20:00:04 <boutil> #startmeeting
20:00:04 <MeetBot> Meeting started Tue Sep 23 20:00:04 2014 UTC.  The chair is boutil. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:04 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
20:00:11 <boutil> Hi everybody!
20:00:17 <zeha> hi
20:00:25 <boutil> Let's start the meeting.
20:00:52 <boutil> Let's have a quick list of topics to discuss:
20:01:01 <boutil> I put on the wiki:
20:01:14 <boutil> Results of mass autopkgtest
20:01:14 <boutil> Approaching freeze
20:01:14 <boutil> package update rallye removal of obsolete packages
20:01:14 <boutil> Quick glance through RC bugs
20:01:20 <boutil> Pending RFS
20:01:30 <boutil> Have you other things to discuss?
20:02:11 <boutil> ok.
20:02:30 <boutil> #topic Results of mass autopkgtest
20:03:16 <boutil> Martin posted on the list the results of the mass autopkg run for all ruby packages
20:03:26 <boutil> The results are pretty good!
20:03:52 <boutil> Thanks Martin for the run, and terceiro for the underlying work
20:04:24 <terceiro> do we have a volunteer to analyse the results? :)
20:05:02 <boutil> I can have a closer look, but help needed.
20:05:21 <boutil> I guess that many tests are failing because of missing build-deps
20:05:27 <terceiro> wow
20:05:38 <terceiro> if it's just that then it's fine
20:05:51 <terceiro> err, more or less. missing build-deps should be RC bugs ):
20:05:53 <boutil> #action boutil will inspect the results of autopkgtest
20:06:00 <terceiro> #link
20:06:04 <boutil> #help needed to inspect those results
20:06:21 <terceiro> #link http://lists.alioth.debian.org/pipermail/autopkgtest-devel/2014-September/thread.html
20:06:29 <tnnn> I'd be happy to help (if I will be able that is...)
20:06:42 <boutil> #link lists.debian.org/20140917152554.GE3724@piware.de
20:06:51 <boutil> #link https://lists.debian.org/20140917152554.GE3724@piware.de
20:06:52 <terceiro> #info see the autpkgtest-devel archives for this month (link above) for previous analyses of test results by the Perl team
20:07:20 <boutil> great, thanks tnnn!
20:07:26 <boutil> #action tnnn will help
20:08:16 <boutil> Should we try to increase (substantively) the rate of success of the tests (by new uploads)?
20:08:17 <terceiro> #info terceiro is available to advice on autopkgtest issues
20:08:42 <boutil> Will the mass autopkgtest be run again? on a timely basis?
20:08:57 <terceiro> boutil: the idea is to whitelist now the packages we know pass
20:09:20 <terceiro> the ones that fail have to be fixed anyway, so they should be uploaded with the fix + the proper Testsuite: control field
20:09:25 <boutil> what exactly means 'whitelist' in this context?
20:09:47 <terceiro> boutil: it means that debci will run tests for them even without Testsuite: control fields
20:10:02 <boutil> ok.
20:10:06 <terceiro> #link http://anonscm.debian.org/cgit/collab-maint/debian-ci-config.git
20:10:07 <hggh> oh, moin!
20:10:14 <terceiro> hell hggh
20:10:17 <terceiro> hello I mean :)
20:10:28 <hggh> I'm late
20:11:20 <hggh> about the auto test, so not every package need to me updated with debian/tests?
20:11:23 <boutil> is there a expected time for this whitelisting to occur?
20:11:52 <terceiro> boutil: at this point I am just waiting for an analysis of the preview run of the ruby packages :)
20:12:25 <boutil> terceiro: ack.
20:12:31 <terceiro> boutil and tnnn one of the outputs of that analysis is a list of packages to be removed from the whitelist at http://anonscm.debian.org/cgit/collab-maint/debian-ci-config.git/tree/ruby.txt
20:13:01 <terceiro> (i.e. we should remove from there the ones that are broken)
20:13:45 <boutil> hggh: the idea is that by default, ruby packages (without Test-Suite field) have a default autopkgtest test suite. It can be overriden by files in debian/tests
20:14:22 <hggh> boutil: ah fine. so standard gems should not have any debian/tests files
20:14:48 <terceiro> yeah
20:14:49 <boutil> it their test suite can work with just the dependencies, then it's enough.
20:14:59 <terceiro> they need to declare `Testsuite: autopkgtest-pkg-ruby` in debian/control
20:15:24 <sbadia> hi
20:15:30 <sbadia> sorry for the delay…
20:15:31 <boutil> if the test suite requires additional packages, then it has to be configured in debian/tests
20:15:36 <boutil> sbadia: hi!
