IRC logs for #baserock for Monday, 2015-11-23

*** gary_perkins_ has quit IRC08:02
*** flatmush has quit IRC08:02
*** nowster has quit IRC08:02
*** CTtpollard has quit IRC08:02
*** nowster has joined #baserock08:02
*** CTtpollard has joined #baserock08:03
*** flatmush has joined #baserock08:03
*** gary_perkins has joined #baserock08:03
*** ctbruce has joined #baserock08:40
straycatpedroalvarez, shall we merge the 15.47 into infrastructure?08:42
straycat*15.47 tag08:42
straycatalso i have no idea what's going on with mason08:44
*** paulwaters_ has joined #baserock08:45
*** tiagogomes_ has joined #baserock08:54
tiagogomes_straycat are you planning to get rid of the branch and merge plugin? I was trying to do the same08:55
*** bashrc_ has joined #baserock09:02
straycattiagogomes_, i think that's definitely the plan yes, so if someone else doesn't do it then i might09:03
*** fay_ has joined #baserock09:04
tiagogomes_straycat getting rid of the plugin is easy once you change the tests to work workspaceless :)09:04
straycatokay cool :)09:07
pedroalvarezstraycat: yes. Do you want me to do it?09:11
tiagogomes_straycat ok. I had a branch to make yarns to not use workspaces. I'll drop it. I can work on the removal of the branch and merge plugin and workspace code09:12
straycatpedroalvarez, no it's okay i have a work in progress branch that i've already made some changes to, some manual fixes will be needed and i need to test them anyway to deploy this gerrit system09:12
straycattiagogomes_, okay that's annoying i didn't know you were doing that, sorry :/09:12
straycattiagogomes_, so far we've been small enough that it's seemed possible to just work on whatever seemed important without duplicating effort09:13
*** franred has joined #baserock09:13
straycattiagogomes_, removal of branch and merge plugin code would be really cool :)09:14
tiagogomes_straycat yep. I'll only started Friday afternoon to fix the tests to see how it goes, so not much time to announce what I was doing09:14
* straycat nods09:16
*** franred has quit IRC09:24
pedroalvarezI'm worried about mason having to rebuild a lot of things after the migration to definitions v709:35
pedroalvarez(this is more clear in https://mason-x86-32.baserock.org/ )09:35
*** rdale has joined #baserock09:38
*** franred has joined #baserock09:39
tiagogomes_straycat, what `git checkout -` does?09:49
pedroalvarezlike `cd -` but with branches?09:50
straycatyes09:51
*** jonathanmaw has joined #baserock09:51
straycattiagogomes_, you could probably remove them since most of the commands say what branch the operation should be performed in, but i just put them in so we restore the original state just in case09:52
tiagogomes_straycat ta09:53
*** franred has quit IRC09:56
*** franred has joined #baserock09:58
pedroalvarezha, i know what's going on09:58
pedroalvarezwe added the following in buildsystem.py:09:59
pedroalvarez            'elif [ -e bootstrap ]; then ./bootstrap; ' +09:59
pedroalvarez            'elif [ -e bootstrap.sh ]; then ./bootstrap.sh; ' +09:59
pedroalvarezbut those are not in DEFAULTS09:59
*** gtristan has joined #baserock10:02
*** ssam2 has joined #baserock10:02
*** ChanServ sets mode: +v ssam210:02
pedroalvarezi don't know how to deal with this :/10:07
pedroalvarezI wonder if ssam2 has any suggestion about it (context: http://paste.baserock.org/efifezokut)10:09
ssam2fix DEFAULTS10:12
ssam2hmm, better fix migrations/initial-defaults-whatever too10:12
pedroalvarezwill do10:12
*** toscalix has joined #baserock10:13
pedroalvarezright: https://gerrit.baserock.org/148510:18
pedroalvarezalso, this one fixes jetson builds: https://gerrit.baserock.org/#/c/1484/10:18
pedroalvarezturns out the migration script was OK, but the migration patch was old, and needed a new version10:19
*** locallycompact has joined #baserock10:20
pedroalvarezThe other fix, (for jetson systems) is needed now that morph complains if there are duplicated chunk names10:20
pedroalvarezhttps://gerrit.baserock.org/148510:27
pedroalvarezgrr10:27
