15:59:39 #startmeeting 15:59:39 Meeting started Thu Aug 6 15:59:39 2015 UTC. The chair is matthieucan. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:59:39 Useful Commands: #action #agreed #help #info #idea #link #topic. 15:59:44 clemux: ping 15:59:47 orestis: ping 15:59:56 pong 16:00:00 pong 16:00:10 #topic next meeting 16:00:19 agree to make a meeting next week before debconf? 16:00:26 yes 16:00:34 on friday? 16:00:44 I'd be strongly enclined for thursday 16:00:49 if you're both available 16:01:05 yeap, coz on friday i ll be in the train for debconf :p 16:01:19 I'll be at debcamp from wednesday, so no problem for me 16:01:40 is 15:00 CEST ok for you? 16:01:43 yeap 16:01:45 yes 16:01:48 clemux: oh cool, didn't know you go there 16:02:03 #agreed next meeting thursday 13 aug, 15:00 CEST 16:02:14 #topic orestis - last week 16:02:39 so file deltas along with some bug fixes got merged (PR#34) 16:02:54 and then the description etc in PR#35 16:03:05 i ll move those from review to merged right? 16:03:13 yes 16:03:58 then there's the api proposal PR#36 which needs some editing 16:04:18 i also PRed the spdx export as i had forgotten when i did it :p 16:04:20 ok, do you miss any reviewing work for this point? 16:04:36 yes, I looked a bit at that 16:04:48 few points need to be discussed later, like the Creator fields 16:04:49 no i think its ok.. i ll edit the proposal tonight but i can start the api by next week 16:05:19 ack for api 16:05:20 yes there are field i wasn't sure what exactly to add.. anyway that is probably the least urgent PR 16:05:28 fields* 16:05:53 ok, I'll write my thoughts about it, but we'll need zack's advice on some uncertain fields 16:06:08 yes maybe we can see this at debconf 16:06:17 for sure! 16:06:20 other points we forgot? 16:06:52 there's the PR#37 for the bug with missing d/copyright files 16:06:57 which you reviewed and updated 16:07:00 i think that's all 16:07:23 oh true, I didn't check your comment for file renaming 16:07:29 should be merged soon I believe :) 16:07:44 ok, great, that week looks productive 16:07:49 can we move on to next week? 16:07:52 yeap 16:08:01 #topic orestis - next week 16:08:21 so 1) is the api 16:08:47 can you prioritise items, according to what should be done before debconf/what would be better done there? 16:08:50 then 2) i think a DB proposal and maybe the plugin to fill relevant info 16:08:53 yes for API, great 16:09:30 hm maybe the thing we talked via email? for being more explicit when there are no patches etc 16:10:01 yes, follow zack's message for that, with an explicit text 16:10:10 yeap i ve taken note for that 16:10:16 what about documentation for /patches? 16:10:31 yes i can also do that 16:10:40 do you prefer doing db relevant things at debconf? 16:10:41 a first shot should be doable by recycling the .txt proposal(s) 16:10:55 true 16:11:06 that's a good point, yes we should do that 16:11:40 ok.. then to summarise: 16:12:18 1) api 2) explicit mention fo versions without patches 3) documentation and templates for patches.d.n 16:12:33 i think that's not going to take me much time though 16:12:42 and i am not sure if i can pick anything else 16:13:02 in that case what about checksum of orig.tar.gz, via the updater as zack suggested via email? 16:13:14 oh 16:13:17 yes but to cache it where? 16:13:17 db is needed for that 16:13:20 yeap :/ 16:13:21 sorry 16:13:47 then: 16:14:02 what about the items in copyright? a few of them seem to be doable quite shortly 16:14:17 verify license links or complete regexes, etc 16:14:35 that's actually in the PR of the license statistics 16:14:50 i have put that trello to remember to check them 16:15:11 in the license statistics PR i did all the regexes for all the licenses ( 16:15:16 (the standard ones) 16:15:29 ok, that's great. So it 16:15:54 the author we said that we won't return it for now, work on the version handling so we avoid to c/p paste that code verywhere 16:15:54 's blocked by #22, and by the update of testdata, which I have no idea how to do (I didn't file the doc) 16:16:00 it's probably postponed to debconf 16:16:34 maybe i can work on the version handling.. but i have no clue how to do it.. but i can try 16:16:44 what is that? 16:17:20 oh ok 16:17:21 for now the version handling is copy pasted every time 16:17:26 just saw the GH discussion 16:17:29 to make the redirects etc 16:17:44 which one? 16:17:50 https://github.com/Debian/debsources/pull/6#issuecomment-110124698 16:18:11 ah yeah ! 