Wednesday, 2010-01-27

[2010/01/27 00:15:49] @ Log started by gepetto
[2010/01/27 00:15:49] @ jamesturnbull joined channel #puppet-dev
[2010/01/27 02:42:54] @ github joined channel #puppet-dev
[2010/01/27 02:42:54] <github> puppet: Martin Englund master SHA1-fcce222
[2010/01/27 02:42:54] <github> First shot at the OpenSolaris pkg(5) provider
[2010/01/27 02:42:55] <github> http://bit.ly/9CrQ6m
[2010/01/27 02:42:57] @ github left channel #puppet-dev ()
[2010/01/27 03:28:20] <masterzen> duritong_: I thought about paralelizing puppetd a long time ago. Unfortunately it is a quite complex task (especially because of relationships and events)...
[2010/01/27 03:28:52] <masterzen> jamesturnbull: is Rowlf imminent? I've seen you pushed some open tickets to Statler (which I suppose is Rowlf successor).
[2010/01/27 03:29:29] <ohadlevy> I'm not sure I want a fast puppet, I mean it would be better if users of the system will not notice puppet (when puppet is actually be doing anything)
[2010/01/27 03:30:12] <masterzen> you can still nice it :-)
[2010/01/27 03:30:22] <jamesturnbull> masterzen: we're preparing the ground
[2010/01/27 03:30:33] <jamesturnbull> masterzen: triming it down to the tickets we're going to include
[2010/01/27 03:30:39] <ohadlevy> masterzen: no i cant, because then it would start the sub processes (e.g. services) with the same nice level
[2010/01/27 03:31:02] <masterzen> jamesturnbull: so I suppose Markus patch for the parser futures won't be in?
[2010/01/27 03:31:55] <masterzen> jamesturnbull: asking that because I was deferring my hashes patch to put it on top of his. But if Markus patch is not pushed, I might as well release mine directly.
[2010/01/27 03:32:26] <masterzen> ohadlevy: you're right, I forgot about this.
[2010/01/27 03:32:43] <jamesturnbull> masterzen: it'll be in
[2010/01/27 03:32:58] <jamesturnbull> masterzen: that and your hashes patch are def. on the list
[2010/01/27 03:33:27] <jamesturnbull> masterzen: can you have a look at some test output for me? it's damn weird the "to_sym" method seems to have gone away - http://pastie.org/796444
[2010/01/27 03:33:38] <masterzen> ohadlevy: anyway, parallelizing means on multiple core you'll get a cpu spike for every cores, instead of a longer cpu spike on only one... I doubt it will be hardly noticeable...
[2010/01/27 03:34:38] <ohadlevy> masterzen: true, but puppet was never cpu bound, if you can copy (or compare) all of the require files at once, this will cut the network latency, which imho has a bigger impact
[2010/01/27 03:35:25] <masterzen> jamesturnbull: to_sym is only for string, but now RDoc sees a Puppet::Environment instead of a string...
[2010/01/27 03:35:43] <jamesturnbull> masterzen: ah
[2010/01/27 03:35:45] <masterzen> jamesturnbull: this is the result of one of Luke's patch
[2010/01/27 03:35:53] <jamesturnbull> masterzen: of course it is
[2010/01/27 03:35:55] <jamesturnbull> :P
[2010/01/27 03:35:59] <masterzen> :-)
[2010/01/27 03:36:19] <masterzen> ohadlevy: you're right again.
[2010/01/27 03:36:26] * masterzen needs his coffe fix to think clearly
[2010/01/27 03:36:30] <masterzen> *coffee
[2010/01/27 03:36:37] <jamesturnbull> I just need to work out where
[2010/01/27 03:36:51] * ohadlevy cant get real coffee around here
[2010/01/27 03:37:47] <jamesturnbull> ohadlevy: singapore was many electrical things and fake dvds but no good coffee
[2010/01/27 03:38:24] <ohadlevy> fake dvds? naa, thats Malaysia
[2010/01/27 03:38:25] <masterzen> jamesturnbull: about the digest ones, I'm puzzled, or maybe you're missing the ruby-openssl extension?
[2010/01/27 03:38:54] <jamesturnbull> masterzen: yeah
[2010/01/27 03:39:00] <jamesturnbull> masterzen: I am puzzled too
[2010/01/27 03:39:29] <masterzen> jamesturnbull: I _think_ openssl ruby extension contains digest in 1.8.6
[2010/01/27 03:39:37] <masterzen> jamesturnbull: would be a shame if not...
[2010/01/27 03:41:12] <jamesturnbull> masterzen: it'd be a big damn problem :)
[2010/01/27 03:48:21] <jamesturnbull> masterzen: so the to_sym in lib/puppet/node/environment.rb is not right?
[2010/01/27 03:52:42] @ Chiku|dc joined channel #puppet-dev
[2010/01/27 03:53:53] @ Quit: jrojas: calvino.freenode.net irc.freenode.net
[2010/01/27 03:54:00] <masterzen> jamesturnbull: I don't know. It all looks like Puppet.settings(:environmnent) is now a Puppet::Node::Environment instead of a string.
[2010/01/27 03:54:00] @ Quit: Chiku: Read error: 54 (Connection reset by peer)
[2010/01/27 03:54:04] @ Quit: bje: Read error: 60 (Operation timed out)
[2010/01/27 03:54:04] <masterzen> jamesturnbull: I think some earlier test you ran in the same session corrupted the environment settings.
[2010/01/27 03:54:09] @ jrojas joined channel #puppet-dev
[2010/01/27 03:54:26] <jamesturnbull> masterzen: yeah the tests run stand-alone - just not in sequence
[2010/01/27 03:54:29] @ bje joined channel #puppet-dev
[2010/01/27 04:15:28] <jamesturnbull> masterzen: so removing that .to_sym results in those tests no longer failing
[2010/01/27 04:25:08] <masterzen> jamesturnbull: looks strange to me
[2010/01/27 04:29:51] <jamesturnbull> masterzen: or not ... I think it's a test ordering issue
[2010/01/27 04:30:26] <masterzen> jamesturnbull: yes I think something leaks a Puppet::Node::Env into the environmnent puppet settings and everything breaks
[2010/01/27 04:33:58] <jamesturnbull> masterzen: and I cannot work out the openssl ones either
[2010/01/27 04:33:59] <jamesturnbull> irb
[2010/01/27 04:34:03] <jamesturnbull> reirb(main):001:0> require 'openssl'
[2010/01/27 04:34:03] <jamesturnbull> => true
[2010/01/27 04:34:03] <jamesturnbull> irb(main):002:0> require 'digest'
[2010/01/27 04:34:03] <jamesturnbull> => true
[2010/01/27 04:34:03] <jamesturnbull> irb(main):003:0> require 'digest/sha1'
[2010/01/27 04:34:06] <jamesturnbull> => true
[2010/01/27 04:34:08] <jamesturnbull> irb(main):004:0> require 'digest/md5'
[2010/01/27 04:34:11] <jamesturnbull> => true
[2010/01/27 04:34:14] <jamesturnbull> irb(main):005:0>
[2010/01/27 04:34:20] <masterzen> jamesturnbull: looks good to me :-)
[2010/01/27 04:34:28] <jamesturnbull> masterzen: yeah so why does it fail??? :)
[2010/01/27 04:36:47] <jamesturnbull> hmmm it seems the tests don't require sha1
[2010/01/27 04:37:17] <masterzen> jamesturnbull: can you try: OpenSSL::Digest.hexdigest("sha1", "this is a test") in irb?