pedroalvarezthanks for reviewing jjardon10:27
pedroalvarezanyone else for  https://gerrit.baserock.org/#/c/1484/ and https://gerrit.baserock.org/1485 ? :) :)10:27
richard_mawseems a little odd that we don't just split the device tree compiler chunk out as we ought to be able to use the same compiler10:29
richard_mawbut I do remember it not being entirely API stable, so I'll +110:29
pedroalvarezthanks richard_maw :)10:31
pedroalvarezthis is being the bugfixing morning :)10:37
pedroalvarezAnother one: https://gerrit.baserock.org/148610:38
richard_maw+2'd10:38
pedroalvarezgood :)10:43
straycatpedroalvarez, the branch i have for the infra is here http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/infrastructure.git/log/?h=baserock/richardipsum/merge-baserock-15.4710:49
straycatif folks would rather i sent the request for merge to the ml then let me know10:50
straycatoh dear, so we are in a situation where the release can be built with 15.34 but cannot build itself?10:52
straycatfor jetson systems that is10:53
pedroalvarezseems I've forgotten how to merge branches10:57
pedroalvarezmerge: baserock/richardipsum/merge-baserock-15.47 - not something we can merge10:57
jjardonId suggest to do a baserock-15.47.1 "bugfix" release10:58
straycatpedroalvarez, eh?10:59
straycatjjardon, i think we probably should11:00
pedroalvarezstraycat: I'm just trying to merge your branch :)11:01
pedroalvarezit looks good, please merge it on master11:01
straycatpedroalvarez, ok merged, i somehow based it a few commits behind latest master, so rebased before merging to avoid another merge commit11:16
pedroalvarez:) ta11:21
pedroalvarezstraycat: any other change needed can go through gerrit11:21
pedroalvarezfeel free to ping me to get reviews on them11:22
* straycat nods11:24
straycatwill do thanks11:24
pedroalvarezstraycat: I'm afraid this one is for you to fix: https://mason-x86-32.baserock.org/log/ae754c67bfc3a278ab9c7e6d8955ebe057e93511--2015-11-23%2011:43:48.log11:45
straycatsigh11:47
straycati blame the reviewers >.>11:47
pedroalvarez:) sorry11:47
pedroalvarezhahah11:47
pedroalvarezwe have this check in mason because it's diffucult to spot11:48
tiagogomes_yikes11:49
tiagogomes_which change introduced that?11:49
straycatthe import deps change11:52
straycatbasically we need to push my local fork to baserock/master of the carton repo on gbo11:53
straycatso, merging http://git.baserock.org/cgi-bin/cgit.cgi/delta/carton.git/log/?h=baserock/richardipsum/allow-set-perl-carton-cpanm-opt into baserock/master and updating the sha should fix that11:58
straycatthe "fork" is actually only a one-line delta on upstream that allows an environment variable to be set11:59
pedroalvarezstraycat: should i fix it?12:00
straycatfix what?12:01
pedroalvarezdo you mind if I call the branch baserock/v1.0.21 ?12:01
straycatthat seems fine to me12:01
pedroalvarezgood :)12:01
straycatpedroalvarez, shall i send the definitions update?12:04
pedroalvarezstraycat: if you have that ready go ahead. I was only a commit msg from that12:06
straycati can send it later once i've seen it build12:10
pedroalvarezok :)12:11
*** petefotheringham has joined #baserock12:37
*** petefotheringham has quit IRC13:12
*** petefotheringham has joined #baserock13:13
*** petefoth1ringham has joined #baserock13:17
*** petefoth1ringham has quit IRC13:21
*** petefotheringham has quit IRC13:21
*** petefotheringham has joined #baserock13:22
*** petefotheringham has quit IRC13:24
*** toscalix has quit IRC13:24
*** petefotheringham has joined #baserock13:28
*** petefotheringham has quit IRC13:31
*** petefotheringham has joined #baserock13:32
*** petefotheringham has quit IRC13:35
*** petefotheringham has joined #baserock13:39
*** toscalix has joined #baserock13:39
*** edcragg has joined #baserock14:15
*** thrace_ is now known as thrace14:54
straycatssam2, can gerrit not use the java stratum that jmacs added?14:54
ssam2sorry, what's the context?15:04
ssam2i've no idea of the answer15:04
ssam2it works with the binary JRE currently15:04
