20:04:36 #startmeeting Development IRC meeting 3 June 2020 20:04:36 Meeting started Wed Jun 3 20:04:36 2020 UTC. The chair is ashimema. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:04:36 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:04:36 The meeting name has been set to 'development_irc_meeting_3_june_2020' 20:04:39 #info Katrin Fischer, BSZ, Germany 20:04:48 #chair cait 20:04:48 Current chairs: ashimema cait 20:04:49 #chair tcohen 20:04:49 Current chairs: ashimema cait tcohen 20:04:50 #chair tuxayo 20:04:50 Current chairs: ashimema cait tcohen tuxayo 20:04:57 #info Tomas Cohen Arazi, Theke Solutions 20:04:57 #info Magnus Enger, Libriotech, Norway 20:04:58 #info Mengü Yazıcıoğlu, Devinim, Turkey 20:05:07 XD 20:05:09 #link https://wiki.koha-community.org/wiki/Development_IRC_meeting_3_June_2020 Agenda 20:05:14 #topic Introductions 20:05:19 #info Owen Leonard, Athens County Public Libraries, Ohio, USA 20:05:24 Please use #info to introduce yourselves 20:05:25 #info David Nind, Wellington, New Zealand 20:05:27 #info Chris Cormack, Catalyst iT 20:05:27 * magnuse relizes who enkidu is and waves in his direction 20:05:36 #info Martin Renvoize, PTFS-Europe, UK 20:05:40 #info tuxayo/Victor Grousset, France 20:05:43 #info Liz Rea 20:05:58 ooh.. hi enkidu, nice to see you here :) 20:06:00 magnuse :) 20:06:06 Hi all 20:06:11 hi enkidu 20:06:25 * magnuse watches nz crimes on television - "brokenwood mysteries" or something like that 20:07:07 #topic Announcements 20:07:25 I have an announcement on behalf of Joubu 20:07:43 #info The 20.05 release happened, just in time to get into May.. delays were caused by some hurdles to get over with packaging. 20:08:10 (announcement about the IPA stocks?) 20:08:20 [off] #info Joubu quit beer 20:08:25 #info Packages should be available very soon, transfering them from the staging repository to live is in progress. 20:08:47 mtj++ :) 20:08:49 mtj++ 20:08:50 Oh no, Joubu without a beer ? 20:09:01 mtj++ 20:09:09 mtj++ 20:09:13 anyone else got any announcements.. 20:09:16 yes 20:09:16 mtj++ 20:09:17 enkidu: just rumours 20:09:17 mtj++ 20:09:38 cait: Oh nice ! I miss the Kohacon days. 20:09:40 We forked koha-gitify into the community Gitlan 20:09:47 Gitlab 20:10:15 :) 20:10:28 and KTD now points there 20:10:33 #info koha-gitify has been forked into the koha-community gitlab organisation 20:10:44 tcohen: it was another in koha's git server before? 20:10:46 community++ 20:10:58 it was in a personal github repo 20:11:02 *was it in koha's git server 20:11:03 ok 20:11:09 #info koha-testing-docker now utilises the community maintained version hosted on gitlab 20:11:16 it was the only stack piece we didn't, as community, control 20:11:26 Nice! 20:11:28 it was on githbub if i recall correctly tuxayo 20:11:46 anyone else got any announcements, or rumours? 20:11:51 not sure how well maintained if has been.. I vaguely recollect the original maintainers have moved on to other pastures? 20:12:23 ashimema that's correct, but have to say anytime we needed anything, our PR were always accepted promptly 20:12:42 good to know 20:13:05 chris hall? 20:13:24 https://github.com/chrisosaurus/koha-gitify 20:13:42 #info Many thanks should be passed to the original maintainers of koha-gitify and their diligent eforts keeping it up to date when pull requests were submitted :) 20:13:44 yeah he works for google now 20:13:52 nice 20:14:05 chris_hall++ :) 20:14:15 chris_hall++ 20:14:46 chris_hall++ 20:14:59 chris_hall++ 20:14:59 & ++ to other maints 20:15:38 as someoen who started out with a classic git installation... things are so much better these days i can tell you :) 20:15:56 community++ 20:15:58 Definitely 20:16:04 gitify was a fantastic evolution 20:16:13 it was the trampolin 20:16:23 :) 20:16:37 filling the gap between deb installs and our dev environments 20:16:45 then kohadevbox, then KTD 20:16:52 all relying on koha-gitify 20:17:03 * ashimema actually still uses a classic git install allot for dev 20:17:08 chris_hall++ 20:17:09 but also leans on k-t-d 20:17:28 (same person, not to get off track, also facilitated the change from the old templating system to template toolkit) 20:17:39 * magnuse needs to get into ktd 20:17:40 I'll wait with ktd until tomas can show me in person again ;) 20:18:01 I'll buy the tickets as soon as they allow me to, cait 20:18:02 ooh, good plan, cait 20:18:10 :) 20:18:14 ktd has really improved 20:18:21 docker itself, has really improved. 