21:58:56 #startmeeting 21:58:56 Meeting started Tue Feb 11 21:58:56 2014 UTC. The chair is boutil. Information about MeetBot at http://wiki.debian.org/MeetBot. 21:58:56 Useful Commands: #action #agreed #help #info #idea #link #topic. 21:59:04 Hi everybody! 21:59:38 hello 21:59:46 Meetbot will be recording logs/minutes of the meeting. 21:59:46 boutil: Error: "will" is not a valid command. 21:59:55 :( 22:00:03 hey 22:00:08 :D 22:00:10 heya 22:00:30 #chair terceiro paulvt hggh 22:00:30 Current chairs: boutil hggh paulvt terceiro 22:00:38 is zeha here? 22:01:28 yes 22:01:31 sorry 22:01:32 :) 22:01:47 I was about to say.. 8min idle 22:01:49 :) 22:02:27 #chair zeha 22:02:27 Current chairs: boutil hggh paulvt terceiro zeha 22:02:37 many chairs 22:02:44 he he :) 22:03:12 there are some more if people want to have a seat 22:03:27 let us start with a list of topics? 22:03:41 agreed 22:03:47 -ruby1.8 removal 22:03:50 just scratch out init-systems ;) 22:03:51 sure, /me proposes look at ruby1.8-removal 22:04:03 -ruby1.9.1 removal 22:04:06 agreed 22:04:09 -ruby2.1 status 22:04:20 -gitlab 22:04:27 -switch to ruby2.0 as default 22:05:00 ok, that is a good start. Maybe some more will come to mind later 22:05:07 #topic ruby1.8 removal 22:05:16 yes. also do we have a time limit? 1h? 22:05:24 1h seems reasonablre 22:05:28 that would be nice :) 22:05:29 k 22:05:31 let's try that, yes. 1h. 22:05:40 all rigth, so about ruby1.8 removal 22:05:53 #link https://release.debian.org/transitions/html/ruby1.8-removal.html 22:05:57 there are still some important-enough packages that need fixing 22:06:06 do we know how many of those pending packages are scheduled for autoremoval? 22:06:10 #link http://bugs.debian.org/cgi-bin/pkgreport.cgi?users=debian-ruby@lists.debian.org;tag=ruby18-removal 22:06:15 terceiro, a lot 22:06:24 but not all of them. 22:06:38 zeha: do we know which of them will not be autoremoved? 22:06:49 ooph, 45+ 22:07:01 graphviz won't be autoremoved 22:08:12 are there packages in the list that the Ruby team would be interested in helping? 22:08:15 ruby-nora, redland-bindings, stfl probably 22:08:28 i think the other packages will go out 22:08:51 are we talking about the list in the second link? 22:09:02 i believe gwolf worked on ruby-bdb, but didn't finish 22:09:36 paulvt, transition tracker page is IMO more helpful for this 22:10:09 rgd. xmms2: I have no idea why it's still listed as bad, and -release couldn't tell me either. somebody please take a look :) 22:10:19 right, some our ours 22:10:21 maybe rubyluabridge? 22:10:39 such as ruby-gnome2, which is mine.. but it should work on >= 1.9.1 AFAICT 22:10:50 another holdup: ruby1.9.1 failed to build on powerpc, preventing migration of ruby-tokyocabinet 22:11:29 #action terceiro will ask for a giveback of ruby1.9.1 on powerpc and hope it builds 22:11:52 there is some work by terceiro and myself on ruby-gnome2. 22:12:13 terceiro: do you plan to finish it, or do you want me to have a look? 22:12:27 boutil: if you could look at it would be nice 22:12:33 or maybe paulvt is interested in having a look? 22:12:37 terceiro: https://lists.debian.org/debian-wb-team/2014/02/msg00007.html 22:12:57 #action boutil fix and upload ruby-gnome2 22:13:04 zeha: ok; so it failed again? 22:13:12 no, nobody cared to do the g-b 22:13:34 ruby-luabridge has no gem2deb, no rubygems.org release, I check if it's build against 2.0 22:14:16 anyway, the situation is much better than 3 weeks ago before the RC bugs. 22:14:46 terceiro: can you remind us when the removal of ruby1.8 is planned? 22:14:52 boutil: I could but I cannot give any timeframe.. it is better for me to give me small and concrete stuff to do with a deadline at the moment 22:14:53 yes; maybe someone could do a 'point out to MIA team' sweep 22:15:43 boutil: I think I said "in 4 weeks" during the sprint 22:15:44 * terceiro checks 22:15:51 (rubyluabrigde is sid-only btw.) 22:16:23 ruby-luabridge does not build with 1.9 22:16:39 hggh: thanks, good to know :( 22:17:25 zeha: cannot find the logs for tokyocabinet powerpc build 22:17:40 it's BD-Uninstallable 22:17:41 how can we have it ? 