Time Nick Message
00:03 tcohen[m] The forjego demo server had configured 'postgres' as the DB image, and they've just jumped from v15 to v16 as default
00:03 tcohen[m] it was not starting correctly
00:03 tcohen[m] just in case anyone is using posgres in docker just beware you need to specify the version to avoid headaches
00:04 tcohen[m] ashimema++# taking care of broken tests
06:38 Joubu @later tell marcelr can I get your opinion on bug 6419 please?
06:38 huginn Joubu: The operation succeeded.
06:56 ashimema bug 6419
06:56 huginn 04Bug https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=6419 enhancement, P5 - low, ---, martin.renvoize, Passed QA , Add customizable areas to intranet home pages
07:59 fridolin Bug 34836 kaboum
07:59 huginn 04Bug https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=34836 critical, P5 - low, ---, oleonard, NEW , OPAC ISBD or MARC view blows up with error 500
08:03 Joubu @later tell tcohen opened bug 34837
08:03 huginn Joubu: The operation succeeded.
08:05 fridolin Seen your messages Joubu thanks
08:05 fridolin i'm not always online so dont hesitate to use email
08:05 Joubu fridolin: you should create a bugzilla search with the 23.05 candidate keyword
08:06 Joubu we are using it to flag bugs to backport
08:26 cait3 yes, that's a helpful one - do we have that in the wiki docs yet? for the rmaints?
08:27 DIGITALBITS[m] <DIGITALBITS[m]> "Indeed" <- LTS Release Tables: I think that the 3rd line of the table should be Nov 2023 and not Nov 2024. Am I wrong?
08:27 fridolin yep noted
08:27 cait3 DIGITALBITS[m]: checking
08:28 cait3 you are rigt, ther eis a mistake, ugh
08:28 cait3 fixed, thx
08:29 DIGITALBITS[m] cait3: Thanks and sorryπ. The LTS matter still cooking in my head. We will have a discussion today about it and am taking my notes.
08:30 * ashimema doesn't really like our notion of LTS at all.. but I've expressed that before
08:31 ashimema the key thing is 'stable' = feature stable NOT bug free.. LTS just means we'll keep supporting it for a bit longer.. if you want 'bug free' you're best option is the oldest of the 'old*stable' options that's available at any given point.
08:33 ashimema and to actually track that.. you'll get the new features more regularly (i.e. every 6 months or so).. though you'll get them at least 18 months after people following branches above you.
08:33 ashimema so you'll get 18 month old, tried and tested features every 6 months
08:34 DIGITALBITS[m] ashimema: > * <@ashimema:matrix.org> doesn't really like our notion of LTS at all.. but I've expressed that before
08:34 DIGITALBITS[m] I think that the time frame is to long considering that every year two versions are launched. If we decide to use LTS we will be using 22.11 that is one year old, and within one year we will have a new LTS.
08:34 ashimema so.. right now.. the most bug free release will be 21.05.. but you'll need to migrate to 21.11 in November as 21.05 goes out of support
08:35 DIGITALBITS[m] Than you have also the Operating System. We run Ubuntu (not LTS for our Koha Lab).
08:35 ashimema I really really wouldn't jump to the LTS as soon as it's released
08:35 ashimema I'd wait at least a year
08:35 ashimema for 12 months of bug fixes
08:36 ashimema a brand new LTS will have waaaay more bugs than a shorter lived oldoldstable
08:36 DIGITALBITS[m] ashimema: Thanks for the advise.
08:36 ashimema or... get a support company to take that risk for you π
08:37 ashimema full disclosure, I work for one of the big support companies π
08:37 ashimema our scheme is one upgrade a year and .11 upgrades only.. usually hitting them at the .06 (with a bunch of people testing early to try and help us wean out the obvious bugs)
08:37 DIGITALBITS[m] ashimema: Than we will have to make a choice... buy the server or pay for the support.
08:38 ashimema yeah, it's always a weighing game.