[2010/01/27 04:38:26] <jamesturnbull> masterzen: NoMethodError: undefined method `hexdigest' for OpenSSL::Digest:Module
[2010/01/27 04:38:37] <masterzen> ouch
[2010/01/27 04:39:02] <masterzen> and what gives OpenSSL::Digest.methods
[2010/01/27 04:39:31] <jamesturnbull> I've required openssl, digest/md5, digest/sha1 and no joy
[2010/01/27 04:39:47] <masterzen> you just need to require 'openssl/digest'
[2010/01/27 04:39:55] <masterzen> and maybe require 'openssl'
[2010/01/27 04:40:11] <jamesturnbull> openssl/digest returns false
[2010/01/27 04:40:13] <masterzen> you don't need to require 'digest' which are the ruby official versions
[2010/01/27 04:43:11] <jamesturnbull> yeah
[2010/01/27 04:43:46] <masterzen> jamesturnbull: ruby 1.8.6 seems to support openssl digest from what I could gather from google (I only have access to 1.8.7 here)
[2010/01/27 04:47:01] <jamesturnbull> masterzen: hmmm so puzzlement as to why it fails now - git biset time
[2010/01/27 04:47:24] <masterzen> jamesturnbull: on what platform are you testing?
[2010/01/27 04:47:53] <jamesturnbull> masterzen: fedora
[2010/01/27 04:48:19] <masterzen> jamesturnbull: isn't there a ruby-openssl or something package to install?
[2010/01/27 04:58:05] <jamesturnbull> masterzen: nope
[2010/01/27 04:58:19] <jamesturnbull> masterzen: surely then require 'openssl' would return false?
[2010/01/27 05:01:05] <masterzen> jamesturnbull: fedora 11, ruby-libs rpm contains: /usr/lib/ruby/1.8/openssl/digest.rb
[2010/01/27 05:01:27] <masterzen> jamesturnbull: so require 'openssl'\n require 'openssl/digest' should work
[2010/01/27 05:02:15] <jamesturnbull> masterzen: and it does
[2010/01/27 05:02:52] <jamesturnbull> masterzen: i'm on fc12
[2010/01/27 05:03:06] <masterzen> jamesturnbull: so in irb, what gives: OpenSSL::Digest.methods ?
[2010/01/27 05:04:43] <masterzen> jamesturnbull: better: OpenSSL::Digest.methods.include?("hexdigest")
[2010/01/27 05:05:03] <jamesturnbull> masterzen: OpenSSL::Digest.methods.include?("hexdigest")
[2010/01/27 05:05:03] <jamesturnbull> => false
[2010/01/27 05:05:34] <masterzen> ouch, hurts badly
[2010/01/27 05:05:37] <jamesturnbull> yeah
[2010/01/27 05:05:49] <jamesturnbull> there's a whole lot of users on 1.8.6 too
[2010/01/27 05:05:50] <masterzen> I think it's time to file a bug
[2010/01/27 05:05:56] <jamesturnbull> on Fedora/CentOS and RHEL
[2010/01/27 05:06:54] <jamesturnbull> tmz: ping
[2010/01/27 05:10:21] <duritong_> ohadlevy: yeah I'm also not sure wether puppet should run fast or not. However I think if you introduce a way to define the max amount of subprocesses (I was thinking about automatically setting to processorcount), then you could also tweak how puppet should behave
[2010/01/27 05:11:03] <duritong_> ohadlevy: I'm just thinking about machines running on multiple cores 8,16,32 where puppet would only use one.
[2010/01/27 05:11:19] <Volcane> runing fast will just make people complain mor about the hit it does on cpu resources
[2010/01/27 05:11:36] <Volcane> my devs are already winging, and my voip clients too :(
[2010/01/27 05:11:49] <duritong_> ohadlevy: but yeah network latency would be another problem. However due to the rest interface downloading multiple files from a nginx fronted master wouldn't affect the master process itself
[2010/01/27 05:11:54] <duritong_> Volcane: yeah :P
[2010/01/27 05:11:56] <Volcane> i want it to run slower and use less resources
[2010/01/27 05:12:55] <jamesturnbull> masterzen: tmz https://bugzilla.redhat.com/show_bug.cgi?id=559158
[2010/01/27 05:12:59] <duritong_> masterzen: it would be a hard task. However the hardest part I see in resources relying on single access such as calls to package managers.
[2010/01/27 05:13:45] <duritong_> masterzen: but as I was talking about independent graph parts, relations shouldn't be a problem, as if there would be any relations the parts wouldn't be independent anymore
[2010/01/27 05:15:45] @ duritong_ is now known as duritong
[2010/01/27 05:16:27] <masterzen> jamesturnbull: AAARRGGGHHH
[2010/01/27 05:16:39] <jamesturnbull> masterzen: ?
[2010/01/27 05:16:53] <masterzen> jamesturnbull: the fedora bug will hurt a lot of people
[2010/01/27 05:17:07] <masterzen> jamesturnbull: sorry for crying loud
[2010/01/27 05:17:20] <jamesturnbull> masterzen: yeah it's a showstopper
[2010/01/27 05:17:31] <masterzen> jamesturnbull: do you think we should fix it on our side?
[2010/01/27 05:19:28] <jamesturnbull> masterzen: we might have too but where does that stop?
