IRC log for #koha, 2021-03-24

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

All times shown according to UTC.

Time Nick Message
00:00 aleisha oh okay !
00:00 the error is ZOOM error 20003 "can't set prefix query"
00:02 dcook I think that usually happens when there's nothing in an index? Admittedly I haven't seen that one myself. Just Googling..
00:02 Let me see if there's a little workaround..
00:02 aleisha hmm interesting
00:03 dcook Let's see..
00:03 Add @attr 14=1 to the very start
00:04 Or rather... @attr 14=1 I guess
00:04 Oh I said that.
00:04 aleisha heh
00:04 dcook That will ignore uninitialized indexes
00:04 aleisha @attr 14=1 @and  @attr 1=Heading-Main  @attr 5=1 @attr 4=6 "fish" @or @attr 1=Subject-heading-thesaurus @attr 3=2 @attr 4=1 @attr 5=1 @attr 6=3 "scisshl" @attr 1=Subject-heading-thesaurus @attr 3=2 @attr 4=1 @attr 5=1 @attr 6=3  "scot" like this?
00:04 huginn aleisha: I'll give you the answer as soon as RDA is ready
00:04 wahanui i already had it that way, huginn.
00:05 dcook yep
00:06 aleisha hmm that gets the same error
00:07 urra[m] joined #koha
00:07 dcook I had a feeling you were going to say that heh
00:08 aleisha wish there was more documentation on the errors!
00:09 dcook https://metacpan.org/release/N[…]ource/lib/ZOOM.pm heh
00:09 Looks like a fairly generic error
00:10 Alas, I should switch to other things, but I'd say your query is OK. It'll be the Perl coding that you need to perfect.
00:10 aleisha thank you for your help!!
00:10 dcook You're very welcome. Always good to review Zebra queries.
00:10 This is custom code now, right?
00:11 aleisha yes it is
00:11 dcook Are you getting your Zconn from C4::Context or creating your own?
00:12 aleisha not creating my own!
00:12 dcook Koha's code is terrible when it comes to working with the ZOOM libraries too btw
00:12 aleisha the query comes through as normal, im just amending it once it gets in
00:12 dcook Ah, good ol' mangling hehe
00:13 aleisha haha also the search stuff is different between biblio and authorities! very hard to troubleshoot
00:13 dcook Yeah, that'll take some digging
00:14 I wish you luck. You can always use new ZOOM::Query::PQF->new() in an eval{} to see at what point the query is causing problems
00:15 Whether it's before or after you alter it
00:15 aleisha okay thank you i'll use that
00:17 koha-jenkins Project Koha_19.11_U18 build #304: ABORTED in 12 min: https://jenkins.koha-community[…]ha_19.11_U18/304/
00:29 Project Koha_19.11_U20 build #202: SUCCESS in 23 min: https://jenkins.koha-community[…]ha_19.11_U20/202/
00:35 Project Koha_19.11_D9 build #308: UNSTABLE in 29 min: https://jenkins.koha-community[…]oha_19.11_D9/308/
00:37 Project Koha_19.11_D8 build #377: SUCCESS in 32 min: https://jenkins.koha-community[…]oha_19.11_D8/377/
00:43 Project Koha_19.11_U18 build #305: ABORTED in 10 min: https://jenkins.koha-community[…]ha_19.11_U18/305/
00:51 Project Koha_19.11_D10 build #146: SUCCESS in 43 min: https://jenkins.koha-community[…]ha_19.11_D10/146/
01:00 Project Koha_19.11_U16 build #61: SUCCESS in 55 min: https://jenkins.koha-community[…]oha_19.11_U16/61/
01:14 stefanberndtsson[m] joined #koha
01:25 aaru_swartz[m] joined #koha
01:34 AndrewFH joined #koha
01:35 koha-jenkins Yippee, build fixed!
01:35 wahanui Congratulations!
