should appear as yourself then12:55
* petefoth does as he is told, for the second time today12:55
ssam2if you still show up as 'storyboard monkey' then something is broken12:56
jjardonthe stories have a "start" when you can "subscribe". What that really means? I will get mail notifications of changes?12:56
ssam2(the story you already made will still have been created by 'storyboard monkey' but new actions should occur under your name)12:56
petefothssam2: I'm now showing up as me, the stories are still credited to Mr Monkey, which is fine by me12:57
petefothssam2: and thanks for getting this working :)12:57
Kinnisonpaulsherwood: thank you for the writeup on declarative definitions, I have referenced it in my writeup about build/deploy time service configuration13:02
* Kinnison now nips off13:02
Zarare: the UI meetup notes, who's sending those to the mailing list? perryl wrote them by hand and Krin announced them13:18
KrinZara, i am intending to do a write-up but i'd also like to pass them around and ensure that everyone remembers thm the same as i do :)13:19
Zaracool, and thanks :) Just wanted to make sure it wasn't a case where we all thought someone else was doing it!13:19
petefothKrin: send them to the list - if anyone disagrees they can disagree there :)13:19
jjardonKrin: send to the list, if there is some mistake the people will crrect them13:20
Krinok, havent goten round to writing them up yet, been going through what IS on the list, but it'll be up before end of day13:20
* SotK is in the process of writing up notes from the licence tracking chat13:22
jjardonred in !  the only error I see is "2015-02-06 12:42:16 Build failed: git-misc" more no more info. Would it be possible to get more logs?13:37
ssam2jjardon: I think the git build didn't actually fail, something else is broken13:38
ssam2when you see 'Progress: Transferring git-misc to shared artifact cache' and *then* 'Build failed: git-misc' it means something went wrong when transferring the artifact to the cache13:39
ssam2should give a better error there really13:39
ssam2 /home on is full again13:39
ssam2we really need to write the tool that cleans out old artifacts.13:40
jjardonah, ok, thanks for looking into this!13:40
ssam2i removed all the system artifacts and there is 26G free now13:40
* jjardon was worried he has broken the build13:40
jjardoncan anyone take a look why the GTK+ mirror is not up-to-date?
ssam2jjardon: the logs show an error from git '    error: src refspec refs/heads/places-sidebar does not match any.'13:59
ssam2one of git's characteristic user-friendly messages13:59
ssam2actually, that's coming frm the .lorry file:
ssam2seems that a while back, we added certain 'allow force pushes to these refs' rules to the .lorry14:01
ssam2these days I think we just allow force-pushing everything, because it's too hard to predict how upstreams will use git14:01
ssam2so, you could fix this by sending a patch for gtk+.lorry that changes the 'refspecs' field to ["+refs/heads/*", "+refs/tags/*"]14:03
tiagogomes_jjardon, you are updating again chunks which I already had updated in a branch :)14:08
jjardontiagogomes_: sorry, I didn't see any branch in definitions repo14:10
jjardonssam2: will do, thanks to looking into this!14:11
*** grahamfinney has quit IRC14:12
*** grahamfinney_ has quit IRC14:12
*** grahamfinney_ has joined #baserock14:19
*** grahamfinney has joined #baserock14:19
*** grahamfinney_ has quit IRC14:22
*** grahamfinney has quit IRC14:24
*** grahamfinney has joined #baserock14:24
*** bashrc has quit IRC14:27
*** bashrc has joined #baserock14:27
*** CTtpollard has quit IRC14:45
*** CTtpollard has joined #baserock14:48
jjardonevery how many minutes is refreshed? Seems it got stuck at 13:36:2715:26
SotKjjardon: its probably in the middle of doing a build15:28
rdale__bother, i just send the wrong patch by accident to the baserock-dev list, so i tried to say 'please ignore' to the wrong email, making the mess even worse. my apologies15:29
pedroalvarezjjardon: it is building 'attr' now15:30
paulsherwoodrdale__: i'm assuming you mean the one that has a /baserock/james/0.10 unpetrify ref?15:31
rdale__yes, that one should be ignored, javier's is fine15:32
jjardonrdale__: :)15:35
rdale__sorry jjardon15:35
jjardonSotK: pedroalvarez sorry for my impatience ;)15:35
jjardonrdale__: np!15:35
pedroalvarezjjardon: you are right, it would be nice to know what is going on15:36
jjardonpedroalvarez: should I open a story in storyboard?15:36
pedroalvarezjjardon: given that this implementation of mason is going to be obsolete... I don't know15:39
ssam2the new Mason will need to satisfy that story too though15:40
ssam2if it does, we get to close the story :)15:40
jjardonwe can create a "Improve CI reports"15:40
jjardonyah,, what Sam said15:40
SotK+1 to a story for better CI reporting :)15:42
ssam2jjardon: I'm not sure your gtk+ patch is correct15:43
ssam2you changed it to use whatever the default value of 'refspecs' is15:43
