19:39:08 <TheSnide> #startmeeting
19:39:08 <MeetBot> Meeting started Wed Dec 17 19:39:08 2014 UTC.  The chair is TheSnide. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:39:08 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
19:39:12 <TheSnide> hi all
19:39:36 <TheSnide> #chair ssm
19:39:36 <MeetBot> Current chairs: TheSnide ssm
19:39:51 <ssm> 'ello
19:40:00 <h01ger> 'ola
19:40:35 <TheSnide> #chair h01ger
19:40:35 <MeetBot> Current chairs: TheSnide h01ger ssm
19:40:50 <ssm> ready for musical chairs?
19:41:05 <TheSnide> welcome for our new meeting on wed ;-)
19:41:15 <h01ger> np: wish you were here
19:41:23 <h01ger> :)
19:42:11 <TheSnide> some news, 2.1.11 is soonish here. i focused on fixing the misskng parts of cgi gfx
19:42:35 <TheSnide> cdef is here, so is aliasing.
19:43:29 * ssm is happy about the return of cdef
19:43:31 <TheSnide> negative on the other hand.... is almost here
19:44:03 <TheSnide> the graph shows ok, but thz legend align in rrd is *hell*
19:44:23 <TheSnide> so, it isn't aligned proerly yet.
19:44:47 <TheSnide> but i don't think it should delay it further.
19:45:09 <TheSnide> ... i also experimented with a multi format html
19:46:57 <TheSnide> (as gfx)
19:47:13 <TheSnide> ... replace the .html by .xml or .json
19:47:30 <TheSnide> it feels much more natural than the munin-api
19:48:02 <ssm> that _is_ a munin api :)
19:48:18 <TheSnide> the data inside is the tree that goes into html::template
19:48:19 <ssm> but not a dedicated one
19:48:41 <TheSnide> very easy to emit. no dup code.
19:49:39 * ssm humbly suggests using the Accepts: requests header  ( Accept: text/html | Accept: application/json )
19:49:41 <TheSnide> will also be in .11
19:50:06 <TheSnide> if no extznsion, yes
19:50:10 <ssm> :)
19:50:16 <TheSnide> if .html then html
19:50:45 <TheSnide> otherwise it feels _very_ diety
19:50:48 <TheSnide> dirty
19:51:06 <ssm> GET /foo.html/nAccepts: application/json ?
19:51:29 <TheSnide> meh
19:52:54 <TheSnide> any news on makzfile++ ?
19:53:24 <ssm> yup
19:53:49 <ssm> #topic Build refactoring
19:53:54 <ssm> Had good progress last week
19:54:05 <ssm> (where "progress" eventually was "delete Makefile and write it from scratch").
19:54:14 <ssm> I feel much better about that… :)
19:54:37 <ssm> Only a bit of cleanup, and  plugin installation to go now, and that could be copied from the old Makefile as soon as I understand how it does that.
19:54:40 <TheSnide> same with cgis ;-)
19:54:49 <ssm> I've not rebased it on 2.1.10 yet, it is still on top of 2.1.9.  Should not be a problem to rebase when I'm done.
19:55:13 <TheSnide> you can evzn mzege it instead of rebase
19:55:31 <ssm> it's the same amount of work, I think, in case of a merge
19:55:35 <ssm> …conflict
19:55:55 <ssm> but it makes a good perl module, installs all the commandline utilities.
19:56:26 <TheSnide> k
19:56:46 <ssm> I've made a Munin::Location (or Munin::BikeShedThisName) module which is written at build time, which contains the location of the configuration file.  From the perl library paths we get the default locations, and from the configuration, all the overrides.
19:57:10 <TheSnide> anyway i don't care much about the *how*. as long as it is in .11 or .12
19:57:31 <ssm> for shell scripts, a "eval $(munin-env)" command should set the necessary variables (load that module, emit shell variables)
19:58:51 <ssm> yes, I hope to get it done for .11, or .12 if you have something release-worthy before I'm done.
20:00:12 <TheSnide> \o/
20:00:47 <TheSnide> ok, having to go.
20:00:53 * ssm has two other topics to mention: "packaging" and "project gpg key"
20:01:07 <TheSnide> you can close the meeting aftzr you leave ;-)
20:01:17 <ssm> k
20:01:20 <ssm> see you :)
20:01:34 <TheSnide> ... staying a couple of min (interested)
20:02:21 <ssm> #topic packaging
20:02:43 <ssm> 2.0.25-1 is in Debian testing, which is nice. That means it'll probably be the version in the next stable.
20:02:49 <ssm> I have packaged 2.1.10-1, but tests in an already complicated munin master setup didn't work well.
20:02:56 <ssm> I'll try a "defaults" configured 2.1.9-1, and upgrade that to 2.1.10-1, and see if I can reproduce the problems.
20:03:08 <ssm> (Vagrant is grinding up a testlab as I type)
20:04:52 <ssm> #topic PGP signed tarballs
20:04:58 <ssm> Tarballs today are signed, which is nice.
20:05:29 <ssm> They are signed with a personal key, which may be a single point of failure, or it may be hard to get that person to sign releases in perpetuity :)
20:05:35 <ssm> It would probably make sense to have a "Project PGP key" for releasing, available on the project server, or available to more than one project memeber.
20:05:43 <ssm> Project members should then sign this project key, so it is a "leaf node" in the tree, er, web of trust.
20:05:53 <ssm> I think that's how it is done in other projects.
20:06:47 <ssm> any thoughts?
20:08:00 <TheSnide> i'm not an expert. but any snide-redundancy is +1
20:08:10 <ssm> git clone TheSnide…
20:09:03 <TheSnide> ssm, can you propose something solid for the next meeting ?
20:09:07 <ssm> "make tar" and "make release" could then do the gruntwork, as long as you have the private key available.
20:09:39 <TheSnide> a signed tag could be enough, with a commit hook
20:10:00 <ssm> #action ssm will propose something solid for the next meeting
20:10:06 <TheSnide> and having a whitelist of sign
20:10:36 <TheSnide> _automated_ ;-)
20:11:08 <ssm> if a signed tag is pushed, and it is a whitelisted signature, the project server will make a signed tarball?
20:12:15 <ssm> with documentation for project members on "log in to the project server, export project public key, sign it, push to keyserver, push to project server"
20:12:19 <ssm> should work
20:12:51 <ssm> Then the project key would benefit from the PGP trust towards each project member.
20:13:09 <TheSnide> yup
20:13:42 <ssm> That'll go into the munin guide, developer's section
20:13:46 <TheSnide> that what i did @work btw
20:14:44 * ssm is out of meeting topics
20:14:57 <TheSnide> perfect ;-)
20:15:44 <TheSnide> end ?
20:15:47 <ssm> I'll keep the meeting open for 15 more minutes, in case
20:15:56 <ssm> or end now?
20:16:04 <TheSnide> ok, i'm leavkng. thx guys
20:16:12 <ssm> TheSnide: cheers.
20:16:40 <TheSnide> ... you can leave open just in case if you don't mind closing it later ;-)
20:16:45 <ssm> #topic waiting for feedback, or a new topic…
20:16:49 <TheSnide> byby
20:31:07 <ssm> #endmeeting