cockpit_weekly_meeting_2016-04-18
LOGS
13:01:06 <andreasn> #startmeeting Cockpit weekly meeting 2016-04-18
13:01:06 <zodbot> Meeting started Mon Apr 18 13:01:06 2016 UTC.  The chair is andreasn. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:01:06 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
13:01:06 <zodbot> The meeting name has been set to 'cockpit_weekly_meeting_2016-04-18'
13:01:13 <andreasn> .hello andreasn
13:01:14 <zodbot> andreasn: andreasn 'Andreas Nilsson' <anilsson@redhat.com>
13:01:17 <stefw> .hello stefw
13:01:18 <zodbot> stefw: stefw 'Stef Walter' <stefw@redhat.com>
13:01:25 <dperpeet> .hello dperpeet
13:01:26 <zodbot> dperpeet: dperpeet 'None' <dperpeet@redhat.com>
13:02:20 <andreasn> #topic Agenda
13:03:05 <andreasn> * list pattern
13:03:17 <dperpeet> * removing jquery from cockpit.js
13:03:25 <dperpeet> * patternfly update
13:03:38 <larsu> stefw: that link 404s for me
13:04:12 <stefw> * Ubuntu PPA packages
13:04:14 <dperpeet> larsu, you're too late, it's not a draft anymore: http://stef.thewalter.net/stop-deploying-packages.html
13:04:33 <larsu> ah, right. was at lunch :)
13:05:51 <andreasn> ok, lets get started then
13:05:54 <andreasn> #topic list pattern
13:06:30 <andreasn> so, I've done some more work on the list pattern. Made some more variantions of them in different situations after feedback from dperpeet
13:06:49 <andreasn> https://raw.githubusercontent.com/cockpit-project/cockpit-design/master/patterns/listing-v2.png
13:06:54 <andreasn> (massive png)
13:07:26 <stefw> nice
13:07:29 <stefw> should we summarize the goals
13:07:31 <stefw> of the change?
13:07:34 <stefw> and what it does?
13:07:34 <mulhern> .hell mulhern
13:07:39 <mulhern> .hello mulhern
13:07:42 <zodbot> mulhern: mulhern 'None' <amulhern@redhat.com>
13:07:56 <andreasn> yes, so in brief:
13:08:31 <andreasn> * it's a way for expanding when needed, navigating to a separate page
13:09:05 <andreasn> * click the item to move to a separate page, showing all the properties of the thing
13:09:24 <andreasn> * click the caret to expand inline, allowing to compare things
13:10:04 <andreasn> * this tries to standardize the different variantions of this pattern that we use in several places already
13:10:29 <dperpeet> * on expansion, the title line shouldn't change much / at all
13:10:32 <andreasn> * checkbox button goes into multi-selection mode
13:11:10 <andreasn> * on multi-select, you're able to do actions on several objects at once
13:11:28 <andreasn> * this also shows how filtering works
13:11:35 <andreasn> I think that's pretty much it
13:11:48 <stefw> cool
13:11:51 <dperpeet> looks great
13:11:55 <mvollmer> .hi mvo
13:11:57 <dperpeet> I have one question though
13:12:02 <mvollmer> .hello mvo
13:12:03 <zodbot> mvollmer: mvo 'Marius Vollmer' <marius.vollmer@gmail.com>
13:12:15 <dperpeet> when navigating to an item and back, what happens to the previous view?
13:12:23 <dperpeet> e.g. applied filters, expansions etc
13:12:34 <dperpeet> mvollmer, you can't be too informal with zodbot, it'll kick you out
13:12:47 <mvollmer> :)
13:12:48 <andreasn> you mean if it resets or stays the same?
13:12:50 <mvollmer> was in a hurry
13:12:51 <stefw> dperpeet, that's a good question
13:12:59 <stefw> how much work do you think it'll be to get it to look exactly the same?
13:13:11 <dperpeet> depends on how we solve the navigation
13:13:19 <stefw> because we do have expansion of an item in the listing as well
13:13:34 <dperpeet> I'm concerned with "losing" time here
13:13:36 <stefw> so for people who really mean business for everything to stay put and just peek into something, they could expand it
13:13:44 <stefw> who's time?
13:13:48 <dperpeet> if I apply a bunch of filters and then click on something, I might lose my time spent on that
13:14:01 <dperpeet> as a user
13:14:01 <stefw> hmmm, yes
13:14:05 <stefw> we could split out the concepts though
13:14:10 <stefw> and say: Filters staying is a requirement
13:14:14 <stefw> sorting staying is a requirement
13:14:25 <stefw> but as far as scrolling, expansion state, and those things ... nice to have?