20:18:36 right.. shall we move on 20:18:40 speaking of which 20:18:42 exaclty 20:18:51 #topic Update from the Release manager 20:19:01 er.. RM is MIA ;) 20:19:17 #info Joubu brewed 20L of IPA, don't count on him for the rest of the week 20:19:20 he said he agreed with whatever you decided 20:19:58 Yes, we can do whatever we want! :D 20:19:59 #info Jonthan is missing this evening, but expressed he has great trust in the team around him 20:20:01 lol 20:20:08 LOL 20:20:16 I like british formality 20:20:27 hehe 20:20:30 hm we could discuss the koha release naming again :P 20:20:30 :) 20:20:49 I like minutes to look half professional 20:20:53 montypython++ 20:21:12 I as very dissapointed to see the first commit this cycle 20:21:26 sooo boring 20:21:48 #topic Updates from the Release Maintainers 20:21:52 rmaints? 20:21:52 rmaints is talljoy, lukeG, hayley 20:22:02 Not anymore! 20:22:16 no, rmaints is lukeG, aleisha, tuxayo 20:22:30 > the first commit this cycle 20:22:33 The one who says «All our codebase are belong to everybody» ? 20:22:43 * oleonard liked that one 20:22:44 rmaints is lukeG, aleisha and tuxayo 20:22:56 * tuxayo too 20:23:13 rmaints? 20:23:13 rmaints is talljoy, lukeG, hayley 20:23:22 forget rmaints 20:23:23 wizzyrea: I forgot rmaints 20:23:37 rmaints is lukeG, aleisha, and tuxayo 20:23:43 rmaints? 20:23:43 lukeG, aleisha, and tuxayo 20:23:49 yay! 20:23:52 wizzyrea++ 20:24:04 hmm.. maybe i missed a commit 20:24:07 I only see 20:24:23 No, make it say "lukeG, aleisha, and tuxayo, I SUMMON THEE!!" 20:24:30 `Koha 20.06 - start of a new dev cycle` 20:24:31 forget rmaints 20:24:31 wizzyrea: I forgot rmaints 20:24:38 lol 20:24:43 #info tuxayo (for 19.05) is learning the RMaint workflow and how to interpret CI results 20:24:48 + NewVersion( $DBversion, undef, 'All our codebase are belong to everybody' ); 20:24:53 rmaints is lukeG, aleisha, and tuxayo, I SUMMON THEE!! 20:25:02 haha 20:25:10 tuxayo we'll talk about it shortly 20:25:12 ah.. nice 20:25:13 :) 20:25:55 #info rmaints are all setup and ready to go.. handover complete :) 20:26:01 * magnuse needs to catch some sleep, even if the sun is high in the sky - have fun #koha! 20:26:07 doubt there's much more to say 20:26:14 enjoy magnuse 20:26:21 bye magnuse 20:26:24 #topic Updates from the QA team 20:26:26 cait 20:26:29 found one test bug, also moved 1 step in the RMaint queue :D 20:26:55 :) 20:27:02 well done tuxayo 20:27:39 oops 20:27:46 I'm worried about 24986 20:27:57 Joubu said when asked: QA ALL THE THINGS! 20:28:01 * ashimema is getting temped by his bed.. 20:28:02 I'd like to forward this message of our RM 20:28:10 :) 20:28:12 otherwise, we are stlil focusing on bugs right now 20:28:20 #info QA ALL THE THINGS :) 20:28:25 I was told tests had to be fixed... but I am not up to date if they have been already 20:28:39 it's been a little busy, will try to catch up soon 20:29:02 i still see a lot of yellow so i guess.. FIX ALL THE TESTS! is stlil on as well 20:29:04 thx 20:29:10 understood 20:29:14 #info bugs are priority for the first month, but there's also a few 20_11_target bugs in the queue I believe which didn't make it end of last cycle and would be nice to do early this cycle. 20:29:46 #info We need to get the test suits passing again... 20:29:49 we've also discussed makeing the items, biblio, biblioitems merge soon 20:30:03 > We need to get the test suits passing again... 20:30:03 At least they pass locally 20:30:11 as this blocks progress in functionality like showing info about deleted items and such in variou splaces 20:30:16 indeed.. I think that's badged up.. if it isn't already we should do so 20:30:17 cancelled records in acq 20:30:28 It's not that centralized even if it's harder without CI 20:30:49 I'm working on jenkings 20:30:52 jenkins 20:30:58 wanted to talk about it next 20:31:05 #topic General development discussion (trends, ideas, ...) 