[10:03:01] <arturo>	 do you know off hand if we have something in puppet.git to apply a patch to a file? instead of just deploying a file
[10:03:08] <arturo>	 I'm just trying not to write such thing myself
[10:09:47] <moritzm>	 what do you want to do? apply a patch to a file which is otherwise shipped from a deb package?
[10:10:01] <arturo>	 moritzm: exactly
[10:10:28] <arturo>	 I will just use an exec { running the patch command I think
[10:11:17] <moritzm>	 we're doing this in Icinga, let me find it
[10:11:45] <moritzm>	 modules/icinga/manifests/init.pp
[10:13:29] * arturo nods
[10:13:33] <moritzm>	 it's not really clean, it breaks some of the assumptions of the dpkg package state and mid term I want to make this impossible (as the same could also be applied maliciously)
[10:13:58] <moritzm>	 but that's nothing which will happen in the near future, other things have priority
[10:14:08] <arturo>	 ack
[10:14:20] <moritzm>	 so feel free to apply this for whatever Openstack migration you're handling :-)
[10:34:58] <_joe_>	 arturo: augeas
[10:35:27] <_joe_>	 that's how you programmatically patch files
[10:35:41] <arturo>	 but that won't work for source code files?
[10:36:02] <_joe_>	 oh sorry
[10:36:27] <_joe_>	 so you're patching the file? just copy the patched file in puppet and install it with file { ...
[10:36:46] <_joe_>	 it's more explicity and allows to properly manage it
[10:37:13] <_joe_>	 but yeah the clean way to do it is to rebuild the deb and use quilt to apply the patch
[10:37:26] <arturo>	 that's how we have been doing it for years (deploying the already patched file)
[10:37:33] <arturo>	 but is a relatively long file, with a very simple patch
[10:37:39] <_joe_>	 don't use an exec please :)
[10:37:55] <arturo>	 and tracking the patched file itself is complex
[10:38:02] <arturo>	 I would rather handle the .patch file
[10:38:17] <_joe_>	 then do as you whish
[10:38:39] <arturo>	 the ultimate goal is to get rid of *any* patching. This is my effort to reduce our custom hack, the first iteration
[10:39:09] <_joe_>	 again, the right way to do it is I think to add your patches to the deb packages
[10:39:22] <_joe_>	 but I might not see some complications
[10:39:47] <_joe_>	 I might suggest, if you plan to reuse this technique, to make this a define with some spec tests
[10:40:05] <_joe_>	 something like openstack::patch_file
[10:40:22] <_joe_>	 or whatever name you prefer
[10:40:39] <arturo>	 I agree
[10:41:23] <arturo>	 will share my code when in gerrit
[15:38:45] <chaomodus>	 o7
[15:40:19] <apergos>	 o8?
[15:46:04] <rlazarus>	 jbond42: can you teach me about the capitalization in that patch? I experimented a little and I can see that it's right, but I don't understand why
[15:46:42] <rlazarus>	 how come puppetdb finds results for "Class/Role::Grafana" but not "role::grafana" even though it's lower-cased in grafana.pp?
[15:47:10] <_joe_>	 rlazarus: when referring to resource, you capitalize them
[15:47:29] <_joe_>	 so foo::bar { 'something': } => Foo::Bar['something']
[15:47:37] <akosiaris>	 https://puppet.com/docs/puppet/latest/lang_data_resource_reference.html
[15:47:42] <_joe_>	 same in puppetdb, it saves resources capitalized
[15:47:51] <_joe_>	 like you refer to them once they're realized
[15:47:54] <akosiaris>	 it's like *p for a pointer dereference in C
[15:48:09] <rlazarus>	 huh, okay
[15:48:17] <rlazarus>	 that is very weird, thanks
[15:48:20] <akosiaris>	 mind bending right? I agree
[15:51:26] <rlazarus>	 I guess it's two weird things, it's weird that the language works that way and it's weird that puppetdb searches by the reference and not the type name
[15:52:37] <rlazarus>	 ... *and* it's weird that puppetdb doesn't just do a case-insensitive match, since that whole setup guarantees it'll never be ambiguous :)
[15:53:16] <rlazarus>	 so there must be something I haven't grokked yet
[15:53:44] <jbond42>	 rlazarus: you will hurt your self if you try to understand all puppets eccentricities ;)
[15:53:53] <jbond42>	 its not you its puppet ;)
[15:54:20] <rlazarus>	 haha okay, fair, I guess it's not the first time I've been warned about that
[16:00:51] <_joe_>	 rlazarus: it's a different thing though
[16:00:57] <_joe_>	 one is the declaration of the resource
[16:01:03] <_joe_>	 the other is the /realized/ resource
[16:01:07] <_joe_>	 don't think of it as a pointer
[16:01:48] <_joe_>	 but yes, remember, always, https://bash.toolforge.org/quip/AVfTAUmefIH_7EDsriqu
[16:02:16] <_joe_>	 rlazarus: also there are interesting parser bugs related to capitalization
[16:06:27] <rlazarus>	 oh?
[16:06:50] <rlazarus>	 I'm fascinated, in the sense of "horrified but I can't look away"
[16:15:14] <_joe_>	 there is some reference in the puppet tree to it
[16:15:20] <_joe_>	 but basically before puppet 5
[16:15:37] <_joe_>	 capitalization rules meant you could have either type foo::bar::baz
[16:15:44] <_joe_>	 or class foo::bar
[16:15:48] <_joe_>	 but not both
[16:15:59] <_joe_>	 or better
[16:16:15] <_joe_>	 you couldn't use type foo::bar::baz inside any class under foo
[16:16:17] <_joe_>	 :P