08:38 ashimema gotta work out which things are most important to you and where your particular risk profile fits
08:39 DIGITALBITS[m] I am particularly very happy with 23.05. Our librarians will start testing next week.
08:44 ashimema π
08:44 ashimema glad you like it
08:45 cait3 the idea was to offer longer support, but your points about most bug-free etc. are explained on the wiki page
08:45 cait3 we are struggling with one update a year
08:46 cait so I'll be gald to have a onger time frame - and we made the LTS overlap 1.5 years to give that time to switch
08:47 ashimema So.. best case scenario I suppose is.. switch from LTS to LTS every 1.5 years half way through the new LTS cycle.
08:57 DIGITALBITS[m] I have another concern. I was reading this article https://wiki.koha-community.org/wiki/System_requirements_and_recommendations
08:57 DIGITALBITS[m] and seems that MariaDB will become the default database engine. Should we in already consider native mySQL.
08:58 ashimema I think we should update our versions table to look more like the ubuntu one's on wikipedia
08:58 ashimema https://en.wikipedia.org/wiki/Ubuntu_version_history
08:59 ashimema bottom of the page.. I like the timeline view too
08:59 ashimema it's very clear then what things mean
08:59 DIGITALBITS[m] Another concern is for newbies (like me) advised to run Elasticsearch instead of Zebra.
09:01 DIGITALBITS[m] ashimema: I was looking to this table yesterday and indeed is bit clear.
09:01 ashimema personally, I think for the ES vs Zebra debate should be much more about specs of the system you want to run and the search features you feel you need.
09:01 ashimema both engines have their benefits and drawbacks
09:02 * ashimema has a soft spot for Zebra still
09:15 Joubu fridolin: 33606 is a major one, it's waiting for 2 months already
09:21 cait Joubu: can you email?
09:21 ashimema bug 33606
09:21 huginn 04Bug https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=33606 major, P5 - low, ---, jonathan.druart+koha, Pushed to master , Access to ERM requires parameters => 'manage_sysprefs'
09:21 cait DIGITALBITS[m]: we use Koha with Debian and always use the DBMS Debian comes with - MariaDB works really well
09:22 cait I think most are probably on MariaDB these days, but I can't be sure as we don't have stats
09:22 cait both are supported I believe
09:23 cait hm ok, I see it's nt listed anymore... maybe because we are not actively testing it, although I see a mysql8 nodeon the dashboard?
09:24 Joubu nope, ping on multi channels is too much energy for me...
09:26 cait someoen remember if we deprecated MySQL formally? The installation guides jsut have a recommendation. https://wiki.koha-community.org/wiki/Koha_on_Debian#MySQL_and_MariaDB
09:28 DIGITALBITS[m] cait: Indeed. And it raises some questions and concerns about implementing the solution with the supported packages or that some packages are discontinued in short time.
09:29 Joubu cait: we are still kind of supporting it. We have jenkins running the test suite on MySQL 8, and lot of users report bugs on My8.
09:31 DIGITALBITS[m] Joubu: Ok. Out DB Team wont be happy to hear this.
09:33 DIGITALBITS[m] In fact I remember in meeting that they asked us about mySQL and MariaDB
09:34 Joubu DIGITALBITS[m]: Why not happy? I am saying we are still supporting it :)
09:35 Joubu when I am saying "report bugs on My8", I mean that we are taking them into account and fix them, instead of ignoring them
09:51 Joubu @later tell mtj do you know why ktd build is failing for trixie?
09:51 huginn Joubu: The operation succeeded.