[2010/01/27 05:20:53] @ Quit: ppickfor: Read error: 113 (No route to host)
[2010/01/27 05:23:53] <duritong> jamesturnbull: no imho this is definately something to be fixed in fedora
[2010/01/27 05:24:32] <duritong> jamesturnbull: let's see what kanarip aka jeroen means about it
[2010/01/27 05:26:44] <duritong> * Wed Aug 26 2009 Tomas Mraz <tmraz@redhat.com> - 1.8.6.369-3 - Rebuild against new openssl
[2010/01/27 05:27:22] <duritong> is this against openssl 1.0 ? which I remember to have read something about that it will break things
[2010/01/27 05:27:56] <jamesturnbull> duritong: yeah currently pawing throygh the sources
[2010/01/27 05:29:20] <jamesturnbull> duritong: and I pinged jeroen
[2010/01/27 05:30:12] * duritong as well
[2010/01/27 05:33:57] <jamesturnbull> I can find the hexdigest method in openssl/digest.rb - this is weird
[2010/01/27 05:36:27] <jamesturnbull> masterzen: irb(main):001:0> require 'openssl'
[2010/01/27 05:36:27] <jamesturnbull> => true
[2010/01/27 05:36:27] <jamesturnbull> irb(main):002:0> OpenSSL::Digest::SHA1.methods.include?("hexdigest")
[2010/01/27 05:36:27] <jamesturnbull> => true
[2010/01/27 05:36:27] <jamesturnbull> irb(main):003:0> OpenSSL::Digest::MD5.methods.include?("hexdigest")
[2010/01/27 05:36:30] <jamesturnbull> => true
[2010/01/27 05:36:32] <jamesturnbull> irb(main):004:0>
[2010/01/27 05:37:38] <masterzen> jamesturnbull: it seems to be available in each sub digest, but not in the digest module itself
[2010/01/27 05:38:23] <jamesturnbull> masterzen: yeah this is making me annoyed :)
[2010/01/27 05:56:57] <jamesturnbull> masterzen: so what's the impact if I change lib/puppet/ssl/base.rb to use OpenSSL::Digest::MD5.hexdigest rather than OpenSSL::Digest.hexdigest?
[2010/01/27 05:57:24] <jamesturnbull> masterzen: okay nothing good :)
[2010/01/27 05:57:29] <masterzen> jamesturnbull: not goot
[2010/01/27 05:57:47] <masterzen> jamesturnbull: the thing is that the fingerprint digest is configurable from a puppet parameter.
[2010/01/27 05:58:06] <masterzen> jamesturnbull: so we need to load the correct digest class
[2010/01/27 05:58:18] <jamesturnbull> masterzen: yeah just worked that out
[2010/01/27 05:58:24] <masterzen> jamesturnbull: maybe the simpler is to call .digest, and then do the hex part ourselves.
[2010/01/27 05:58:24] <jamesturnbull> sorry a little slow tonight
[2010/01/27 05:58:59] <masterzen> jamesturnbull: I'll do it tonight. I'm pretty busy today, I'm in the middle of a release for an internal scrum sprint...
[2010/01/27 06:00:06] <jamesturnbull> masterzen: should this be fixed downstream? if we recreate every method...
[2010/01/27 06:00:20] <masterzen> jamesturnbull: yes, this sucks :-(
[2010/01/27 06:00:29] <jamesturnbull> masterzen: ah - sorry to bother you - I'll go away :)
[2010/01/27 06:00:38] <masterzen> jamesturnbull: no it's ok :-)
[2010/01/27 06:00:50] <masterzen> jamesturnbull: I'm procrastinating...
[2010/01/27 06:01:40] <jamesturnbull> lol
[2010/01/27 06:01:49] <jamesturnbull> I've written off my evening as a result
[2010/01/27 06:02:05] <jamesturnbull> trust me to assume Luke's code would work when I merged it :)
[2010/01/27 06:30:37] @ stahnma_ joined channel #puppet-dev
[2010/01/27 06:31:45] @ Quit: stahnma: Read error: 104 (Connection reset by peer)
[2010/01/27 06:32:35] @ stahnma_ is now known as stahnma
[2010/01/27 07:12:33] @ Quit: jrojas: calvino.freenode.net irc.freenode.net
[2010/01/27 07:12:47] @ jrojas joined channel #puppet-dev
[2010/01/27 08:38:39] @ Quit: rocket: calvino.freenode.net irc.freenode.net
[2010/01/27 08:38:46] @ Quit: masterzen: calvino.freenode.net irc.freenode.net
[2010/01/27 08:38:52] @ Quit: ch: calvino.freenode.net irc.freenode.net
[2010/01/27 08:38:54] @ ch joined channel #puppet-dev
[2010/01/27 08:38:54] @ rocket joined channel #puppet-dev
[2010/01/27 08:38:54] @ masterzen joined channel #puppet-dev
[2010/01/27 12:27:45] <rocket> anyone here able to help me debug a spec file?
[2010/01/27 12:28:14] <rocket> my new type and providor work, but when running the associated spec I get failures and I am not sure why ..
[2010/01/27 12:28:34] <rocket> which I attribute to my limited understanding of the test framework ..:/
[2010/01/27 12:28:43] <rocket> pointers where to read more would be helpful :)
[2010/01/27 12:32:07] <nasrat> pointer to a gist/github/pastie would help
[2010/01/27 12:32:38] <rocket> nasrat: pastie coming ...
[2010/01/27 12:34:40] <rocket> http://pastebin.com/m2833bd30
[2010/01/27 12:43:26] @ ppickfor joined channel #puppet-dev
[2010/01/27 12:52:23] <rocket> nasrat: does that help any?
[2010/01/27 12:58:27] <nasrat> sorry caught up at work
[2010/01/27 12:59:00] <nasrat> it looks as if it's complaining about your expectations not being met which is usually a sign you're not calling the mocked objects in the real code
[2010/01/27 12:59:12] <nasrat> do you have the code under test too
[2010/01/27 12:59:40] <nasrat> I need to head off in a sec though
[2010/01/27 12:59:44] <rocket> I thought so .. but perhaps not?
[2010/01/27 13:00:16] <rocket> how would I make sure of that?
[2010/01/27 13:00:44] <rocket> I have added the file I am building for test cases at the bottom of the pastie
[2010/01/27 13:06:46] <nasrat> ok sorry have to head off
[2010/01/27 13:06:48] <nasrat> may be back later
[2010/01/27 13:54:39] @ jes5 joined channel #puppet-dev
[2010/01/27 14:18:30] @ mqr joined channel #puppet-dev
[2010/01/27 14:25:58] <masterzen> jamesturnbull: are the failing tests you show me on master or on 0.25.x, or testing?
[2010/01/27 14:26:05] @ Quit: fsweetser: Remote closed the connection
[2010/01/27 15:04:57] <jamesturnbull> masterzen: master
[2010/01/27 15:05:29] <rocket> grrr ... me bumps head against wall .. what am I doing wrong with these rspec tests ...
