IRC log for #koha, 2023-03-06

← Previous day | Today | Next day → | Search | Index

All times shown according to UTC.

Time Nick Message
03:36 Oak joined #koha
07:15 cait joined #koha
07:22 cait joined #koha
07:31 thibaud_g joined #koha
07:32 cait joined #koha
07:36 cait joined #koha
07:54 alex_ joined #koha
08:02 lds joined #koha
08:03 cait joined #koha
08:16 Oak joined #koha
08:16 liliputech joined #koha
08:16 liliputech hi koha
08:36 magnuse \o/
08:49 cait good morning #koha
09:06 marcelr joined #koha
09:06 marcelr hi #koha
09:11 i am getting 404s back from rest api in current master; does that ring a bell ?
09:13 ashimema Morning
09:16 marcelr hi
09:16 stuff like this: [2023/03/06 09:13:28] [ERROR] Warning: Could not load REST API spec bundle: /parameters/authorised_value_id_pp/in: /oneOf/0 Not in enum list: body./parameters/authorised_value_id_pp/in: /oneOf/1/oneOf/0
09:16 thats recently pushed probably?
09:17 cait there have been 3 bugs for AV API
09:17 recently pushed (I had listed them in the QA mail)
09:17 marcelr it block the whole api now
09:17 cait I was not involved, so only thought it was good news I could add to the mail ;)
09:18 marcelr i'll try now applying them
09:18 cait do you mean reversing?
09:19 Joubu: is hea still cleaned up regularly? numbers are high! (just a little suspicious :) )
09:19 marcelr you say that there is a fix ?
09:19 cait 21.863 today, 19910 in January
09:19 you asked about somethign recently pushed - i said 3 bugs were pushed about AV
09:19 because i had listed them in the email
09:19 I am not aware of a fix
09:19 marcelr ok misunderstanding
09:20 i hoped for a pending fix
09:22 cait yeah sorry :(
09:24 marcelr well, just commenting out authorised_value_id_pp in the swagger.yml is quick and dirty fix to overcome the complete block
09:24 the author of authorised values should fix the rest
09:35 cait maybe put a note on one of the bugs?
09:40 PedroAmorim[m] good morning
09:40 tcohen: I believe add_pagination_headers are required on the mojo controller because it adds X-Total-Count header, and that is required for count method (see http-client.js)
09:46 cait 2 more demos on 22.11 now :) https://koha-community.org/demo/
09:47 ashimema we keep threatening to add one too
09:47 hmm
09:48 I don't see schema validation issues for av marcelr..
09:48 I do for erm eholdings though.
09:48 marcelr on top of bug 32983 ?
09:48 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32983 enhancement, P5 - low, ---, jonathan.druart+koha, Pushed to stable , Use REST API route to retrieve authorised values
09:49 marcelr there is something missing or wrong with authorised_value_id_pp
09:49 ashimema wow.. those spec are a bit odd
09:50 hmm
09:50 I just pulled master and validated the schema
09:51 marcelr could it also depend on Mojo versions and such ?
09:55 bug 32981 is causing trouble already
09:55 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32981 enhancement, P5 - low, ---, jonathan.druart+koha, Pushed to stable , Add REST API endpoint to list authorised values for a given category
09:55 marcelr the api is ok when you go before the commit of bug 32981
10:06 ashimema bug 33145
10:06 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=33145 normal, P5 - low, ---, martin.renvoize, Needs Signoff , Invalid specification for ERM routes
10:06 ashimema to fix the erm stuff
10:06 weird that there's issues in 32981
10:06 I can have a little look
10:07 are you on a dev install or is this in ktd?
10:07 marcelr dev install
10:08 ashimema ok
10:08 I don't have one of them anymore.. I was a long standing holdout using both ktd and a traditional dev setup...
10:08 but I moved local dev and went ktd only recently
10:08 let me have a little play
10:10 TriveniChandriki[m] marcelr: Dev installation any problem is there
10:11 ashimema did oyu say all api endpoints were 404 marcelr?