20:15:41 <hggh> ok fine, I was trying to get it working with ruby-colorize
20:15:58 <terceiro> autopkgtest will add all build-deps to the test deps
20:16:03 <terceiro> except debhelper and gem2deb
20:16:09 <terceiro> plus gem2deb-test-runner
20:16:18 <boutil> ah! I missed that
20:16:53 <terceiro> boutil: yeah I added some intelligence there :)
20:17:31 <boutil> terceiro: should we add the Testsuite: autopkgtest-pkg-ruby ourselves, or is it added automatically by gem2deb/autopkgtest/black magic?
20:17:49 <terceiro> boutil: we need to add
20:17:56 <terceiro> in dh-make-ruby
20:17:58 <terceiro> or by hand
20:18:25 <terceiro> boutil: I actually already committed a change to gem2deb
20:18:27 <terceiro> for that
20:18:32 <terceiro> I guess I will make an upload
20:18:39 <boutil> #info: we need to add "Testsuite: autopkgtest-pkg-ruby" to packages for which autopkgtest passed (without modification)
20:18:52 <hggh> Testsuite: autopkgtest-pkg-ruby
20:18:57 <terceiro> boutil: no we don't
20:19:00 <boutil> terceiro: maybe we can discuss gem2deb a bit later. I'd like to add a functionality.
20:19:07 <hggh> is already available in the templates of gem2deb
20:19:26 <terceiro> I mean, we should add them, but the ones already whitelisted  will be run anyway
20:19:33 <terceiro> boutil: sure
20:19:37 <terceiro> (re gem2deb)
20:20:29 <terceiro> do we have anything else on this topic?
20:20:40 <boutil> terceiro: so we need to add it in the long term, but the whitelisted packages will be run anyway right?
20:20:50 * boutil is confused
20:20:50 <terceiro> boutil: exactly
20:20:56 <boutil> ok.
20:20:58 <hggh> I got it :)
20:21:15 <boutil> #info we need to add it in the long term, but the whitelisted packages will be run anyway.
20:21:32 <boutil> other questions about autopkgtest?
20:21:54 <boutil> let's jump then to another topic
20:21:55 <hggh> perhaps we need some information how to use it in quick for ruby-* packages
20:21:59 <terceiro> #info we need to add the `Testsuite: autopkgtest-pkg-ruby` control field in the long term
20:22:09 <hggh> in the wiki
20:22:32 <hggh> #action hggh will add information to wiki page
20:22:46 <boutil> hggh thanks :)
20:23:02 <terceiro> thanks!
20:23:07 <boutil> #topic The freeze is approaching
20:23:17 <boutil> We have about 1 month left
20:23:23 <boutil> (i'm a bit pessimistic)
20:23:35 <terceiro> boutil: what are your concerns?
20:23:56 <boutil> a lot of packages are a bit/a lot outdated
20:23:58 <terceiro> #info terceiro should upload redmine soon; waiting for ruby-request-store to pass NEW
20:24:24 <boutil> some are so old that they should go away
20:24:32 <boutil> (like ruby-opengl?)
20:24:52 <terceiro> I don't know; it's hard to judge stuff I don't use
20:25:06 <terceiro> is it broken?
20:25:47 <boutil> there is a bug report about ruby-opengl crashing with 2.1. It is working with 1.8 and 1.9.1 apparently.
20:26:03 <hggh> yeah, if you so many packages that are upstream dead and unmaintained
20:26:10 <boutil> It was already not easy to bring it into shape for wheezy...
20:26:50 <terceiro> new upstream 	ruby-opengl 	New version available: 0.61.0
20:26:55 <boutil> my point is that we probably should put some effort into updating packages.
20:27:12 <terceiro> #link http://udd.debian.org/dmd/?email1=pkg-ruby-extras-maintainers%40lists.alioth.debian.org
20:27:21 * terceiro nods
20:27:25 <boutil> 0.61.0 is in fact an empty gem, requiring another implementation (using ffi)
20:27:42 <boutil> I don't know how much it is compatible
20:29:03 <boutil> before wheezy's freeze, we organize some packaging parties on IRC with paulvt (once a week) were we would update package and he would sponsor the uploads
20:29:31 <hggh> good idea.
20:29:59 <terceiro> we can do that, but I think we should stop 2 weeks before the freeze to let some time for testing and fixed
20:29:59 <boutil> It may be difficult to organize worldwide, but for Europe, I could try to organize something of this type.
20:30:01 <terceiro> fixes
20:30:26 <hggh> perhaps we should start a mass rebuild 3w before the freeze?
