IRC log for #koha, 2023-04-14

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

All times shown according to UTC.

Time Nick Message
03:28 bag joined #koha
03:59 aleisha joined #koha
04:12 bag_ joined #koha
06:09 marcelr joined #koha
06:10 marcelr o/
06:18 magnuse \o/
06:26 reiveune joined #koha
06:26 reiveune hello
06:45 alex_ joined #koha
06:51 thibaud_g joined #koha
06:58 lds joined #koha
07:27 paulderscheid[m] Morning #koha
07:29 ashimema please see my last attempt on bug 33490 and the corresponding comment Joubu
07:29 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=33490 normal, P5 - low, ---, jonathan.druart+koha, Signed Off , Agreements - Filter by expired results in error
07:29 ashimema interested in your take
07:34 cait joined #koha
07:34 MarkHofstetter joined #koha
07:46 Joubu ashimema: will have a look soon!
07:46 ashimema Cheers
07:46 It's a tiny patch
07:46 Just had a caveat
07:49 marcelr anyone willing to test bug 33360 ?
07:49 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=33360 enhancement, P5 - low, ---, m.de.rooy, Needs Signoff , SendQueuedMessages: Improve limit behavior and add domain limits
08:11 ashimema so.. is that about the to address marcelr?
08:12 i.e. only sent TO upto 50 gmail addresses in 1 minute?
08:12 that sort of limiting?
08:13 I've not had such issues yet myself.. interested to know the use cases or if I'm reading the code right/wrong
08:31 AnkeB joined #koha
08:49 marcelr ashimema: yes --limit will apply now to the volume sent (max x messages) and the domain limit applies to the number sent in a given period
08:49 i noticed in the past that gmail was deferring mail because of volume
08:50 and recently i saw while using this patch that yahoo was doing it too, since i didnt add a limit for yahoo
08:50 ashimema but it's about the receiver or the sender.. i couldn't quite work that out;.
08:50 interesting
08:51 marcelr it is about the recipient domains
08:51 ashimema so it is about the recieving domain
08:51 coolios..
08:51 that is interesting
08:51 marcelr sending to gmail, which is about 40% of our patrons
08:51 e.g.
08:51 ashimema how did you spot it?
08:51 * ashimema is wondering if we've just been ignorant of it happening here?
08:51 marcelr reading exim logs is a hobby :)
08:51 ashimema 😂
08:52 marcelr no, i had limits too on my container and oom-killer followed them too
08:52 so panic log entries etc
08:52 ashimema very interesting.. thanks
08:52 well.. I have the patch applied here.. will have a little test now
08:52 the code looks pretty readable
08:52 marcelr great
08:52 it runs on my 21.11 already btw
08:53 ashimema yeah, i saw you're note
08:53 I wondered if we aught to add some default settings?
08:53 or at least describe how best to work out what settings are sensible?
08:53 marcelr i think personally that we should add the limit parameter to the cron file at least
08:53 about 100 per run
08:53 and leave domain limits to the sysadmin
08:54 but i didnt dare to enforce them here
08:54 the frequency of 15 mins is a bit outdated too
08:54 i would say 5 mins ?
08:54 ashimema yeah, sounds reasonable
08:54 marcelr running proc msq
08:55 they could go on a separate report
08:58 ashimema yeah..
08:58 though I'd love to see them too
09:00 marcelr np, if this gets further, i add them
09:00 but changing defaults tend to trigger discussion :)
09:00 ashimema this could add a fair bit of DB traffic during the cron run
09:01 did you notice any issues in that regard?
09:01 marcelr i optimized code for it
09:01 we only count if we come across that domain
09:01 and we only look for a patron if needed
09:02 not calculating to_address over and over again
09:02 and if you just not add them, there is no change
09:03 ashimema coolios
09:03 marcelr but the advantage is really not to flood some mail servers triggering deferrals, frozen messages etc
09:03 ashimema good to see you've thought about it from that perspective
09:04 AnkeB joined #koha
09:04 magnuse marcelr++
09:06 marcelr: do you think 33360 will solve bug 30013, or is that another thing, where Koha is talking to an external smtp and not sending email directly?
09:06 huginn 04Bug https://bugs.koha-community.or[…]_bug.cgi?id=30013 enhancement, P5 - low, ---, koha-bugs, NEW , Limit/throttle/retry SMTP connections
09:07 marcelr hmm didnt see that one before, let me look
09:08 magnuse yes the --limit should help you rightaway and the domain limits would be added on top
09:08 i think we can close that one as a duplicate
09:09 magnuse ok, very cool!
