IRC log for #koha, 2022-01-20

All times shown according to UTC.

Time S Nick Message
00:04 koha-jenkins Yippee, build fixed!
00:04 koha-jenkins Project Koha_Master_D12 build #64: FIXED in 34 min: https://jenkins.koha-community[…]ha_Master_D12/64/
00:16 koha-jenkins Project Koha_Master_D10 build #514: SUCCESS in 54 min: https://jenkins.koha-community[…]a_Master_D10/514/
00:19 koha-jenkins Project Koha_Master_D11_MDB_Latest build #775: SUCCESS in 1 hr 2 min: https://jenkins.koha-community[…]1_MDB_Latest/775/
00:22 lukeg joined #koha
00:50 tuxayo @later tell mtj did my 2 emails arrived in the last 48h?
00:50 huginn tuxayo: The operation succeeded.
01:02 koha-jenkins Project Koha_Master_D11_MDB_Latest build #776: SUCCESS in 42 min: https://jenkins.koha-community[…]1_MDB_Latest/776/
05:11 JBoyer_ joined #koha
05:48 jeff joined #koha
06:02 lmstrand joined #koha
07:18 alex_a joined #koha
07:18 marcelr joined #koha
07:18 marcelr hi #koha
07:18 alex_a Bonjour
07:21 marcelr o/
07:32 reiveune joined #koha
07:32 reiveune hello
07:33 marcelr hi reiveune
08:02 marion joined #koha
08:04 cait joined #koha
08:06 cait1 joined #koha
08:31 marcelr @Joubu around
08:31 huginn marcelr: I'll give you the answer as soon as RDA is ready
08:33 Joubu yes
08:33 marcelr Joubu: we still have some glitch in Auth.pm with 2fa
08:33 * cait1 waves
08:33 marcelr hi cait1
08:34 marcelr reported it on the bug itself
08:35 Joubu I will have a look a bit later
08:35 marcelr great
08:36 marcelr if I find it already, will let you know ;)
08:36 marcelr just demonstrates the ugliness of changing Auth.pm
08:39 cait1 one day those price formatting erros will be my end... spent way too much time on bug 29911 yesterday
08:39 huginn Bug https://bugs.koha-community.or[…]_bug.cgi?id=29911 normal, P5 - low, ---, koha-bugs, NEW , Default amounts for fee types display with comma in some cases
08:45 Joubu marcelr: I do recreate. That's crazy. It's considered logged in :-/
08:46 marcelr yeah
08:54 marcelr but related to the changes
08:57 Joubu marcelr: the behaviour on the UI is coming from the patches, but the 'bug' exists in master.
08:57 marcelr yes, but feels like we cant ignore it ?
09:03 marcelr we did
09:08 magnuse marcelr++ Joubu++
09:09 ashimema morning
09:09 marcelr hi magnuse ashimema
09:10 marcelr two factor agitation
09:11 ashimema mm, sounds a bit scary
09:11 matts :à)
09:11 ashimema I hadn't had enough time to really dig into it
09:11 ashimema Auth.pm is scary
09:11 ashimema really glad to see you looking at it marcelr
09:12 marcelr and dont forget Search.pm and holds; the big three :)
09:13 Joubu bug is in master...
09:13 ashimema Circulation.pm is also kinda scary
09:13 marcelr yeah but i dont wanna know lol
09:13 ashimema hehe
09:17 cait1 i don't like this
09:19 marcelr better know than not know
09:22 ashimema indeed
09:28 cait1 true
09:39 ashimema https://www.ptfs-europe.com/vacancies/
09:45 marcelr when i am moving to the UK, i let you know ;)
09:47 ashimema 🙂
09:47 ashimema hehe
09:51 cait joined #koha
09:57 cait1 joined #koha
10:29 ashimema your code doesn't make sense to me Joubu
10:29 ashimema $userid is set from $session->param('id').. then you subsequently check they're equal
10:31 marcelr ok i concluded the same now
10:31 marcelr on the report
10:32 alex_a joined #koha
10:32 * ashimema added a comment to the bug
10:34 ashimema the comment in there doesn't entirely make sense either
10:34 ashimema 'see if we have a valid session cookie already. However, if a userid paramenter is present...`
10:34 ashimema I don't see how a userid param is ever checked for being present
10:34 ashimema am I missing something obvious?
