16:01:40 <andreasn> #startmeeting Cockpit public meeting 2014-11-10
16:01:40 <zodbot> Meeting started Mon Nov 10 16:01:40 2014 UTC.  The chair is andreasn. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:40 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:01:50 <andreasn> #meetingname Cockpit
16:01:50 <zodbot> The meeting name has been set to 'cockpit'
16:02:04 <andreasn> #chair  stefw andreasn mvollmer
16:02:04 <zodbot> Current chairs: andreasn mvollmer stefw
16:02:12 <andreasn> #topic welcome
16:02:17 <andreasn> .hellomynameis andreasn
16:02:19 <zodbot> andreasn: andreasn 'Andreas Nilsson' <anilsson@redhat.com>
16:02:31 <mvollmer> .hellomynameis mvo
16:02:32 <zodbot> mvollmer: mvo 'Marius Vollmer' <marius.vollmer@gmail.com>
16:02:56 <mvollmer> ok, agenda.
16:02:56 <stefw> .hello stefw
16:02:57 <zodbot> stefw: stefw 'Stef Walter' <stefw@redhat.com>
16:03:09 <andreasn> anyone else?
16:03:29 <mvollmer> * multi dash status
16:03:32 <andreasn> seems note
16:03:34 <andreasn> #topic Agenda
16:03:53 <andreasn> * multi dash status
16:04:04 <andreasn> * Fedora 21 Server branding
16:04:15 <andreasn> * translation work
16:04:21 <mvollmer> * web site update
16:04:37 <mvollmer> * atomic?
16:04:40 <andreasn> yeah
16:05:07 <andreasn> sounds good
16:05:20 <mvollmer> * road map, as always
16:05:34 <andreasn> sounds like a good agenda, lets run with thta
16:05:36 <andreasn> that
16:05:39 <andreasn> stefw said he need to leave soon
16:05:41 <mvollmer> stefw, do you want to start?
16:05:44 <andreasn> so lets start with his item
16:05:51 <stefw> sure, the translation stuff ...
16:05:53 <andreasn> #topic translation work
16:06:02 <stefw> so unfortunately, we found another blocker for modularizing real code
16:06:11 <stefw> i was working through modularizing the journal
16:06:19 <stefw> and remembered that we've been putting off translation
16:06:50 <stefw> while reworking code for the modularization, we're cleaning up various things, and this should probably be one of them
16:07:15 <stefw> not only making sure that it's possible to do sane translations for our packages
16:07:28 <stefw> but also tagging stuff correctly as we rework it to put it in a package
16:07:39 <stefw> calling the right internationalization functions, etc.
16:07:44 <stefw> so today i've been working on figuring that out
16:07:51 <stefw> it's coming together well
16:07:59 <stefw> i've figured out expanding pluralization
16:08:11 <stefw> which ended up being pretty easy, using javasript eval()
16:08:24 <stefw> i've also ditched the perl based po2json
16:08:27 <stefw> because i don't like perl :)
16:08:35 <stefw> and moved to a nodejs based one
16:08:45 <stefw> talked with mvollmer about this
16:08:53 <stefw> we want to not require nodejs to build from a tarball
16:09:01 <stefw> but it'll pretty much be required for building from git
16:09:07 <stefw> we already use it for phantomjs
16:09:24 <stefw> in addition RHEL does not have perl-Locale-PO
16:09:35 <stefw> so this does solve several build issues
16:09:40 <stefw> that i've run into there
16:09:49 * stefw thinks that's it
16:10:15 <mvollmer> how does JS eval() work in this context?
16:10:31 <mvollmer> translators write JS and we evaluate it?
16:10:31 <stefw> the gnu plural form is in a C type mini-language
16:10:35 <stefw> that works perfectly with javascript
16:10:45 <stefw> as long as we trust our PO sources, we should be good
16:10:46 <mvollmer> oh.