pedroalvarezmy guess is that it wasn't used given that that java stratum was created 4 months later15:04
straycattiagogomes_, i meant to say that i think https://gerrit.baserock.org/#/c/929/ may have removed a little too much, we probably still want to test temporary build branches since it's used for local-changes15:05
straycati was working on something to partially restore that but it's incomplete15:05
straycatsome of those yarns were removed because they used edit, they could instead be rewritten to not use edit.15:07
tiagogomes_straycat maybe, but as we have a BuildBranch class, maybe we could unit test it instead?15:07
straycati don't know, for now since it's dead i wouldn't -1 removing it, but expect that we may end up restoring/rewriting some of the implementations15:09
tiagogomes_it will be always on git history15:09
straycatindeed15:10
tiagogomes_as a fun fact, morph currently doesn't delete temporary branches when not using workspaces :)15:10
* persia continues to think morph should fail loudly when it encounters uncommitted local changes in definitions15:11
straycatyes well i guess none of that is tested15:11
straycatpersia, it's user-defined at the moment15:11
straycatif you want that behaviour then you set --local-changes=ignore (which is what i tend to use at least)15:12
persiaYes.  Just in the context of removing code ... :)15:13
* straycat nods15:13
*** petefotheringham has quit IRC15:13
* tiagogomes_ likes that morph includes uncommitted changes by default15:13
*** petefotheringham has joined #baserock15:14
* persia thinks it supports workflows with invrequent commits, leading to lack of traceability15:14
straycatotoh, i get that it can be annoying to have to commit (and push) every little change, i think leaving it up to the user is actually okay for now15:15
persiaWhy do you have to push?15:17
* persia would very much not support any workflow suggesting push on every change, as that encourages `push -f`, which is bad15:17
straycatwithout local changes it's using the remote15:18
persiaRight then.15:18
persiaThat clearly should be disentangled.  I want a way to make a commit locally and be able to work from that.  Without that, the distributed nature of git is not helping.15:19
persiaAnd if we're not being distributed, then all the fussing with git isn't helping: we could just use tools that maintain a common store.15:20
straycat*nod* we'll get there i think :)15:21
ssam2I agree that Morph should be able to read the files from the repo the user is working in, instead of only looking in its locally cached version of that repo15:22
ssam2you might still need to push on every commit if you are using a remote builder15:22
ssam2which you probably need to do if you are building for non-x8615:22
persiaThat needs sorting.15:23
persiaI should be able to pass the build information in such a way that it doesn't require me to muddy the central infrastructure.15:23
persiaOtherwise every possible spelling mistake that got test-built ends up being stored on the internet forever, which doesn't sound appealing at all.15:24
tiagogomes_what is the purpose of the get-chunk-details command ?16:43
tiagogomes_seems useless to tme16:44
tiagogomes_s/tme/me16:44
straycatit seems to print a chunk's details to the screen :)16:46
ssam2was requested by a user16:46
straycatit does at least show what your repo alias gets resolves to16:47
* tiagogomes_ would like to know what was on the mind of that user16:47
straycat*resolved16:47
straycatand your ref16:47
straycatit does seem a little pointless16:50
tiagogomes_to me it seems a bit more than little16:51
ssam2the idea was so they could get the real URL for a chunk, i believe16:51
tiagogomes_mm, the alias are still hardcoded in morph right16:52
straycatyes, but you can also set them in the config i think16:52
richard_mawwhich is unfortunate, since it ought to belong in the definitions16:53
tiagogomes_that's what I was thinking16:53
tiagogomes_why didn't go into the DEFAULTS?16:53