10:38 Joubu nope
10:38 Joubu the patch is wrong
10:40 marcelr ashimema: we are saying the same but talk other language
10:40 ashimema hehe
10:40 ashimema yup
10:40 ashimema I realised once I re-read your comment
10:42 ashimema I think perhaps it was actually working as intended
10:42 marcelr no
10:42 ashimema the cookie could be for an anonymous session
10:42 marcelr therefore
10:43 marcelr we should check for a true value
10:43 ashimema if there are no permissions required for the view.. then the anon session returning OK is correct.. isn't it.
10:43 ashimema so the test in haspermission is what fails you
10:43 Joubu modified patch attached
10:43 marcelr hmm
10:43 ashimema check what for a true value.. id?
10:44 marcelr yes like the new patch indeed
10:44 marcelr if we would return ok for anon, elsewhere a wrong conclusion is made
10:44 ashimema I understand the new patch..
10:45 ashimema but we need to dig deeper to understand if check_cookie_auth should return 'ok' for anon sessions when there are no permissions required to view the resource
10:45 Joubu yes the question is, should we return "ok" or not for anon? Here we are going to fail
10:45 ashimema yup
10:45 Joubu but the sub is saying ""ok" -- user authenticated; C<$sessionID> have valid values."
10:45 Joubu anon is not auth :)
10:45 ashimema yowsers.. check_cookie_auth is used all over the place ☹️
10:45 marcelr public api perhaps
10:46 ashimema well.. it's 'authentication' vs 'authorization' again isn't it
10:46 ashimema we're mixing the two in here.. again!
10:46 marcelr they are in one module
10:46 marcelr its a big mess
10:46 marcelr oops
10:46 ashimema yup
10:46 Joubu here we generate a new session for anon, that's wrong as well
10:46 ashimema my point is that that is confusing 😉
10:47 marcelr which i understand at least
10:47 ashimema yeah
10:47 ashimema I agree
10:47 ashimema once you have an anon session open.. you should keep that anon session unless you explicitly upgrade it to a user session (by logging in)
10:48 marcelr lets rewrite the whole thing ;)
10:48 marcelr but right
10:48 ashimema perhaps we need to just run the api test suite against this patch.. see if anything breaks
10:49 ashimema that's assuming it's tested properly
10:49 ashimema lol
10:49 marcelr there was not even one test for this sub
10:49 marcelr direct test
10:49 ashimema mm
10:49 ashimema lots of indirect one's I reckon
10:50 ashimema but it would be nice to have a bunch of direct ones
10:50 Joubu I think we need to explicitely return "anon" and deal with that in the caller
10:50 Joubu to make things explicit
10:50 ashimema in Koha/REST/V1/Auth.pm we have
10:51 marcelr depends on the impact volume
10:51 marcelr we return expired if there is no session; thats theoretically wrong too?
10:51 ashimema ` $user = Koha::Patrons->find( $session->param('number') ) unless $session->param('sessiontype') and $session->param('sessiontype') eq 'anon'`
10:51 Joubu V1/Auth is weird, it's dealing with anon. It's like Tomas felt there was something wrong already
10:51 ashimema that hints to me that tomas thought about this case
10:51 ashimema exactly
10:51 ashimema haha
10:53 ashimema hmm.. expired sort of makes sense I think in that case marcelr..
10:53 ashimema it means the session has gone away on the server even if the client still has a cookie
10:53 ashimema so it is 'expired' at the server..
10:53 ashimema it overlaps with our timeout expired stuff though of course..
10:54 ashimema we could be more explicit in our return I suppose..
10:54 ashimema though I can't think of a nicer name
10:54 ashimema expired vs timed out
10:56 marcelr ok np
10:56 marcelr just theory
10:57 Joubu lol we are already generating a new sessionID everytime we hit a page at the OPAC if we are not logged in...
10:57 Joubu so we shouldn't deal with that here
10:57 ashimema I feel like this routine should always be passed a flag to check.. it's more check_authorization than check_authentication
10:57 ashimema yuk
10:58 marcelr you know now why that file is so large
10:58 marcelr thx googlebot
10:58 ashimema indeed
10:58 ashimema I've contemplated working on a new auth module for years
10:59 ashimema but never had the time/guts to go for it.