20:30:32 <boutil> terceiro: sure. the goal is not to break the whole archive :)
20:30:59 <terceiro> boutil: I know :)
20:31:05 <hggh> :)
20:31:08 <boutil> #idea: organize "packaging parties" on IRC, where one can do live RFS
20:31:20 <terceiro> rails and redmine, and everything else that uses Gemfile's can be sensitive to new versions of dependencies
20:31:29 <boutil> yes...
20:32:11 <boutil> Is there an easy way to test somehow using rubygems-integration metadata, that dependencies at the gem level are okayish?
20:33:12 <boutil> I am thinking for example of ruby-mime-types, which has a new major version waiting for upload.
20:33:45 <boutil> If we could have a way to at least smell that something may break without having to rebuild all reverse deps, that would be great
20:34:38 <terceiro> boutil: that's what ci is for
20:35:01 <terceiro> the updated redmine package has a smoke test for each supported DB :)
20:35:04 <boutil> but it is somehow too late, because the problematic package is already in the archive.
20:35:19 <terceiro> right
20:35:47 <terceiro> I intend to add experimental ASAP, but not in time to help up right now
20:36:16 <boutil> experimental could indeed be the place to break stuff :)
20:36:31 <boutil> that is what it's for, right?
20:37:31 <terceiro> yes
20:37:49 <boutil> so let's say we try to update a max of stuff before October 22.
20:38:32 <boutil> #info try to update a maximum of packages before October 22.
20:38:40 <terceiro> should work
20:39:31 <boutil> other questions/comments about the freeze?
20:39:48 <hggh> no
20:39:57 <boutil> let's move to the next topic
20:39:59 <terceiro> not from me
20:40:08 <boutil> #topic quick glance through the RC bugs
20:40:28 <boutil> #link http://udd.debian.org/dmd.cgi?email=pkg-ruby-extras-maintainers%40lists.alioth.debian.org
20:40:46 <boutil> chef is quite in a bad state...
20:41:12 * hggh is using puppet
20:41:15 <terceiro> it's not the chef core though
20:41:39 <hggh> what is about capistrano?
20:41:53 <boutil> ruby-kramdown is missing ruby-prawn-table (gem which has been split from ruby-prawn, and entered the archive today)
20:42:37 <hggh> "Capistrano 3 is very much not backwards-compatible with capistrano 2. Perhaps
20:42:40 <hggh> that needs to be taken into account."
20:43:19 <boutil> but if you have to chose between deprecated but compatible, or supported but incompatible...
20:44:03 <hggh> second one.
20:44:13 <terceiro> me 2
20:44:28 <boutil> idem
20:44:40 <hggh> we are using capifony, that is an addon to capistrano, I can have a look
20:44:58 <hggh> and sommer is gone, I should have more time again :)
20:46:21 <boutil> the list is quite long, thanks to the mass rebuild by lucas
20:46:39 <boutil> we shouldn't have said that we were quite fine wrt to RC bugs :)
20:46:46 <hggh> #action hggh is looking at capistrano
20:47:17 <boutil> questions/comments about RC bugs?
20:47:34 <terceiro> I think there's not much to discuss
20:47:37 <boutil> #topic pending RFS
20:47:43 <terceiro> what we need is to go, investigate, and fix
20:47:56 <boutil> is there someone here with pending RFS?
20:47:57 <terceiro> giving priority for the stuff that is out of testing
20:48:11 <hggh> boutil: have no open RFS
20:48:15 <boutil> (I guess no)
20:48:31 <boutil> #topic gem2deb
20:49:02 <boutil> terceiro: you were saying that you wanted to add some commits to gem2deb?
20:49:03 <tnnn> (If nothing bad happens I will have one by the weekend - my first one so some advice might be needed).
20:49:32 <terceiro> boutil: it's already in
20:49:33 <boutil> tnnn: no problem, send an email to the list, and we'll answer ;)
20:49:41 <terceiro> the tip of the master  branch
20:49:58 <tpot> i have some RFS’s but boutil is taking good care of me
20:50:04 <terceiro> that changes the Testsuite: field to autopkgtest-pkg-ruby
20:50:12 <boutil> terceiro: ok
20:50:16 <boutil> tpot: :)
20:50:21 <terceiro> boutil: I wanted to have that in the archive RSN
20:50:33 <terceiro> boutil: do you have pending commits of your own?
20:50:51 <boutil> terceiro: I would like to add also something which is in the extconf-options branch
20:51:06 <boutil> it misses yet some documentation
20:51:22 <terceiro> boutil: do you have an ETA?