16:10:59 <stefw> http://localization-guide.readthedocs.org/en/latest/l10n/pluralforms.html
16:11:10 <stefw> see that link for pluralization form examples
16:11:31 <stefw> russian is fun
16:11:44 <stefw> previously we weren't doing correct pluralization
16:11:49 <stefw> and that would have bit us later
16:12:15 <stefw> mvollmer, i'll be working on one gettext domain for all of cockpit for now
16:12:30 <stefw> we get some cache winnings, even though we're loading strings that each iframe doesn't need
16:12:30 <mvollmer> welsh... :-)
16:12:49 <stefw> anything with a mod % operator is gold
16:13:05 <mvollmer> so we import translations into our sources and then ship them all in the rpms, right?
16:13:13 <stefw> yes
16:13:21 <stefw> i'm putting them in the base module for now
16:13:26 <mvollmer> ok.
16:13:31 <stefw> because i'm already reinventing enough stuff here
16:13:37 <stefw> but we can move them later
16:13:38 <stefw> if necessary
16:13:44 <stefw> it doesn't need to be part of the public API, in any case
16:13:53 <stefw> there are some new localization functions that are part of the public api
16:14:01 <mvollmer> how do you like the 'translatable="yes"' markup?
16:14:02 <stefw> ie: loc = cockpit.locale(bundle)
16:14:08 <stefw> mvollmer, that works pretty well
16:14:16 <stefw> obviously it doesn't work for plurals
16:14:17 <mvollmer> it's verbose, no?
16:14:21 <stefw> but that's not a surprise
16:14:32 <stefw> mvollmer, if we want to switch to something more succinct i wouldn't be against it
16:14:42 <mvollmer> let's think about it.
16:14:48 <stefw> ok, i gotta run
16:14:56 <mvollmer> right now, I think we get free scanning from gettext
16:15:01 <mvollmer> stefw, ok, take care!
16:15:12 <andreasn> see you tomorrow!
16:15:18 <andreasn> ok, next topic then
16:15:27 <mvollmer> yep
16:15:35 <andreasn> #topic multi dash status
16:15:38 <mvollmer> right
16:15:45 <mvollmer> coming together very nicely
16:15:53 <andreasn> yes, I agree
16:15:55 <mvollmer> I'll split it into multiple pull requests.
16:16:08 <andreasn> sounds good
16:16:26 <mvollmer> and let's attack PCP correctly this time.
16:16:44 <andreasn> will the reboot messages get into this pull request, or will that come later?
16:17:38 <mvollmer> so after the basics in #1455, i want to tackle PCP for long term monitoring, with a stable, cross-OS API via cockpit-bridge (maybe D-Bus, maybe not)
16:17:47 <mvollmer> need to discuss this with stef
16:18:04 <mvollmer> after that come events.
16:18:06 <andreasn> just so the Events part don't get hanging too long
16:18:37 <andreasn> all right
16:18:38 <mvollmer> ah, and before PCP come avatar and color choosers.
16:18:43 <andreasn> right
16:19:06 <andreasn> I'll be testing this branch out some more
16:19:10 <andreasn> but looks very good already
16:19:15 <mvollmer> can't say how long this will take, but it's the immediate short term road map for me, so maybe two weeks or so.
16:19:25 <andreasn> ok
16:19:40 <andreasn> ok, next topic?
16:19:51 <mvollmer> wait
16:19:56 <andreasn> sure :)
16:20:32 <mvollmer> one sub-topic is probably a bit of bug hunting to get more reliable error reporting when there are network problems etc.
16:20:48 <mvollmer> ok, done. :-)
16:21:04 <andreasn> right, lets discuss that more tomorrow with stefw
16:21:22 <mvollmer> and fche2 if he is available.
16:21:25 <andreasn> yeah
16:21:36 <andreasn> #topic Fedora 21 Branding
16:21:53 <andreasn> so, I did some work on Friday on https://bugzilla.redhat.com/show_bug.cgi?id=1161775
16:22:05 <fche2> hey guys
16:22:08 <andreasn> still work in progress, but hope to be done with it this week
16:22:12 <andreasn> hi fche2!