ssam2I don't know where the default value of 'refspecs' is documented15:44
rdale__it's in the trove reference manual - i was reading that  this morning15:44
ssam2but if it defaults to ["refs/heads/*", "refs/tags/*"] then the .lorry still won't work, it'll break the first time someone force-updates a branch or tag15:44
ssam2ah, cool15:44
ssam2it says 'Default is all.'15:45
ssam2but I don't think 'all' is even a valid value for a git refspec15:45
ssam2'man git push' doesn't say anything about it15:46
ssam2I think you should set refspecs to ["+refs/heads/*", "+refs/tags/*"] so that all force-updates are allowed though15:46
*** CTtpollard has quit IRC15:46
pedroalvarezssam2: hm... I think it will fail to push whatever needs a force push, but everything else will be ok. This is only a problem with some upstream projects where master and tags are rebased.15:47
ssam2ah, ok15:48
ssam2so force-pushed refs would not get updated, but it wouldn't actually raise errors and cause the whole repo to stop being updated15:48
pedroalvarezI think15:49
ssam2this needs to be clearer in the trove reference manual15:49
ssam2but I can't do that until I actually understand how it works :/15:49
pedroalvarezas a thought: should we create branches to anchor commits used in definitions on those repos where we allow force pushing?15:50
pedroalvarez(I hope 'anchor commits' makes sense)15:51
ssam2i'm planning on a command for Morph which creates 'anchor tags' for every ref involved in the build of a given system15:51
paulsherwoodi'm just a bit worried we may flood repos with tags... before realizing that's a bad idea, and be unable to clean them up because policy15:52
pedroalvarezso, whenever you  do a release, and you want it to be reproducible over the years, you do that to create a branch in every repo used?15:52
ssam2yes, that's possible15:52
ssam2it might be that we decide not to use it on yet15:52
ssam2a customer has asked for this to use in their trove, and hopefully it'll be useful for the baserock project too15:53
pedroalvarezyeah, it can be really useful for some people15:53
jjardonssam2: +100015:54
*** CTtpollard has joined #baserock15:57
*** mdunford has quit IRC15:59
paulsherwoodi'm not against it, i just hope folks try it out properly and get feedback from realistic situations so we don't actually damage user experience....15:59
paulsherwoodfor example
paulsherwoodour default Tag view shows upstream tags, which is nice. would be a shame to flood those off the page with temporary or user/specific tags16:00
ssam2I agree that we need to be careful about that16:00
ssam2if I was writing a policy for this, I'd say we only use the command when making releases of the baserock reference systems16:00
ssam2so there'd be at most 1 new tag a week16:00
ssam2but that still means we'd lose upstream tags16:01
paulsherwoodwithin 10 weeks :)16:01
pedroalvarezI wouldn't use it in g.b.o :)16:01
ssam2one possibility is we could do this in a different git server16:01
ssam2great minds think alike :)16:01
pedroalvarezbut I think is a useful thing to have for some user16:11
pedroalvarezis there a way to anchor commits without creating noise?16:11
pedroalvarez(tags or branches)16:11
ssam2we could create 'hidden' refs16:15
ssam2like refs/anchor/foo16:15
ssam2but there might be corner cases with using those16:16
ssam2worth investigating16:16
*** mdunford has joined #baserock16:26
KinnisonIt'd also need checking on mirror tactics etc16:38
SotKIn which project should I put the CI story on storyboard? baserock/infrastructure maybe?16:41
tiagogomes_jjardon, !! vala now?16:49
tiagogomes_tiagogomes_ jjardon I was planning to update that one, but thanks for doing it for me :)16:49
SotKjjardon: I was going to update FUSE soon too :P16:50
jjardontiagogomes_: :) you should submit patches upstream as soon as possible16:50
jjardonSotK: ^16:50
tiagogomes_jjardon the patches are not ready.16:50
SotKjjardon: :)16:51
*** CTtpollard has quit IRC16:54
tiagogomes_jjardon, I updated in a branch gdb, make, lzo, file, gdbm, texinfo, libexpat, btrfs-progs, strace16:54
*** mariaderidder has quit IRC17:01
*** tiagogomes_ has quit IRC17:05
*** gary_perkins has quit IRC17:07
*** petefoth has quit IRC17:07
*** tiagogomes_ has joined #baserock17:07
*** petefoth has joined #baserock17:08
*** gary_perkins has joined #baserock17:09
*** ssam2 has quit IRC17:22
*** inara has joined #baserock17:23
*** mauricemoss_ has quit IRC17:23
*** flatmush has quit IRC17:23
*** Krin has quit IRC17:23
*** bashrc has quit IRC17:23
*** mauricemoss_ has joined #baserock17:24
*** bashrc has joined #baserock17:24
*** flatmush has joined #baserock17:24
*** Krin has joined #baserock17:24
*** jonathanmaw has quit IRC17:29
jjardontiagogomes_: branch in definitions? Send it to the list!17:50
*** bashrc has quit IRC17:58
*** inara has quit IRC17:58
tiagogomes_jjardon baserock/tiagogomes/armv8l6417:59
*** Krin has quit IRC17:59
*** inara has joined #baserock18:02
jjardontiagogomes_: nice18:03