22:17:42 so no logs 22:17:44 #action terceiro start dialogue with RT about actually removing ruby1.8 from jessie ASAP 22:17:49 boutil: ^ 22:18:18 so, let's try to save some that seem important to us, and let the other packages break by the removal? 22:18:20 paulvt, btw, I NMU'ed gnoemoe2, which lists you as an Uploader; it seems the original Maintainer disappeared? 22:18:44 zeha: yes, I saw it.. he has 22:18:52 I actually also have on my list to port it to Gtk+3 22:19:01 the interface to Ruby was fairly simple IIRC 22:19:21 maybe I can port that too 22:19:28 boutil, I think we need to fix graphviz in any case, -release won't remove that 22:19:30 who wants to NMU graphviz? 22:19:33 (the original Maintainer is now actually a co-author of gedit btw) 22:20:12 my last graphviz build try failed, so not me please :-) 22:20:13 terceiro: I had I look at graphviz on the sprint, the easy solution was removing the ruby bindings 22:20:24 i think it would build with 1.9 or 2.0 22:20:29 hggh: but that's an option in all cases no? 22:20:31 but the package is annoying to build 22:20:40 zeha: ack 22:20:41 I mean.. non-Ruby-libs packages that is 22:20:48 graphviz is a terrible swig binding 22:21:20 the almost always are 22:21:34 removing the ruby binding could be a last resort action. 22:21:41 for work I've been struggling with libproxy's binding (not built in Debian fortunately, or that also would've been a concern) 22:21:46 an NMU to remove a binary package is too brutal 22:22:19 #action NMU graphviz 22:22:28 ^ needs a volunteer :) 22:22:45 ok. Something else to add about ruby1.8? 22:22:53 I don't think so 22:23:12 #topic ruby2.0 as default / ruby2.1 status 22:23:23 2.1 is stuck in NEW 22:23:24 yes please, what is the status 22:23:42 let's poke super paultag :) 22:23:47 or luca 22:23:57 for 2.0-default, we're waiting for the test rebuild 22:24:00 is there a blocker to switch the default to 2.0? 22:24:04 ah, ok. 22:24:23 everybody could help testing by installing 'ruby' from experimental 22:24:29 ah! 22:24:30 yep 22:24:31 (and see what breaks locally) 22:24:36 ah... about graphviz. there is already a package named ruby-graphviz. so libgv-ruby should be renamed to ruby-gv 22:24:47 #info one can install ruby from experimental to test ruby2.0 as default 22:24:56 done! 22:25:12 btw, side-question, there is no ri2.0 and ri2.1 anymore? 22:25:26 there is 22:25:31 hggh: if we fix the package but keep its old name, I think it is ok... 22:25:31 ruby2.0: /usr/bin/ri2.0 22:25:36 $ which ri2.0 22:25:36 /usr/bin/ri2.0 22:25:51 boutil, agreed, I'd also keep libgv-ruby in this case. no point of going to NEW for this 22:25:52 #action hggh will try to build graphviz with ruby 1.9 22:26:06 yes, but the ri-data was separate 22:26:23 paulvt: yep I think you need ruby2.0-doc 22:26:34 ri1.9.1 contained the ri data for binary ri1.9.1 (which I've found to confuse users) 22:26:37 great, that's better 22:26:39 hggh: try changing it to the default ruby whatever it is 22:26:42 #action zeha will ask -ftp to get 2.1 out of NEW 22:26:45 hggh: wrt graphviz 22:26:56 terceiro: yep, that was also on my mind :) 22:27:13 hggh: cool, than next time we can just request a binnmu 22:27:22 i think that's it for 2.0/2.1? 22:27:30 pretty much I think 22:27:47 terceiro: about rubygems-integration, the content seems not symmetric in ruby1.9 /ruby2.0. Is that normal? 22:27:59 boutil: what you mean? 22:28:06 (yay, my most important application works with "ruby" from exp" 22:28:27 (maybe I'll ask rubygems-integr. question later when I collect info...) 22:28:59 do we know an ETA for the test rebuild results? Has deiv already scheduled it? 22:29:14 I don't know really 22:29:24 deiv is testing against 2.1? 