16:18:41 ok, let's keep it for this week in case you have some time, I might be around to help anyway 16:18:50 so I think we're good to go? 16:18:56 yes great :) 16:19:09 let's move on 16:19:14 #topic clemux - last week 16:19:23 how's the async going? 16:19:39 so, the garbage collector stage is done, but I need tests to be sure 16:19:50 great! 16:19:50 the update_metadata stage is also done 16:19:57 needs tests as well? 16:20:00 yes 16:20:05 ok 16:20:12 I spent the rest of the week working on the tests 16:20:31 I still have issues with them :/ 16:20:35 for the other stages? 16:21:01 do you mean issues for testing, or issues for the stages themselves? 16:21:23 issues for testing 16:21:46 I have modified assert_db_schema_equal (https://github.com/Debian/debsources/blob/master/debsources/tests/test_updater.py#L85) for comparing specific tables 16:22:03 so that we can have an assertion for each stage/task 16:22:14 that sounds good 16:22:28 I intend to test each plugin separately, etc. 16:22:47 but for now, the "actual" list of objets is empty 16:23:02 even though the test database is correctly filled with data from the async updater 16:23:31 you mean you fill the db, and can't retrieve its content? 16:23:36 yes 16:23:43 I'll commit the code and ask for help 16:23:54 good idea 16:24:02 hard to say without diving in the code 16:24:05 yup 16:24:23 do some stages work or is it the same for all? 16:25:13 it's the same for all tables 16:25:34 anyway, I'll show you the code and explain the problem better :) 16:25:45 ok, let's do that! 16:26:00 anything else to add for last week? were there other stages scheduled? 16:26:19 stats and charts were scheduled, but I haven't got around to do them 16:26:59 if you're blocked by the tests, send code/shout for help as soon as possible 16:27:12 yup 16:27:24 besides that, I've worked on improving the worfklow of celery tasks in the updater 16:27:37 what changed? 16:27:52 for example, currently the update_suites main task is called in the extract_new stage 16:28:09 this is annoying for unit tests (I need to mock out the update_suites task) 16:28:20 and means we can't disable stages as in the sync updater 16:28:35 I still need to do some research, but I think I've found out how to do it 16:28:46 I'll submit a PR for the async documentation once I'm sure 16:29:11 (I'll add that to trello right after the meeting) 16:29:12 ok, that goes in the category "fix tests" then? 16:29:35 it's more of a general improvement 16:29:50 ack 16:29:53 anything else? 16:29:56 that's it :) 16:30:05 #topic clemux - next week 16:30:18 do you think we can expect to merge the async branch at debconf? 16:30:20 yes! 16:30:41 awesome! so let's list precisely the blockers and needed steps? 16:31:18 1) finish the tests, current blocker: problem with the fixture/test db comparison 16:31:27 2) update_stats 16:31:32 3) update_charts 16:31:56 4) "task workflow improvements" (I can't find a better name) 16:32:20 not necessarily in that order I presume? 16:33:07 that order would be best (for 1,2,3 at least) 16:33:39 ok, then it looks like a week plan, doesn't it? :) 16:33:52 I've added 4 on trello with a better name: "make tasks independant from each other" 16:34:03 matthieucan: yup 16:34:50 independent tasks look great for async! :) 16:34:51 after that (and fixes to the likely bugs found with the tests), it'll be time for merging and testing on sourcesdev 16:35:23 do you expect a smooth merge or will we fight with conflicts? 16:35:55 sourcesdev uses the same db as s.d.n, I don't know if we should do it this way 16:35:56 there will be conflicts with orestis' changes to the statistics stages (iirc) 16:36:01 matthieucan: oh, right 16:36:11 (they are not merged) 16:36:29 (ah, yes, indeed) 16:37:12 so we'll see that when needed I guess! 16:37:22 can we move on? incoming week is clear? 16:37:25 yup 16:37:30 cool :) 16:37:35 #topic misc 16:37:43 anything to add or discuss? 16:37:53 matthieucan: will you be at debcamp? 16:38:10 clemux: no, I'll arrive saturday 15 16:38:17 oh, ok 16:38:59 nothing else to add :) 16:39:04 ah yes me neither :p 16:39:09 alright! 16:39:15 thank you for the meeting 16:39:16 #endmeeting