[2010/01/27 15:05:47] <rocket> http://pastebin.com/m2833bd30
[2010/01/27 15:21:52] @ lak joined channel #puppet-dev
[2010/01/27 15:27:54] <rocket> lak! :)
[2010/01/27 15:28:22] <rocket> lak: I have a type and provider that seems to work .. :p
[2010/01/27 15:28:31] <lak> rocket: sweet :)
[2010/01/27 15:28:43] <rocket> lak: however i am trying to backfill the unit tests and I am running into problems
[2010/01/27 15:29:05] <rocket> lak: running the code without the unit tests seems to work fine .. so I am not sure what I am doing wrong
[2010/01/27 15:29:08] <rocket> http://pastebin.com/m2833bd30
[2010/01/27 15:29:58] <rocket> eg I can create a repository, modify it, delete it etc .. but my rspec tests are failing .. I am trying to copy the package unit tests as they are most closely related
[2010/01/27 15:30:27] <rocket> any pointers would be greatly appreciated .. I have been reading the website for hints that I may have missed but am coming up short
[2010/01/27 15:31:34] <lak> rocket: i'll try to look at it later, gotta run to lunch :/
[2010/01/27 15:31:45] @ Quit: lak:
[2010/01/27 16:07:32] <jrojas> lunch heh.
[2010/01/27 16:58:40] <jrojas> is marcus in here at all ?
[2010/01/27 17:09:07] <jamesturnbull> jrojas: meet mqr - mqr meet jrojas
[2010/01/27 17:09:13] <jrojas> hehe
[2010/01/27 17:09:17] <jrojas> thanks
[2010/01/27 17:09:25] <jrojas> jamesturnbull: have you looked at 3114 recently?
[2010/01/27 17:09:57] <mqr> Hello.
[2010/01/27 17:10:02] <jrojas> not sure if it fixes your issue on OSX but its working for the most part for me... just trying to find the proper place to set Process.maxgroups
[2010/01/27 17:10:33] <mqr> I'm thinking either util or monkeypatch.
[2010/01/27 17:10:49] <mqr> It's annoying that we have to set it at all.
[2010/01/27 17:10:57] <jrojas> agreed,
[2010/01/27 17:11:14] <jrojas> what about in Puppet::Util::SUIDManager's initgroups ?
[2010/01/27 17:11:46] <mqr> I just answered you on the ticket re: 0.25.4; if there aren't any new issues with 0.25.4rc3 we need to push it as 0.25.4
[2010/01/27 17:12:04] <jrojas> ah k
[2010/01/27 17:12:10] <jrojas> btw this is interesting http://timetobleed.com/5-things-you-dont-know-about-user-ids-that-will-destroy-you/
[2010/01/27 17:12:14] <mqr> That would make the maxgroups an immediate 0.25.5rc1
[2010/01/27 17:12:39] <mqr> Will that work for you?
[2010/01/27 17:13:03] <jrojas> yeah, ill probably patch my rpm for 0.25.4 in the meantime when its released and go forward with that.
[2010/01/27 17:15:02] <mqr> That link in regards to #3107?
[2010/01/27 17:15:03] <gepetto> mqr: #3107 is http://projects.reductivelabs.com/issues/show/3107 "Puppet - Bug #3107: Execute() in util.rb does not execute setuid properly on AIX - ReductiveLabs.com"
[2010/01/27 17:15:26] <jrojas> i found it when looking for other related things to #3114
[2010/01/27 17:15:27] <gepetto> jrojas: #3114 is http://projects.reductivelabs.com/issues/show/3114 "Puppet - Bug #3114: Puppet crashes after run - ReductiveLabs.com"
[2010/01/27 17:15:33] <jrojas> thought it was an interesting read
[2010/01/27 17:18:06] <jamesturnbull> mqr: so I can understand this - we're happy to push .4 now?
[2010/01/27 17:18:16] <mqr> I've also been researching to try to figure out why there's even a maxgroups limit in Process to begin with. No leads so far.
[2010/01/27 17:18:28] <jamesturnbull> mqr: and we'll put the fix to 3114 into .5?
[2010/01/27 17:18:42] <jamesturnbull> mqr: because I'm happy to go with a release now
[2010/01/27 17:19:27] <jrojas> mqr: http://linux.about.com/library/cmd/blcmdl2_getgroups.htm
[2010/01/27 17:19:52] <jrojas> mqr: its possible there is an error somewhere down in that line, and a -1 is returned and ruby goes boom?
[2010/01/27 17:19:54] <mqr> jamestuenbull: let's go down our checklist: what does Hudson thing of 0.25.4rc3? Are there any unresolved bugs created by it or that meet the pre-inclusion criteria? Have people tested it & what's the feedback?
[2010/01/27 17:19:57] @ github joined channel #puppet-dev
[2010/01/27 17:19:57] <github> puppet: James Turnbull master SHA1-04842ef
[2010/01/27 17:19:57] <github> Fixed test error message.
[2010/01/27 17:19:58] <github> http://bit.ly/bnjGfY
[2010/01/27 17:20:00] @ github left channel #puppet-dev ()
[2010/01/27 17:21:54] <jamesturnbull> mqr: no new broken tests other than the ones already broke
[2010/01/27 17:22:28] <jamesturnbull> mqr: the only open ticket is #3083
[2010/01/27 17:22:29] <gepetto> jamesturnbull: mqr: #3083 is http://projects.reductivelabs.com/issues/show/3083 "Puppet - Bug #3083: Failed to retrieve current state of resource: undefined method `closed?' for nil:NilClass - ReductiveLabs.com"
[2010/01/27 17:22:53] <jamesturnbull> mqr: a few people have tested .. not a lot .. but two or three and it seems to okay ... *shrugs* is that enough? *shrugs* again
[2010/01/27 17:25:05] <jamesturnbull> jrojas: you've done some rc3 testing right?
[2010/01/27 17:25:10] <jamesturnbull> tmz: rc3?
[2010/01/27 17:25:29] <jrojas> yeah, nothing broken on my end
[2010/01/27 17:25:39] <jrojas> except for my Process.maxgroups issue
[2010/01/27 17:25:43] <jrojas> :)
[2010/01/27 17:25:53] <jamesturnbull> jrojas: which is mine too but it's too big a chnage to make without an rc4
[2010/01/27 17:25:55] <mqr> I'm sitting in a room full of people who just used it for training.
[2010/01/27 17:26:03] <mqr> No new issues uncovered here.
[2010/01/27 17:26:17] <jrojas> awesome.
[2010/01/27 17:26:53] <jamesturnbull> mqr: OSU?
[2010/01/27 17:28:39] <jamesturnbull> okay so I am thinking we're go for a release then
[2010/01/27 17:28:42] <jamesturnbull> mqr: you concur?