22:29:31 2.0 22:29:38 ok 22:30:02 anything to add on this topic? 22:30:32 let's move to the next one... 22:30:35 +1 22:30:42 #topic ruby1.9 removal 22:31:01 have we an tracker for 1.9 removal already? 22:31:06 we don't 22:31:10 I can ask for one 22:31:13 we have an old snapshot 22:31:16 http://people.debian.org/~zeha/rubyremoval/html/ruby191-removal.html 22:31:16 should we start filing bugs? 22:31:32 #link http://people.debian.org/~zeha/rubyremoval/html/ruby191-removal.html 22:31:35 #action terceiro ask for a ruby1.9.1-removal tracker on release.debian.org 22:31:37 just to give people an idea 22:31:49 I think we should 22:32:00 yes, having it on release would be good 22:32:09 first we should disable gem2deb to build with 1.9? 22:32:10 but can we do this, this fast? 22:32:20 severity important for now; for the packages that are hardcoding ruby1.9.1 somehow 22:32:21 I've been still mainly using 1.8 at work (we're on Precise) 22:32:37 1.9 needs to go before we freeze 22:32:38 :) 22:32:42 so better do it now 22:32:50 .. so I don't know about the situation 22:32:57 #action file important bugs against packages hard-coding ruby1.9.1 in dependencies 22:33:00 2.0 and 1.9.1 is identicaly? or at least inclusive? 22:33:09 paulvt, to fill you in, jessie should be 2.1-only 22:33:15 yes, I read that 22:33:22 everything else won't be viable security-wise 22:33:36 and 2.0 has only minor changes from 1.9 22:33:57 so if it builds/works with 1.9, it likely builds/works with 2.0/2.1 22:34:12 most packages on the tracker needs a rebuild with newer gem2deb that does not build for 1.9 22:34:17 yes, I've seen that somewhere too.. but no real-life experience with it.. pure Ruby and extension-wise 22:34:31 ideally gem2deb should use ruby-all-dev 22:34:38 zeha: it will 22:34:39 but i haven't found time to make this work yet 22:35:00 I have been working on gem2deb in the last days, I can also change that 22:35:03 I am also assuming 1.9.1-remove is/will be blocked until ruby with 2.0 as default enters unstable 22:35:06 is there a timeline for that? 22:35:12 paulvt: ASAP 22:35:16 terceiro: great work btw 22:35:26 ok, so testing is done 22:35:29 we just need to know the size of the breakage with ruby2.0 as default 22:35:52 right! 22:36:23 and let's break some things in unstable.. our Ruby stuff has been too nice for unstable :P Python and Perl have wrecked the place far more often ;) 22:36:33 no, but seriously 22:36:42 we did manage to break unstable often enough :) 22:36:51 ? 22:37:18 ruby-debian dropped 1.8 binaries, and apt-listbugs broke for some users, etc. 22:37:40 (and other minor stuff, like 'ruby' not being installable) 22:37:44 hmm, ok, so... ruby with 2.0 as default moves to unstable ASAP 22:37:59 then ruby1.9.1 removal is started concurrently with the 1.8 removal 22:38:16 but it shouldn't be that bad given the small differences between 2.0 and 1.9.1 22:38:23 I was thinking of having small cgi scripts serving lists of packages with some properties, like build-depending explicitly on ruby1.9.1, like the one with packages maintained by the team. 22:38:27 ... summarizing 22:38:38 would you find that handy for following bug filings, etc? 22:39:18 paulvt: that's the transition tracker 22:39:29 except it doesn't separate by maintainer 22:39:31 http://people.debian.org/~zeha/rubyremoval/html/ruby191-removal.html updated 22:39:53 zeha: how do you do that? :) 22:40:18 ouch 22:40:19 you can install ben from unstable 22:40:24 it 'mostly' works 22:40:30 ah 22:40:54 ok... but we would need a list with only a selection on Build-Depends matching ruby1.9.1 22:41:35 yes because now everything that uses gem2deb is in there 22:41:40 to know packages that are really problematic, against those needing only a rebuild against newer gem2deb 22:42:13 libruby1.9 is matched in the depends line, if it is added by gemdeb for arch: any packages 22:42:37 yeah 22:42:49 yeah 22:42:54 i'll try that (but ben will need a few minutes) 22:43:05 zeha: could you provide another list to complement with just the match on build-depends to have an idea? 22:43:48 zeha: also the Good: line could be just "not matches ruby1.9" 22:44:09 next topic? 