13:00:51 <andreasn> #startmeeting Cockpit Weekly Meeting 2016-05-09
13:00:51 <zodbot> Meeting started Mon May  9 13:00:51 2016 UTC.  The chair is andreasn. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:51 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
13:00:51 <zodbot> The meeting name has been set to 'cockpit_weekly_meeting_2016-05-09'
13:01:23 <andreasn> .hello andreasn
13:01:23 <zodbot> andreasn: andreasn 'Andreas Nilsson' <anilsson@redhat.com>
13:01:44 <dperpeet> .hello dperpeet
13:01:45 <zodbot> dperpeet: dperpeet 'None' <dperpeet@redhat.com>
13:01:49 <harish__> .hello harish__
13:01:50 <zodbot> harish__: Sorry, but you don't exist
13:02:01 <stefw> .hello stefw
13:02:03 <zodbot> stefw: stefw 'Stef Walter' <stefw@redhat.com>
13:02:09 <larsu> .hello larsu
13:02:10 <harish__> .hello harish
13:02:15 <zodbot> larsu: larsu 'Lars Uebernickel' <lars@uebernic.de>
13:02:18 <zodbot> harish__: harish 'Harishwar' <harishwar91@yahoo.co.in>
13:02:41 <andreasn> #topic agenda
13:02:44 <dperpeet> harish__ you need to have a registered nickname with the server
13:03:23 <stefw> * Per package javascript builders
13:03:23 <harish__> okay. im currently on a bnc. thats y i think
13:04:34 <mvollmer> .hello mvo
13:04:35 <zodbot> mvollmer: mvo 'Marius Vollmer' <marius.vollmer@gmail.com>
13:04:43 <dperpeet> * docker page rework
13:05:46 <harish__> * timer
13:06:27 <andreasn> sounds good. All right
13:06:41 <andreasn> #topic Per package javascript builders
13:07:03 <stefw> larsu and I have been discussing the package builders
13:07:04 <stefw> https://github.com/cockpit-project/cockpit/issues/4384
13:07:10 <stefw> we're putting some notes in that issue there ^^
13:07:22 <stefw> still need to write down the goals
13:07:43 <stefw> but essentially the idea that it should be possible to use webpack (or something similar) to build the bundles in a module
13:08:44 <mvollmer> what can webpack do?
13:08:48 <stefw> one of the real problems we have is that the JSX transformation broke the
13:08:49 <mvollmer> minification?
13:09:04 <stefw> way that we allow people to link to a directory and refresh it when things change
13:09:29 <stefw> mvollmer, there's various plugins (some default) that do things like that
13:09:30 <dperpeet> webpack is a monitoring system, right? no in-browser transformation
13:09:39 <stefw> the interesting thing about webpack is that it's rather focused compared to grunt or gulp
13:09:48 <stefw> the latter two are essentially 'make' replacements
13:10:04 <stefw> webpack has a monitoring setup yes, so when files change it can turn around and rebundle them
13:10:24 <dperpeet> nice
13:10:51 <stefw> so as outlined in the bug we're thinking of having a well known subdirectory that the bundled, minified, output goes into ready to be served
13:11:00 <dperpeet> it would be nice if we could use this out of the box, i.e without having to set up more stuff after checking out cockpit
13:11:17 <stefw> yeah that would be nice
13:11:22 <dperpeet> subdirectory of the build dir, right?
13:11:26 <stefw> obviously linking it into place
13:11:28 <dperpeet> because I don't touch my source tree
13:11:45 <cockpitbot> 2 tests failed - http://fedorapeople.org/groups/cockpit/logs/master-2ef2d0b4-verify-rhel-7/log.html
13:11:51 <stefw> dperpeet, that's a good question ... but for people not using autogen.sh or make there would be no build dir
13:12:16 <stefw> but it could setup in such a way
13:12:20 <dperpeet> ok
13:12:31 <dperpeet> at the very least it shouldn't overwrite files that are in the source tree
13:12:35 <stefw> that if you do run autogen.sh then all the package output or dist directories follow suit on the build dir
13:12:38 <stefw> dperpeet, yup definitely not
13:12:42 <dperpeet> for example, I sync from my dev machine to the target machines
13:13:00 <andreasn> same for me
13:13:13 <dperpeet> then I have no objection
13:13:43 <dperpeet> do I still need to log out and back in for the changes to take effect?
13:13:45 <dperpeet> of cockpit
13:14:10 <dperpeet> not having to do that would be a major boon
13:14:23 <andreasn> and only reload?
13:14:24 <stefw> well once you've enabled dev mode for a given package, no further logging out and logging in is necessary
13:14:47 <stefw> well there'll be something (probably webpack) running that regenerates things as they change
13:14:50 <stefw> and then yes, you can just reload
13:14:58 <dperpeet> great
13:15:28 <stefw> ok, well we can probably move on
13:15:31 <andreasn> yep
13:15:33 <stefw> there'll be tons more to talk about this as we try it out
13:15:46 <andreasn> #topic Docker page rework
13:16:14 <larsu> working on it :)
13:16:39 <larsu> need some design input :)
13:16:47 <andreasn> sure
13:16:55 <andreasn> is there a branch I can try already?