[2010/01/27 17:28:49] * jamesturnbull is arse covering a little
[2010/01/27 17:28:56] <mqr> PSU
[2010/01/27 17:29:11] <jamesturnbull> mqr: yeah that's what I meant
[2010/01/27 17:29:16] <mqr> What are your thoughts on #3092?
[2010/01/27 17:29:16] <gepetto> mqr: #3092 is http://projects.reductivelabs.com/issues/show/3092 "Puppet - Bug #3092: No entries means no security? - ReductiveLabs.com"
[2010/01/27 17:29:35] <mqr> I'm going through tickets to see if we've let anything slip by.
[2010/01/27 17:29:51] <jamesturnbull> mqr: if we do anything else - it's another rc
[2010/01/27 17:29:59] <jamesturnbull> mqr: and we're running out of time for Ubuntu
[2010/01/27 17:30:12] <jamesturnbull> brb
[2010/01/27 17:33:30] <jrojas> mqr: added my quick patch to #3114
[2010/01/27 17:33:31] <gepetto> jrojas: mqr: #3114 is http://projects.reductivelabs.com/issues/show/3114 "Puppet - Bug #3114: Puppet crashes after run - ReductiveLabs.com"
[2010/01/27 17:40:52] <mqr> Ticket to think about #3092, #3083.
[2010/01/27 17:40:52] <gepetto> mqr: #3092 is http://projects.reductivelabs.com/issues/show/3092 "Puppet - Bug #3092: No entries means no security? - ReductiveLabs.com"
[2010/01/27 17:40:53] <gepetto> mqr: #3083 is http://projects.reductivelabs.com/issues/show/3083 "Puppet - Bug #3083: Failed to retrieve current state of resource: undefined method `closed?' for nil:NilClass - ReductiveLabs.com"
[2010/01/27 17:41:11] <mqr> #3092 I don't understand.
[2010/01/27 17:41:12] <gepetto> mqr: #3092 is http://projects.reductivelabs.com/issues/show/3092 "Puppet - Bug #3092: No entries means no security? - ReductiveLabs.com"
[2010/01/27 17:41:21] <mqr> #3083 I can't reproduce.
[2010/01/27 17:41:22] <gepetto> mqr: #3083 is http://projects.reductivelabs.com/issues/show/3083 "Puppet - Bug #3083: Failed to retrieve current state of resource: undefined method `closed?' for nil:NilClass - ReductiveLabs.com"
[2010/01/27 17:42:16] <mqr> Also #3041, which is a little of both.
[2010/01/27 17:42:17] <gepetto> mqr: #3041 is http://projects.reductivelabs.com/issues/show/3041 "Puppet - Bug #3041: undefined method `name=' for #<Puppet::FileServing::Content when retrieving a plug-in - ReductiveLabs.com"
[2010/01/27 17:44:00] <masterzen> #3092, I think we insert (certainly my fault) some default allow all rules when we don't have [modules] or [plugins], except it triggers _when_ we have empty [modules]/[plugins]
[2010/01/27 17:44:01] <gepetto> masterzen: #3092: #3092 is http://projects.reductivelabs.com/issues/show/3092 "Puppet - Bug #3092: No entries means no security? - ReductiveLabs.com"
[2010/01/27 17:47:35] <mqr> I'm still not understanding the issue.
[2010/01/27 17:50:32] <mqr> Ah. Editing the text of the ticket so that the middle of the first paragraph wasn't obscured helped.
[2010/01/27 17:50:56] @ jes5 left channel #puppet-dev ()
[2010/01/27 17:50:57] @ jes5 joined channel #puppet-dev
[2010/01/27 17:52:25] <mqr> So is 3092 an sufficiently significant issue? I'm thinking no.
[2010/01/27 17:56:13] <masterzen> mqr: it can be a kind of security issue: you think you're protected but you're not.
[2010/01/27 17:57:33] <jes5> it seems like an edge case to me
[2010/01/27 18:00:34] <mqr> I'm going to bump this as a design issue; the assumption in the present system is that plugins and mounts are accessible by default, and this would be a significant behavior change for existing installs.
[2010/01/27 18:01:03] <masterzen> I introduced the issue in 6b254e
[2010/01/27 18:01:20] <masterzen> #2753
[2010/01/27 18:01:21] <gepetto> masterzen: #2753 is http://projects.reductivelabs.com/issues/show/2753 "Puppet - Bug #2753: fileserver.conf allow/deny directives not honored for [modules], [plugins] - ReductiveLabs.com"
[2010/01/27 18:01:42] <masterzen> which itself originate #2424
[2010/01/27 18:01:43] <gepetto> masterzen: #2424 is http://projects.reductivelabs.com/issues/show/2424 "Puppet - Bug #2424: File server can't find module in environment - ReductiveLabs.com"
[2010/01/27 18:41:14] <jrojas> hmm i dont think that the patch for the Process.maxgroups is working
[2010/01/27 18:58:31] <mqr> jrojas: why not..
[2010/01/27 18:58:32] <mqr> ?
[2010/01/27 19:00:03] @ lak joined channel #puppet-dev
[2010/01/27 19:01:34] <mqr> jrojas: is it exhibiting the same errors, or something new?
[2010/01/27 19:02:04] <jrojas> same error
[2010/01/27 19:02:17] <jrojas> i think asuser is called before initgroups in some cases
[2010/01/27 19:02:37] <jrojas> mqr: im testing Process.maxgroups=1024 in util.rb outside of any method
[2010/01/27 19:05:53] <jrojas> mqr: strace is showing getgroups() returning gids properly
[2010/01/27 19:06:08] <jrojas> and the clients havent died at all
[2010/01/27 19:09:44] <mqr> let me look at something...
[2010/01/27 19:13:08] <mqr> jrojas: try putting it as the first line of lib/puppet/util/monkey_patches.rb
[2010/01/27 19:14:04] <jrojas> above module RDoc ?
[2010/01/27 19:15:23] <mqr> Sure. Or at the very end. Just not inside.
[2010/01/27 19:19:02] <mqr> jrojas: got to log out. Will be back on in a few hours, or communicate via the ticket.
[2010/01/27 19:19:35] @ Quit: mqr:
[2010/01/27 19:33:24] <jamesturnbull> lak: ping - same branch is fine as long as you have a commit comment somewhere that mentions it
[2010/01/27 19:33:36] <jamesturnbull> lak: the test leakage is hihgly frustrating
[2010/01/27 19:33:56] <lak> i think that's actually a bug, or maybe two
[2010/01/27 19:34:20] <lak> i'll look at it, but i'm traveling and thus of limited attention
[2010/01/27 19:36:32] <jamesturnbull> lak: yeah understand - and this is different from normal how? :)
[2010/01/27 20:04:09] @ Quit: lak:
[2010/01/27 20:16:42] @ mqr joined channel #puppet-dev
[2010/01/27 20:22:01] <mqr> jrojas: I'm back.