10:00 Joubu paulderscheid[m]: https://koha-community-org.vercel.app - I am not seeing the koha logo (top left) when opening in a new private window (tried to disable the dark reader FF add-on)
10:01 gk_ Koha 23.05 Package Installation on Ubuntu 22.04; Update fails with error: Upgrade to 23.05.00.002 [15:01:45]: Bug 21983 - Deleted biblio handling on ILL ERROR: {UNKNOWN}: DBI Exception: DBD::mysql::db do failed: Duplicate key name 'illrequests_bibfk' at /usr/share/koha/lib/C4/Installer.pm line 741
10:01 huginn 04Bug https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=21983 blocker, P5 - low, ---, tomascohen, Pushed to oldstable , Better handling of deleted biblios on ILL requests
10:04 Joubu gk_: see
10:04 Joubu bug 34276
10:04 huginn 04Bug https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=34276 blocker, P5 - low, ---, jonathan.druart+koha, Pushed to oldstable , upgrading 23.05 to 23.05.002
10:24 oleonard Hi #koha
10:26 ashimema morning oleonard
10:26 ashimema I just stuck your name on a fun little bug
10:26 ashimema well.. enhancement
10:36 paulderscheid[m] joubu, I'll look into it. Most likely because dark theme is presumed to be the default. Need to change that.
10:41 cait got to say... first impression is nice
10:41 cait although i got immedaitely lost in the labyrinth of lost books :D
10:42 ashimema lost books?
10:43 Joubu 404
10:43 ashimema lol
10:44 cait https://koha-community-org.vercel.app/support/paid-support/continent/
10:51 paulderscheid[m] Always a reminder that the images need to be optimised .. a lot :D
10:51 cait i thought the lost books were not a bad fit :)
10:51 cait but better get rid of the 404 of course
10:53 paulderscheid[m] The idea came from a language model. Prompted for "funny 404 messages in the context of libraries"
10:53 paulderscheid[m] Too lazy to be original ^^
11:16 tcohen[m] Good morning
11:24 cait hi tcohen
11:34 tcohen[m] hi cait
12:18 oleonard If something is pushed to 21.11 is that the end of the line? Can be marked fixed?
12:18 ashimema or 'needs documentation'
12:19 ashimema the rmaint should be setting 'resolved fixed' or 'needs documentation' depending on what it is.. in most cases if it's made it that far through it is most likely a 'resolved fixed'
12:19 ashimema but.. historically.. Katrin has done a tidy up each cycle as it's often just been left as 'pushed to old*stable'
12:22 oleonard I keep an eye on my bugs for no good reason, just curiosity, and when I see something from, say May 2023 I wonder if I should just call it fixed.
12:30 tcohen[m] We won't revert things for the lack of documentation, but we need patch authors or sponsors to document the features they introduce
12:30 tcohen[m] I'm not sure who is looking after needs_documentation bugs
12:31 ashimema they're listed on the dashboard to do list
12:32 ashimema the docs team picks them off
12:32 ashimema and devs should be trying to too
12:32 ashimema when switching something away from needs documentation you get recognised on the dashboard too
12:33 ashimema anyone can make the decision that a needs documentation actually is already documented well enough and call it resolved fixed
12:33 ashimema that's the current workflow as I understand it.
12:34 Joubu you should require the documentation (release note text) before the push, it's more effective than expect people to follow-up once their stuff is pushed..
12:36 ashimema documentation is not release note text
12:36 ashimema documentation means a submission to the manual
12:37 ashimema but yes.. I'd also love to see release note text more strict
12:38 Joubu Hence the parenthesis ;) That's the same but it will be nice to always have them. If there is nothing in the release note text it should mean there is nothing to document.
12:38 Joubu that's *not* the same*
12:47 paulderscheid[m] Hi guys, wasn't there a page to generate API keys for client credentials somewhere in the UI or have I dreamed that?
12:48 ashimema it's under the user you wish to give access to
12:49 paulderscheid[m] Aha!
12:50 ashimema it's a bit funky π
12:50 paulderscheid[m] It makes sense for the flow implemented.
12:51 ashimema mm
12:52 * ashimema goes on a tab killing spree
12:56 Joubu (force-pushing misc4dev, sorry about that!)
13:14 tcohen[m] Joubu: thank you
13:14 tcohen[m] Joubu++
13:27 caroline Does someone know how to add translation context in the new _("Order") format?
