19:06:29 #startmeeting QA Team Meeting 19:06:29 Meeting started Tue Sep 3 19:06:29 2013 UTC. The chair is cait. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:06:29 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 19:06:29 The meeting name has been set to 'qa_team_meeting' 19:06:43 #topic introductions 19:06:43 #info wahanui, a bot that has become sentient 19:06:53 ok, quick poll - who is around? :) 19:07:02 #info Katrin Fischer, BSZ 19:07:08 #info Brendan Gallagher 19:07:29 #info Paul Poulain, BibLibre 19:07:30 #info Kyle M Hall, ByWater Solutions 19:07:47 #topic Agenda 19:07:55 #info Jonathan Druart, BibLibre 19:07:59 ah :) 19:08:14 ok, today's agenda is on the wiki 19:08:14 (sorry for being late) 19:08:19 #info Agenda: http://wiki.koha-community.org/wiki/QA_team_meeting,_3rd_September_2013 19:08:31 we are not as many as I had hoped for 19:08:44 but maybe we can still discuss a few things and gather some ideas 19:09:08 is there something missing on the agenda? or something we should leave out for today? Discuss first? 19:10:05 cait = joubu & I have discussed of what we putted on the agenda, you added a few things that I agree with, so the agenda is OK for me 19:10:06 do i need to bribe you all with cookies? 19:10:11 ;-) 19:10:19 looks good to me! 19:10:20 #info Ruth Bavousett 19:10:32 I'm here, with cookies! 19:10:43 :) 19:10:50 ok, i think we got another agenda item 19:10:57 ruth has offered to help out with QA 19:11:09 druthb++ 19:11:22 #topic QA team 19:11:32 khall: i think that was a yes? :) 19:11:42 absolutely! 19:11:52 Not sure how much time I'll have, but I'll try to throw a few in, particularly on the UNIMARC patches that are needing signoff, and other things that Her Caitness needs me to do. 19:11:53 druthb++ (yes for me as well 19:11:54 ) 19:11:55 any more opinions? 19:11:59 I can test SIP - but all SIP2 patches in agenda are from khall so I can't… perhaps I can schedule sometime with cait to help QA those? 19:12:16 bag: how about those ldap related patches? 19:12:19 bag: that would be helpful - i have figured out how to set up the sip server by now :) 19:12:28 I will do the LDAP one. I can test that :) 19:12:45 great :) 19:12:51 let's move a bit slower so i can keep up 19:12:54 heh 19:12:59 i will note that we have ruth on the qa team now 19:13:24 #info druthb will join the QA team to help out qa'ing 19:13:29 bag & cait & others = in the previous release, the rule for QA was "anyone from the QA team, even someone involved in the patch or signoff, can QA once the patch is more than 1 month old". afaik, this rule has not been cancelled 19:13:46 no it has not, but the old ones are pretty complicated 19:13:57 so i think for those it's a bit difficult 19:14:00 let me st a new topic 19:14:05 #topic QA priorities 19:14:09 ok 19:14:12 we have bag for LDAP? 19:14:20 oh paul_p I did not know that 19:14:35 bag meeting already useful then ;-) 19:14:38 #action bag offered to QA LDAP patches - starting with bug 9299 19:14:39 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=9299 major, P3, ---, nunyo, Signed Off , for loop in Auth_with_ldap.pm requires an extended patron attribute to be set or LDAP logins fail 19:15:06 #action bag also offered to help cait test SIP2 patches 19:15:10 ;) 19:15:19 nor did I. I assume QA'ing my own original patch would be a bad idea still? But QAing something where I only wrote a followup would be ok, assuming it's older than a month. 19:15:21 ok, and druthb would offer to look at UNIMARC? 19:15:28 khall: definitely 19:15:29 yis. 19:15:33 and it's an emergency rule 19:15:46 and not for huge patches 19:16:09 #action druthb will try to take a look at UNIMARC patches 19:16:19 i am going to add you as a QA contact to those later? 19:16:23 what equals a HUGE patch? 19:16:37 bigger font? :P 19:16:48 bug 8015 is quite huge if you look at the list of patches there 19:16:49 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=8015 new feature, P4, ---, kyle.m.hall, Signed Off , Add MARC Modifications Templates 19:16:54 i think big new features count as huge 19:17:09 * druthb hands cait the wiffle bat, to use as needed on rowdy Californians. 19:17:17 i think it would be better if you get worried about a bug to contact other qa team members who can take a look first 19:17:33 sounds like a good idea 19:17:58 hi all 19:18:02 hi mtj 19:18:31 I know currently a lot of things are stuck in the queue 19:18:35 * mtj reads the scrollback... 