20:31:09 all your tochen :) 20:31:28 #info Test suite, proper environment to test stable branches 20:31:43 lets do your first tcohen.. I think it's more important 20:31:46 The last couple weeks, we've been fighting packaging, and the OSs we support 20:32:12 some problematic libraries involve the API (i.e. Mojolicious, OpenAPI, YAML, etc) 20:32:32 making things work on all the supported Debian/Ubuntu versions 20:32:34 has been a pain 20:32:36 mtj++ 20:32:49 (CI for tests is like an addiction, we see the signs of lack of it quickly ^^) 20:32:54 mtj++ 20:33:05 There are nasty problems with DBD::mysql (regressions) 20:33:21 mtj++ 20:33:21 tcohen++ 20:33:21 The new matrix of test envs is great! 20:33:36 So.. is the jenkins change that we're now avoiding cpan for installing any deps 20:33:47 that cause some tests to fail. I consider those failures minor, as they are mostly problems with representation of floats 20:34:00 i.e. relying more strictly on the koha-common package pulling in all the right things? 20:34:00 BUT 20:34:06 rigt 20:34:22 this morning I branched KTD for 19.11 20:34:29 and cleaned all the things it installs 20:34:37 > the jenkins change 20:34:37 Only in Jenkins or in koha-testing-docker in general? 20:34:37 excellent 20:34:37 darn tootin' it is. 20:34:39 it will now *only* install koha-common and dependencies 20:34:48 and just a few cpan things required for tests 20:34:57 this way we can be sure about things 20:35:15 Joubu considered we should discuss this somehow as it is an overhead 20:35:24 I belive it is fairly 'light' 20:35:38 specially if we finish building the .deb building pipeline 20:35:41 i must admit I still find the control file syntax weird.. surely test-depends should be what it sounds like 20:35:44 and all gets automated 20:35:54 but that's a topic for another day.. 20:36:42 right 20:36:47 I agree the .deb building pipeline is key to this.. and has been the missing element for the whole 12 months of my time as RM 20:36:51 well.. 20:36:54 So, my plan is to make all the stable branches tests 20:37:00 run on version-spcific KTD 20:37:08 great! 20:37:19 tcohen++ 20:37:26 so 19.11 will be run against koha/koha-testing:19.11-stretch 20:37:28 the full pipeline has always been missing.. but during my tenure with the loss of mirko we lost an important part of it which was working 20:37:44 and I'd say rmaints should do their job probably on that images as well 20:37:45 tcohen++ 20:38:13 what do you mean by rmaints do their job properly on those images? 20:38:21 once the packages for 20.05 arrive I'll branch KTD for that as well 20:38:28 > and I'd say rmaints should do their job probably on that images as well 20:38:28 Good, that's how I understood it. 20:38:30 watch the debian.control bugs and ensure they test and push patches that land there 20:38:40 rmaints: ask me whatever you need about jenkins or the pipelines we are setting 20:38:48 and ask me the changes you want 20:38:55 (for example, if you want to test on another OS) 20:39:42 > (for example, if you want to test on another OS) 20:39:43 are there missing OSes that we support? 20:39:50 another OS? Do we have so much people to work on? 20:40:17 Or maybe you mean like OS+the DBMS 20:40:54 * ashimema thinks we're now well covered for combinations really 20:41:02 I'm talking about debianderivative like Ubuntu 20:41:06 or newer Debian 20:41:14 but if you want to set another OS 20:41:22 I could guide or help if needed 20:41:27 we could in theory test a few more combinations.. but I'm not entirely sure we need to 20:41:49 having said that.. we do know the DBD versions can make a difference.. so perhaps we should be. 20:42:00 we should 20:42:04 I was thinking the stable branches might need MariaDB testing. 20:42:06 But maybe the angle should be: should MySQL remain the default DBMS? 20:42:28 > we do know the DBD versions can make a difference 20:42:28 What do you mean? 20:42:46 https://github.com/perl5-dbi/DBD-mysql/issues/212 20:42:50 * ashimema opened a bug asking for DB versions in HEA to help asses this 20:42:59 it got mostly shot down 20:43:24 i think if it was just major version number 20:43:27 thats ok 20:43:39 5.5 not 5.5.10~dsfg etc 20:43:39 I found a bug on DBD::mysql, that got fixed in v4.042 20:43:45 I wanted to try to move this bug forward. It didn't seemed blocked. 