10:59 marcelr some even submitted stuff
10:59 marcelr but it never made it
10:59 ashimema yup
10:59 ashimema it's all the crazy edge cases
11:00 marcelr the code should have been split earlier
11:00 marcelr cas, shibboleth etc
11:00 marcelr authentication vs authorization like you said etc
11:02 marcelr afk for a bit now, when I come  back i test your new Auth module Joubu lol
11:09 ashimema there's so much duplicate code in here ☹️
11:11 Joubu I did some cleaning recently
11:11 Joubu bug 28785
11:11 huginn Bug https://bugs.koha-community.or[…]_bug.cgi?id=28785 enhancement, P5 - low, ---, jonathan.druart+koha, RESOLVED FIXED, Code in C4::Auth::checkauth is copy pasted
11:21 ashimema indeed
11:21 ashimema be nice to keep going on that 🙂
11:34 khall_ joined #koha
11:35 oleonard Hi all
11:56 oleonard One more Flatpickr bug to go before all jQueryUI datepickers will be gone: Bug 29228
11:56 huginn Bug https://bugs.koha-community.or[…]_bug.cgi?id=29228 enhancement, P5 - low, ---, oleonard, Needs Signoff , Use Flatpickr on offline circulation page
11:56 oleonard Offline circ is scary I guess
12:22 marion joined #koha
12:29 marcelr joined #koha
12:29 marcelr hi #koha
12:49 oleonard ashimema: still around?
12:51 alex_a joined #koha
13:05 oleonard Plugin searching not working as it should, I think.
13:40 khall joined #koha
13:42 Brooke joined #koha
13:42 Brooke 0/
13:44 oleonard Hi Brooke
13:44 Brooke Hi there!
13:55 Dyrcona joined #koha
13:59 marie-luce joined #koha
14:22 gooble_gobble joined #koha
14:31 lukeg joined #koha
14:34 ashimema back oleonard
14:34 ashimema will take a look at your last flatpickr next
14:35 ashimema just looking at a bug from nugged first 😉
14:41 khall_ joined #koha
14:52 oleonard ashimema: I was testing 29787 and got inconsistent search results, I wondered if it was something you'd run into
14:52 oleonard Some searches fail with a big error... others return results from one repo where there should be results from multiple
14:53 ashimema ah, I didn't go as far as testing it
14:53 oleonard Bug 29787
14:53 huginn Bug https://bugs.koha-community.or[…]_bug.cgi?id=29787 enhancement, P5 - low, ---, tomascohen, Signed Off , Add plugin version to plugin search results
14:53 ashimema prior to this patch I've tended to find the searches fairly reliable
14:55 oleonard For example:  I would expect a search for "marc" to return results from more than just ByWater
14:55 ashimema humm.. the tests for holds put vs patch are terrible ☹️
14:55 ashimema agreed.
14:55 ashimema it's been a while since I've played with the feature I'm afraid oleonard
14:56 oleonard No problem, I'll file a bug.
14:56 ashimema cheers
14:57 KBaldwin joined #koha
14:58 KBaldwin Hello All
14:58 KBaldwin Would anyone be able to help me with a report I'm trying to formulate?
14:58 oleonard Ask, and perhaps someone will answer
14:59 KBaldwin I'm trying to find/create a cataloging report that pulls duplicate EANs/UPCs (the 024 field) from our collection. So far, I've only found reports for duplicate ISBNs
15:06 caroline Any European french speakers around? I have a terminology question :)
15:07 oleonard KBaldwin: The duplicate ISBN report should be able to be modified, but you'll have to look at the "extractvalue" process described in the SQL reports wiki
15:11 KBaldwin oleonard: Let me see if I can find that particular report in the library. I played around with a couple, but my SQL knowledge is very limited
15:27 KBaldwin This is what I have, but it's not generating anything. Do you know what I'm missing?