13:28 caroline I usually just replace the word by [% tp('noun', 'Order') | html %]
13:28 cait is this the JS one?
13:28 caroline can I do the same with _("abc") ?
13:28 caroline I think it's JS, but I can't be sure not knowing the language
13:29 oleonard The underscore is always for JS
13:29 oleonard But I don't know the answer to your question
13:31 Joubu caroline: __p()
13:31 Joubu see koha-tmpl/opac-tmpl/bootstrap/js/basket.js: showCartUpdate( __p("Bibliographic record", "The item has been added to your cart") );
13:31 Joubu other functions are defined in koha-tmpl/intranet-tmpl/js/i18n.js
13:31 caroline cool thanks Joubu++ !
13:31 caroline So it's two underscores?
13:31 Joubu yes
13:32 oleonard caroline: Two underscores if you're working with a .js file. One underscore if you're in a template
13:32 caroline oof, ok... I think this is a template file
13:33 caroline result += '<li><a href="/cgi-bin/koha/acqui/showorder.pl?ordernumber=' + encodeURIComponent(row.order_id) + '" class="previewData">' + _p("noun","Order") + '</a></li>';
13:33 caroline this is good?
13:35 caroline ok yes I think it worked. Thanks Joubu++ and oleonard++
13:36 caroline oh yes, another question... in an HTML attribute?
13:36 caroline <a href="/cgi-bin/koha/acqui/neworderempty.pl?frameworkcode=[% frameworkcode | uri %]&breedingid=[% breeding_loo.breedingid | uri %]&booksellerid=[% booksellerid | uri %]&basketno=[% basketno | uri %]" class="chosen" title="Order" data-action="order"><i class="fa fa-plus"></i> [% tp('verb', 'Order') | html %]</a>
13:37 caroline the title="Order"
13:38 Joubu caroline: you need 2 underscores here.
13:38 Joubu _p() does not exists
13:38 Joubu _() exists, __() exists, then the other i18n function defined in js/i18n.js are all prefixed with 2 underscores
13:42 oleonard I supposed we could use [% tp() %] instead of _() in inline JS?
13:43 caroline so, if I understand correctly, if it's a .js file = 1 underscore; if it's a .tt file and it's in a JS block = 1 underscore; if it's in a .tt file inside an HTML tag as an attribute= 2 underscores
13:43 oleonard If it's a js file = 2 underscores
13:44 caroline I'm writing this in my cheatsheet...
13:46 caroline so the only expection is if it's in a JS bloc in a .tt file
13:46 caroline exception
13:47 caroline hm, I don't think it worked
13:48 * oleonard is doing some tests
13:49 caroline argh, if I put one underscore, the page doesn't load, if I put 2 it doesn't translate
13:50 cait did you break it up?
13:50 caroline I definitely broke it, but how do you mean?
13:51 cait i meant the oexamle with the + ... + ... but I have not beenf ollowing closely, better listen to oleonard
13:52 pastebot "oleonard" at 127.0.0.1 pasted "caroline: I think if you need context in a JS file you can use this kind of construction" (1 line) at https://paste.koha-community.org/26445
13:53 oleonard That seems to work for me functionally and for translation
13:53 oleonard It would be great to hear from jajm whether there is any reason not to do that
13:55 caroline ok yes this works for me too!
13:55 jajm oleonard: not to do what ?
13:55 oleonard use [% tp() %] instead of _() in inline JS?
13:57 oleonard So that we can give translators context, etc, which aren't possible with _()
13:57 jajm there's always the risk that the translation breaks JS syntax, but if escaped properly it should be ok (do we have a TT filter for that ?)
13:57 cait jajm: is matts around maybe?
13:58 oleonard jajm: Hm... true, but isn't that true with either method?
13:58 jajm we can use __p instead (a JS function)
13:58 oleonard jajm: In an template?
