cockpit_weekly_meeting_2017-09-04
LOGS
13:04:15 <andreasn1> #startmeeting Cockpit weekly meeting 2017-09-04
13:04:16 <zodbot> Meeting started Mon Sep  4 13:04:15 2017 UTC.  The chair is andreasn1. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:04:16 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
13:04:16 <zodbot> The meeting name has been set to 'cockpit_weekly_meeting_2017-09-04'
13:04:20 <andreasn1> .hello andreasn
13:04:21 <zodbot> andreasn1: andreasn 'Andreas Nilsson' <anilsson@redhat.com>
13:04:38 <dperpeet> .hello dperpeet
13:04:39 <zodbot> dperpeet: dperpeet 'None' <dperpeet@redhat.com>
13:04:44 <pitti> .hello martinpitt
13:04:45 <zodbot> pitti: martinpitt 'Martin Pitt' <martin@piware.de>
13:07:05 <andreasn1> #topic Agenda
13:08:21 <andreasn1> * NFS client
13:09:29 <mvollmer> .hello mvo
13:09:30 <zodbot> mvollmer: mvo 'Marius Vollmer' <marius.vollmer@gmail.com>
13:10:01 <mvollmer> * Out-of-tree cockpit "apps"
13:11:10 <andreasn1> ok, let's kick this off
13:11:19 <andreasn1> # topic NFS Client
13:11:38 <andreasn1> #link https://github.com/cockpit-project/cockpit/wiki/Feature:-NFS-client
13:11:49 <andreasn1> #link https://github.com/cockpit-project/cockpit/pull/7580
13:12:12 <andreasn1> UI is in pretty good shape, some small things to work out still
13:12:18 <mvollmer> i have started writing tests
13:12:33 <mvollmer> the usual race condition debugging...
13:14:33 <mvollmer> there will be some followups, like a dedicated options ui
13:14:41 <mvollmer> a checkbox for rw/ro, etc
13:15:25 <mvollmer> related unrelated: I have a hard time making sense of backtraces from phantomjs
13:17:54 <andreasn1> all right, next topic?
13:18:01 <mvollmer> yep
13:18:58 <andreasn1> #topic Out-of-tree cockpit Apps
13:19:08 <pitti> we just had that
13:19:31 <pitti> err, ignore me
13:20:03 <mvollmer> so I want to package cockpit-app-freeipa in an acceptable way
13:20:18 <mvollmer> which means copying how cockpit itself is packages, I guess
13:20:43 <mvollmer> for Fedora: include sources and results of webpack in source rpm
13:21:07 <mvollmer> build the srpm then just shuffles the webpack results into the binary package without any real building
13:21:20 <mvollmer> for Debian: I don't know.  Pitti? :-)
13:21:42 <dperpeet> mvollmer, are you keeping your effort here in sync with the starter kit?
13:21:53 <dperpeet> it might make sense to extend the starter kit for debian also
13:22:02 <mvollmer> yes, sure
13:22:07 <dperpeet> https://github.com/cockpit-project/starter-kit
13:22:32 <pitti> mvollmer: should be similar -- we already put the result of webpack'ing into our release tarballs, so that downstream packaging can make use of it
13:23:13 <mvollmer> pitti, does Debian use the "cache" tarball that we create during "make dist"?
13:23:27 <mvollmer> it has all the build-time node things, like webpack itself
13:23:37 <pitti> mvollmer: no, not right now; it's not needed for building the debs, just if you want to rebuild the webpack output
13:23:47 <mvollmer> okay
13:24:03 <pitti> mvollmer: that said, it's possible to add additional tarballs to the source package if necessary
13:24:31 <pitti> aside from requiring a massively bigger copyright file, checking/fixing licenses in those, etc :)
13:24:54 <mvollmer> well, personally, I think building a package should always start from the sources, but that doesn't seem to be feasible in our case
13:25:29 <pitti> it's a pragmatic middle-ground
13:25:36 <mvollmer> so this is really about doing the minimal thing that makes the distros happy
13:25:49 <pitti> and some tools or our build system would need to be changed
13:26:03 <dperpeet> the starter kit doesn't use any tarballs from Cockpit
13:26:07 <pitti> so that it doesn't try to talk to the network for "npm install" (and other commands) when building, and always just takes the cached output
13:26:59 <pitti> but I don't understand how all that is related to building -freeipa?
13:27:28 <pitti> if it's another subpackage built by cockpit srpm/dsc there should not be any change, and if it's a separate project (which is most probably your goal) it needs its own build system anyway
13:27:34 <mvollmer> it's a cockpit "plugin", and it's build just like cockpit itself
13:28:08 <mvollmer> pitti, yes, but I want to give it a build system that is acceptable to the distros
13:28:19 <mvollmer> and I figure that cockpit has one that is acceptable
13:28:22 <mvollmer> so I copy it
13:28:33 <mvollmer> i just didn't understand the debian part
13:29:50 <pitti> mvollmer: I suppose we can discuss this outside the meeting, i'm happy to answer questions (but it should generally not be that different from the spec)
13:30:10 <mvollmer> yeah, that's good enough for me. :)
13:30:11 <pitti> in both cases we use the pregenerated webpack output
13:30:31 <pitti> so that is acceptable to both the Fedora-ish and the Debian-ish distros, if that is the root of your question
13:31:04 <pitti> it's indeed not really desirable, but as builds must be able to happen offline it's the only practical way right now
13:31:20 <mvollmer> yes
13:31:47 <mvollmer> using the cache tarball makes patching easier
13:32:09 <mvollmer> but doesn't change anything fundamentally
13:32:27 <mvollmer> like avoiding to use stuff from outside the distro
13:33:58 <mvollmer> okay, done! :)
13:34:31 <andreasn1> all right, I guess that was it
13:34:34 <andreasn1> #endmeeting