10:11 I can't replicate that here
10:11 marcelr There is obviously something wrong or missing in the AV stuff for API. Yes, i got 404s all over the place.
10:11 ashimema I get none
10:11 freshly checked out master
10:11 marcelr do you see errros in plack-api
10:12 i guess no problems in KTD will not prevent lots of trouble in production
10:12 cait coudl it be dependencies after all?
10:12 marcelr yeah might be
10:13 i very recently did an upgrade on my container
10:14 ashimema JSON::Validator 5.08 and Mojolicous::Plugin::OpenAPI 5.05 are the current recommended versions
10:14 may be worth a quick check of those.
10:14 marcelr i do
10:15 JSON::Validator (5.08 )
10:15 * ashimema is confused by the versions table on the about page now...
10:15 marcelr Mojolicious::Plugin::OpenAPI (5.05 )
10:15 these are fine
10:15 ashimema what's the in brackets vs own column version numbers mean?
10:15 humm.. that's very strange then
10:15 and you say commenting that one API route fixes it for you..
10:16 marcelr What about: Mojolicious (8.12 ) 9.22
10:16 ashimema so it's not an apache directive gone array
10:16 * ashimema had all sorts of apache issues with dev installs and the API..
10:16 ashimema yeah, that's the same as I have for Mojolicious
10:16 marcelr yeah i checked one apache rewrite rule too in apached-shared-intranet
10:17 ashimema Very odd indeed
10:17 marcelr but since the issue is cleared when reverting, thats not the problem
10:17 ashimema Worst kind of bugs heisenbugs
10:18 Yeah, i agree
10:18 marcelr and the Mojolicious version ?
10:19 libmojolicious-perl/now 9.90+really9.22-koha1 all [installed,local]
10:19 libmojolicious-plugin-openapi-perl/now 5.99+really5.05-koha1 all [installed,local]
10:19 hmm'
10:20 5.99 really 5.05
10:21 i would say we have a configuration issue in the swagger files that is exposed with some higher dependency
10:32 ashimema we have the same versions of everything you've listed so far ☹️
10:33 marcelr well the best road now is fix the AV stuff recently pushed then
10:36 let me try bug 33145 then
10:36 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=33145 normal, P5 - low, ---, martin.renvoize, Needs Signoff , Invalid specification for ERM routes
10:37 marcelr yes ashimema thats the fix i want !
10:40 I dont understand why it fixes a problem with the authorised_value_id but it seems so
10:41 ashimema ho.. now that is even more odd
10:41 glad it resolves it for you
10:41 wish I knew how/why though.. haha
10:41 I'm actually really surprised the api doesn't appear broken for everyone right now without bug 33145
10:41 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=33145 normal, P5 - low, ---, martin.renvoize, Needs Signoff , Invalid specification for ERM routes
10:42 ashimema I'd have thought that would lead to a crash in ::OpenAPI with invalid spec errors as a whole
10:42 but it appears we don't validate the spec as a whole somehow
10:43 marcelr yes confusing but something like the wrong error message here ?
10:44 but they are linked: Bug 32983: ERM - Retrieve AVs from an endpoint
10:44 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32983 enhancement, P5 - low, ---, jonathan.druart+koha, Pushed to stable , Use REST API route to retrieve authorised values
10:46 marcelr weird
10:48 ashimema do you have any plugins installed marcelr?
10:49 looking at the code we actually disable validation on the first try to load the schema...
10:49 marcelr what kind of plugin
10:51 ashimema koha plugin
10:51 marcelr no nothing
10:51 ashimema any that may include some api routes
10:51 doh.. OK
10:51 that's a red herring too then
11:22 tcohen hole #koha
11:23 ashimema: me skip validation of the spec on 'each' plugin on a first run. If it fails we load them one by one again to catch the faulty one
11:24 this is because, most of the time, plugins don't break the API and the penatly for checking one by one is not acceptable. If it fails, checking will be done and slowly
11:24 heh
11:28 ashimema makes sense
11:29 any idea how our faulty spec for the erm stuff got through?