15:27 KBaldwin SELECT biblionumber, ExtractValue(metadata,'//datafield[@tag="024"]') AS EAN/UPC   FROM biblio_metadata  WHERE ExtractValue(metadata,'//datafield[@tag="024"]')>1
15:29 oleonard The first problem is that "EAN/UPC" can't have a slash
15:33 KBaldwin Okay
15:36 KBaldwin The EAN/UPC didn't seem to make a difference. That was just what I wanted to call the column. Perhaps the extractvalue is coming from the wrong source? I adjusted for the tag being controlled rather than fixed, but something is still off
15:37 KBaldwin Currently: SELECT biblionumber, ExtractValue(metadata,'//c​ontrolfield[@tag="024"]') AS EAN   FROM biblio_metadata  WHERE ExtractValue(metadata,'//co​ntrolfield[@tag="024"]')>1
15:37 oleonard ExtractValue(metadata,'//datafield​[@tag="024"]/subfield[@code="a"]')
15:39 oleonard Still doesn't work, but I'm trying to figure out the pieces
15:41 marion caroline I'm french. I'm not sure I can help you but at least I can try :)
15:41 caroline I was just wondering if "Téléphone mobile" was an acceptable term in europe for Cell phone
15:42 caroline I think the more common term is "portable", but here in Québec that means laptop
15:43 oleonard KBaldwin: Try SELECT GROUP_CONCAT(biblionumber SEPARATOR ', ') AS biblionumbers, ExtractValue(metadata,'//datafield​[@tag="024"]/subfield[@code="a"]') AS `EAN` FROM biblio_metadata GROUP BY `EAN` HAVING COUNT(`EAN`)>1;
15:44 marion caroline Yes it works :) Mayby "Téléphone portable" is more natural to me but "Téléphone mobile" is perfectly fine too
15:46 caroline thx! As long as there is no confusion... it's for an option in a survey so I wanted to make sure everyone understands the same thing :)
15:46 * ashimema starts to contemplate hackfest
15:47 ashimema damn.. they no longer do direct flights from my local airport ☹️
15:49 KBaldwin oleonard: With what you gave me, I'm getting a list of biblio numbers, but there is nothing in the EAN column, and the biblionumbers have no 024 associated with them when I searched
15:50 oleonard :/
15:51 KBaldwin oleonard: XD I appreciate you trying to figure this out with me. I was about to throw my computer out the window.
15:52 KBaldwin Do I maybe have to add a JOIN function?
15:55 caroline you shouldn't need to since metadata and biblionumber can both be found in biblio_metadata
15:57 caroline can you try SELECT biblionumber, ExtractValue(metadata,'//datafield​[@tag="024"]/subfield[@code="a"]') AS `EAN` FROM biblio_metadata WHERE biblionumber = XXXX; (replace XXXX by a biblionumber you know has a value in 024$a
15:58 caroline I'm curious to see if it returns the correct value
16:00 KBaldwin caroline: Hi, Yes it was able to pull a specific biblionumber with its EAN
16:00 caroline ok good, so the ExtractValue part is correct
16:02 caroline oh maybe add "WHERE `EAN` IS NOT NULL" between biblio_metadata and GROUP BY
16:06 KBaldwin BRB, going on lunch break . Thank you both!
16:07 khall joined #koha
16:10 oleonard KBaldwin: I think this works:
16:10 oleonard SELECT GROUP_CONCAT(biblionumber SEPARATOR ', ') AS biblionumbers, ExtractValue(metadata,'//datafield​[@tag="024"]/subfield[@code="a"]') AS `EAN` FROM biblio_metadata WHERE (ExtractValue(metadata,'//datafield​[@tag="024"]/subfield[@code="a"]') != '' AND ExtractValue(metadata,'//datafield​[@tag="024"]/subfield[@code="a"]') IS NOT NULL ) GROUP BY `EAN` HAVING COUNT(`EAN`)>1;
16:24 AndrewFH joined #koha
16:30 reiveune bye
16:30 reiveune left #koha
16:37 marion_ joined #koha
16:43 KBaldwin oleonard: It works! Thank you, thank you, thank you! You've made my week!
16:44 oleonard Excellent!
16:44 KBaldwin left #koha
16:47 cait1 hm i really need to look into group concat
16:48 caroline oleonard++
16:48 oleonard caroline++
16:48 cait1 oleonard++ caroline++
16:57 caroline oleonard: do you mind if I add it to the wiki for you?
16:57 oleonard Oh not at all, thank you
16:59 AndrewFH joined #koha
16:59 caroline For future reference https://wiki.koha-community.or[…]ary#Duplicate_EAN.2FUPC_.28024.24a.29
17:00 caroline I just wanted to make sure we don't lose it in the logs
17:32 cait1 left #koha
17:54 marion joined #koha
18:29 AndrewFH joined #koha
19:40 khall joined #koha
19:54 khall_ joined #koha
20:31 cait joined #koha
21:00 fridolin joined #koha
23:02 cait joined #koha
23:20 cait joined #koha

| Channels | #koha index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary