IRC logs for #baserock for Friday, 2015-04-03

*** pedroalvarez has quit IRC01:00
*** pedroalvarez has joined #baserock01:01
*** ChanServ sets mode: +v pedroalvarez01:01
*** jjardon has quit IRC04:47
*** jjardon has joined #baserock04:50
*** pacon has quit IRC06:28
*** pacon has joined #baserock06:28
*** pacon has quit IRC06:32
*** pacon has joined #baserock06:32
*** pacon has quit IRC07:07
*** pacon has joined #baserock07:08
*** pacon has quit IRC07:19
*** pacon has joined #baserock07:20
*** wdutch_ has joined #baserock07:44
*** pacon has quit IRC09:22
*** pacon has joined #baserock09:23
*** rdale has joined #baserock10:49
* straycat tries to do something useful with the branch morph series11:24
straycatOkay, so let's leave the branching for now, and leave unversioned == version 011:34
straycatthe main thing I'm after here is getting this bug fixed, and the fix kinda got lost in the noise11:34
straycatpedroalvarez, ping11:45
paulsherwoodwhich is the bug, again?11:48
straycatyou can put a path to a file that doesn't exist in your morph field and morph will just fetch the morph from the chunk repo or do build system detection, without giving any kind of warning11:49
paulsherwoodah11:51
*** rdale has quit IRC12:12
*** pacon has quit IRC13:08
straycat! [remote rejected] HEAD -> refs/publish/master/straycat/fix-invalid-chunk-morph-ref-bug (no new changes)13:46
straycathmmm13:46
* straycat proposes https://gerrit.baserock.org/207 and https://gerrit.baserock.org/20813:59
jjardonstraycat: not sure you have to add a new definitions version; I actually consider the fact that morph doesn't warn you / don't fail when a reference morph file doesn't exist a bug14:45
straycatit is a bug, but if we fix it without supplying a conditional and a version update then we could theoretically break someone's definitions15:22
straycati'm playing devil's advocate here because i originally argued this shouldn't introduce a new version of definitions, but consensus seemed to be that it should15:23
paulsherwoodi don't think it should introduce a new version of definitions either15:24
paulsherwoodi wonder if i believed the opposite before? :015:25
richard_mawit _is_ a change15:26
richard_mawif it could break someone's definitions, it's a version change15:27
doffmHow do I find the baserock version inside a baserock system?15:30
doffmDoes baserock have a release file?15:30
paulsherwoodrichard_maw: only because of a bug in morph + a bug in their definitions15:32
rjekDo we still have an /etc/lsb_release?15:32
rjekI mean /etc/os-release15:32
paulsherwoodyes but with no version number15:33
paulsherwooddoffm: closest thing is the meta files in /baserock i think15:33
doffmpaulsherwood, rjek: Thanks. I'll just skip version number. Check os-release.15:34
paulsherwooddoffm: not sure os-release helps you much?15:34
richard_mawpaulsherwood: So what if it is because of a bug? It is _still_ a version break, and we can only justify changing the behaviour without bumping the version if we are sure nobody will notice15:46
straycatIt's a fair enough reason to introduce a new version, I don't really mind either way, just want the fix ;)15:53
jjardonrichard_maw: if it breaks is because the user definitions is broken. Its a but not a definitions format change16:07
jjardonIts a bug*16:07
* straycat proposes https://gerrit.baserock.org/20916:08
jjardonstraycat: thanks for teaching me proper python :)16:12
straycatglad you like it :)16:19
paulsherwoodstraycat: that looks very elegant :)16:35
straycatthank you :)16:38
jjardonstraycat: I commented in https://gerrit.baserock.org/#/c/207/ Do you think makes sense?16:39
straycatAhh yes, I didn't think that the version file itself might change16:41
straycatbah, http://sprunge.us/PGQM is pretty vague and seems to surprise the folks in #virt18:36
straycatmaybe morph shouldn't assume a default NIC_CONFIG for kvm deployments18:37
straycatno i guess it's fine as it is, if you want to deploy to a bridge then you should specify a NIC_CONFIG for that, there's no reason to make everyone specify NIC_CONFIG18:45
jjardoninvestigating how to do the builds faster, I wonder if someone has already taken a look to icecream? https://github.com/icecc/icecream20:26
rjekjjardon: ISTR people looking at it 2 years or so ago, and dismissing as complex and not significantly better than distcc21:14

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!