11:41 oleonard joined #koha
11:48 tcohen haven't seen the bug yet
11:52 liliputech got a red build on Ubuntu22... somehow installation fails, is there a way I can reproduce this scenario (install koha on U22) with ktd to get a better chance of fixing things? (maybe package missing or out-of-date on U22)
11:52 (speaking for 21.11, about to release 21.11.18)
11:54 tcohen ashimema, marcelr: this was clearly working
11:54 and the changelog on api/ doesn't tell much
12:23 marcelr joined #koha
12:23 marcelr back
12:28 alex_ joined #koha
12:31 marcelr tcohen: now i checked without 33145 again, restarted and i dont see 404s as api response, but this is very weird
12:33 tcohen the erm API has a problem
12:33 I don't reproduce the issue
12:35 I mean, I'm hitting endpoints using Postman without errors
12:35 I'm glad this bug showed up anyway
12:35 because it might be the reason I feel the API docs haven't updated correctly recently
12:35 ashimema Agreed
12:36 tcohen QAing the patches
12:36 ashimema I couldn't replicate 404's
12:36 tcohen for pushing
12:36 and we need to add an extra test for the spec
12:36 ashimema I spotted the spec issue using swagger-cli
12:36 tcohen using the swagger-cli tool
12:37 ashimema But yeah, I couldn't work out why a) things just work for us but failed weirdly for marcel
12:37 tcohen Joubu: maybe it is time to run the test suite, and then separate tests before killing the container
12:37 so
12:37 ashimema Or b) why the internal validation doesn't catch a bad spec at start uo
12:37 Agreed
12:38 tcohen launch ktd -> docker run misc4dev/run_tests.pl -> docker run cypress -> docker run swagger-cli
12:38 -> kill ktd
12:42 thank you marcelr and ashimema for raising this and providing a fix so quickly
12:42 I guess it is time to start pushing features
12:42 :-D
12:43 hammat joined #koha
12:57 lds joined #koha
12:59 ashimema D
13:01 oleonard Thanks for the weekend QA cait
13:01 cait you are welcome
13:01 the remaining patches have my SO lines...
13:07 oleonard Only about 25 breadcrumb wrapper patches to be written XD
13:08 tcohen XDXD
13:13 lukeg joined #koha
13:24 mtj hi team
13:25 * mtj reads back on the api/swagger chat
13:28 mtj ooh, i like the swagger-cli idea, for testing
13:28 ashimema me too
13:28 don't know if I'll get a moment to code it up though
13:28 tcohen mtj: marcelr had a momentary general API failure, and in the meantime a real bug was found. Which is not breaking the API it seems
13:30 mtj nice
13:34 liliputech joined #koha
13:59 marcelr tcohen: any problems with 32472 ? for master this is a no-op actually
13:59 fixing a POD line
13:59 for 21.11 it is interesting
14:00 tcohen ok
14:00 wasn't aware
14:00 marcelr np
14:00 tcohen the queue is really high for pushes
14:01 marcelr if SO gets lower, PQA get higher..
14:01 tcohen those 21.11 backports got dangerous with the 'scalar context only' move
14:02 marcelr this one isnt
14:02 it is before that move
14:02 22.05 and 22.11 have the scalar context
14:03 tcohen my interpretation is 21.11 got a straight backport
14:03 marcelr sure?
14:03 tcohen looking at it I assumed it is related to bug 32472
14:03 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32472 major, P5 - low, ---, m.de.rooy, Passed QA , [21.11 CRASH] The method Koha::Item->count is not covered by tests
14:03 tcohen wasn't aware
14:03 oh
14:04 funky pastes on ubuntu
14:04 bug 29844
14:04 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=29844 enhancement, P5 - low, ---, tomascohen, RESOLVED FIXED, Remove uses of wantarray in Koha::Objects
14:04 marcelr sub search {     my ( $self, $params, $attributes ) = @_;      if (wantarray) {
14:04 i still see wantarray
14:04 tcohen right 21.11 didn't get 29844
14:04 marcelr thats the reason for crashing
14:04 tcohen 22.05 did
14:05 so people are (probably) backporting things that 'look correct' but due to bug 29844 actually break
14:05 marcelr yeah that could be very dangerous
14:06 tcohen this is particularly complex to catch if it is happening in the controllers, which don't have automated tests
14:06 I'd say those backportes need a 'backport QA' step
14:06 marcelr my example came even from a module
14:07 but wasnt caught
14:07 just via production logs
14:07 tcohen it means then we don't cover that with tests
14:09 marcelr should we change status now to move it into 21.11 queue ?