16:22:20 <andreasn> #chair fche2
16:22:20 <zodbot> Current chairs: andreasn fche2 mvollmer stefw
16:22:29 * fche2 prefers a stool
16:22:36 <mvollmer> :-)
16:22:45 <mvollmer> #stool fche2
16:23:11 * fche2 still has some deliverables in the queue for you guys (and others), in the form of a much simpler api for use by the cockpit-agentd widget
16:23:22 <mvollmer> fche2, great!
16:23:25 <andreasn> I think we should have an easy way to switch out the logos for other distributions too, but we'll tackle that seperatley
16:23:45 <andreasn> I think there is a issue open about that already
16:23:48 <andreasn> ok, done re branding
16:24:09 <mvollmer> intermission for fche2?
16:24:21 <andreasn> yes
16:24:27 <andreasn> what should I call that topic?
16:24:29 <mvollmer> fche2, we ware just saying that now is the time for us to get serious about PCP.
16:24:32 <mvollmer> *were
16:24:42 <andreasn> #topic PCP
16:24:47 <mvollmer> (fche2, we have a meeting here :-)
16:24:56 <fche2> ok
16:25:23 <mvollmer> so now I think we know what we really want from PCP
16:25:43 <fche2> out with it
16:25:55 <mvollmer> which is long term monitoring behind a pretty abstract API that we can implement on all our supported OSes.
16:26:28 <mvollmer> so, we need to sort out background monitoring
16:27:26 <mvollmer> that API should also do real-time monitoring, with higher frequencies.
16:28:32 <mvollmer> should be right in the comfortzone of PCP, no?
16:28:42 <fche2> ok.  so viewing pcp as an api, an app such as cockpitd can do all those things directly now, in the sense that pcp is a cross-platform api, which a client can use to poll data as (in)frequently as desired,
16:29:02 <fche2> (the simpler api under construction will turn those into few-liners rather than few-dozen-liners)
16:29:25 <mvollmer> does pcp have it owns background data gatherer?
16:29:33 <fche2> viewing pcp as a service, where another process does background monitoring, logging, that too is there
16:29:41 <mvollmer> right
16:29:49 <fche2> (and those logs can be read by pcp clients via the api or command line tools, now)
16:30:42 <mvollmer> ok, I guess we just have to figure out the details and best practices.
16:30:44 <fche2> so in a way all you need is a configuration of these things that works for you, and perhaps a bit of RPC where pcp doesn't already
16:31:17 <fche2> (that RPC angle comes in when one's talking about logging across the network)
16:31:43 <mvollmer> fche2, sounds great!  I'll get back to you in a few days, ok?  :-)
16:31:52 <fche2> certainly.
16:32:04 <mvollmer> let's continue the meeting.  thanks!
16:32:08 <fche2> so far I was focusing on a little bit of the overall angle, but if y'all are ready to talk big picture, whenever, sure
16:32:22 <mvollmer> yep
16:32:26 <andreasn> thanks fche2!
16:32:30 <andreasn> ok, next topic
16:32:31 <fche2> np, any time
16:32:36 <andreasn> #topic website updates
16:32:50 <andreasn> so I filed a bunch of issues I hope to get to this week https://github.com/cockpit-project/cockpit-project.github.io/issues
16:32:58 <andreasn> did I miss anything we discussed?
16:34:23 <mvollmer> maybe videos?
16:34:44 <andreasn> some screencasts would be great yeah
16:34:46 <mvollmer> my biggest question is how we approach Atomic
16:34:59 <andreasn> no idea
16:35:12 <mvollmer> there is project atomic, fedora atomic, and centos atomic.
16:35:34 <mvollmer> we can just point to project atomic, or directly to fedora or centos.
16:35:42 <andreasn> what of them include cockpit?
16:35:48 <mvollmer> all
16:36:14 <andreasn> right. Any of them have a more up to date version than any other?
16:36:26 <andreasn> Newer cockpit = better
16:36:32 <mvollmer> well, project atomic is mostly about the tools, like rpm/yum; Fedora Atomic and CentOS Atomic are the actual OSes.
16:36:58 <mvollmer> don't know what they have or how they update.