13:14:27 <stefw> what do you think?
13:14:35 <dperpeet> hm
13:15:04 <dperpeet> we can work around this by saying we model navigation to one item by adding a filter to select that one item
13:15:27 <dperpeet> but the question is, does the ui use the little back button or show that?
13:16:00 <stefw> if we do go down this road, then we should make sure the filtering and such operatiens update the address bar
13:16:04 <stefw> so the user can bookmark that state
13:16:07 <dperpeet> we don't need to hash thish out now
13:16:08 <stefw> and can use the browser navigation, etc
13:16:10 <dperpeet> yeah
13:16:19 <stefw> but it's good to set bounds around what this means and how much work it is
13:16:21 <dperpeet> let's just keep that in mind when implementing this
13:16:37 <dperpeet> I'm willing to give this new design a spin with the services page
13:16:43 <andreasn> sounds good
13:16:44 <dperpeet> unless someone else needs to pick it up first
13:17:04 <dperpeet> I can rework the troubleshoot react template to match this and add filtering
13:17:16 <stefw> wasn't the docker page up for design?
13:17:20 <stefw> or refactoring?
13:17:38 <dperpeet> sure, but that needs a code refactor as well
13:17:39 <andreasn> the docker redesign kind of blocks the scap scanning
13:18:03 <dperpeet> I can make a pr for the template
13:18:09 <andreasn> cool
13:18:10 <dperpeet> and docker work can be based on that
13:18:27 <andreasn> sounds good
13:18:30 <stefw> i wonder if it makes sense to work together with lars on this
13:18:40 <stefw> and use the docker stuff as the point where we first implement this and try it out
13:18:47 <stefw> what do you think?
13:18:53 <dperpeet> also a possibility
13:19:14 <dperpeet> getting the services stuff I've already done off my plate was a personal goal, but I'm willing to set that back some more
13:19:26 <dperpeet> larsu, what are your docker page plans?
13:19:44 <stefw> well obviously up to the contributors ... just an idea
13:19:56 <dperpeet> or how about this: I rework the setroubleshoot template to be a bit more generic and look like this, without the filtering
13:19:58 <stefw> usually easier to do such work in one place and get it right, and then expand to other ones
13:20:09 <dperpeet> that should be done pretty quickly
13:20:42 <dperpeet> and needs to happen anyway if the pattern is to be made available to other packages
13:20:50 <github> [cockpit] petervo closed pull request #4224: test: Properly prune filenames (master...vmimages) https://git.io/vwq8c
13:21:32 <dperpeet> lars can start with refactoring the docker page to split the controller stuff out
13:21:34 <dperpeet> so he won't block
13:21:47 <dperpeet> I'll coordinate with lars
13:21:50 <dperpeet> end of topic on my end
13:21:59 <andreasn> all right, next up
13:22:06 <dperpeet> either way, I'll make sure this goes into the generic template
13:22:09 <stefw> nice
13:22:16 <andreasn> sounds like a good plan
13:22:24 <andreasn> #topic removing jquery from cockpit.js
13:22:41 <dperpeet> stefw made this happen
13:22:47 <stefw> It's one of the things that just got merged as part of the internal API stabilization in Cockpit
13:22:53 <stefw> It uses real promises now
13:23:00 <stefw> but they're still jQuery compatible
13:23:07 <stefw> so not much of the other code in cockpit needed to change
13:23:22 <stefw> and this is intended to future proof stuff, since various parts of cockpit are no longer using jQuery
13:23:28 <dperpeet> thanks for doing this
13:23:49 <github> [cockpit] petervo closed pull request #4225: tuned: Return early when event is empty (master...dialog-fix) https://git.io/vwq8K
13:23:49 <dperpeet> we'll benefit most when using smaller react components
13:23:59 <dperpeet> since they usually don't need jquery either
13:24:29 <dperpeet> stefw, you may have to amend your using dbus in cockpit examples
13:24:49 <stefw> does it need changes?
13:24:55 <stefw> that does remind me that the documentation does need some more updates for this
13:24:56 <dperpeet> I'm wondering if you're using $ somewhere
13:25:20 <stefw> should be fine no?