09:09 marcelr the current --limit is already a throttle
09:10 ashimema indeed
09:13 cait not changing the frequency please
09:13 the frequency is because of the CHECKIN/CHECKOUT messages
09:14 or we need to split and keep them at another freequence
09:14 5 minutes is too low - you'll end up with mutiple messages per checkin/checkout session too often
09:15 the notice gets genarted with first transaction and added to until sent
09:15 marcelr thats true
09:15 thats why we have sendalert to bypass the frequency
09:16 and several SendQueue calls with one message_id
09:16 probably we should have to solve the checkin digest thing itself?
09:17 cait maybe, but either way it needs a little work before we can shorten it
09:17 we coudl also schedule separate runs now I thinik - you can include/exclude notice codes
09:17 marcelr no worries, it is not changed currently, just thinking about opening a report for it
09:18 cait that would work "right now"
09:18 maybe the checkin/checkout could just be generated by a separate job every x minutes? it might delay sending a little.. but otherweise shoudl be fine
09:19 marcelr yeah something across those lines
09:19 ashimema: did you intend to change status to SO ? thx for testing !
09:19 davewood i could add this feature but im new to the process. what do I do. :) https://bugs.koha-community.or[…]_bug.cgi?id=17387
09:19 huginn 04Bug 17387: enhancement, P5 - low, ---, gmcharlt, NEW , Add an undelete feature for items/biblios
09:19 ashimema just have
09:20 got called into a call
09:20 marcelr np
09:20 ashimema++
09:20 davewood did you see general pages on our wiki about participating in development?
09:21 it would be great if you join, but fairly adding this feature as first thing sounds quite ambitious
09:22 Joubu PedroAmorim[m]: something "fun" going on with filters on 33490
09:22 ashimema: ^
09:22 The code should be extremely easy and we end up with something complicated. There is something wrong.
09:23 ashimema yeah.. I found it a little odd
09:23 in a call right now.. will have another look in a minute
09:29 davewood marcelr: any particular concern you are having?
09:32 Jenny joined #koha
09:33 davewood https://koha-community.org/get[…]d/for-developers/ links to "A really excellent post about Koha’s bug workflow by Owen Leonard of ACPL" which is dead ... http://www.myacpl.org/koha/?p=555 ... This site has been archived or suspended.
09:36 cait maybe it has been moved, always loved oleonards blog
09:36 I think it's also linked from the website, manual and other places
09:37 marcelr davewood: no, just warning you not to get disappointed when you first meet the signoff and qa process
09:38 it may not always be that quick as we would like it to be
09:38 expectation management
09:38 davewood thats part of the reason why I want to add this feature, to get to know the process for future endevours
09:38 marcelr davewood++
09:39 davewood s/endevours/endeavours/
09:40 cait usuall I'd recomment to start with a typo fix or similar for the first runt hrough the workflow :)
09:41 but I think you will be fine
09:46 davewood I did stuff before ... just forgot to follow up ... https://bugs.koha-community.or[…]_bug.cgi?id=30349
09:46 huginn 04Bug 30349: enhancement, P5 - low, ---, davewood, Needs Signoff , Cleanup bulkmarcimport.pl
09:51 mtj hi paulderscheid[m], did you want me yesterday?
09:52 paulderscheid[m] Yeah, don't know whether there's something wrong with my setup but the default detection of arm64 doesn't seem to work on my end for koha-testing-docker.
09:52 But I'm happy to go over my setup again to rule any errors out.
09:56 mtj ah, ok
09:57 paulderscheid[m]: what 'ktd up' error did you get?
09:58 paulderscheid[m] I can tell you in 10 mins
09:58 mtj currently, we are supporting bullseye and focal only, on master branch
09:58 paulderscheid[m] Just have to submit a bugfix real quick
09:59 mtj (still need to add ./Dockerfiles for other OS')
10:01 i did test ok on AWS system, where `uname -m` is aarch64
10:06 marcelr bye #koha
10:13 ashimema stupid question Joubu.. do we even need the by_expired param?
10:13 we only really seem to use the max_expiration_date
10:14 the by_expired seems superflous
10:16 ignore me.. I think I understand
10:23 Joubu yes, we could remove it and deal with it the other way around I guess (tick when the date is there)
10:23 ashimema I think leave as is for now
10:23 I'm reading your most recent patch
10:23 wondering about setting things to ""
10:23 Joubu but with the current behaviour it (should) allows to pass a date that will prefill the input, even if we don't filter by expired (how useful it is, I don't know!)