19:18:50 that's part of the reason for the meeting... I think we need to try and lower the numbers 19:18:57 because the deadlines are getting really close 19:19:04 I started to qa 8015 and submitted a lot of followups, so I would prefer others pair of eyes look at it. But it looks very good to me! 19:19:13 bug 8015 19:19:13 bug 8015 is quite huge if you look at the list of patches there 19:19:42 Joubu: I am going to try - but i don't have a lot of spare time right now - september is busy month here, which is a problem 19:19:44 cait OTOH, without galen pushing, reducing QA queue won't be successfull (too bad gmcharlt is not around...) 19:20:08 paul_p: gmcharlt wanted to be here, but i think he has been travelling the last few days and only comes back today 19:20:29 I think even RM deserves some days off? :) 19:22:06 no days off for the RM! 19:22:15 i know that developers are frustrated about their things not going in 19:22:18 it's understandable 19:22:39 but we only can change that by working together - i can't do a lot more than i have been doing the last few months - so I really need some help here to get things moving 19:22:42 how many passed QA are still waiting to be pushed? 19:22:46 we currently have 63 bugs set to Passed QA 19:22:47 #info Mirko Tietgen, lurking 19:23:07 hi drojf 19:23:14 dashboard? 19:23:14 dashboard is at http://dashboard.koha-community.org/ 19:23:19 * bag sends caffeine to gmcharlt 19:23:20 bag: the numbers are there, in color :) 19:23:22 or is it over already? :) 19:23:33 cait I am THE dense sometimes :P 19:23:50 i think the one waiting to be pushed are not the problem, those numbers tend to get lower pretty fast when the RM is in action 19:24:17 and we can help out with giving things a second sign off there too 19:24:35 i think the queues we need to discuss are waiting for qa and needs sign-off 19:25:24 agreed 19:26:40 I would reall ylike to try and get the number of bugs down in that queue first - because those get transported down into the maintained versions 19:26:43 cait = I agree we have to speak of the to-qa queue, but the pile of passed QA patches contains 10 patches that are pending since june. So I don't share your optimism about the QA queue being lowered when the RM pushes. But let's speak of the to-qa queue 19:26:47 and we should have them as stable as possible 19:27:14 paul_p: I know some of those are quite big 19:27:55 that's a sensible proposition. Is there anyway to filter bugzilla for just signed-off bugs? I haven't found one. 19:28:21 hmm do some of the to-qa patches depend on "passed QA" to be pushed? 19:28:34 khall: not sure i understand you right : ) but thequeues on he dashboard are linked :) 19:28:35 cait I already have said that many times, but my conviction is that "push asap enhancement, and take time to test/check them before release" is the most efficent way to go 19:28:54 yeah, but noone is going to test 19:28:58 cait (an off topic question once again...) 19:29:01 noone is going to take that time 19:29:14 in my experience :) and i have said that before 19:29:25 I'm afraid I agree with both cait and paul_p ; ) 19:29:57 if noone takes the time before they go oin - it's unlikely someone will once they do 19:30:07 pulling things out that have proven problamtic after being pushed has proved quite hard 19:30:14 and getting bugs fixed often as well 19:30:30 but we will always disagree on that I think :) 19:30:31 cait: I can search for all signed-off patches easily, but I'm not aware of any way to filter by "Importance" 19:30:45 doesn't signoff and QA help eliminate some of that problem though? isn't that the point? 19:30:46 cait = i'd like to leave this question for later, and go back to agenda. But that's a topic for hackfest, probably... 19:31:25 khall: you can add severity and priority to the list as columns, the option is at the bottom of the bugzilla search result list 19:31:26 agreed, we could discuss this for hours 19:31:30 khall: andsort by those 19:31:35 thanks cait! 19:31:46 bag: yep i think that is the point :) 19:31:59 ok, i think we are not ´moving here... so moving on seems like a good idea 19:32:14 #topic passing QA 19:33:06 we have seen a lot of work on unit tests lately :) that's great 19:33:08 kenza++ again 19:33:17 i also find them really helpful when doing QA 19:33:32 and I think we should encourage them the best we can 19:34:04 I think currently we have been asking for tests for new subs and changes to existing ones 19:34:10 that seems to work quite well 19:35:01 agreed. 