01:35 koha-jenkins Project Koha_19.11_U18 build #306: FIXED in 40 min: https://jenkins.koha-community[…]ha_19.11_U18/306/
01:42 tubaclarinet[m] joined #koha
01:44 oleonard joined #koha
02:12 AndrewFH joined #koha
02:25 manjukaliany[m] joined #koha
02:59 aleisha joined #koha
06:01 MarkHofstetter joined #koha
06:24 chriss joined #koha
07:06 MarkHofstetter1 joined #koha
07:18 severine_q joined #koha
07:18 severine_q good morning #koha :)
07:26 reiveune joined #koha
07:27 reiveune hello
07:27 wahanui privet, reiveune
07:27 paul_p__ joined #koha
07:30 alex_a joined #koha
07:30 alex_a Bonjour
07:36 did joined #koha
07:53 fridolin joined #koha
07:56 cait joined #koha
07:57 cait1 joined #koha
08:27 lds joined #koha
08:32 alex_a_ joined #koha
08:57 severine_q joined #koha
09:21 MarkHofstetter1 Good morning, is it/should it be possible to add a guarantor to an existing patron, or is this forbidden by design? thx
09:23 cait1 no it's possible
09:23 but it depends on the categories and such
09:24 if you are offered the guarantor fields or not
09:28 MarkHofstetter1 I mean there may be very good reasons not to be able to change that, but it may be not userfriendly
09:57 cait1 but you can change the guarantor :)
09:57 it's probably another reason you don't see it
09:57 have you opened the record for editing?
10:04 MarkHofstetter1 yes
10:28 fridolin joined #koha
10:33 eythian ashimema: sorry, wasn't around yesterday
10:33 cm joined #koha
10:33 cm hi
10:39 ashimema no worries eythian
10:39 I was going to ask if I was remembering rightly that your a bit of an email guru..
10:40 I think..
10:40 or maybe it was something else
10:40 lol.. I've totally forgotten what it was
10:40 how are you anyways eythian
10:41 eythian ashimema: I know the occasional thing about it :) and yeah, I see you added me to a bug about it, but it looks like it was resolved anyway.
10:41 ashimema: not too bad, just generally bored :)
10:45 MarkHofstetter joined #koha
10:49 shaffendi joined #koha
10:49 shaffendi hi hello
10:49 anyone here
11:06 khall_ joined #koha
11:21 AndrewFH joined #koha
11:36 khall joined #koha
11:36 oleonard Hi #koha
11:36 * oleonard finished testing, signing off, and following-up Bug 27846 and boy are my arms tired
11:36 huginn Bug https://bugs.koha-community.or[…]_bug.cgi?id=27846 enhancement, P5 - low, ---, wainuiwitikapark, Needs Signoff , Accessibility: Staff Client - Breadcrumbs should be more accessible
11:36 MarkHofstetter joined #koha
11:37 oleonard 4 follow-ups still waiting for signoff.
11:52 ashimema nice one oleonard
11:52 that bug was already on my list
11:54 kidclamp ashimema++
11:54 ashimema pleasure kidclamp :)
11:56 kidclamp oleonard++
11:56 making me take responsibility for my code
12:01 ashimema oleonard
12:01 wahanui oleonard is probably happy for ashimema to write the release script
12:01 ashimema `and the last breadcrumb has bold text` in the test plan
12:01 I don't see that.. did you?
12:01 oleonard Oh actually I was going to comment on that... That's the one thing I would change
12:02 I don't like the bold (which I do see)
12:02 ashimema oh really
12:02 it doesn't go bold for me
12:02 chrome?
12:02 wahanui rumour has it chrome is fine fox123
12:03 oleonard Haha then I broke something
12:03 ashimema lol
12:03 oleonard Give me a moment to look into it
12:03 ashimema but.. you're saying you don't think it should be boldended?
12:03 I don't think it was prior to this was it?
12:03 oleonard It wasn't
12:04 ashimema and I don't see the accessibility guideline saying it should be either
12:04 koha-jenkins Project Koha_20.11_U20 build #60: SUCCESS in 31 min: https://jenkins.koha-community[…]oha_20.11_U20/60/
12:04 ashimema must admit.. I think it's a shame we loose the chevron
12:04 I liked that display queue
12:07 oleonard I'm not sure if that change was accessibility-related or not
12:07 * ashimema wonders if we could add that back but use aria-hidden on it.