13:58 caroline I tried __p() and it didn't translate
13:59 jajm oleonard: iirc this should work yes
13:59 Joubu it's in messages-js.po
13:59 caroline The translation tool picked it up because it was in the .po file, but when I "installed" the translation it was not translated in the page
14:00 jajm caroline: did you check in messages-js.po ?
14:00 caroline i checked in message.po
14:00 caroline let me try again
14:01 matts cait, hi, around indeed
14:01 oleonard Why are the po files separate?
14:01 cait hey
14:02 jajm oleonard: so that the JS file containing translations (and being downloaded by the browser) contains only strings used in JS (so it's not too big)
14:02 caroline oleonard, no idea... everything is slowly migrating to messages.po
14:02 cait matts: pm
14:03 Joubu messages.po != messages-js.po
14:04 caroline I know, but staff-prog and opac-bootstrap is losing to messages.po (not message-js.po). I don't know why we have the other ones still if every new thing is going in messages.po
14:04 Joubu messages-js.po contains the strings sent to the client. They are coming from the i18n.js functions
14:05 Joubu messages.po are the strings that are not replaced using the `translate install` process. The strings won't appear in the localized translate. There are coming from the pl and TT methods [% p() %] and those in the I18N.pm perl module
14:07 oleonard So *-messages.po has strings that are processed as the template is processed
14:07 Joubu the strings retrieved automatically by the script will still be put in staff-prog.po: inside the HTML tags
14:07 Joubu yes, oleonard
14:08 oleonard That makes sense. It's too bad we can't have some kind of meta-po file which includes all strings so that translators never have to translate something twice.
14:08 Joubu hence the perf discussion we had on the "wrap all the strings inside TT tags"
14:08 Joubu bug
14:08 Joubu maybe with the move to Weblate? :)
14:09 caroline it might be logical for developers, but for translators (or maybe just me, I don't want to speak for everyone else) messages.po containing both opac and intranet stuff doens't make sense
14:09 Joubu well, using the weblate's UI ofc
14:10 cait they are still not that big, but with the files growing it gets more anoying
14:11 Joubu open a bug? :)
14:17 caroline it seems it doesn't work in the HTML attribute
14:17 caroline one underscore https://snipboard.io/pgQcDw.jpg
14:17 caroline two underscores https://snipboard.io/t2ZCTM.jpg
14:18 caroline this is admittedly the less annoying of the ones I found though
14:18 cait two should only be needed when you are in a .js file
14:19 Joubu no, not if you are using __p()
14:19 Joubu it's always 2 underscores!
14:19 cait do we have documentation for all that?
14:20 cait (you knew that was coming... right?)
14:20 Joubu _() has one underscore, that's all, all others have 2
14:20 Joubu yes we have :D
14:20 caroline I saw a bug about one vs two, but I didn't what it was about
14:21 caroline *didn't know
14:24 oleonard I guess this page could be updated to include information about _() and __() https://wiki.koha-community.org/wiki/Internationalization_options_in_templates
14:28 caroline I attached a patch with oleonard's option to bug 34834
14:28 huginn 04Bug https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=34834 enhancement, P5 - low, ---, koha-bugs, NEW , Add translation context for "Order"
14:28 caroline I'll try again with the (I don't remember how many) underscores and messages-js.po after my meeting
14:29 oleonard caroline: Keep adding underscores until it works :D
14:33 cait oleonard++ caroline++ Joubu++ matts++
14:34 oleonard matts did secret helping
14:35 cait he did indeed
15:13 reiveune bye
15:41 caroline can I be in on the secret?
15:52 caroline Back from my meeting. I tried _p("noun","Order") and __p("noun","Order") in parcel.tt without success
15:53 caroline So I kept my patch as is, if someone wants to look bug 34834
15:53 huginn 04Bug https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=34834 enhancement, P5 - low, ---, koha-bugs, Needs Signoff , Add translation context for "Order"
16:55 oleonard-away caroline: Don't forget to run the qa tool on your patches before you submit them
17:57 * oleonard curses having lost his carefully-written test plan
18:03 caroline oh no! :((
18:50 ashimema So, are you requesting we split messages.po and messages-js.po into messages-intranet.po, messages-opac.pl, messages-intranet-js.po and messages-opac-js.po?