[2010/01/27 21:10:02] @ github joined channel #puppet-dev
[2010/01/27 21:10:02] <github> puppet: Brice Figureau master SHA1-46f9d00
[2010/01/27 21:10:02] <github> Fix #3117 - cert fingerprinting uses a method not available in ruby <= 1.8.6
[2010/01/27 21:10:02] <gepetto> github: #3117 is http://projects.reductivelabs.com/issues/show/3117 "Puppet - Bug #3117: Cert fingerprinting fails with ruby <= 1.8.6 - ReductiveLabs.com"
[2010/01/27 21:10:03] <github> http://bit.ly/9HjIBJ
[2010/01/27 21:10:05] <github> puppet: James Turnbull master SHA1-70c71c5
[2010/01/27 21:10:07] <github> Fixed Rails database tests
[2010/01/27 21:10:09] <github> http://bit.ly/c8NK6e
[2010/01/27 21:10:11] @ github left channel #puppet-dev ()
[2010/01/27 21:11:47] <jamesturnbull> mqr: I'm still here
[2010/01/27 21:11:53] <jamesturnbull> mqr: but I have no life so ...
[2010/01/27 21:12:20] <jamesturnbull> mqr: did we reach any .4 conclusions?
[2010/01/27 21:24:32] <tmz> jamesturnbull: Sorry, I've not had much time to give it a lot of testing. The little I did on rc3 hasn't turned up anything (but it's worth noting that I'd have said the same of 0.25.3 with broken cron resources ;).
[2010/01/27 21:24:36] <mqr> jamesturnbull: We're good as far as I can determine. Dan & Teyo gave me another 11th hour heart-stopper (#3118) but it turns out to be dusty (same behaviour back to 0.24.8 at least, probably longer).
[2010/01/27 21:24:37] <gepetto> mqr: jamesturnbull: #3118 is http://projects.reductivelabs.com/issues/show/3118 "Puppet - Bug #3118: External node tool definitions appear to be cumulative - ReductiveLabs.com"
[2010/01/27 21:50:13] <jamesturnbull> mqr: yeah I could have told you that buit didn't have time to update ticket
[2010/01/27 21:50:33] <mqr> Teyo seemed very surprised.
[2010/01/27 21:50:44] <mqr> The class seemed split 50/50
[2010/01/27 21:51:03] <mqr> Actually 25/25/50, with half not having thought of it.
[2010/01/27 21:52:11] <jamesturnbull> mqr: I was looking at it recently as my book mentioned it
[2010/01/27 21:52:18] <jamesturnbull> mqr: and I couldn't remember why I tohught that
[2010/01/27 21:53:58] <jamesturnbull> mqr: okay let's release
[2010/01/27 21:54:44] <mqr> Sounds good. I'll check back in an hour or two to discover why we shouldn't have :)
[2010/01/27 21:57:41] <jamesturnbull> mqr: thaks for that
[2010/01/27 21:58:32] <mqr> Hey, mildly amusing cynicism is just one of many fine services I offer.
[2010/01/27 21:59:01] <mqr> Could I interest you in a little sarcasm, perhaps?
[2010/01/27 21:59:53] @ nigelk joined channel #puppet-dev
[2010/01/27 22:00:10] <jamesturnbull> sorry - not used to Americans who have mastered it ... can i have a dollop of irony with my snide please?
[2010/01/27 22:00:25] <jamesturnbull> nigelk: have you been following that #3113
[2010/01/27 22:00:25] <gepetto> jamesturnbull: nigelk: #3113 is http://projects.reductivelabs.com/issues/show/3113 "Puppet - Bug #3113: When importing a manifest puppet needs to chill - ReductiveLabs.com"
[2010/01/27 22:00:26] <nigelk> get a dog up ya
[2010/01/27 22:00:34] <jamesturnbull> nigelk: hmm #3114
[2010/01/27 22:00:35] <gepetto> jamesturnbull: nigelk: #3114 is http://projects.reductivelabs.com/issues/show/3114 "Puppet - Bug #3114: Puppet crashes after run - ReductiveLabs.com"
[2010/01/27 22:00:43] <nigelk> jamesturnbull: looking now... but i'm having major issues with 0.25.4rc3
[2010/01/27 22:00:47] <nigelk> linux seems fine
[2010/01/27 22:00:51] <nigelk> Mac does not....
[2010/01/27 22:00:54] <jamesturnbull> nigelk: be curious to see your thoughts on OSX
[2010/01/27 22:01:04] <jamesturnbull> actually its get a big black dog up ya
[2010/01/27 22:01:09] <nigelk> http://pastie.org/798208
[2010/01/27 22:01:11] <nigelk> heh
[2010/01/27 22:01:42] <nigelk> jamesturnbull: I don't quite understand #3114
[2010/01/27 22:01:43] <gepetto> nigelk: jamesturnbull: #3114 is http://projects.reductivelabs.com/issues/show/3114 "Puppet - Bug #3114: Puppet crashes after run - ReductiveLabs.com"
[2010/01/27 22:01:52] <nigelk> oh I was looking at 3113... :)
[2010/01/27 22:02:57] <jamesturnbull> nigelk: 3114 is that auser bug
[2010/01/27 22:03:20] <nigelk> only happens when you're not running as root on the clients?
[2010/01/27 22:03:30] <jamesturnbull> mqr: ^^
[2010/01/27 22:04:46] <nigelk> http://pastie.org/798211 <- I don't think I'm seeing this problem on OS X
[2010/01/27 22:05:50] <jamesturnbull> nigelk: I see it
[2010/01/27 22:06:58] <nigelk> really? with stock ruby?
[2010/01/27 22:07:03] <nigelk> 10.5 or 10.6?
[2010/01/27 22:07:10] <jamesturnbull> 10.6
[2010/01/27 22:07:14] <jamesturnbull> on my MBP
[2010/01/27 22:07:17] * nigelk checks 10.6
[2010/01/27 22:07:25] <jamesturnbull> ruby-1.8.7 p72 I think
[2010/01/27 22:08:48] <jamesturnbull> nigelk: I am about to release .4 btw
[2010/01/27 22:09:01] <nigelk> erk
[2010/01/27 22:09:13] <nigelk> did you see my pastie? I'm trying to find anyone who has 0.25.4rc3 working with OS X
[2010/01/27 22:09:15] <ohadlevy> jamesturnbull: can you make sure that the annoying report messages are not in warn level
[2010/01/27 22:09:32] <jamesturnbull> ohadlevy: did you log a ticket?