10:23 ashimema my js is rusty..
10:24 but it somehow feels wrong to set things to an empty string instread of explicitly making them null or undefined.
10:26 Joubu the flatpickr plugin complains otherwise IIRC
10:26 ashimema ah, OK
10:26 in which case.. it makes sense
10:35 paulderscheid[m] mtj: Essentially it's pulling the wrong images for linux/amd64
10:35 Look... (full message at <https://matrix.org/_matrix/med[…]cGSHtfykSqlBQeBcI>)
10:36 uname -m gives arm64
10:39 Maybe I can give some more info..
10:40 The way I got around it is to just run the compose file explicitly... (full message at <https://matrix.org/_matrix/med[…]JJmslyXCloMOVbAJz>)
10:55 mtj hi paulderscheid[m], whats your KOHA_IMAGE set to?
10:56 paulderscheid[m] I had it at master-bullseye
10:56 mtj hmm, ok - that sounds good
10:57 paulderscheid[m] Just tested it with master as well, same error
10:58 mtj bash -x ./bin/ktd up
10:58 pastebot "mtj" at 127.0.0.1 pasted "arm64" (12 lines) at https://paste.koha-community.org/4814
10:59 mtj paulderscheid[m]: you get something similar? ^
10:59 + ARCH=arm64v8
10:59 + KOHA_IMAGE=master-bullseye-arm64v8
11:00 paulderscheid[m] Cool, I have to memorize that :D
11:00 I see the issue now.
11:00 KOHA_IMAGE is unset
11:01 mtj export KOHA_IMAGE="master"
11:01 pastebot "paulderscheid[m]" at 127.0.0.1 pasted "KOHA_IMAGE is unset" (7 lines) at https://paste.koha-community.org/4816
11:02 paulderscheid[m] But why is it unset, though?
11:02 rhen976 joined #koha
11:02 mtj export | grep KOHA_IMAGE
11:03 rhen976 left #koha
11:03 rhen976 joined #koha
11:03 perplexedtheta hi
11:05 paulderscheid[m] mtj: Doesn't return anything. This is really weird. Is this not populated through .env?
11:06 <mtj> "export KOHA_IMAGE="master"" <- This obviously works, btw. ^^
11:09 mtj paulderscheid[m]: i dont bother with .env file myself... but yes, afaik  ^
11:11 aah, i think docker itself uses the .env file
11:11 ...but that means that bin/ktd doesnt :/
11:14 aah well.. thats all fixable
11:14 paulderscheid[m] Are there any side effects to adding source ${KTD_HOME}/.env for the script?
11:15 I mean within bin/ktd?
11:15 mtj pass, ive never done it
11:15 try and report back, xx
11:17 paulderscheid[m] Works
11:17 A little overkill though
11:20 mtj if KOHA_IMAGE is not set, we should tell bin/ktd to get it from .env
11:23 paulderscheid[m] I think I've found a solution for that
11:25 alex_ joined #koha
11:28 mtj perhaps send a patch?, i think others will encounter your problem
11:29 tcohen hola #koha o/
11:30 paulderscheid[m] I will
11:31 Can I just PR on koha-testing-docker?
11:31 tcohen the problem with sourcing it...
11:31 paulderscheid[m] It overrides a lot of ENV vars
11:31 mtj tcohen: you cant override it?
11:31 tcohen is that it will override any override you have using env
11:31 like, yeah
11:31 mtj yeah, snap
11:31 paulderscheid[m] KOHA_IMAGE=$(grep '^KOHA_IMAGE=' ${KTD_HOME}/.env | cut -d '=' -f2)
11:31 tcohen I use KOHA_IMAGE=<whatever I need> ktd up
11:32 paulderscheid[m] It would be cool if KOHA_IMAGE just defaulted to what's in .env if unset.
11:33 mtj yeah, i like that idea ^
11:34 or defaulted to master?
11:34 # if master or unset
11:34 if [ "$KOHA_IMAGE" == "master" ] || [ -z "$KOHA_IMAGE" ] ; then
11:34 KOHA_IMAGE="master-bullseye"
11:34 fi
11:35 tcohen I'm not sure I follow entirely
11:35 because that's exactly what .env is used for
11:35 this smells
11:35 paulderscheid[m] Yeah, but we need that var for set_arch
11:35 tcohen ah!