22:44:18 aye 22:44:23 terceiro, do you know how to say that in ben syntax? 22:44:29 !~ ? 22:44:47 #topic gitlab packaging? 22:44:53 #topic gitlab packaging 22:45:09 there hasn't been much progress lately 22:45:25 #link http://people.debian.org/~boutil/gitlab/gitlab_deps20140128.pdf 22:46:05 I haven't done precise stats, but I think we are about 75% through 22:46:14 (maybe more). 22:46:19 that's awesome 22:46:28 huge task indeed 22:46:51 on a side-note, wasn't there also a diaspora effort of about the same size, or did that die out? 22:46:54 Some packages are almost ready in the repository, but Daniel Marti didn't show up for a while on the list 22:47:17 paulvt: diaspora packaging is still +/- active 22:47:27 who is running that? 22:47:35 praveen 22:47:36 I reckon they run into similar issues 22:47:37 ok 22:47:40 #link http://people.debian.org/~boutil/diaspora/diaspora_deps20140128.pdf 22:47:49 perhaps I could grab some of the gitlab deps 22:47:52 about the same state 22:47:57 indeed 22:48:13 I think they are both quite nice applications to have 22:48:20 same progress / same current activity 22:48:38 paulvt: ack 22:48:40 there is a dedicated channel #debian-diaspora 22:48:55 yep 22:49:08 once we have redmine, gitlab and diaspora in debian it will be really cool 22:49:13 yes 22:49:19 yep 22:49:21 there is a lot of related ITP filed, but no packages around. 22:49:32 one would need to prioritize that. 22:49:37 #action hggh will have a look at the gitlab deps 22:49:55 prioritize packaging those "dead" ITPs 22:50:07 someone please give a debian account to hggh :) 22:50:23 terceiro: I have some open rfs :P 22:50:28 hggh: I know! :) 22:50:34 so! 22:50:35 everybody knows :) 22:50:58 hggh: we should forward all of those to your AM 22:51:03 lol 22:51:08 not I, but I've been a bit lost in mail and stuff... I see PET is working again 22:51:09 :D 22:51:09 .oO my status is "waiting for AM to confirm" 22:51:18 whois is your AM? 22:51:30 paulvt: nhandler 22:51:33 PET is working again, and DMD provides a nice list. 22:51:56 I will send an email to the list to try to revive gitlab packaging 22:52:14 is it just packaging issues? 22:52:21 or also gems that are modified within the project and stuff? 22:52:41 there's also that 22:52:46 #action boutil mail debian-ruby@ about gitlab packaging initiative 22:52:49 but I'd say it's only a handfull of packages 22:52:52 paulvt: i think the modified issue has gone down with the newest gitlab version 22:52:59 good 22:53:08 lol, still a lot of missing tags in PET I see :) 22:53:22 next topic! 22:53:24 pet is somewhat broken i think 22:53:36 #action boutil mail debian-ruby@ about the state of the team repo 22:53:59 next topic beeing... 22:54:08 5min left 22:54:19 ahhaha! 22:54:37 I think that was the last topic 22:54:44 yes 22:54:47 wow ok 22:54:58 random thought: if anybody wants to look at 'fix "rails new" to work ootb', please do ;-) 22:55:03 so I made some mental note about something, but forgot 22:55:13 zeha: it doesn't work? 22:55:16 so about rubygems-integration: now we have are supposed to have an all/ directory? 22:55:17 zeha: the fix in stuck in NEW 22:55:23 ooph 22:55:25 terceiro, ah, indeed 22:55:36 + lots of dependencies for 4.0 22:55:53 yes! can we talk about rubygems-integration a bit 22:56:00 I built a package recently, and couldn't find it in the result... 