[2010/01/27 22:09:34] <mqr> jamesturnbull: what was the 0.25.4rc3 doesn't work on OSX that just went by.
[2010/01/27 22:09:36] <jamesturnbull> nigelk: oh I see
[2010/01/27 22:09:47] <ohadlevy> I didnt, because there was already a ticket about it
[2010/01/27 22:10:13] <ohadlevy> I think
[2010/01/27 22:10:19] <nigelk> with a Mobile Account, Process.groups works fine for me on 10.6
[2010/01/27 22:10:27] <nigelk> checking local accounts
[2010/01/27 22:10:27] <jamesturnbull> ohadlevy: which ticket?
[2010/01/27 22:11:18] <nigelk> works fine on a mobile account too. This is a 10.6.3 seed though, so it's not vanilla 10.6.2
[2010/01/27 22:11:30] <ohadlevy> I'm searching.. the one with the pson warning messages
[2010/01/27 22:11:54] <mqr> I'm not pushing #3114 for 0.25.4 because 1) it's not new, 2) there's an easy patch
[2010/01/27 22:11:56] <gepetto> mqr: #3114 is http://projects.reductivelabs.com/issues/show/3114 "Puppet - Bug #3114: Puppet crashes after run - ReductiveLabs.com"
[2010/01/27 22:11:56] <ohadlevy> the actual error I'm talking about is after a puppetd -t
[2010/01/27 22:11:58] <ohadlevy> warning: Value of 'preferred_serialization_format' (pson) is invalid for report, using default (yaml)
[2010/01/27 22:12:43] <mqr> The issue with 3114 will only come up if you are running with more that 32 process groups.
[2010/01/27 22:12:57] <mqr> It's a ruby limitation that has been there since 1.8.1
[2010/01/27 22:13:47] <mqr> There's an easy patch (see the ticket) and I'm considering it the first fix for 0.25.5
[2010/01/27 22:14:02] <jamesturnbull> ohadlevy: it's still a warning
[2010/01/27 22:14:19] <ohadlevy> jamesturnbull: IMHO it should not even be a debug message
[2010/01/27 22:14:56] <ohadlevy> why would any user care about it as a warning?
[2010/01/27 22:15:42] <mqr> ohadlevy: If you request one format and don't get it, that's supposed to generate a warning, by our message guidelines.
[2010/01/27 22:16:14] <mqr> Ask for A, when A isn't possible, you get B, and a warning.
[2010/01/27 22:16:17] <jamesturnbull> nigelk: so you're saying .4rc3 is failing for you to get back o the showstopper three minutes after I built the damn packages... :)
[2010/01/27 22:16:26] <nigelk> hah
[2010/01/27 22:16:27] <ohadlevy> mqr: well... i expect the same amount of noise like all of to_s method warning in debug mode
[2010/01/27 22:16:36] <nigelk> so I've had almost no time to touch puppet 0.25 for weeks
[2010/01/27 22:16:46] <nigelk> I'm not sure I even looked at 0.25.2 or 0.25.3 at all
[2010/01/27 22:17:15] <jamesturnbull> nigelk: anyone would think a nation-state hacked your enterprise and stole all your IP
[2010/01/27 22:17:20] <nigelk> indeed.
[2010/01/27 22:17:21] <jamesturnbull> nigelk: between your slacking off
[2010/01/27 22:17:38] <mqr> nigelk: is your problem with 0.25.4rc3 the 3114 (EINVAL on Process.groups)?
[2010/01/27 22:17:40] <jamesturnbull> nigelk: so that pastie?
[2010/01/27 22:17:47] <nigelk> I believe I've had at least 3 days this year where I was at work for less than 14 hours
[2010/01/27 22:18:00] <nigelk> mqr: no, jamesturnbull was just distracting me :)
[2010/01/27 22:18:01] <jamesturnbull> that sucks
[2010/01/27 22:18:02] <ohadlevy> nigelk: I think you should change jobs ;)
[2010/01/27 22:18:11] <nigelk> I am! that's the problem :)
[2010/01/27 22:18:17] <jamesturnbull> new job?
[2010/01/27 22:18:25] <nigelk> yeah. big shuffling going on.
[2010/01/27 22:18:29] <nigelk> new job at same company.
[2010/01/27 22:18:47] <mqr> nigelk: Yeah, you should work for us. Guarantied at least four days a year < 14 hrs.
[2010/01/27 22:19:10] <ohadlevy> mqr: why would a user want to care about that warning message? esp that now the user dont know if he using pson or yaml etc?
[2010/01/27 22:19:10] <nigelk> heh. This isn't normal. I usually have quite good quality of life.
[2010/01/27 22:19:27] <nigelk> http://pastie.org/798208
[2010/01/27 22:19:30] <nigelk> that's the thing troubling me
[2010/01/27 22:19:44] <nigelk> 0.25.4rc3 on Hardy server, hardy clients same version are fine
[2010/01/27 22:19:47] <nigelk> mac clients are not.
[2010/01/27 22:20:37] <mqr> Ugh.
[2010/01/27 22:21:06] <mqr> nigelk: is there a ticket for this?
[2010/01/27 22:21:39] <nigelk> not yet :) I wanted to see if there was something obvious before I threw that over the wall
[2010/01/27 22:21:41] <mqr> It looks like it's getting the raw to_s sent to it, and trying to interpret it as PSON. Or something.
[2010/01/27 22:22:00] <nigelk> I think I'm also getting #2532 even though no-one else was able to reproduce it....
[2010/01/27 22:22:01] <gepetto> nigelk: #2532 is http://projects.reductivelabs.com/issues/show/2532 "Puppet - Bug #2532: ca_server/ca_port config settings overrides server/masterport config settings - ReductiveLabs.com"
[2010/01/27 22:22:05] * jamesturnbull removes .4 packages from wiki page
[2010/01/27 22:23:16] <mqr> nigelk: What's puppet:///security/screensaver/EngineDefaults.plist look like?
[2010/01/27 22:25:24] <nigelk> a pretty standard OS X plist file... lemme pastie it
[2010/01/27 22:25:39] <mqr> nigelk: could you update the ticket on 2532 to give some hints on how to reproduce it?
[2010/01/27 22:25:57] <nigelk> http://pastie.org/798231
[2010/01/27 22:26:04] <nigelk> mqr: I will absolutely.
[2010/01/27 22:26:30] <nigelk> I've been completely swamped and have let all my bugs lapse, which is entirely my fault
[2010/01/27 22:26:45] <mqr> Also, what's the resource that's generating the EngineDefaults.plist look like?
