15:04:46 #startmeeting horizon 15:04:47 Meeting started Wed Apr 29 15:04:46 2020 UTC and is due to finish in 60 minutes. The chair is amotoki. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:04:49 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:04:51 The meeting name has been set to 'horizon' 15:05:01 anyone around? 15:05:34 hi 15:06:08 e0ne: hi 15:06:49 today is a start of the holiday week in my side, so I have less topics. 15:07:12 #topic announcements/notices 15:07:45 next week is a week of final RCs and intermediary releases 15:08:00 we will cut at least one release to import translations. 15:09:14 the holiday week in Japan continues till Wed next week, but I will check reviews occasionally. 15:09:39 anything other to share? 15:10:05 a few words about PTG planning 15:10:09 #link https://etherpad.opendev.org/p/horizon-v-ptg 15:11:16 I requested 3 timeslots for us 15:11:22 I failed to find time to add topics last week 15:11:37 thanks for proposing the timeslots 15:11:48 The last one on Wednesday is to cover some discussions in US timezone if needed 15:11:59 I'll send a mail to openstack-discuss@ 15:12:31 the main idea is to cover everything on Mon-Tue 15:12:46 I haven't checked my conflict with other projects and my availability depends on the schedule as a whole 15:12:47 3 hours should be enough 15:13:26 amotoki: I tried to not conflict with neutron 15:13:53 feel free to ping me if we need some other or more timeslots 15:14:42 e0ne: thanks for your consideration 15:15:58 I'll add few more topics to discuss later this week 15:16:43 hopefully me too :) 15:16:49 :) 15:17:58 moving on 15:18:09 #topic On-Demand agenda 15:18:39 I am skipping others and moving to the on-demand agenda section this week :) 15:19:04 I have a small topic on Django 3.0 15:19:19 I tried Django 3.0 a few weeks ago 15:19:30 and noticed django-pyscss can be a blocker 15:19:48 horizon support would be simple 15:20:32 what is the issue with django-pyscss? 15:20:35 we may hit a same problem on django-pyscss as what we had in pyscss 15:21:04 django-pyscss depends on a module which was dropped in Django 3.0 15:21:33 :( 15:21:39 django.utils.six was dropped 15:21:53 maybe it's time we switched to node tools 15:22:13 rdopiera: +1 for it in the long-term perspective 15:22:22 rdopiera: do you mean webpack or some? 15:23:37 I don't know, really, all I know is that most of the web tools are originally created for node.js (except for the ones that used to be for ruby), and that is the best supported version of them 15:23:58 we already require node for some of our tests 15:24:22 btw, I've got django-pyscss2 fork :) 15:24:23 yeah most tooling around JS world depends on node.js 15:24:50 we are fighting an uphill battle by trying to stick to python 15:25:10 webpack-based setup will be awesome! but I'm afraid how muxh efforts it will require 15:25:20 s/muxh/much 15:25:45 IIRC, legacy packaging like rpm does not allow us to use multiple versions in a single system and that's the background of xstatic modules, right? 15:26:42 or maybe npm wasn't good enough so xstatic-* was created 15:27:49 amotoki: not quite, xstatic was to track dependencies using python tools, they still don't ship multiple versions 15:28:30 I mean, on Debian xstatic is just a thin shim pointing to where the javascript library is installed in the system 15:28:53 the package itself doesn't contain javascript code 15:29:26 e0ne: speaking of effort, is there a plan to get to a supported version of angular? 15:29:32 what I am afraid is rpm or deb packaging does not allow us to use mixed versions of JS library. If project A requires JS module verX and project B requirs JS module verY, I am not sure they can handle it. 15:29:44 rdopiera: re-write everything 15:29:55 amotoki: we can't handle it right now either 15:30:13 let's discuss the possibility of npm usage during the PTG 15:30:20 e0ne: I was recently backporting a bunch of CVE patches to a version of jquery that works with horizon 15:30:35 e0ne: but we will need to upgrade to a supported version sooner or later 15:30:43 +++ 15:30:56 problem is, our angular doesn't work with jquery 3 15:31:25 so, do we fork it? 15:31:37 we have many problems which affect each other :( 15:32:37 what I hit in Django30 is just a piece of our issues. 15:33:10 so in the long run, either we fork and maintain our own angular, or we do the rewrite to something supported, and hope it will still be supported by the time we finish 15:33:38 or create our own framework 15:33:51 like we discussed some time ago during the ptg 15:33:57 either way sounds a tough road to us.... 15:34:19 it looks like better to explore our possibility before PTG for productive discussion in the pTG 15:35:24 amotoki: +1 15:37:02 I remember that on some of the previous PTGs we also discussed the possibility of backporting the changes to the old Python views and ditching the angular views 15:38:57 which would solve a lot of our problems, in my opinion 15:39:01 rdopiera: I am not sure we discussed an option to throw away angularjs impls explicitly. 15:39:28 at least we agree that we have no priority on angularjs migration 15:39:36 I think it was in Dublin 15:39:56 yeah, maybe we will get back to it again on the PTG 15:41:02 anyway I totally agree that we need to discuss the future of mixed usage of django and angular impls 15:43:01 anything to discuss? 15:43:09 nothing from me 15:43:20 it was a good brainstorming discussion. let's add note to the ptg etherpad. 15:44:11 okay, let's wrap up the meeting 15:44:13 thanks all 15:44:18 #endmeeting