12:07 ashimema the bold or the right pointing quote
12:07 oleonard The dividers
12:08 ashimema it's part of the guideline that's linked
12:08 they're borders instead of characters
12:08 idea being a border isn't anounced by a screen reader
12:09 oleonard Is CSS content announced?
12:09 ashimema but yeah.. I reckon we can achieve the same by adding aria-hidden to them
12:13 AndrewFH joined #koha
12:13 oleonard I have corrected my CSS patch so that it doesn't undo the bold breadcrumb
12:14 koha-jenkins Project Koha_20.11_U16 build #61: SUCCESS in 38 min: https://jenkins.koha-community[…]oha_20.11_U16/61/
12:14 ashimema okies
12:14 right.. I'm going for a quick walk
12:14 back shortly.. then I will continue here
12:24 koha-jenkins Project Koha_20.11_U2010 build #55: SUCCESS in 51 min: https://jenkins.koha-community[…]a_20.11_U2010/55/
12:27 Project Koha_20.11_D10 build #76: UNSTABLE in 54 min: https://jenkins.koha-community[…]oha_20.11_D10/76/
12:31 Project Koha_20.11_D11 build #90: SUCCESS in 27 min: https://jenkins.koha-community[…]oha_20.11_D11/90/
12:40 Project Koha_20.11_D9 build #78: SUCCESS in 1 hr 6 min: https://jenkins.koha-community[…]Koha_20.11_D9/78/
12:56 marie-luce joined #koha
12:56 cait1 i think i would like it non-bold :)
12:56 oh
12:57 scroll first!
12:58 koha-jenkins Project Koha_20.11_U18 build #56: SUCCESS in 1 hr 25 min: https://jenkins.koha-community[…]oha_20.11_U18/56/
13:00 ashimema It was made bold by the patchset.. oleonard then broke that so he fixed it back
13:00 but the question still stands.. should it be bold or not
13:00 cait1 i think as we have the heading on the page that is prominent
13:00 oleonard I didn't want to reverse the intention of the original patchset without talking about it first
13:00 cait1 i would not bold it
13:00 it seems a double up
13:00 ashimema thoughts on '/' vs '>' cait1 ?
13:01 cait1 hm >
13:01 we have also used that in documentation and i think in the manual.... and i like the visual
13:01 ashimema me too
13:01 cait1 i feel it's easier to read
13:02 not sure if there are any general recommendations we shoudl follow, but that would be my opinion right now :)
13:02 ashimema they changed it to '/' for accessability..
13:02 but
13:02 it's not an actual '/'
13:03 it's a css border that's been tilted forward
13:03 cait1 hm
13:03 ashimema the idea being.. as it's a graphical element rather than a character it won't get anounced by a screen reader
13:03 cait1 but you could use css to add a > too
13:03 couldn't you?
13:03 if it's just amatter of it not being read out lout
13:03 ashimema I think we can achieve the same with our > char, but adding in an 'aria-hidden' attribute on it.
13:04 cait1 or that
13:04 ashimema you can indeed add > with css
13:04 but I think if it's added as a character it still gets read out
13:04 unless we mark it somehow not to be
13:06 ooh..
13:06 Dyrcona joined #koha
13:06 ashimema it turns the text black too at the moment
13:06 oleonard.. was that the case before your patches too?
13:06 bold and black
13:06 I think I expected bold but still blue
13:11 koha-jenkins Project Koha_20.11_D11 build #91: SUCCESS in 39 min: https://jenkins.koha-community[…]oha_20.11_D11/91/
13:20 cait1 they were black i think
13:21 ashimema so, black but not bold
13:21 cait1 yes?
13:23 * ashimema checks out master and check
13:23 ashimema so it was
13:29 dersmon joined #koha
13:29 dersmon hi!
13:31 ashimema oh I see..
13:31 caroline so what is the final decision on the breadcrumbs?
13:31 ashimema the current page is now an anchor
13:32 got anyone that uses a screenreader caroline?
13:32 caroline no :(
13:33 maybe we can ask our accessibility advocate?