13:17:04 <dperpeet> I'm reworking the dialog
13:17:12 <dperpeet> and marius is touching on the storage parts
13:17:34 <dperpeet> I tried to get as many of stef's relevant changes to the css classes in as I could
13:17:39 <larsu> andreasn: no, not yet sorry. I'm wiring up client.js with the react components I've been working on
13:17:50 <larsu> which proves to be a bit more work than I thought it would be
13:17:56 <larsu> and it's really not in a usable state right now
13:17:56 <dperpeet> larsu, what components are you working on?
13:17:57 <andreasn> larsu: just let me know when you have something you want me to try
13:18:03 <dperpeet> so we don't duplicate effort
13:18:21 <dperpeet> I'm moving select to base1 and reworking some of the ux issues we have with the dialog and sliders
13:18:24 <larsu> dperpeet: right now the container and image listings
13:18:31 <larsu> I also have some of the tabs done
13:18:34 <dperpeet> ok, so no new standard components at this point
13:18:35 <larsu> just not wired up yet :)
13:18:40 <andreasn> larsu: any particular design thing that is blocking right now?
13:18:52 <larsu> dperpeet: nope, but I could really use the detail view thing you've been working on
13:19:03 <andreasn> like, if there is something I should dive in and flesh out today
13:19:08 <larsu> andreasn: yes, but I'm sure more will come up. Maybe we could have a chat some time this week?
13:19:15 <andreasn> larsu: sure, sounds good!
13:19:21 <larsu> stefw is here and we're working on the bundling stuff today
13:19:24 <dperpeet> larsu, do you mean the rolled out tabs?
13:19:26 <larsu> wed or so would be better
13:19:29 <larsu> dperpeet: yes
13:19:46 <dperpeet> larsu, are you blocking on that? it moved way down the list :)
13:19:49 <larsu> andreasn: the one issue I had today is that I don't like the empty state, which would be two empty tables
13:19:56 <larsu> dperpeet: nope
13:20:04 <larsu> way too much to do right now
13:20:07 <larsu> :)
13:20:08 <dperpeet> ok
13:20:14 <dperpeet> ping me if it does become more pressing
13:20:21 <larsu> of course ;)
13:20:43 <andreasn> larsu: ah. I'll look into that
13:21:15 <dperpeet> EmptyState is more for cases where the page / large section is empty, right?
13:21:21 <dperpeet> not if a little sub-component is empty
13:22:05 <larsu> yes, I think so
13:22:23 <dperpeet> ok
13:22:24 <andreasn> we have tiny empty states in some places, such as for iscsi in Storage
13:22:32 <andreasn> with a little bit of helper text
13:22:37 <andreasn> could be worth looking into
13:23:18 <andreasn> ok, next up?
13:24:35 <andreasn> #topic Timers
13:25:02 <harish__> last week i found out timers properties could not be set via dbus.
13:25:16 <harish__> so we can file an RFE this week
13:25:26 <harish__> to get a function for major properties of timer like calendar and monotonic ones
13:25:30 <andreasn> against systemd?
13:25:35 <harish__> yes
13:26:09 <harish__> this week i can work around make a playground app for creating and setting timers/services
13:26:25 <dperpeet> but there's a workaround, right? create files on disk
13:26:27 <harish__> and also i like look into how to do modification of timers as well.
13:26:43 <dperpeet> so future work doesn't block on the rfe
13:27:49 <harish__> we can always hard code values to timer file.
13:28:20 <andreasn> anything unclear about the wireframes at this point? Apart from the dropdown (that I seems to have forgot what it was for)?
13:28:34 <harish__> so there is a way to implement it. but having dbus method is the correct approach
13:28:48 <cockpitbot> 1 tests failed - http://fedorapeople.org/groups/cockpit/logs/master-2ef2d0b4-verify-fedora-atomic/log.html
13:29:18 <dperpeet> nice
13:29:21 <harish__> andreasn, i think we can look into that once i get started with the work. mockup helped me get an idea about the wrk
13:29:27 <dperpeet> you're working on the wiki page right now, correct?
13:29:44 <harish__> yes. i hope file rfe by today or tomorrow
13:29:50 <dperpeet> great
13:30:00 <harish__> here is the link to blog post i made for last weeks work https://medium.com/@harishanand95/getting-started-with-gsoc-2016-with-cockpit-619f7e61bebc#.gn4bvwyri
13:30:07 <dperpeet> https://github.com/cockpit-project/cockpit/wiki/Feature:-Systemd-timers
13:31:09 <dperpeet> nice
13:31:13 <harish__> okay. i will modify wiki contents today.