11:35 paulderscheid[m] And that's before .env is read I think
11:36 tcohen I thought our if construct was correctly giving a feck about KOHA_IMAGE
11:36 mtj tcohen: bin/ktd doesnt read .env,  only docker
11:36 tcohen and forcing master-bullseye-arm64v8
11:38 mtj set_arch() assumes that KOHA_IMAGE is set, but it may not be
11:38 * tcohen reads back
11:38 tcohen in the irc logs
11:48 magnuse davidnind++
11:57 * cait waves
12:05 paulderscheid[m] I actually think I found a flaw in the set_arch func
12:09 tcohen I feel like we should only mess with KOHA_IMAGE if we detect the arm arch
12:12 paulderscheid[m] Me too.
12:12 I think we have to export KOHA_IMAGE in the set_arch func because the var is limited in scrope to bin/ktd. I could be wrong, though. But when exporting the func finally works as expected.
12:14 pastebot "paulderscheid[m]" at 127.0.0.1 pasted "This is how I modified it based on the recommentations" (17 lines) at https://paste.koha-community.org/4826
12:14 tcohen you are right
12:15 paulderscheid[m] The case thing was just for future architectures. I actually pasted the wrong one ^^
12:19 pastebot "paulderscheid[m]" at 127.0.0.1 pasted "Here, this is the right one" (18 lines) at https://paste.koha-community.org/4828
12:20 tcohen I agree with that paulderscheid[m]
12:21 paulderscheid[m] Do we just submit PRs on ktd?
12:26 mtj paulderscheid[m]: yep, looks good to me
12:36 paulderscheid[m] I have created a merge request on ktd
12:44 kidclamp joined #koha
12:46 tcohen paulderscheid[m]++
12:46 mtj++
12:46 paulderscheid[m] mtj++
12:46 tcohen++ :D
12:58 Dyrcona joined #koha
13:04 ashimema paulderscheid++
13:11 tcohen mtj: not enough disk space on the arm runner
13:37 caroline Joubu, ashimema I apologize for yesterday. I'm not usually this resistant to change (I don't think) :/
13:42 My spell/grammar checker let me know this morning that my style of writing was "ancient" lol I think I'm getting old?
13:46 oleonard-away Are those two previous messages related caroline? :D
13:47 caroline Kind of? I mean, even my spell checker is telling me I need to change my ways
13:47 oleonard I'm always saying, "there goes caroline again, lecturing us about how life is just shadows on the wall of a cave..."
13:47 caroline It is!! I'm telling you!
13:49 Now I feel like the old cat lady in the simpsons who's always muttering and throwing cats (although I would never throw a cat)
13:51 oleonard caroline: That's okay, I'm "Old Man Yells at Cloud"
13:51 cait oleonard: what happened to your blog?
13:52 oleonard I took it down when I was having problems with our Wordpress installation. I'll try to revive it even if it is a historical artifact at this point.
13:58 lukeg joined #koha
13:59 ashimema LOL
14:29 JonTech Hello, I am running into some errors when running the upgrade schema command. Error are here: https://imgur.com/a/njtHz4k
14:29 I am moving from Koha 3.16 to latest version
14:32 tcohen you forgot to load the DB?
14:32 or it failed
14:37 cait oleonard: i think if you don#t want to... we should at least clean up the places it#s linked to - just let me know
15:04 ... the places that link to it...
15:04 time for the weekend
15:04 bye all!
15:05 cait left #koha
15:17 ashimema blimey marcelr.. that bug went through quick 🙂
15:22 reiveune bye
15:22 reiveune left #koha
15:56 oleonard I guess I should learn what item groups are for.
16:04 I wonder if there's a purpose for item groups outside the context of placing item group holds
16:18 PedroAmorim[m] tcohen: around?
17:32 caroline oleonard, I think right now that's pretty much it
17:54 cait joined #koha
18:06 tcohen PedroAmorim[m]: :-D
18:07 oleonard: exactly, right now it's only so any item on a group can fill an item-group-level hold
18:15 caroline Not sure it will be used by current Koha libraries, since it was not possible, probable all of them have volumes separated on distinct records, but it is a functionality in other ILS
18:37 tcohen this would be expression-level holds
18:37 good subject to discuss a friday evening
18:57 have a great weekend y'all \o

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

koha1