13:33 is it henry?
13:33 ashimema indeed
13:35 oleonard In order to use "area-hidden" we'd have to add the separators as an html element
13:35 ashimema in a quick test here.. simply adding them via css did the trick
13:35 oleonard The advantage of the CSS solution is that it works just by adding the <li> for the breadcrumb
13:36 ashimema the screen reader didn't pick them up
13:36 oleonard ashimema: using css content?
13:36 ashimema yup
13:36 I'm just adding my SO line now and adding the followup
13:36 oleonard What I found in my google searches was that screen readers sometimes do read them
13:36 So it's not really an option
13:36 ashimema but I'm going to ask for a second check with the screen reader they're using at catalyst
13:36 just to make doubly sure
13:37 ack..
13:37 oleonard There are CSS alternatives, though, so we just need to find the right thing http://jsfiddle.net/kevinkirchner/DQruj/
13:38 (that example uses markup, but I'm sure we can find something right)
13:40 ashimema good call
13:45 were you going to have a go at that oleonard..?
13:45 oleonard I will do some tests later today. Gotta disappear for a while.
14:23 Michael joined #koha
14:23 MichaelKuhn joined #koha
14:25 MichaelKuhn Hi - I'm looking for a way to download the most current (but unreleased) version of "Email.pm" with alle the changes made via bug 26705. Is it possible?
14:25 huginn Bug https://bugs.koha-community.or[…]_bug.cgi?id=26705 major, P5 - low, ---, tomascohen, Passed QA , System preference NoticeBcc not working
14:31 cait1 MichaelKuhn: yu can browse the repository via git
14:31 https://git.koha-community.org
14:31 i am sure you can download it from there
14:31 but i'd double check the patch for changes made to other files
14:32 https://git.koha-community.org[…]ter/Koha/Email.pm
14:35 cait joined #koha
14:36 MichaelKuhn Hallo Katrin - da war ich schon, aber da befindet sich offenbar nur die bereits veröffentlichte Version... ich habe nun die Änderungen händisch eingepflegt (seit gestern einige Male hin und her...). Trotzdem danke!
14:36 cait MichaelKuhn: ah i see why
14:36 it hasn't been pushed yet
14:36 MichaelKuhn Weitere Änderungen gab es nur in der Datei "Email.t", welche aber offenbar nur für interne Test benötigt wird und auf dem Koha-System selber nicht vorhanden ist.
14:37 cait vorhanden ist sie
14:37 MichaelKuhn Yes, I thought so - not pushed
14:37 cait aber sie wird im laufende Betrieb nicht ausgeführt
14:37 MichaelKuhn But that's why I asked :)
14:37 cait ja, da war mein Denkfehler - hatte es nicht gleich gesehen
14:37 MichaelKuhn Email.t habe ich aber auf dem Koha-System nicht gefunden?
14:37 Gut... ich brauche sie ja auch nicht
14:38 cait2 joined #koha
14:38 cait2 disconnect, vpn, sorry
14:39 vielleicht wurde sie mal umbenannt oder verschoben, generell müssten die .t Dateien mit den Unit tests aber auch volsltändig in den Tarballs bzw. Packages sein
14:39 evtl, anderer Pfad
14:40 MichaelKuhn Ich habe eben nochmal mit "find / -name Email.t -print" gesucht - keine Datei "Email.t". Aber wenn sie natürlich einen anderen namen hätte kann ich sie nicht finden...
14:46 * ashimema schoolboy german isn't up to following that.
14:51 teertha joined #koha
14:59 cait2 ashimema: discussing if the Email.t is in packages or not
14:59 i don't think we are leaving files out, do we?
15:08 teertha kia ora #koha
15:10 My name is Teertha this the first time i am getting into koha.
15:11 working as a junior librarian i am facing some problem in serial module.
15:12 can anyone please help me understand the purpose and functionality of Display Order in managing frequencies of issue?
15:13 @seen cait2
15:13 huginn teertha: cait2 was last seen in #koha 13 minutes and 56 seconds ago: <cait2> i don't think we are leaving files out, do we?