[2010/01/27 22:26:54] <mqr> Understandable.
[2010/01/27 22:26:58] <nigelk> hah
[2010/01/27 22:27:02] <nigelk> it's all the plists that are failing
[2010/01/27 22:27:11] <nigelk> I think you're onto something... I don't use plists on my linux boxes at all :)
[2010/01/27 22:28:02] <nigelk> http://pastie.org/798233 <- manifest
[2010/01/27 22:28:30] <nigelk> inside a class in a module fwiw
[2010/01/27 22:29:46] <mqr> So here's a thought; what happens if you use preferred_serialization_format=yaml?
[2010/01/27 22:30:20] * nigelk tests
[2010/01/27 22:30:22] <mqr> My thought: some character sequence in that XML is being "special" to something and needs to be escaped.
[2010/01/27 22:30:26] <nigelk> yep.
[2010/01/27 22:30:47] <nigelk> actually.
[2010/01/27 22:30:56] <nigelk> more than just plists are doing it.
[2010/01/27 22:31:03] <mqr> Ok
[2010/01/27 22:31:04] <nigelk> and I'm still getting the same errors
[2010/01/27 22:31:16] <mqr> Same errors with yaml?
[2010/01/27 22:31:25] <nigelk> yes.... which is odd
[2010/01/27 22:31:36] * nigelk restarts everything
[2010/01/27 22:31:56] <mqr> Another lovely theory shot down in its tracks by an ugly fact.
[2010/01/27 22:32:14] <nigelk> invalid option: ---preferred_serialization_format
[2010/01/27 22:32:25] <nigelk> hah. so we have to put that in the conf file ....
[2010/01/27 22:32:51] <nigelk> and now we're cooking with gas.
[2010/01/27 22:33:27] <nigelk> yep. files transferring fine now.
[2010/01/27 22:33:29] <mqr> Or use two '-'s maybe?
[2010/01/27 22:33:35] <nigelk> ...
[2010/01/27 22:33:42] <nigelk> thank you... :)
[2010/01/27 22:34:00] <nigelk> and now you see why I'm not putting in actual bug reports right now ;)
[2010/01/27 22:34:07] <mqr> No problem. That's what I'm here for (well, not according to my wife)
[2010/01/27 22:34:21] <mqr> I'd call that an actual bug.
[2010/01/27 22:35:03] <mqr> And it should be reproducible: transferring that file with pson fails.
[2010/01/27 22:36:03] <nigelk> you think it's just transfers? or a wider problem on OS X?
[2010/01/27 22:36:03] <mqr> So, jamesturnbull (and anyone else who cares to weigh in) does that warrant a 0.25.4rc4, or are we getting a good start on 0.25.5rc1?
[2010/01/27 22:36:28] <mqr> I think it's just transfers (though, admittedly, I'm still guessing)
[2010/01/27 22:37:19] <mqr> Specifically, I think that there's something in that file that pson is choking on, so it comes out like a transporter malfunction in a bad SciFi movie. Not pretty.
[2010/01/27 22:37:37] <mqr> Fix the transport, file comes over ok and life is good.
[2010/01/27 22:37:41] <jamesturnbull> mqr: then I'd say thats rc4
[2010/01/27 22:38:14] <nigelk> lemme try to see what the non-plists are choking on
[2010/01/27 22:38:16] <mqr> Pass on Ubuntu?
[2010/01/27 22:39:04] <mqr> nigelk: yes, that could be very interesting. Either prove me wrong or narrow down the possible character sequences. Either way, very useful.
[2010/01/27 22:40:33] <nigelk> hrm. choking on this too... http://pastie.org/798247 which is a reasonably standard shell script.
[2010/01/27 22:40:51] <nigelk> PS. *FUCK* Huawei and their spurious route setting on 3G cards
[2010/01/27 22:41:55] <jamesturnbull> that's an ... interesting ... shell script
[2010/01/27 22:42:11] <nigelk> :)
[2010/01/27 22:42:28] <nigelk> so these cards tend to set spurious routes on -net 10
[2010/01/27 22:42:40] <nigelk> which completely screws anyone using them internally on say a VPN connection
[2010/01/27 22:43:03] <nigelk> hrm. even my /etc/shells is doing it.
[2010/01/27 22:44:10] <mqr> I've got to run for a few hours starting about 12 minutes ago. When I get back I'll try to set up a quick iteration on the failing files to try to find the problematic character sequence(s) or disprove that theory.
[2010/01/27 22:44:36] <mqr> And I've been bitten by Huawei (didn't get the t-shirt though).
[2010/01/27 22:44:43] <mqr> ttfn
[2010/01/27 22:44:53] <nigelk> thanks mqr
[2010/01/27 22:45:40] <nigelk> ok. it's choking on a completely vanilla /etc/shells that works on Ubuntu clients
[2010/01/27 22:45:55] <nigelk> will get something together once I've grokked how the pson stuff works
[2010/01/27 22:48:03] <jamesturnbull> nigelk: thanks - I'll halt release
[2010/01/27 22:48:06] * jamesturnbull sighs
[2010/01/27 22:48:14] * jamesturnbull needs a cigarette
[2010/01/27 22:48:49] <nigelk> Someone wandered in with your security book the other day jamesturnbull :)
[2010/01/27 22:48:54] <nigelk> they were complaining it's not available on kindle
[2010/01/27 22:49:25] <jamesturnbull> nigelk: oh cool - I am a bit embarassed about that book to be honest
[2010/01/27 22:49:41] <jamesturnbull> nigelk: it was in my very grumpy days when I refused to package things
[2010/01/27 22:49:49] <nigelk> their main comment was that they liked the overall approach
[2010/01/27 22:49:53] <nigelk> i haven't read it myself... :)
[2010/01/27 22:50:05] <jamesturnbull> it was designed to harden bastion hosts
[2010/01/27 22:50:21] <jamesturnbull> with application services - like email. ftp, imap etc
[2010/01/27 22:52:57] <nigelk> yep
[2010/01/27 22:54:16] <jamesturnbull> dog I was an arroagnt little bastard back then ... all of five years ago :P
[2010/01/27 23:01:28] <nigelk> ok. time to go home. i'll try to do some more testing.
[2010/01/27 23:01:35] <nigelk> oh. I should run the unit test suite on OS X.... :)
[2010/01/27 23:01:42] @ Quit: nigelk:
[2010/01/27 23:57:54] <mqr> jamesturnbull: I'm back. Sounds like nigelk thinks it's something OS X specific (not my PSON theory)?

Generated by irclog2html.py 2.6 by Marius Gedminas - find it at mg.pov.lt!