19:35:09 agreed 19:35:34 +1 19:35:41 cool :) so we can agree! :) 19:35:48 +1 19:36:17 @agreed ask for unit tests when new routines get added to the API and existing ones are being changed 19:36:18 cait: I'll give you the answer just as soon as RDA is ready 19:36:22 side question = do we have a "qa rule" page on the wiki. I think we should 19:36:33 #agreed agreed ask for unit tests when new routines get added to the API and existing ones are being changed 19:36:45 no, we haven't one yet - but i think it's a good idea 19:36:54 yes that would be nice 19:37:03 #action add a page with information about QA to the wiki 19:37:05 I agree too, even if a lot of modules/routines are not easy at all to test without having a global unit tests file for this module 19:37:08 QA manager, you take care of creating it ? 19:37:19 i thik the unit test one could maybe go on the coding guidelines as well - if the RM agrees 19:37:48 i can 19:38:21 Joubu can you give more details about " a lot of modules/routines are not easy at all to test without having a global unit tests file" ? Could we quickly add an "empty" unit test file for all modules ? 19:38:29 looks like we all agree about the unit tests :) 19:38:34 something like a framework 19:39:01 Joubu: is the problem creating the data? 19:39:34 cait: I'd say that's usually the most difficult part for me. 19:39:49 i think it got easier using transactions now 19:40:01 i have tried my hands on some and it's seems a lot easier than mocking 19:40:13 but i am still stuck on a patch where i need to mock a cgi object 19:40:15 ... 19:40:16 the biggest problem I see with requiring unit tests for exisiting subroutines are some are quite large and do far too much 19:40:23 C4::Circ, C4::Auth, C4::Acq don't have unit tests for all routines. It will be though to modify some routines in these modules. 19:40:38 tough 19:40:51 i thik we can still make exceptions in some cases, but without trying we won't get good examples to copy from 19:41:04 i am iwlling to make exceptions for bugs for example 19:41:18 if the solution can be tested well using a test plan 19:41:33 agreed. I think new subs need a strict requiremtn for unit tests, existing code a yes with case by case exemptions 19:42:23 i think maybe differentiate between adding new features and fixing bad behaviour there too 19:43:18 I'd like to postpone the discussion about the new module 19:43:27 yes. We don't want to hold up bugfixes any more than we have too 19:43:45 I'd like to get an opinion of gmcharlt on the possiblity of dbix for 3.14 19:44:52 that would be great 19:45:03 it would be 19:45:24 Joubu: i am sorry, i know that#s your topic, i will try to follow up on that 19:46:46 ok to move on to the next topic? should I add something more to the minutes? 19:47:13 I asked Galen to give his opinion on bug 10363, but I didn't have any feedback 19:47:14 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=10363 enhancement, P5 - low, ---, jonathan.druart, In Discussion , There is no package for authorised values. 19:47:41 Joubu: i will try to look after that one 19:48:33 which brings me to the next topic :) 19:48:42 #topic Ideas to improve throughput 19:49:38 cloning Joubu and the rest of the team 19:50:07 bag: not sure i can let that count :) but .... oh well 19:50:45 #idea clone the QA team :) 19:50:52 something else? :) 19:50:59 falling back on the 1 month rule would indeed improve throughput 19:51:12 but in the wrong way. 19:51:16 I think QAers should not failed QA a patch for minor things (some tabs, etc.) but rather submit a followup themselves 19:51:33 Joubu: I agree 19:51:38 also capitalization maybe 19:51:49 small things, not changing logic 19:51:59 no, of course 19:52:21 #idea QA team can provide follow ups for small problems (tabs, typos, capitalization...) 