15:16 caroline teertha: it's to control the order in which the frequencies appear in the drop-down menu
15:16 0 is the top position
15:16 so If there is a frequency that you use a lot, you can put 0 as display order and it will appear at the top so you don't have to look for it
15:18 teertha @caroline: so basically jusr sorting and kind of bookmarking for quick access?
15:18 huginn teertha: downloading the Perl source
15:19 teertha @huginn forget perl source
15:19 huginn teertha: downloading the Perl source
15:19 caroline it's just for sorting
15:24 teertha caroline: thanks :)
15:38 did joined #koha
15:45 did joined #koha
15:46 MarkHofstetter1 joined #koha
15:48 fridolin joined #koha
15:59 caroline Do any of your clients share one copy of an ebook for many sites?
16:00 One of my clients is looking at how to indicate that an ebook "item" is the property of multiple branches
16:01 she says right now they create one "item" per branch but she says it doesn't reflect the actual number of items
16:34 oleonard caroline we do that but then we alter the display so that it doesn't look like an item
16:35 https://search.myacpl.org/cgi-[…]&limit=ccode:DNLD
16:58 fridolin joined #koha
16:59 fridolin left #koha
17:11 cait2 oleonard: haen't been on your catalog in a while - looks really pretty
17:11 oleonard Thanks!
17:12 cait2 we usually try to avoid items...
17:12 putting the link in the record with a link text
17:12 and then also moveit to holdings with jquery if people want to
17:12 or otherwise highlight it visually
17:13 ok, i gotta go outside for a little bit, bye all
17:14 cait2 left #koha
17:18 reiveune bye
17:18 reiveune left #koha
19:01 thd joined #koha
19:59 oleonard No Joubu eh
20:00 thd Only monitor of Joubu
20:00 cait joined #koha
20:00 oleonard Oh I missed him on my list
20:01 Hi cait
20:01 Not sure who all is here for a meeting
20:02 caroline I'm vaguely around... I had an action from last meeting (because I wasn't there lol!)
20:02 thd oleanard: Do you have any experience running multiple versions of PHP on the same webserver in different virtual hosts?
20:02 mkuhn joined #koha
20:02 cait i'd be her
20:02 e
20:02 mkuhn #info Michael Kuhn, Switzerland
20:02 cait i tihnk it hasn't been started yet?
20:03 mkuhn Hello cait, long time I didn't see you here :)
20:03 oleonard thd: No I don't
20:03 cait new nick? ;)
20:03 mkuhn well, old nick
20:04 cait I think Joubu mentioned he would be away for some time end of month
20:04 not sure if that is now already
20:04 * cait tries the bettlejuice trick
20:04 cait Joubu Joubu ... Joubu
20:04 thd I was trying to enable current old version and fully updated versions of the wiki to run on the same server.
20:06 mkuhn If he was called Louis you could call for Louisssssssss (as in Jackie Brown)
20:07 cait it might ont be working
20:08 mkuhn @thd: was ist successful? I run my own Mediawiki software, but only one version at once (because of dependencies etc)
20:08 cait do we have a lot of new stuf fon the agenda? i am going ot check
20:08 huginn mkuhn: downloading the Perl source
20:08 mkuhn I'm here because of the wiki thing
20:08 cait https://wiki.koha-community.or[…]ing_14_April_2021
20:08 we don't have new thing slisted
20:08 but maybe you 2 could talk off-meeting and share on the mailing list?
20:09 mkuhn Caroline Cyr-La-Rose wrote the wiki question is often on the Development IRC meetings agenda, she mentioned this date today
20:10 thd After the last meeting I discussed with Joubu how he had missed character encoding and MediaWiki support issues in transfering data between Postgres and MySQL.
20:10 caroline glad to know my email got some attention :D
20:11 mkuhn I'm running Mediawiki with MariaDB (Mysql), never tried Postgresql
20:11 thd I had basically solved the wiki issue at the end of 2019 but 2020 killed my time.
20:12 cait i'll be off for a while - been a long day (training again) cya later all
20:12 thd Yesterday, I had my first COVID-19 vaccine so I am becoming safer and just dropping everything to support whatever peple want to do with the wiki.