18:50 ashimema To split between intra and OPAC again?
18:52 caroline is it not possible to have the strings stay in staff-prog and opac-bootstrap?
18:53 caroline I understand the distinction of js vs non-js, but I'm talking about the non-js ones
18:56 DIGITALBITS[m] Well it looks like MariaDB will become the default database engine, the path indicates that. They are testing an environment with a failover of mySQL. if we want to go live now i bellieve we better start with MariaDB, to avoid the switch in a short time.
18:58 caroline ashimema, actually, I don't mind either way. I just find it confusion that there is a messages.po AND and opac-bootstrap/staff-prog, if that makes sense. Like why are some opac strings in messages.po and some in opac-bootstrap. It doesn't make sense to me
18:58 caroline *confusing
19:00 ashimema That's a technical issue
19:00 oleonard Doesn't it also confuse the statistics about % of OPAC translated?
19:02 caroline I guess it would. I hadn't looked at it that way. I mostly look at % of completion per file, messages.po being smaller than the other files, it's usually pretty easy to keep it at or close to 100%
19:08 Joubu caroline: please open a bug if it's not done yet. It would be interesting to have a look at it when/if we move to weblate.
19:08 caroline ok I will
19:08 caroline I hope the move to weblate is more when than if :)
19:11 ashimema Sorry, was doing a dad taxi run
19:11 caroline :)
19:12 ashimema Where's the performance discussion around wrapping all strings joubu?
19:13 ashimema I thought it would be in bug 20988
19:13 huginn 04Bug https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=20988 enhancement, P5 - low, ---, julian.maurice, BLOCKED , [OMNIBUS] Internationalization: wrap all translatable text inside t() calls
19:13 ashimema But I'm not seeing it
19:14 oleonard-away Maybe bug 26392
19:14 huginn 04Bug https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=26392 enhancement, P5 - low, ---, julian.maurice, Patch doesn't apply , Wrap all translatable strings in opac-main.tt
19:39 Kolvir Is there a way to add the first 3 letters of the author's last name, MARC 100 to the end of 082a using a modification template? I've tried it with regex, but can't get it to work. 'copy and replace 100 a to field 082 a regex s/^\D{3}' but I can't figure out what to put in the replacement box. I either get the whole field or the field without the first 3 letters.
19:43 Kolvir I've got to reboot, if someone has a suggestion, please post it and I'll check the log.
20:05 ashimema Pass
20:05 ashimema Nope, still can't find a discussion around performance of wrapping all strins
20:05 ashimema Strings
20:05 ashimema Lots of debate over syntax to use and tokens and things . But I think those are all mute
20:11 ashimema Found the perf stuff at last.
20:12 ashimema It is back in 20988.. near the bottom
20:12 ashimema Unresolved though
21:43 kivilahtio Hi! I took a look at using Websockets from a Koha-plugin. It worked like spiffy, gliffy and spindizzy when I was testing it with Mojololicious. I deployed a new websocket route from a Koha-plugin's REST endpoint Mojolicious Controller handle, and worked like a charm.
21:43 kivilahtio Also apache2 supports Websocket proxying in the latest version, baked in the mod_proxy_http
21:43 kivilahtio Too bad the plack/PSGI-middleware doesn't.
21:43 kivilahtio A solution would be to run Mojolicious as it's own separate process. I think we ran it like that in Finland when the REST API was being incubated.
21:43 kivilahtio It would be cool to have websockets with Koha, now that the Vue.js has also been introduced to the core.
21:43 kivilahtio We could do some cool stuff with Koha.
23:44 dcook @later tell kivilahtio Sounds good to me
23:44 huginn dcook: The operation succeeded.