20:43:58 (bug about HEA & DBMS) 20:44:16 and people find the fix got reverted in a few newer versions 20:44:21 (Buster's, for example) 20:44:31 lets ove it forward tuxayo.. I have more time for things like that now 20:44:38 only so many battles I could fight as RM 20:45:15 sounds sane to me rangi.. and all we really need I think 20:45:20 yep 20:45:50 ooh.. rangi.. did you see I added a followup pull request on dashboard.. 20:45:52 > 5.5 not 5.5.10~dsfg etc 20:45:52 That was the next step for me. Try to find a reliable way to parse at least MySQL and MariaDB version strings. 20:45:58 ashimema: yep merged and deployed 20:46:08 awesome, thanks 20:46:15 apologies for messing up :( 20:47:10 no worries 20:47:10 #info rmaints contact tcohen about jenkins and environments their branches should be tested against 20:48:42 * ashimema is really happy to see Jenkins growing so much... it really improved last couple of cycle and the trend is only getting better :) 20:49:19 shall we move on tcohen? 20:49:27 yes 20:49:42 #info Can we/Should we deprecate MySQL 5.5 support? 20:49:56 yes please 20:49:57 haha 20:50:07 we have a few challenges to overcome if we want to continue to support it 20:50:14 dcook could argue if he has some old RHEL 20:50:17 bug 10517 for instance 20:50:19 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=10517 normal, P5 - low, ---, rudolfbyker, Needs Signoff , koha-restore fails to create mysqluser@mysql_hostname so zebra update fails 20:50:19 Which OS has it by default? 20:50:22 but I'm sure there are some others 20:50:43 D8 did, D9 does if you upgraded from D8 I believe 20:51:23 do we have some statistics about OS and DB versions around the world? 20:51:27 Is D8 deprecated? For 20.11 20:51:37 I found some case where you could certainly end up with it on D9 without really trying to.. when upset me. 20:51:40 it is EOL 20:51:49 D8 is deprecated already 20:52:11 crap.. I forgot to add specific mention of that in the release notes 20:52:22 20.05 supports it, but not deliberately.. 20:52:23 https://wiki.debian.org/DebianReleases 20:52:38 it seems stretch is EOL soon LOL 20:52:50 everybody should switch to Ubuntu 20.04 hahaha 20:53:04 the idea is to formally drop D8 support ith 20.05.00, 19.05.12, 18.11.06 (i.e. the set of releases that just happened) 20:53:34 but.. we could do it with the next set of point releases.. which will actually correspond to when D8 is officially dropped by Debian themselves 20:53:36 > I found some case where you could certainly end up with it on D9 without really trying to 20:53:39 Sound reasonable to ask for a migration to MariaDB or a latter MySQL version. 20:54:12 what is the mysql 5.5 correspondent mariadb version? 20:54:44 5.5 I think 20:55:25 5.5, Mysql5.6 is MariaDB 10.0 I remember 20:55:34 ^ that's right 20:55:41 stretch ships 10.1 20:55:42 personally.. I think we should bump to minimum of MariaDB 10.1 20:56:02 that's the default if you install from fresh in D9.. 20:56:07 it is 20:56:32 and MySQL 5.7 20:56:40 pretty sure you only get mysql5.5 if you upgraded from D8 or explicitly ask for MySQL instead of MariaDB 20:56:54 indeed 20:57:21 lets vote for deprecation 20:57:30 Is MariaDB 10.1 == MySQL 5.7 ? 20:57:30 Are there know limitations that the MySQL side would cause? 20:57:34 ++ 20:57:40 5.7 -> 10.2 20:57:56 great! 20:58:21 Anyone has an info or argument before voting? 20:59:05 these days you have to jump through some hoops to install MySQL if you really want it.. all distros I've worked with of late use MariaDB even if you install the mysql-server package it's just a symlink 21:00:21 Thanks, I had this impression also 21:00:45 #startvote Should we officially deprecate MySQL5.5 support (upping minimum SQL versions to MariaDB 10.1/MySQL 5.7? Yes, No, Abstain 21:00:45 Begin voting on: Should we officially deprecate MySQL5.5 support (upping minimum SQL versions to MariaDB 10.1/MySQL 5.7? Valid vote options are Yes, No, Abstain. 