20:13 mkuhn I just saw someone would like to chnage the software so I was asking myself what is exactly the goal of it?
20:14 thd mkuhn: So you have not tried multiple versions of PHP on the save server in different virtual hosts.
20:14 mkuhn Will, for example, a stricter regime be implemented with the new wiki regarding the addition of new articles, their quality, their currentness and their linking to other articles?
20:14 Will some or all data be transferred to the new wiki? Will all the URL change so nothing is findable anymore und their known URL? (except thru the Wayback machine)
20:14 thd mkuhn: No less strict.
20:14 mkuhn: Let me explain the history.
20:15 mkuhn: When the ecomony went bad last time circa 2009 there was a community problem with LibLime where people in charge were especially worried about sharing with the community.
20:15 mkuhn My main concern would be that some data will be lost, all the URL will change etc. so people (like me) won't find things anymore
20:16 thd mkuhn: I put enormous effort into ensuring that nothing will be lost.
20:16 caroline mkuhn: I think the plan is to still have the current wiki available, but archived
20:16 cait yes
20:16 caroline the info that is still relevant would be transferred to the new wiki
20:16 cait it will just change to ol-dwiki or something
20:17 i really think w eneed to loose some stuff :)
20:17 mkuhn If the data and the URL of the current data will be kept, I will be happy
20:17 cait Joubu suggested redirecting/forwarding for th emost prominent pages - sql reports etc.
20:18 thd At the least I intend to fix the archive so that the archive will not be running out of date software and can move forward over time.
20:18 mkuhn Maybe the new wiki could just be called else, so the old URLs could stay. I find it VERY annoying when large site change their URLs and you can't find anything anymore (like the Geman National Library did, for example, and many others, that should know better)
20:19 cait I think if we make sure we have the better and easier to find information in a new wiki, i think that would be better
20:19 but we will figure it out
20:19 i actually wanted to get some reading in :) so leaving now for real
20:19 mkuhn But this would mean there must be a stricter regime... otherwise the new wiki will end up liek the current one
20:20 thd mkuhn: We ended up with what we had not appreciated was only experimental support for Postgres when choosing that database for MediaWiki.
20:20 mkuhn: Any wiki will have outdated content over time.
20:21 mkuhn Sure, but then I don't see no need for a change? If nothing will change?
20:21 thd mkuhn: What is import is to be able to find what is current.
20:21 We need to test various options.
20:21 mkuhn Yes, agree... but this usually needs a stricter regime...
20:22 caroline mkuhn: the need to change is that the mediawiki version we run is out of date and it's too complicated to upgrade it
20:22 (from what I understand)
20:22 mkuhn Even with outdated and old stuff in it, the current wiki is THE Koha source, for me... and sometimes even old/outdated stuff is useful
20:23 Hm, okay... it is really a very old version
20:23 caroline I'm talking about the software itself, not the information
20:23 thd mkuhn: The problem we arrived at by accident was in 2009 in the wake of lack community cooperation the old wiki went down and the untested test implementation of MediaWiki became the only thing.\.
20:23 It is not too complicated to upgrade.
20:24 I may have an upgraded copy at this time next week.
20:25 It was complicated because it definitely required a lot of work.
20:25 mkuhn This would be great, in my case I was always able to upgrade my mediawiki versions (though I didn't ever wait so long)
20:25 thd The problem was never waiting too long.
20:25 mkuhn I see it is Mediawiki 1.16, uhuuh
20:25 thd The problem was choosing the wrong database, Postgres, to test.
20:26 mkuhn OK, I don't see that from the outside, as I said I always used MariaDB (Mysql before)
20:27 caroline Along with the software upgrade, I think the couple of us who had volunteered to clean up the wiki were a bit too daunted to clea it up, hence the suggestion to open a new one and just transfer what we know is correct and go from there
20:27 mkuhn 1.16 was released in 2010
20:28 thd mkuhn: Postgres seemed a more standards compliant choice and we had not found the documents stating that Postgres was not well supported.
20:29 In any wiki we may need to add some tag to pages marking those as current and a different tag to other pages marking those as outdated.
