IRC logs for #baserock for Saturday, 2015-09-26

pedroalvarezHm.. ci.morph broken (see mason)07:38
pedroalvarezAnd no more morph edit07:38
pedroalvarez:/ I will have to fix the license check in definitions.git07:38
paulsherwoodcan i help?07:39
pedroalvarezI'll send a quick patch for ci.morph that you can review :)07:40
pedroalvarezhey tlsa :)07:50
tlsaI thew this together last night:
tlsait's not actually monitoring what's happening on the system07:50
pedroalvarezThis should put mason green:
pedroalvareztlsa: hey, that's nice07:50
tlsajust a demo of what the UI could look like07:50
pedroalvarezI guessed that (it's not live :P)07:51
pedroalvarezhave you found the way to get all the information needed from buildbot?07:52
paulsherwoodtlsa: fab. do you hate rounded corners?07:52
tlsait produces json, which could be parsed and used to generate the page07:52
paulsherwoodpedroalvarez: merged07:53
pedroalvarezpaulsherwood: ta!07:53
paulsherwoodi like the font choice... reminds me of Blake's Seven :)07:53
tlsapaulsherwood: now with rounded corners07:54
paulsherwoodmaybe a teesy bit smaller rounds?07:58
tlsameh, it's triavially changed, the important thing is somone would need to hook it up so it shows whats currently happening07:59
paulsherwoodi wonder how best to show the parallel tracks. buildbot waterfall is actually quite confusing for me08:00
paulsherwoodmaybe it could just be lots of those sets-of-four blocks. provide an identifier/label for each08:02
tlsaI was going to do something like:08:02
tlsa[Firehose]----[Build x86]----[Deploy]----[Test]08:03
tlsa           \08:03
tlsa            `-[Build arm]08:03
paulsherwoodah, ok. and if n targets, n of those branches?08:03
tlsabut I did not have enough time to spend on it, and am on something else now08:03
paulsherwoodi like the ideas08:03
tlsaWell, I'd hae collapsed it to just show one "Build" block unless you click on it to "open" it, or unless the ones under "build" have mixed pass/fail states08:04
paulsherwoodeven better08:06
paulsherwoodthat would be pretty cool, actually.08:06
* pedroalvarez imagines a UI full of boxes and lines, representing the complexity of CIAT but with some buttons to filter per architecture, firehose sets, and more things to see simpler views of it08:07
pedroalvarezmore things are broken from the gnome system patch08:12
pedroalvarezI guess jjardon will want to lorry m4-common08:13
* SotK tries to make sense of BuildBot's JSON API09:16
tlsaSotK: You can get whether they're idle or not from:
tlsaand you can get the number of the last cached build09:20
tlsaso for "2. x86_64 Build", that's 2409:20
tlsaso you can construct this URL from that:
tlsaand from there you can extract whether or not it was "successful"09:22
SotKtlsa: brilliant, thanks!09:30
SotKIs there a way to figure out how the pipeline is arranged beyond guessing based on the names?09:58
* SotK has to stop for a bit now, I shoved some javascript into tlsa's demo UI from earlier:
paulsherwoodcool :-)10:27
tlsaSotK: nice :)10:34
pedroalvareztlsa: now that you are around, would you be so kind of giving me access to the ciat machine? :)10:34
tlsapedroalvarez: the machine on aws?10:35
pedroalvareztlsa: yes, please :)10:35
tlsalet me see10:36
* paulsherwood thinks that maybe kbas could serve git repos to other ybd instances too, save everything hitting g.b.o10:43
*** zoli__ has joined #baserock11:05
pedroalvarezthanks tlsa!11:06
tlsaSotK: that won't set the "active", class yet, if I read it right11:06
tlsapedroalvarez: yw11:06
*** SotK_ has joined #baserock11:34
tlsahi SotK_11:34
SotK_tlsa: correct, it'll be simple to do though11:34
* SotK_ had to go out before he could do it11:35
tlsayou can remove the ",400" from the font URL11:36
tlsaI decided not to use that weight for the body text; it looked nasty11:36
*** SotK_ has quit IRC11:41
pedroalvarezjjardon: thanks for fixing m4-common, although it may fail because it has .gitmodules12:30
jjardonI wonder if  it would be better to improve morph/ybd to make the translation automatically instead us patching all the repos that have .gitmodules12:40
jjardonAny idea if this has been discussed before?12:40
paulsherwoodi don't think it has, jjardon12:44
paulsherwoodyou're suggesting replace with upstream:submodule-name by default ?12:46
* jjardon just realize some of the gnome chunks are probably using gitmodules as well12:46
jjardonpaulsherwood: yes12:46
paulsherwoodseems sensible to me12:48
pedroalvarezI remember some discussions to add some functionality to morph to solve this problem13:21
jjardonmmm, Im getting a weird error here:
jjardonany idea what I can be doing wrong?13:26
jjardonmmm, it happen with current master as well13:27
jjardonoh, this is probably related with some of the recent changes in morph13:29
SotKIf a build is in progress, should the colour of the box indicate the success/failure of the previous build?14:08
tlsait should always show the result of the previous build14:08
pedroalvarezjjardon, paulsherwood: regarding replacing url's, we can't do that for all of them, given that we haven't been consitent  when creating lorries14:08
tlsaand if there's an active build in progress, set the active class14:08
pedroalvarezjjardon, paulsherwood: although we could use lorries.git to get that info I guess14:09
pedroalvareztlsa: so, if building, and last build successful, green and the in-progress indicator?14:10
tlsaand if last build failed, and its now doing a new build: red + flashing14:10
SotKtlsa: done :)14:14
SotK(not the failed part actually, is there an example of a failed build somewhere?)14:14
tlsalet me see14:15
tlsahas "failed"14:16
tlsaif the builder has no history (say its just been restarted), then neither the pass nor fail class should be set :)14:20
pedroalvarezthere is another status I believe: interrupted14:20
tlsa(that should show the box as gray, rather than red/green)14:20
tlsapedroalvarez: that would show as gray too14:21
* pedroalvarez nods14:21
tlsaantway, noce work SotK :)14:21
pedroalvarezindeed, looking good14:21
tlsaSotK: does it keep refreshing automatically?14:22
SotKtlsa: not yet14:22
SotKany opinons on a sensible refresh rate?14:22
tlsa1 min?14:23
SotKseems sensible14:26
tlsa'fail': !step.lastBuild.success  maybe needs to be  'fail': !step.lastBuild.failed14:27
SotKgrr, I updated angularjs and now the lastBuild isn't working properly14:57
tlsaSotK: if you update to the CSS from it should improve the spacing a bit15:23
SotKtlsa: thanks!15:28
SotKshould be auto-refreshing the data once per minute now :)15:41
*** pedroalvarez has quit IRC18:03
*** pedroalvarez has joined #baserock18:04
*** ChanServ sets mode: +v pedroalvarez18:04
pedroalvareztlsa: so, was something like this your idea to add support to specify users?
*** _longines has quit IRC19:16
pedroalvarezI've decided to do it that way, although the wait_for_ssh also needs an user to do the actual check (which is ssh in and run true)19:39
pedroalvarezI guess that can be replaced with something else19:39
pedroalvarezmaybe check that there is something in the port 2219:41
*** zoli__ has quit IRC21:34
*** zoli__ has joined #baserock22:01
*** zoli__ has quit IRC22:05
*** zoli__ has joined #baserock22:06
*** zoli__ has quit IRC22:53
*** zoli__ has joined #baserock23:49

Generated by 2.14.0 by Marius Gedminas - find it at!