16:37:19 <mvollmer> what about making a page with all the ways to get cockpit that we know of?
16:37:31 <andreasn> yeah, sounds good
16:37:39 <mvollmer> and link to that from the front page?
16:38:09 <andreasn> sure. Can you file an issue about that?
16:38:14 <mvollmer> yes.
16:38:18 <andreasn> just so I don't forget when I work through them
16:38:38 <mvollmer> it's related to #12.  should I change that?
16:39:11 <andreasn> yeah, just leave a comment
16:39:41 <mvollmer> ok
16:39:52 <andreasn> I mean, the main goal is to try out Cockpit, but having a choice between the kind of OS, atomic or more regular will be good
16:40:38 <andreasn> but yeah, lets continue discussing on the issue tracker
16:41:00 <andreasn> next topic is Atomic
16:41:04 <andreasn> #topic Atomic
16:41:48 <mvollmer> stefw managed to get Cockpit running via a container, awesome stuff.
16:41:54 <andreasn> yeah, that's cool
16:42:06 <andreasn> is it fully functional?
16:42:17 <mvollmer> no
16:42:32 <mvollmer> I think we can call it "Cocker". :-)
16:42:46 <andreasn> tss
16:43:10 <andreasn> is it in master yet?
16:43:12 <mvollmer> anyway, it works well enough and personally I see it more as a discussion starter
16:43:18 <mvollmer> yes, it's in master.
16:43:18 <andreasn> yes
16:43:21 <andreasn> great
16:43:29 <andreasn> how do I try it out?
16:43:43 <mvollmer> ./tools/make-docker
16:44:03 <mvollmer> and then some magic docker run command.
16:44:13 <andreasn> neat
16:44:42 <andreasn> do we need to discuss anything about the Atomic features on the roadmap, or should we wait with that, since other things are coming up?
16:44:51 <mvollmer> right now, both Fedora and CentOS atomic include Cockpit as rpms, but maybe they want to stop doing that.
16:45:22 <andreasn> Fedora Atomic and CentOS Atomic, right?
16:45:26 <mvollmer> but in any case, Cockpit in Atomic needs to be better tailored to it.
16:45:50 <mvollmer> to reduce unwanted dependencies, and to improve UX.
16:45:56 <andreasn> yes
16:46:16 <mvollmer> this will come with/after the modularization has been completed.
16:46:34 <andreasn> ok, so lets wait with those feature discussions a bit then
16:47:43 <andreasn> onto roadmap?
16:47:55 <mvollmer> that should be the last topic
16:48:00 <mvollmer> do we have anything else?
16:48:06 <andreasn> #topic roadmap
16:48:23 <mvollmer> ok
16:48:34 <mvollmer> andreasn, how do you feel about the patternfly update?
16:48:37 <andreasn> I'll add website updates to the roadmap
16:48:51 <andreasn> it needs to be done, but I haven't started tackling it yet
16:49:19 <andreasn> but should be doable within 4 weeks
16:49:41 <andreasn> we should add PCP there too
16:49:49 <mvollmer> yes
16:49:50 <andreasn> what should that point be called?
16:51:19 <andreasn> "PCP intergration?"
16:51:25 <mvollmer> "Cross-OS API for long term monitoring of resources and important events."
16:52:10 <andreasn> added that and the website updates
16:53:16 <mvollmer> ok
16:53:22 <andreasn> ok, open floor next I guess
16:53:26 <andreasn> #topic open floor
16:53:36 <andreasn> anything else for todays meeting?
16:53:39 <andreasn> I don't really have anything
16:53:57 <mvollmer> andreasn, what about removing Navigation redesign from the road map and putting multi-server dashboard makeover on it?
16:54:20 <andreasn> yes, sounds good
16:54:49 <mvollmer> are you going to edit the wiki?
16:55:22 <andreasn> I did already
16:55:31 <andreasn> reload :)
16:55:43 <mvollmer> yes! :-)
16:56:02 <andreasn> all right, that's all for today then
16:56:08 <andreasn> #endmeeting