20:29 mkuhn I understand, Postgres was hyped by some, but it never ever got where Mysql/MariaDB is now
20:31 This could be done with a header in each article - but someone would have to maintain these tags
20:31 thd MySQL AB took the deliberate choice of not following some standards to pressure companies into buying proprietary licenses.
20:31 mkuhn: A header would be created magically by an extension.
20:32 mkuhn: Yes maintaining the tags is important.
20:32 mkuhn Yes, but still it is good software - of course I prefer MariaDB
20:32 Maintanining is more regime than now
20:33 thd With appropriate tagging even minimally for current searches could automatically avoid out of date content.
20:34 mkuhn What exactly is out of date? Some people are still using Koha 3.22 or older, as I see in the mailing list
20:35 thd mkuhn: You correctly identify but one of many reasons that old content should be well preserved.
20:35 mkuhn Sometimes I am very happy to find old information about not so current versions that people ask me about
20:35 Yes, as I said - if everything is preserved with the current/unchanged URL I am happy with everything else
20:36 OK, and I admit - I almost never go into the history of wiki pages, so that content could be deleted (as far as i am concerned)
20:37 thd mkuhn: The problem is that we have issues such as multiple pages with installation instructions which are not distinguished as to which applies to the current version and uninformed users can become confused.
20:38 mkuhn I know that problem, but I can only see a solution with that by a stricter regime... more maintenance...
20:38 Who will do that? Who will want to do that?
20:38 thd mkuhn: People use different names for the same function and one does not see what the other has done.
20:38 caroline mkuhn: are you volunteering :D
20:38 thd mkuhn: It can all be done with tagging pages.
20:39 mkuhn Only if I had the time... or only for a defined set of pages...
20:39 thd The number of pages is relatively small.
20:39 caroline it's the same for everyone... no one has time to work on this, hence the current state
20:40 mkuhn Yes, different names for different functions... I struggle with that problem even in the Koha software itself - where it is even worsened by inconsistent translations
20:41 thd The bigger problem as ashima stressed to me a few years ago is migrating the database so that extensions which support faceted tags/categories will even work.
20:43 mkuhn So this means newer mediawiki software supports a better way of tagging? I don't know about that... I still use categories and links in my own installations, ols skool
20:44 thd mkuhn: No MySQL based MediaWiki allows Semantic MediaWiki and many other extensions to run which fail command line tests under Postgres.
20:45 mkuhn OK, I have never implemented Semantic MediaWiki - sounds great (if there is someone who understands how to do it :)
20:46 There are sooo many extensions for Mediawiki - and I have tested and used a lot...
20:46 thd mkuhn: In circa 2009, when the old Dokuwiki went down with LibLime non-cooperation with the community we had not tested enough to discover that command line checks for some things would fail with MediaWiki under Postgres.
20:47 mkuhn What do you mean with "command line checks"?
20:49 thd mkuhn: When installing extensions there can be various command line procedures to ensure that everything is running correctly for some extensions to validate the installation.
20:50 mkuhn: Such checks are not universal but depend upon the extension manager to write them.
20:51 mkuhn: We do not necessarily need Semantic MediaWiki even if it would be nice but we need extensions to generally work and not merely the subset which happens to work fine under Postgres.
20:52 mkuhn I'm fine with going away from Postgres
20:53 thd mkuhn: MediaWiki documentation circa 2008 2009 showed some commitment to database agnosticism but it was unsustainable for the MediaWiki community as a whole and never became anything other than an experiment.
20:53 mkuhn Extensions generally do work if apllied to the correct version - the problem with some extensions is they are sometimes abandoned
20:53 thd mkuhn: I have had success modifying extensions to work in more recent versions.
20:55 mkuhn I remember someone wanted to write support for Postgres when using Koha - it never got anywhere. In my experience it was never a good idea to support everything. Because that always means someone has to maintain it and that's why it always ends the same way...
20:56 thd mkuhn: One advantage in favour of possibly using Dokuwiki is that the software is much less complicated.  However, there is also less help to be had from the much smaller Dokuwiki community fixing broken extensions for which the functionality is a core feature of MediaWiki.