19:52:42 joubu, agreed, thats what i usually do 19:52:50 i have found it's easier to qa big patches when the developer is around to ask questions 19:53:14 i have done that a few times and it was quite productive 19:53:15 mtj: yes, me too 19:53:20 cait & joubu = that's what I made when I was RM and QAed patches. small things like capitalization or things like that are easy to fix & should be by the QAer 19:53:24 i think we see a similar effect during the hackfests too 19:53:25 Joubu: in cases where it would take longer to do the follow-up, I suggest just failing it. That was my rule during 3.12. If it's faster to fix than ask for a fix, fix it. Otherwise, kick it back and ask for a fix. 19:54:37 #idea set up a time with the developer to work on their patches - so you can ask questions 19:54:43 maybe have a QA day or sprint? like we scheduled this meeting - or something similar to Global bug squashing day 19:55:28 GBSD might help with that - we could be around looking at patches of developers that are on IRC that time, give feedback more directly 19:55:45 ah, bag is typing faster :) 19:55:59 it's the coffee - not me 19:56:01 i was thinking maybe fold it into gbsd - but we could also do a special one 19:56:16 I really love the idea of setting up a time to discuss of a given bug. Something regular, like "every friday, at 6AM GMT" 19:56:20 special_qa_sprint_day++ 19:56:28 would it be possible that people dedicate some time to that idea? 19:56:33 if we set up a date? 19:56:50 if the date is set - I can clear my schedule with enough notice 19:56:51 6 AM GMT? And here I thought that the French had given up on torture. 19:56:59 #idea have a QA sprint day 19:57:23 i guess it would have to be a global friday - we are really far spread around the globe 19:58:06 this time is late for europe, but actually the best option we have with the qa team widely spread 19:58:30 should we try and set a date for a qa sprint? 19:58:49 would be best to have RM present - maybe fridays aren't the best for him? let's pole the RM 19:58:58 pole? 19:59:05 poll 19:59:06 druthb it was just an example. and we could have various options. 19:59:07 ah 19:59:28 maybe we could suggest a date 19:59:37 paul_p: :P 19:59:40 and see if that works 19:59:51 cait: let's do it and see what happens 19:59:53 what about.... 20:00:00 * cait searches for her calendar 20:00:08 how about a wednesday? 20:00:09 a wednesday is fine for me 20:01:00 wednesdy woudl work for me 20:01:03 It would be really great if wahanui could qa patches... 20:01:05 what about 11th? 20:01:22 +1 20:02:01 +1 20:02:14 +1 20:02:21 alright back to training for me - cya later 20:02:21 you know that there is the regular meeting too? 20:05:18 no, but might not be a bad thing 20:05:18 i mean it doesn't necessarily interfere 20:05:18 hm depends if we want to set a time too or make it a global one 20:05:18 we could also make it thursday 20:05:18 or friday 20:05:18 12th or 13th 20:05:18 any preferences? 20:05:42 12th could be possible for me. 13th (my) morning won't be possible 20:06:12 that's for 12th? 20:06:27 #agreed 12th September suggested date for QA sprint 20:07:24 mtj: ok for you too? sorry i might have been too fast 20:07:48 ok, any more ideas? 20:07:48 ok for me too :) 20:08:13 would someone be willing ot set up a wiki page for it? 20:08:33 i am travelling all day tomorrow and it would be nice to get that out and announced 20:09:00 i can do the wiki page today 20:09:08 mtj++ 20:09:09 mtj: awesome! 20:09:09 That'll be $1 for the awesome jar, cait 20:09:12 mtj++ 20:09:19 mtj++ ! 20:09:22 #action mtj to create a wiki page for the first QA sprint 20:09:56 * druthb puts .76 Euro in the jar for cait. 20:10:11 thx druthb 20:10:48 do we want to set up another meeting? 20:10:56 or see how the qa sprint works out first? 20:11:19 cait I don't have a preference 20:11:28 cait both are possible 20:12:20 so let's see how the qa sprint goes 20:12:37 i#d love if as many qa people as possible coudl dedicate sometime 20:12:48 maybe send a note to the list or put on the wiki page when you can be around 20:13:04 working on qA 20:14:11 cait, ill ping you when the wiki is done 20:14:23 thank you mtj 20:14:26 and thx for volunteering 20:15:09 maybe you could add a section for a little schedule? let's see if we can make use of being globally spread? 20:15:27 ok, i am going to close the meeting now 20:15:44 #agreed new meeting to be determined after QA sprint day 20:15:48 #endmeeting