14:10 tcohen liliputech: please push 32472 ASAP, thanks
14:19 marcelr bug 32990
14:19 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32990 enhancement, P5 - low, ---, m.de.rooy, Signed Off , Possible deadlock in C4::ImportBatch::_update_batch_record_counts
14:19 marcelr wonder if we should change enh here though
14:19 but it doesnt matter anyway
14:20 no backport
14:24 lukeg cait about?
14:33 liliputech tcohen: working on it
14:40 oleonard ashimema around?
14:45 ashimema hi
14:46 oleonard Sorry I think I just answered my own question!
14:47 breadcrumb wrapper patches should be 'minor' instead of 'enh' because backporting them helps *other* patches get backported, not because an eventual Bootstrap upgrade will be backported
14:59 ashimema correct 🙂
15:00 they don't really lead to a noticeable change to the end user (yet), but they do clean up templates so we felt for long term maintenance (especially as in theory 22.11 is the LTS) they were sensible to backport.
15:01 cait lukeg: on the phone - in a little bit
15:01 ashimema sometimes I find our bz terms a bit hard to work with.. really they're an 'architectural enhancement' I would say.. but in reality people reading the maintenance release notes don't want to see many enhancements.. they just want clear bugfixes.
15:02 I wouldn't expect them to get backported any further than the current stable, 22.11
15:10 caroline_catlady hi everyone! I hope Monday has been treating you all well so far!
15:17 oleonard Hi caroline. If I may, I would like to promote Bug 33077
15:17 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=33077 enhancement, P5 - low, ---, oleonard, Needs Signoff , Improve ease of translating template title tags
15:17 caroline oleonard, no prob, I'll take a look at it
15:33 caroline joined #koha
15:34 JesseM joined #koha
16:33 oleonard jajm around?
16:33 ashimema "It's another manic Monday.... I wish it could be Sunday..."
16:34 cait ashimema: thief :)
16:35 thibaud_glt joined #koha
16:35 ashimema lol
16:35 😄
16:37 kidclamp joined #koha
16:37 * kidclamp waves
16:39 * cait waves back
16:40 * tcohen waves
16:40 * cait sang it first
16:40 kidclamp it's not a competition :-)
16:41 ashimema 😄
16:41 cait Koha's next superstar?
16:42 * ashimema is starting to wish he'd not spent his weekend doing safety checks at the high ropes facility he volunteers at this weekend...
16:42 can barely lift his arms to the keyboard
16:43 ashimema hanging upside down at 40ft up torque wrenching up cable clamps is hard work
16:43 tcohen we should do karaoke in MArseille
16:44 cait I've never done Karaoke... I am in
16:45 caroline I love karaoke! And I suck at signing, so it's great for other people to feel good about themselves XD
16:45 oleonard caroline are you going to Hackfest?
16:46 caroline no... :(
16:46 oleonard Me neither :(
16:46 Saving up for Finland
16:47 caroline will you be there? awesome!
16:48 * tcohen is planning on going there too
16:49 caroline yay!
16:55 oleonard, regarding 33077, is the last one supposed to be
16:55 msgid "%s %s Confirm deletion of city %s %s %s Cities %s %s %s %s "
16:56 oleonard No, I'll take a look
16:56 emlam joined #koha
16:58 cait hi emlam :)
16:58 emlam hi cait!
17:01 cait don't think we don't see you :)
17:03 emlam haha, I usually keep an IRC window open...sometimes I have something to say, sometimes it's just useful to learn from what others are chatting about :)
17:08 tcohen oh, I ctrl+c too late
17:08 pushed without signing....
17:08 so no one will be sure it was me
17:08 oleonard caroline: I don't get the same result!