20:56 mkuhn However, I'm not exactly sure If I could help in some way, except in sharing opinions...
20:57 About Dokuwiki I would use it, but I sure wouldn't start installing, configuring, LEARNING it...
20:58 thd mkuhn: Supporting multiple databases should be easy if everything is abstracted to be written in a database agnostic manner and not specially written for each database.
20:59 mkuhn The crucial thing here is the should, I guess. I heard that so many times but it never really happened. In the end one of the software pieces is king and the rest is released - untested...
21:00 thd mkuhn: Dokuwiki like any software has some serious flaws which had frustrated me greatly.  The biggest problem was really people were very poor at remembering to put a tag or a category on a newly created page.
21:01 mkuhn That's human, it seems. That won't change in Dokuwiki, mediawiki, wahteverwiki. This can only be solved through regime - and people don't like that either...
21:02 For my part I wouldn't even be unhappy if nothing changes with the current wiki (but I see the technical problem of using software from 2010, old as the hills).
21:02 thd mkuhn: Dokuwiki has a tag extension which for some period allowed finding "orphan" untagged pages.  However, that feature is reported to have been broken for a long period now.
21:03 mkuhn In mediawiki you can find untagged pages in various ways, through the spcial page - no extensions needed
21:04 You can list all uncategorised categories, uncategorised pages, uncategorised templates etc etc
21:04 thd mkuhn: I had maintained an extension which forced people to choose a category for new pages but that was unpopular because the procedure needed for adding a new category slowed people down.
21:04 mkuhn also orphaned files, orphaned categories, pages, templates etc
21:05 thd mkuhn: Exactly, finding "orphans" is a core feature of MediaWiki but not Dokuwiki.
21:05 mkuhn Yes I understand - people don't like regimes, they like to do what they want...
21:05 thd In Dokuwiki most of the pages people had ever created had been lost.
21:06 mkuhn: The regime should never stand in people's way when they just want to create some content.
21:06 mkuhn As far as I'm concerned, Dokuwiki is not the way to go
21:07 thd mkuhn: It is fine to let people create stuff and clean it later.
21:07 mkuhn However, if the regime stands not before creating content, then at least the content should be maintained AFTER it has been created
21:07 But how does that? People don't like to do it
21:08 who does that? People don't like to do it
21:09 Also some people have very weird ideas how to "create content" (copy&paste for example)
21:09 thd mkuhn: Yes, but the best tools for moving forward with MediaWiki or even maintaining an archive of old stuff will not be available if it cannot be demonstrated upgraded to the current version with no data loss and migrated to MySQL.
21:11 mkuhn: I am going to go back to proving that the migration works without the data loss shortcut which Joubu took after going to the post office now.
21:11 mkuhn As I see in https://wiki.koha-community.or[…]i/Special:Version there are not so many extensions
21:12 But I never updated 1.16 to 1.35 - what a jump
21:13 Btw some extensions are completely out of date, unmaintained, even archived
21:13 thd mkuhn: The core software should work find for upgrading in one jump.  One merely sees each of the incremental and unavoidable error messages at once.
21:14 mkuhn: https://www.mediawiki.org/wiki[…]sion?_In_one_step,_or_in_several_steps?
21:15 mkuhn Mediawiki 1.5 ?? Brrrr
21:16 "you can upgrade in one step, from your old version to the latest stable version" sounds good - then get rid of old extensions... and of course the database thing
21:17 thd mkuhn: MediaWiki 1.16 should have no problems.  Semantic MediaWiki which never worked may have to be dropped and re-added.
21:18 mkuhn: The database thing is a solved problem for over a year but I need to check my work and write some automated tests for verification.
21:18 * thd goes to the post office
21:19 mkuhn I'm now soon going to bed (it's 10 PM over here), so have a nice evening!
22:13 tuxayo Shoots, I missed the meeting although I was free
22:17 caroline tuxayo: don't worry, it didn't really happen
22:17 just a lot of dicussion about the wiki
22:17 (informal)
22:17 tuxayo Good, things happened ^^
23:17 did joined #koha

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

koha1