22:56:09 #topic rubygems-integration 22:56:10 boutil: yes, I uploaded rubygems-integration 1.5 yesterday with support for "all" 22:56:11 boutil: the creating pdf for gitlab stops at 28.01, it only creates a new pdf if some changes or cronjob stopped? 22:56:24 http://people.debian.org/~zeha/rubyremoval/html/ruby191-bd.html this is probably the "ugly" set 22:56:29 boutil: the corresponding gem2deb chnage will be uplooaded tonight most probably 22:56:35 hggh: it's a bug in the way graphs migrate to my homepage. 22:56:38 will fix it. 22:56:47 terceiro: ah ok. 22:56:50 boutil: is the source available? 22:57:05 zeha: the ugly set is small enough 22:57:12 indeed! 22:57:20 but almost none are "ours" 22:57:25 hggh: give you the link later when I find it. 22:57:29 haha that are also the ugly one from the ruby1.8er removal 22:57:42 libzip-ruby, ruby-filesystem. -.- 22:57:47 no surprises there ;) 22:57:56 about rubygems-integration: there is an issue with the specs of packages declaring binaries 22:58:10 boutil: what issue? 22:58:25 I remember that paulvt filed a bug of this kind for rake. is that correct? 22:58:33 boutil: 22:58:38 ...yes 22:58:47 #710814 22:58:48 ? 22:59:01 the binary field is usually relative to the gem structure. 22:59:16 I've been using rubygems(-integration ) to develop on my sid machine and be able to deploy on wheezy 22:59:20 on squeeze* 22:59:25 boutil: right 22:59:38 when the specs are moved to /usr/share/rubygems-integration, the binary is not where it is looked for. 23:00:01 right, and that is something that "we" break 23:00:03 IIRC rubygems-integration does not say anything about bin_path 23:00:04 should we manually patch the gemspecs to fix the path? 23:00:05 terceiro: indeed 23:00:26 boutil: https://gitorious.org/debian-diaspora/gemdeps/source/e6640d4603661ec606397f1e69dbb67a0d6e48a1: I think that's the source. 23:00:59 hggh: indeed :) 23:01:48 http://people.debian.org/~boutil/diaspora/README found it here 23:01:49 boutil: or patch the stub 23:01:56 * boutil opened the rake gemspecs 23:02:53 I don't know the solution now, but we probably want to make it point to /usr/bin instead, and do that in rubygems-integration 23:03:02 hmm no, that's hard if you would install a X and Y version of a gem with a binary stub 23:03:05 so the 'executables' should be understood by rubygems for specs in /usr/share/rubygems-integration to be in /usr/bin, and not in their /usr/share/rubygems-integration directory 23:03:20 yeah 23:04:43 other misc topics? 23:04:47 another random thought... it would be nice somehow if gem list could show somehow which are Debians versions 23:04:56 +1 23:05:08 oh +1 23:05:18 at some point I'm coninuously using gem list and apt-cache policy to find out what the hell is going on 23:05:21 yeah me too, patches welcome :) 23:05:39 I've looked into this a bit, but it goes pretty deep 23:05:50 there is also the policy thing... 23:05:53 maybe rubygems-integration could provide something similar 23:06:13 to show "the status of the integration" and then we don't have to dive into Rubygems 23:06:23 I need to go out now, maybe we can add the policy to the agenda of the next meeting? 23:06:29 agreed 23:06:30 ok. 23:06:40 +1 23:06:41 boutil: thanks for organizing! 23:06:42 I need to go to bed too 23:06:47 shall we fix roughly a day for the next meeting? 23:07:00 let's say four weeks from now 23:07:01 so if it's at 23:00 I can attend on Mondays/Tuesdays 23:07:06 ack 23:07:10 works for me 23:07:15 fine for me 23:07:20 (at 22:00 UTC I mean) 23:07:21 ok, let's stay with tuesday then 23:07:40 see you (if not before) on March 11, 22:00 UTC. 23:07:46 ack 23:07:53 cool 23:07:53 I'll try before 23:07:56 thanks everyone 23:07:59 I'd also like to reiterate: 23:08:16 if you have a small task for me with a given deadline... pass it on.. otherwise I cannot do much atm 23:08:29 ruby-dbus needs help 23:08:34 (such as a RFS) 23:08:38 oh? 23:08:39 * terceiro needs to go o/ 23:08:43 #endmeeting