13:25:29 <stefw> the goal here isn't to move away from jQuery
13:25:31 <stefw> but to allow that to happen
13:25:37 <dperpeet> ok, let's just keep in mind that $ might work in fewer places now
13:25:39 <dperpeet> when inspecting
13:26:10 <stefw> i don't think that's the case
13:26:18 <stefw> previously cockpit.js would expose less functionality when jQuery wasn't loaded
13:26:27 <stefw> but i guess there may have been side effects?
13:26:45 <stefw> now it exposes the same amount of functionality with or without jQuery
13:27:02 <dperpeet> but packages might not have jquery available now
13:27:12 <dperpeet> nothing bad, just notable
13:28:42 <andreasn> next up?
13:29:07 <andreasn> #topic Patternfly update
13:29:40 <dperpeet> Cockpit updated to PatternFly 3.2.0
13:29:56 <dperpeet> from 2.8.0
13:30:05 <dperpeet> so we can use some of the newer things / css from tehre
13:30:07 <dperpeet> there
13:30:15 <andreasn> like the new, dark sidebar
13:31:32 <andreasn> I went over every page and I couldn't see any regressions, but if anyone do, file issues
13:31:37 <andreasn> #topic Ubuntu PPA packages
13:32:03 <larsu> ya, there's a PR here: https://github.com/cockpit-project/cockpituous/pull/30
13:32:13 <larsu> needs another round of review
13:32:38 <stefw> and is there another commit that needs to be "un-reverted"?
13:32:57 <stefw> i'd like to try this out today before the 0.103 release tomorrow
13:33:44 <dperpeet> I agree, we need to test this before 0.103
13:33:44 <larsu> yep, the one that builds the non-native package, which landed in cockpit this morning
13:34:34 <larsu> stefw: did you actually revert that? Can't find the revert commit
13:34:40 <stefw> aha
13:34:42 <stefw> i didn't push the revert
13:34:44 <stefw> just the resulting container
13:34:49 <stefw> not sure if taht's good or bad
13:34:53 <stefw> but i can just reset --hard HEAD here
13:35:05 <larsu> ya, that should work
13:35:11 <stefw> ok, will try this out
13:35:14 <stefw> thanks larsu
13:35:49 <larsu> stefw: 8d0a19f6 is the one you need
13:36:00 <github> [cockpit] stefwalter opened pull request #4229: shell: Fix machine port troubleshoot styling (master...edit-machine-port-styling) https://git.io/vwmCP
13:37:00 <andreasn> #topic Open Floor
13:37:07 * mvollmer jumps up
13:37:17 <mvollmer> docker storage setup?
13:37:28 <github> [cockpit] stefwalter opened pull request #4230: shell: List correct port in failure message (master...ssh-running-on-port) https://git.io/vwmCp
13:37:35 <mvollmer> alright
13:37:43 <andreasn> sure
13:37:43 <larsu> stefw: let me know if you need any help with the ppa (you definitely need to verify the key I uploaded)
13:37:47 <mvollmer> I was going a bit back and forth, mostly back
13:37:59 <stefw> larsu, can you send me the key id?
13:38:02 <mvollmer> the new plan is now to add API to the "atomic" command to support our use case directly
13:38:10 <stefw> and we can do a verify thingy over voip or something?
13:38:17 <mvollmer> dan has approved that in principle
13:38:30 <mvollmer> docker-storage-setup is too low-level to be our API
13:38:40 <larsu> stefw: no, it's the key you already uploaded. cockpituous@gmail.com has a mail with instructions
13:38:54 <larsu> I couldn't do it because I don't have the private key
13:39:03 <mvollmer> i have the new code working in a stand-alone binary, pretty much
13:39:07 <mvollmer> next is to use it in the UI
13:39:10 <mvollmer> and bang on it
13:39:26 <mvollmer> there are some subtle cases about who wipes what
13:39:41 <mvollmer> and I have to test overlayfa and loop-back as well
13:40:26 <stefw> that's really cool mvollmer glad to see it's helping make the underlying tool more coherent
13:40:39 <mvollmer> yeah
13:41:02 <mvollmer> with the new plan, the ugliness is well hidden and can be improved between "atomic" and "d-s-s"
13:41:16 <mvollmer> but let's hear it first from the atomic guys
13:41:33 <mvollmer> d-s-s is a weird beast
13:41:46 <mvollmer> and it's best not to expose it as an API or interface, I'd say.
13:42:22 <mvollmer> over
13:43:01 <andreasn> anything else, or should I close the meeting?
13:44:11 <andreasn> #endmeeting