| *** gtristan has quit IRC | 02:45 | |
| *** gtristan has joined #baserock | 03:36 | |
| *** gtristan has quit IRC | 07:21 | |
| *** gtristan has joined #baserock | 07:34 | |
| *** toscalix has joined #baserock | 07:39 | |
| *** jonathanmaw has joined #baserock | 08:14 | |
| *** locallycompact has joined #baserock | 09:33 | |
| *** bwh has left #baserock | 10:05 | |
| radiofree | hello | 10:27 |
|---|---|---|
| radiofree | is it easy to setup a trove that *isn't* a mirror of gbo, but has the same git repos | 10:27 |
| radiofree | we just copy the lorry files right? | 10:27 |
| locallycompact | yes | 10:28 |
| pedroalvarez | note, it has to be powerfull to lorry some svn | 10:28 |
| pedroalvarez | and | 10:28 |
| pedroalvarez | some lorries may be pointing to dead repos | 10:28 |
| pedroalvarez | repos/urls/tarballs etc | 10:28 |
| pedroalvarez | and | 10:28 |
| pedroalvarez | probably other reasons :) | 10:28 |
| pedroalvarez | and possible failures | 10:29 |
| radiofree | ok, thanks | 10:29 |
| *** gtristan has quit IRC | 11:10 | |
| *** gtristan has joined #baserock | 11:23 | |
| paulsherwood | radiofree: the issue is, if you do that, any conversions will not have the same SHAs in their histories | 11:38 |
| paulsherwood | that may be ok for your usecase, or not | 11:38 |
| locallycompact | really? any? | 11:39 |
| *** gtristan has quit IRC | 11:40 | |
| pedroalvarez | not any | 11:40 |
| paulsherwood | all | 11:40 |
| paulsherwood | so for example python-setuptools | 11:41 |
| locallycompact | always different? | 11:41 |
| paulsherwood | definitions has 0aa6a4de5931d02876428388678802db2371fd37 in python2-core.morph. | 11:42 |
| pedroalvarez | I know that not all the conversions are going to be different, only cvs, and probably svn (and maybe old tarballs) | 11:42 |
| paulsherwood | if a new trove is created that mirrors directly from http://svn.python.org/projects/sandbox/ that sha won't exist | 11:43 |
| paulsherwood | locallycompact: pedroalvarez may be right that some conversions may be ok (hg? bzr?), but in any case, radiofree would need to take this into account | 11:48 |
| pedroalvarez | i think is a crazy idea :) | 11:48 |
| locallycompact | I'm semi-interested to know what *isn't* deterministic | 11:48 |
| locallycompact | and why | 11:48 |
| paulsherwood | pedroalvarez: which part is crazy? | 11:49 |
| pedroalvarez | re-lorrying from scratch everything | 11:49 |
| paulsherwood | locallycompact: lorry just runs various converters - some are deterministic, some aren't | 11:50 |
| locallycompact | aright | 11:50 |
| locallycompact | lorrying from scratch is pretty much pointless | 11:50 |
| locallycompact | although | 11:50 |
| locallycompact | in the past we've unhooked the upstream and pulled in the lorries from the upstream in order to disconnect mirroring | 11:51 |
| locallycompact | but that was when everything was git | 11:51 |
| paulsherwood | pedroalvarez: well, radiofree could start by making a trove as 'snapshot' of current gbo repos... which would give it the same content... but if thereafter it starts lorrying direct from upstream, those SHAs diverge | 11:51 |
| locallycompact | if there are non deterministic converters I don't know how that would react | 11:51 |
| pedroalvarez | locallycompact: to do that with non-git repos you will need a copy of the working area | 11:51 |
| locallycompact | right | 11:52 |
| pedroalvarez | but what paulsherwood said is true | 11:52 |
| locallycompact | but you still need the working area | 11:52 |
| * paulsherwood tries hard to make sure that everything he says is true :) | 11:53 | |
| pedroalvarez | if the plan is to use baserock definitions, I doubt lorrying from upstream will work | 11:53 |
| rjek | paulsherwood: Try saying "This statement is false" | 11:53 |
| paulsherwood | rjek: i would never say that :) | 11:54 |
| pedroalvarez | that would be like: http://i.makeagif.com/media/7-07-2015/0lUqF0.gif | 11:54 |
| paulsherwood | pedroalvarez: maybe the plan is to start from definitions, and diverge. i'm not sure there are many organisations the require openstack + genivi + trove :) | 11:55 |
| rjek | pedroalvarez: The sound isn't working on that gif | 11:55 |
| pedroalvarez | rjek: have you tried installing the gifaudio drivers? | 11:55 |
| rjek | Nope, that's probably it | 11:56 |
| paulsherwood | download some more ram, while you're at it | 11:56 |
| pedroalvarez | paulsherwood: of course, but upstream will still provide upgrades to some components | 11:56 |
| paulsherwood | yup | 11:57 |
| pedroalvarez | but yes, I understand | 11:57 |
| *** CTtpollard has quit IRC | 11:58 | |
| locallycompact | You can also do both of these things together | 11:58 |
| locallycompact | downstream the trove to a different namespace and lorry as normal | 11:58 |
| paulsherwood | that would be my recommendation. then if gbo were to disappear or fail, re-point the downstream trove to real upstreams at that point | 12:00 |
| *** CTtpollard has joined #baserock | 12:00 | |
| pedroalvarez | yup, that sounds like the best idea | 12:02 |
| *** tiagogomes has quit IRC | 12:35 | |
| *** tiagogomes has joined #baserock | 12:48 | |
| paulsherwood | http://wiki.baserock.org/gbo | 14:07 |
| *** anahuelamo has quit IRC | 15:14 | |
| *** anahuelamo has joined #baserock | 15:21 | |
| *** faybrocklebank has quit IRC | 15:21 | |
| *** CTtpollard has quit IRC | 15:32 | |
| *** gtristan has joined #baserock | 16:12 | |
| *** jonathanmaw has quit IRC | 16:46 | |
| *** CTtpollard has joined #baserock | 17:01 | |
| *** richard_maw has quit IRC | 17:03 | |
| *** richard_maw has joined #baserock | 17:10 | |
| *** ChanServ sets mode: +v richard_maw | 17:10 | |
| *** locallycompact has quit IRC | 17:22 | |
| *** CTtpollard has quit IRC | 18:17 | |
| *** toscalix has quit IRC | 19:47 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