17:09 caroline :/
17:10 Ah, I think I found it... I was looking in the wrong file it's moved from staff-prog to messages...
17:11 oleonard Oh yeah that breadcrumb wrapper markup is waiting for an update from another patch. That'll fix the mess you refer to above
17:12 tcohen oleonard: I'm confused by 'Part 3' and such
17:12 they don't seem to require a particular order in which they are applied, right?
17:12 oleonard tcohen: They're separate and not dependent batches
17:13 caroline ok that's the breadcrumb, I hadn't realized... The line numbers didn't match, so I kind of knew I was lost. Testing for real
17:13 oleonard tcohen: I just didn't want to submit a giant patch
17:13 caroline: Sorry, I'm not sure why some of the line numbers are wrong
17:14 caroline no they're not wrong, it's just that in the staff-prog, I think it's only the headers now
17:14 I will get there!
17:14 tcohen what is the right terminology
17:15 Authorities? Authority records?
17:15 caroline tcohen, accoring to terminology page, it's authorities
17:15 authority record
17:15 sorry
17:15 Use "authority record", not "authority".
17:15 https://wiki.koha-community.org/wiki/Terminology
17:16 oleonard But we refer to the "module" as "Authorities"
17:16 tcohen this is for the API, so maybe not Koha-centric audience
17:16 the API docs
17:16 will need a 'section'
17:16 for authority record-related endpoints
17:16 should it be Authorities or Authority records
17:17 the same goes for biblios
17:17 caroline In that case, both are good I think? https://authorities.loc.gov/
17:17 LOC calls them "Authorities"
17:17 tcohen https://api.koha-community.org/
17:17 our API docs already call them biblios, so changing that would be a separate bug, if ever required
17:17 and if Authorities is correct, nothing to be done then
17:18 we will now have CRUD for authorities
17:18 \o/
17:18 caroline with the indexes its biblios and authorities so I think that should work for API too
17:27 cait emlam: but also seeing your sign-offs - emlam++
17:27 bye #koha :)
17:27 cait left #koha
17:34 caroline oleonard, it works super well when you're not a confused mess like me :)
17:34 Don't hesitate to CC me on any translation bugs
17:35 oleonard I really appreciate your time looking at it! I hope that method ends up being useful everywhere
17:40 Dyrcona joined #koha
19:34 magnuse_ joined #koha
19:37 hammat_ joined #koha
19:40 hammat joined #koha
19:40 MoOmar[m] joined #koha
19:40 jeffhub[m] joined #koha
19:40 TriveniChandriki[m] joined #koha
19:40 magnuse joined #koha
19:40 wrmslibrarian[m] joined #koha
19:40 ShopGeek[m] joined #koha
19:40 eric1122[m] joined #koha
19:55 cait joined #koha
19:59 cait joined #koha
20:10 MoOmar[m] joined #koha
20:10 jeffhub[m] joined #koha
20:10 TriveniChandriki[m] joined #koha
20:10 wrmslibrarian[m] joined #koha
20:10 ShopGeek[m] joined #koha
20:10 eric1122[m] joined #koha
20:48 cait emlam: still around?
20:48 emlam yes - what's up?
20:48 cait just wondering: are you testing using usandboxes?
20:48 or a dev env?
20:48 emlam KTD
20:49 cait ah ok
20:49 emlam why?
20:49 cait and i think you are adding your sign-off line in the git bz attach step maybe?
20:49 i noticed that you add comments with your sign-off line, bu tdon't actually add it to the patches
20:49 so something off in the process
20:50 when you are using ktd, the process usually is:
20:51 git bz apply bug
20:51 test test test...
20:51 git so <number of patches>
20:51 git bz attach -e <bug number> HEAD~<number of patches>..
20:54 the last reuploads the patch files with your added sign-off line
20:56 does that make sense?
20:58 emlam Thanks! I haven't been using git so; I've been using git commit --amend -s ...which I now see is only supposed to work for bugs with one patch
20:58 cait ah yes, that only adds it to the last patch
20:58 emlam I'll make sure to use so next time :)
20:58 cait make sure to also use the git bz attach
20:58 maybe you did a git bz edit?