tiagogomes_jjardon do you now how can I update the config.{guess,sub} scripts when they are not in the top directory of the project18:05
jjardontiagogomes_: they should be generated by autoreconf, they should normally be in the git tree18:06
jjardonshould not*18:07
*** inara has quit IRC18:13
tiagogomes_jjardon nope, most projects have them in the git tree18:14
jjardontiagogomes_: The one in the tree are not needed if you are bootstrapping the project (which is what you want to do if you compile from git) config.guess and config.sub are created by automake (which is called in the appropiate order by autoreconf). The fact that some project include them in the git tree doesnt mean they are needed.18:19
jjardonSome project include them in case you dont want to bootstrap, but use ./configure directly. (similar as if they have the contents of the tarball in the git tree)18:21
pedroalvarezjjardon: if you are wondering: Feb 06 18:21:20 mason-x86-64[15374]: 2015-02-06 18:21:20 Progress: Transferring xorg-lib-libXdamage-misc to shared artifact cache18:21
*** mauricemoss_ has quit IRC18:22
jjardonpedroalvarez: nice!18:22
*** rdale__ has quit IRC18:24
tiagogomes_jjardon autotools is a pita, most projects have already have a configure script and makefiles even in git; I need to update config.guess and config.sub scripts but autoreconf rarely works18:39
jjardontiagogomes_: or you use, so you dont need config.sub/guess or you use configure and you dont have to change the config.guess/sub18:40
jjardonevery project should build with autoreonf, if not its a bug18:41
tiagogomes_jjardon configure calls config.guess and config.sub18:41
tiagogomes_jjardon, then I can assure you that almost every project has a bug. The thing here is, automake doesn't have backwards compatibility18:42
jjardontiagogomes_: yes, autoreconf call the toold that generate configure, config.guess and config.sub18:42
tiagogomes_jjardon, yep, but autoreconf almost never succeeds18:43
jjardontiagogomes_: thats probably a bug18:44
jjardontiagogomes_: with what packages are you have problems?18:44
tiagogomes_which gives me two choices, or either I download the scripts from here:; or either I copy them from /usr/share/automake1.418:44
jjardontiagogomes_: both are bad ideas18:44
tiagogomes_the first introduces a delta -> meh; the second breaks when we update automake -> meh18:44
tiagogomes_I agree18:44
jjardontiagogomes_: so with what packages are you having problems so I can take a look?18:45
*** inara has joined #baserock18:45
jjardontiagogomes_: 3 option is to fix the package if needed and sent the patch upstream ;)18:45
tiagogomes_jjardon cvs-tarball, distcc, ncurses, groff, openssh18:46
jjardoncvs-tarball? why you need that for?18:47
*** grahamfinney has quit IRC18:48
tiagogomes_jjardon it is on devel system18:48
* jjardon looking18:48
jjardontiagogomes_: ok, if its a tarball you should use ./configure18:50
*** grahamfinney has joined #baserock18:51
tiagogomes_jjardon again, config.guess and config.sub are tool old, they don't recognize aarch6418:51
*** edcragg has quit IRC18:52
jjardontiagogomes_: oh! you should bootstrap then18:54
*** inara has quit IRC18:54
tiagogomes_jjardon, I know! but our version of automake doesn't understand some old's and ':(18:55
*** grahamfinney has quit IRC18:55
jjardontiagogomes_: I fear you have to be more specific about the problems you are having ;)18:56
jjardontiagogomes_: do you have any logs around?18:56
*** tiagogomes_ has quit IRC19:00
jjardonmmm, we are using tarballs in most of those packages: first step would be to use the git repo. Also openssh mirror is not up-to-date19:01
*** HoloIRCUser1 has joined #baserock19:02
jjardonwould be ok this patch to update the openssh lorry?:
*** lachlanmackenzie has quit IRC19:24
*** gary_perkins has quit IRC19:31
*** HoloIRCUser3 has joined #baserock19:37
*** HoloIRCUser1 has quit IRC19:38
*** HoloIRCUser1 has joined #baserock19:43
*** HoloIRCUser3 has quit IRC19:46
*** inara has joined #baserock19:50
*** inara has quit IRC19:55
*** inara has joined #baserock19:59
*** inara has quit IRC20:04
*** inara has joined #baserock20:07
*** inara has quit IRC20:16
*** inara has joined #baserock20:26
*** inara has quit IRC20:43
*** HoloIRCUser1 has quit IRC21:02
paulsherwoodjjardon: i think that would lead to different refs vs the cvs convert we have been using?21:03
paulsherwoodmaybe best to call it openssh-git for this, stop lorrying the cvs?21:04
*** tiagogomes_ has joined #baserock21:24
*** tiagogomes_ has quit IRC21:29
paulsherwood2015-02-06 20:13:26 [Build 82/209] [systemd] Elapsed time 00:29:5921:42
paulsherwoodseems like a lot longer than last time i checked21:43
*** zoli__ has joined #baserock21:47
*** zoli__ has quit IRC21:53
*** zoli__ has joined #baserock21:53
*** zoli__ has quit IRC22:34
*** zoli__ has joined #baserock22:35
*** zoli__ has joined #baserock22:36

Generated by 2.15.3 by Marius Gedminas - find it at!