21:00:45 Vote using '#vote OPTION'. Only your last vote counts. 21:01:02 #vote Yes 21:01:13 #vote Yes 21:01:14 #vote Yes 21:01:16 #vote yes 21:01:57 For with Koha version BTW? 21:01:58 *Which 21:02:10 #vote Yes 21:02:15 #vote Yes 21:02:16 onwards 21:02:21 20.11 21:02:22 Like for Debian ("retroactive drop") or less strict 21:02:30 Thanks tcohen 21:03:07 I would say probably retroactive (to keep your tasks as rmaints simple) 21:03:42 people should really be upgrading debian at this point and as part of that expecting/requirement them to upgrade mysql doesn't seem terrible to me.. 21:03:52 We are not taking advantage of Joubu's absense laissez-faire message 21:04:02 but.. Im happy to be overruled on that one.. i's not a process I've had to do much of.. 21:04:19 jessie goes out of support Very Soon Now 21:04:19 we could do with a sysops view on that 21:04:23 > as part of that expecting/requirement them to upgrade mysql doesn't seem terrible to me.. 21:04:23 ++ 21:04:42 any more for any more.. 21:04:45 3 21:04:52 2 21:04:53 1 21:04:54 deprecating Debian 8 21:04:54 . 21:04:58 implies somehow 21:04:59 #endvote 21:04:59 Voted on "Should we officially deprecate MySQL5.5 support (upping minimum SQL versions to MariaDB 10.1/MySQL 5.7?" Results are 21:04:59 Yes (6): davidnind, wizzyrea, ashimema, tuxayo, tcohen, enkidu 21:05:08 excellent 21:05:08 darn tootin' it is. 21:05:32 any more general topics chaps? 21:05:35 or shall we move on 21:05:39 Where should/are these things (Debian, DBMS support) documented? 21:06:29 > any more general topics chaps? 21:06:29 Can we have MariaDB by default in dev&test environment then? 21:06:36 wiki pages somewhere.. the website on the install details page and in every release note 21:06:40 #info Joubu volunteers to announce MySQL 5.5 deprecation 21:06:45 perhaps the coding guidelines too 21:06:54 nice touch tochen 21:06:57 tcohen.. even 21:07:01 > Joubu volunteers 21:07:01 he he 21:07:15 I will update the section in the manual 21:07:25 thanks davidnind 21:07:34 Joubu is quietly preparing to FQA my patches, for sure 21:07:45 #action davidnind will update the manual with new minimum requirements updates 21:07:56 #topic Review of coding guidelines 21:08:07 i don't believe there's anything to review 21:08:09 so... 21:08:15 right 21:08:25 #topic Set time of next meeting 21:08:32 should we file a bug for KTD to default to other DB version then? 21:08:36 davidnind: great, there is a section in the manual :D 21:08:40 nevermind 21:08:55 > should we file a bug for KTD to default to other DB version then? 21:08:55 Ok, I can do that 21:09:08 deffo tcohen 21:09:23 I'd say each OS it's shipped version 21:09:29 and MySQL 8 21:09:33 * ashimema is bowing out.. brain is fading fast now 21:09:35 as we do already 21:09:40 go to bed! 21:09:46 agreed tcohen 21:09:53 > I'd say each OS it's shipped version 21:09:53 Yes, I was thinking about that. 21:11:25 > and MySQL 8 21:11:25 So every matching MariaDB by default + an optional way to have MySQL 8 to offer a cutting edge option. 21:11:36 right 21:11:38 ashimema 21:11:38 it has been said that ashimema is literally just writing up the release anouncement ;) 21:11:38 Like with ES 7 21:11:41 Good 21:11:44 yup 21:11:48 time for next meeting 21:12:09 June 17th? 21:12:42 general meeting is on 17th too 21:13:10 how did we end up conflicting again 21:13:18 mwah 21:13:35 it's a 3pm.. we could do 2 and force ourselves to finish on time 21:13:47 or follow the general meeting 21:13:50 or change day 21:13:50 ashimema: that doesn't look right 21:14:02 > how did we end up conflicting again 21:14:02 Months are a bit more than 4.0 weeks 21:14:11 indeed 21:14:22 same day, another time? 21:15:39 Do people usually manage to have 2 meeting in the same day? 21:15:57 without having trouble fitting them with the rest. 21:16:00 3 or 4 21:16:14 I'd say same day another time 21:16:17 and make them fit 21:17:40 ok, so one hour before the general meeting? 21:19:23 13 or 15 UTC.. which is 14 or 16 UK, 10 or 12 BA, 21:19:40 13 works for me.. 21:20:08 I think that fit more timezones (except Wellington.. sorry rangi, davidnind) 21:21:16 #info Next meeting: 17 June 2020, 13 UTC 21:21:18 #endmeeting