20:59 when you are done, the uploaded file shoudl show your name under attachments in bugzilla
21:00 emlam I...think I've always done a git bz attach. But I don't think I've always used the correct options for multiple patches on that, either
21:00 cait hm, not sure then, but I am sure we can figure it out
21:02 emlam thanks!
21:03 caroline cait++ #solving problems left and right
21:03 cait sure! happy to help if you get there next time (I am around a lot)
21:03 emlam cait++
21:03 cait caroline: and never catching up with all of them!
21:04 davidnind emlam: here are my notes for signing off using KTD https://gitlab.com/-/snippets/[…]-sign-off-a-patch
21:04 cait davidnind++
21:05 emlam whoa, that's awesome!
21:05 davidnind++
21:05 should I redo the signoffs on any bugs that I signed off on previously that haven't moved on yet?
21:07 cait you could try on one if you want
21:07 if I do QA on them, I can also fix it, but it might make it a little easier on the team if you could help
21:12 also making sure you get the right number of patches in the release notes :)
21:12 you've been testing a lot!
21:14 emlam hmm...turns out I'm not seeing any that I signed off on that haven't already been QA'd or pushed to master. I would imagine redoing those now would make things messier?
21:14 cait yeha, the ones already pushed I woudl not touch
21:15 i thought I saw one (just reading through today's bug changes) one sec
21:15 emlam there's 30928, but it needs to be rebased
21:17 cait the search feature in my mail client is failing me
21:20 bug 32964
21:20 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32964 enhancement, P5 - low, ---, koha-bugs, Signed Off , OPACResultsMaxItemsUnavailable description is misleading
21:20 cait bug 32970
21:20 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32970 enhancement, P5 - low, ---, koha-bugs, Signed Off , Allow export of batch item modification results in background jobs
21:22 emlam Those ones aren't me - I'm Emily Lamancusa
21:22 cait bug 33150
21:22 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=33150 trivial, P5 - low, ---, koha-bugs, Signed Off , Add specific message for renewal too_soon situation
21:22 cait ooooh!
21:22 I am so sorry
21:22 you are right
21:23 emlam haha no worries! I am not in fact certain that I've been doing multi-patch signoffs correctly either, so still useful feedback!
21:23 cait might be a little too late here
21:23 I am so sorry
21:24 emlam it's fine! :)
21:26 cait ... now that this has been cleared up - where are you from emily?
21:26 emlam MCPL in Maryland, US
21:26 and you're from Germany, right?
21:27 cait nice :)
21:27 yes
21:30 Konstanz, right at the Swiss border and with a nice lake :)
21:30 emlam ooh, nice :)
21:31 cait yes, but can't wait for spring, we get a lot of greyness in winter
21:32 bag joined #koha
21:32 caroline oh emlam did your colleague do a presentation at kohacon?
21:33 there was someone from maryland (I think MCPL too) who did a presentation on data visualisation, it was super interesting
21:33 emlam I think so! That was before I started (I've only been working here since October), but that sounds right
21:34 caroline it was in september, so yes, before your time :)
21:34 cait it's always great to have more librarians/people working in libraries join us here
21:35 emlam :)
21:38 cait time to leave for me - good night all :)
21:39 emlam good night!
22:25 cait joined #koha
22:33 cait joined #koha
22:37 ShopGeek[m] joined #koha
22:42 eric1122[m] joined #koha
22:42 TriveniChandriki[m] joined #koha
22:42 wrmslibrarian[m] joined #koha
22:43 MoOmar[m] joined #koha
22:46 jeffhub[m] joined #koha
23:04 tuxayo Hi :)
23:04 Does anyone know what to run after making/appling changes to vue pages? Like when testing bug 32924 for example. yarn build and restart_all and even reset_all didn't do the trick for me. I had to restart my whole ktd to see the new checkbox.
23:04 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=32924 enhancement, P5 - low, ---, jonathan.druart+koha, Passed QA , Filter agreements by logged in librarian

← Previous day | Today | Next day → | Search | Index

koha1