13:31:16 <andreasn> I added some previous art of other timer UIs on the other page, or was it this page?
13:31:30 <harish__> no. i found it in trello
13:31:31 <cockpitbot> 4 tests failed - http://files.cockpit-project.org/logs/master-2ef2d0b4-verify-fedora-23/log.html
13:31:42 <harish__> https://trello.com/c/1B2lZViZ/74-timers-and-cron
13:31:47 <dperpeet> andreasn, you added it there, but it was lost during a revision
13:31:54 <andreasn> ah, I can add them again
13:31:57 <dperpeet> harish__ will add that back in :)
13:32:03 <harish__> oh okay
13:32:10 <andreasn> I only for FreeNAS and Easycron
13:32:27 <andreasn> I think there is a ton of others out there
13:32:34 <cockpitbot> 2 tests failed - http://fedorapeople.org/groups/cockpit/logs/master-2ef2d0b4-verify-fedora-testing/log.html
13:32:46 <andreasn> of varying quality :)
13:33:17 <harish__> oh. i will look into it.
13:33:21 <andreasn> harish__: feel free to ping at any time with any design questions. I'll try and reply as soon as I see it
13:33:31 <harish__> thanks andreasn
13:33:45 <harish__> that's it from me.
13:33:50 <harish__> and thanks dperpeet
13:34:05 <andreasn> #topic Open Floor
13:34:23 <dperpeet> we need to get https://github.com/cockpit-project/cockpit/pull/4382 merged soon
13:34:29 <dperpeet> master is broken in debug mode
13:34:34 <dperpeet> since modal dialogs can't open
13:34:51 <dperpeet> this affects tuned and the playground currently
13:35:16 <dperpeet> also, stef has been working the cockpit "ct" into css classes
13:35:21 <dperpeet> (namespacing what we use)
13:35:35 <dperpeet> so if you run intro strange issues, you may have found something that slipped through the cracks
13:35:38 <andreasn> I think it's fine to go in. Unless someone else wants to do a review as well
13:36:10 <mvollmer> works for me
13:36:17 <mvollmer> i can merge
13:36:28 <dperpeet> thanks
13:37:53 <github> [cockpit] mvollmer pushed 1 new commit to master: https://git.io/vre1f
13:37:53 <github> cockpit/master 5b3e9a1 Dominik Perpeet: tuned: Fix dialog css class when visible...
13:38:01 <mvollmer> something re docker storage, maybe
13:38:10 <mvollmer> story goes on
13:38:20 <mvollmer> I just removed the dependency on storaged... :-)
13:38:33 <andreasn> how does it work now?
13:38:41 <andreasn> what do you depend on instead?
13:38:44 <mvollmer> adding things to "atomic storage" seems to be agreed upon, implementation further down is still open
13:38:58 <mvollmer> andreasn, straight udev via pyudev
13:39:03 <andreasn> ah, ok
13:39:13 <mvollmer> i copied the logic from storaged
13:39:26 <mvollmer> udev properties are a zoo
13:39:47 <andreasn> full of funky animals?
13:39:56 <mvollmer> yeah, better kept behind bars
13:39:58 <andreasn> funky creatures
13:40:19 <mvollmer> and lots of smelly sh*t, if you are in a foul mood :-)
13:40:32 <andreasn> tss
13:40:47 <dperpeet> mvollmer, it would be nice if you could split some of that work into separate pull requests
13:40:54 <dperpeet> make it a bit less monolithic
13:41:08 <mvollmer> i can try
13:41:22 <mvollmer> but right now it's mostly adding two or three files
13:41:34 <mvollmer> not sure what the benefit is of splitting that up
13:41:40 <dperpeet> updating translation files is a good candidate :)
13:41:48 <mvollmer> it's not changing a lot of existing code
13:42:02 <dperpeet> well, right now 1700 lines changed
13:42:12 <dperpeet> so the benefit is having the review be manageable :)
13:42:18 <mvollmer> so one PR for breaking translations, the other for fixing it again? :-)
13:42:35 <dperpeet> I'm not saying you have to, just consider it
13:42:39 <mvollmer> sure
13:43:02 <dperpeet> like this one https://github.com/cockpit-project/cockpit/pull/4350/commits/f9eed487c608f26f4bc6654a44773c8cfc853abc
13:43:53 <mvollmer> i can split out changing the storage overview box in the existing page
13:44:13 <dperpeet> if it's not too much work, that might be nice
13:46:58 <andreasn> anything else?
13:47:02 <cockpitbot> 5 tests failed - http://fedorapeople.org/groups/cockpit/logs/master-2ef2d0b4-verify-fedora-24/log.html
13:47:02 <andreasn> or should I close the meeting?
13:48:28 <andreasn> sounds like it
13:48:31 <andreasn> #endmeeting