IRC logs for #baserock for Tuesday, 2015-08-11

*** paulw has joined #baserock06:50
paulsherwoodi've just watched ssam's baserock talk at
paulsherwoodi guess now would be a good time to put ybd on g.b.o officially?07:17
paulsherwood(and adopt gerrit for changes to it?)07:18
paulsherwood(unless there are other plans...07:18
*** jonathanmaw has joined #baserock07:51
tiagogomesout of curiosity, under a delta or baserock/baserock prefix?07:58
*** mike has joined #baserock08:01
*** mike is now known as Guest6764508:02
*** bashrc has joined #baserock08:02
paulsherwoodi would be happy with either. revies/merges have been relatively easy-going on github ... but moving to g.b.o. woudl be a sensible time to formalize how it's reviewed, and i think the contributors would be happy to adopt baserock's gerrit review process if we agree it should be a baserock/baserock project08:04
*** mdunford has joined #baserock08:06
tiagogomesI think this boils down to whether or not ybd should be officially part of the Baserock tool set08:36
*** CTtpollard has quit IRC08:38
paulsherwoodagreed... what do you think?08:39
*** CTtpollard has joined #baserock08:41
*** edcragg has joined #baserock08:58
pedroalvarezI've been recently investigating a project, with different tools with different names to do the same thing, and it was very confusing08:59
tiagogomespaulsherwood, I think that having two tools to do the same job will cause confusion to the users about which tool to use. It will also increase the maintenance burden. If ybd grows up to be as feature complete as morph, then it could replace the latter as the official tool to use. But for the time being I would prefer it to have a delta prefix09:01
*** rdale has joined #baserock09:03
paulsherwoodok by me09:11
*** edcragg has quit IRC09:16
*** edcragg has joined #baserock09:21
*** ssam2 has joined #baserock09:28
*** ChanServ sets mode: +v ssam209:28
pedroalvarezssam2: hey, good talk :)09:35
ssam2i didn't realise you were watching!09:36
ssam2but thanks  :)09:36
paulsherwoodssam2: what was the feedback like at the event?09:36
ssam2oh, that's the europython talk!09:36
ssam2i thought you  meant GUADEC for a second :)09:36
* De|ta stops looking for GUADEC videos09:37
ssam2feedback was mostly good, lots of discussion of related projects09:37
ssam2e.g. a couple of people thought integration into Packer would be amazing09:37
paulsherwoodhow many loc is that? :)09:38
ssam2we discussed NixOS and OpenStack's diskimagebuilder too09:38
ssam2paulsherwood: Packer, or the proposed integration? it has a plugin API so probably not too hard09:38
paulsherwooderk, written in Go? :)09:39
* ssam2 updates
*** Guest67645 has quit IRC09:45
*** ssam2 has quit IRC09:51
pedroalvarezpaulsherwood: so, I take that you want ybd to stay in GitHub instead of moving it to g.b.o baserock/ and gerrit?10:02
paulsherwoodpedroalvarez: i don't mind, but tiagogomes said above he would prefer it to stay as a delta project10:04
paulsherwoodi'm happy to go wihchever route is best for the community.10:04
pedroalvarezThe only thing I don't want is a confusing documentation :)10:05
paulsherwoodpedroalvarez: i've held off documenting ybd so far...but in any case its documentation will be much simpler than morph's (it has far fewer options)10:08
paulsherwoodpedroalvarez: do you have a preference?10:08
paulsherwood(delta or baserock/baserock)10:09
richard_mawCan we defer this a couple of weeks? I'll be in a position to have an opinion by then10:10
pedroalvarezok by me :)10:10
paulsherwoodi was mainly reacting to the talk mentioning that ybd is on gbo. the current repo there is deprecated10:11
pedroalvareztiagogomes: would you +1 if I fix the "import warnings" thing?11:54
pedroalvarezI'll fix it anyway11:56
tiagogomespedroalvarez yep11:57
pedroalvarezhere we are:
pedroalvarezI've also sent a new version of this one:
*** rdale has quit IRC13:35
pedroalvarezKvm deployment failed with the 'rawdisk-partitioning' changes13:44
pedroalvarezI've created a list of things that we should test for that change to be merged:
tiagogomescool, that indeed needs lots of testing before being considered to be merged13:47
*** rdale has joined #baserock14:06
*** lachlanmackenzie has joined #baserock14:52
persiaJust to add my opinion to the mix: I'm of the opinion that ybd should either be a core baserock project (managed by baserock, and not at github), or not lorried and used as a potential source of implementation ideas for build tools.  I don't have a strong opinion as to which is the right model.15:01
paulsherwoodpersia do you mean 'not lorried and just used' or 'not lorried and not used'?15:04
*** paulw has quit IRC15:15
persiapaulsherwood: Not lorried and not used: we need an integration tool we control.15:33
persiaThat can be ybd, or not be ybd: I don't have an opinion there, but I don't like using something that we don't control with our development processes.15:37
paulsherwoodcontrol in open source seems like an odd approach, but fine :)15:37
*** sambishop has joined #baserock15:37
paulsherwoodwe seem to be using *quite a lot of stuff* that is not baserock/baserock :)15:38
paulsherwoodbut in any case as i've said i'm happ for ybd to move under baserock processes if folks want to do that15:39
*** rdale has quit IRC15:39
*** rdale has joined #baserock15:41
nowsterdoes "core-openssl" exist?   systems/chef-system-x86_64-container.morph15:44
pedroalvarezhm.. wait15:46
nowstercore does not currently have anything that produces a "core-openssl" artifact15:46
persiapaulsherwood: There is a difference between using a compiler and using a tool that performs what we consider a core function.  To me, the wrong choice would be to have baserock/baserock/morph and delta/ybd/, as the functionality is nearly identical, and from past discussions, the projects seem to compete.15:48
persiaBut if we have a tool for a core function that we cannot use our process to adjust, then I think we need to step back and look at why we declare that function core: perhaps the function doesn't belong in baseock at all (in which case, delta/ is fine).15:49
persiaIf that function *does* belong in baserock, or is, as some have argued, the primary function of baserock, than the baserock project isn't very interesting if it does not develop that tool.15:49
nowsterpaulsherwood: having thought on this a bit more, the "splits" in strata are more akin to macros which simplify system definitions.15:56
paulsherwoodnowster: and are 'splits' in chunks, the same?15:57
nowsterno, because they operate on files15:57
nowstermeta-splits in strata operate on chunks15:57
paulsherwoodomg. wht is a meta-split?15:59
nowstera group of splits15:59
paulsherwoodhow obscure can we make these names?15:59
richard_mawpaulserwood: it's not a concept we currently have16:00
nowsterThe most odd is the build-essential-minimal one which selects the fhs structure, busybox binary, and only part of glibc, and nothing else out of build-essential16:00
paulsherwoodnowster: others are telling me this is broken. maybe let's drop it? :)16:24
*** jonathanmaw has quit IRC16:29
*** mdunford has quit IRC16:48
*** flatmush has quit IRC16:59
*** bashrc has quit IRC16:59
*** flatmush has joined #baserock17:02
*** edcragg has quit IRC17:07
*** lachlanmackenzie has quit IRC17:24
pedroalvarezoh, I was going to set up an organization in Docker to try to push a Baserock Docker system, and I found that there is already one :)21:51
pedroalvarezI've been thinking about creating a system with an user that has docker configured as its user shell, so that everytime you log in you enter in a fresh docker container. This in combination with something like shellinabox might be cool as a "try baserock" thing21:56

Generated by 2.15.3 by Marius Gedminas - find it at!