20:51:43 <boutil> it would allow to pass options to extconf.rb via a dh_auto_install addistional option
20:51:55 <boutil> I think you can go ahead with your change.
20:52:19 <terceiro> ok
20:52:20 <boutil> It's nothing really we would need *for* jessie itself
20:52:44 <boutil> terceiro: if you could have a quick look at some point, I'd like to have your feedback
20:52:59 <terceiro> boutil: I can
20:53:05 <terceiro> we can have a second upload later
20:53:24 <terceiro> there is another feature I will probably need for work, which is support for installing rails engines in a sane way
20:53:57 <boutil> that would be great, indeed
20:54:16 <boutil> it sometimes the source of file conflicts
20:55:16 <boutil> is /usr/lib/ruby/vendor_ruby/rails/generators/<packagename> a canonical path to put them?
20:55:30 <terceiro> it's not
20:56:08 <terceiro> rails generators assume they are installed via rubygems with their own private $LOAD_PATH
20:56:13 <terceiro> $LOAD_PATH component
20:56:36 <terceiro> it's just the way it is, I don't have hopes of changing
20:56:38 <boutil> do you have in mind a solution that would avoid moving/patching code manually?
20:56:50 <terceiro> installing them as if they were actual gems
20:57:01 <terceiro> to /usr/share/rubygems-integration/...
20:57:17 <terceiro> and it's not only generators
20:57:30 <terceiro> some rails engines have an rails app structure with app/ db/ etc
20:57:37 <terceiro> so there is not really another wy out
20:57:59 * boutil sigh
20:58:40 <boutil> next topic?
20:59:02 <boutil> #topic ruby-policy (again!)
20:59:30 <boutil> #action sbadia boutil will gather and work on the ruby policy, for real
20:59:37 <terceiro> lol
21:00:03 <hggh> we should have done that by the spring back in january ;)
21:00:04 <boutil> caitlin gave some input, which we should take into account.
21:00:09 <terceiro> we'll propose it's inclusion in the debian-policy repo right?
21:00:25 <hggh> yes, i think so
21:00:46 <boutil> yes, but first we need to have it ready :)
21:01:19 <boutil> let's make it a release goal for Jessie + 1
21:01:40 <terceiro> ok :)
21:01:58 <boutil> #topic Ruby sprint
21:02:30 <hggh> terceiro is right to do a sprint after the release of jessie
21:02:36 <hggh> so we can break more stuff
21:02:41 <hggh> ah fix it!
21:02:59 <boutil> agreed.
21:03:28 <boutil> the freeze period doesn't give too much freedom to move...
21:03:44 <boutil> I hope the freeze won't las forever
21:03:50 <boutil> *last
21:03:50 <terceiro> it won't
21:04:04 <terceiro> this time the really broken stuff is already  not in testing
21:04:10 <boutil> thanks autoremoval!
21:04:17 <terceiro> I have hopes of a very quick freeze
21:04:25 <terceiro> but don't quote me on that ;-)
21:04:52 <boutil> so we can make (vague) plans for Spring then
21:04:55 <hggh> hm as a sysadmin I hope for a longer freeze *duck* because
21:05:14 <hggh> but for my debian hat on, I have a short freeze
21:05:46 <hggh> boutil: the paris meetup is in another city?
21:05:59 <boutil> it's in Lyon
21:06:31 <terceiro> we can always go the Paris anyway before or after the miniDebConf
21:06:37 <terceiro> since we have IRILL
21:06:47 <boutil> (it's 2h by train from Paris, and about the same (even closer) from Geneva)
21:07:02 <terceiro> piece of cake
21:07:16 <terceiro> it would be better to have a friendly organization in Lyon already
21:07:23 <terceiro> but 2h by train is totally doable
21:07:25 <boutil> IRILL is a great place (and they have free coffee)
21:07:32 <hggh> :)
21:07:45 <hggh> only terceiro needs again a long flight :)
21:08:17 <boutil> maybe Mozilla offices in Paris can also host (sylvestre is now working there)
21:08:49 <boutil> there's room at my university too.
21:08:58 <boutil> so in Paris, it can definitly happen
21:09:26 <terceiro> hggh: all the logic options I will have my flying to Europe anyway
21:09:43 <boutil> I'll contact people in Lyon to see if they can host sprints as satellite events of the mini-debconf
21:09:51 <terceiro> cool
21:10:02 <boutil> other topics you want to bring up?
21:10:44 <boutil> ok, let's stop here for today
21:10:47 <boutil> thanks!
21:10:53 <boutil> #endmeeting