tiagogomes_if they were there, it would weakness further the existence of the get-chunk plugin16:54
straycatif repo aliases are that confusing, we could also simplify it by just having an alias for the trove url, and leave the repo section of the url in tact, so instead of upstrea:perl you have trove:delta/perl16:54
tiagogomes_I never founded the aliases to be confusing16:54
tiagogomes_finded16:55
tiagogomes_no still not16:55
tiagogomes_argh16:55
* tiagogomes_ gives up16:55
straycatthe 'baserock' can be, i originally thought that was hardcoded to git.baserock.org, not $trove:baserock/baserock/<repo>16:55
straycatso if it was trove:baserock/baserock/<repo> that might be clearer? idk16:56
* edcragg found the aliases confusing16:56
edcragg'upstream' is a very vague alias16:56
tiagogomes_Well,  the mapping upstream -> delta can be confusing16:58
SotKare the aliases not build tool configuration rather than something that belongs in DEFAULTS?17:00
tiagogomes_I think that they should belong to DEFAULTS as they are used in definitions17:01
ssam2Having default ones in definitions.git makes sense, perhaps17:03
ssam2but they are a mechanism to separate the URLs in definitions from the real URLs17:03
* persia also found aliases confusing, and likes the use of "trove:delta/perl" or similar17:05
straycatpedroalvarez, is the jetson build/deployment with the bsp change looking okay?17:10
SotKI think I'd prefer that to putting default translations of the aliases in definitions17:11
ssam2persia: i thought you hated troves :-)17:11
persiassam2: I do, but that lets me then complain about useless-use-of-the-word-trove, and move to "git-server:delta/perl" :)17:11
pedroalvarezstraycat: yep17:11
persia(or local-mirror:delta/perl)17:12
pedroalvarezstraycat: is this for fixing the release?17:12
straycatpedroalvarez, okay, i think we should probably make a new branch for a point release of 15.47.1 by cherry picking http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/commit/?id=7d7c21db688eb73ecfd5c426fc5491a0e068e72c onto baserock/release/baserock-15.4717:12
straycatthen tag that17:13
pedroalvarezstraycat: we have had a similar situation in the past17:15
pedroalvarezwith the 15.1917:16
pedroalvarezI'm sad we didn't create a branch for tags 15.191 and .217:17
pedroalvarezs/191/19.1/17:17
*** ctbruce has quit IRC17:17
straycatso you're happy with that suggestion then?17:17
pedroalvarezI was going to suggest to cherrypick the patch on the same release branch17:17
pedroalvarezbut I'm h appy with the suggestion, yes17:18
pedroalvarezI don't think we need to release the systems this time though17:18
straycatokay, i don't mind, branches are cheap though17:18
pedroalvarezthey are! :)17:18
straycatyes i don't think we need to release the systems either17:19
straycatokay i'll do that then17:19
* tiagogomes_ sends https://gerrit.baserock.org/#/q/status:open+project:baserock/baserock/morph+branch:master+topic:rip-branch-and-merge17:20
tiagogomes_I'll have a look in the wiki and do the necessary updates17:21
straycatpedroalvarez, does http://git.baserock.org/cgi-bin/cgit.cgi/baserock/baserock/definitions.git/log/?h=baserock/release/baserock-15.47.1 look correct to you?17:34
pedroalvarezit does17:35
straycatgreat, i'll tag it then17:35
pedroalvarezthanks straycat17:36
*** tiagogomes_ has quit IRC17:43
*** franred has quit IRC17:51
straycatno problem, hopefully that's sorted now :)17:52
*** edcragg has quit IRC17:57
*** edcragg has joined #baserock17:58
*** jonathanmaw has quit IRC18:00
*** locallycompact has quit IRC18:01
*** bashrc_ has quit IRC18:04
*** rdale has quit IRC18:13
*** ssam2 has quit IRC18:40
*** gtristan has quit IRC18:50
*** edcragg has quit IRC19:03
*** toscalix has quit IRC19:12
jjardonHi!, any idea when mason will be updated? seems its failing at the moment after the recent upgrades19:26
*** locallycompact has joined #baserock21:34
pedroalvarezjjardon: it has been upgraded. We had some failures but everything has been fixed now21:53
jjardonAh! Nice!21:55

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!