IRC log for #koha, 2022-10-10

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

All times shown according to UTC.

Time Nick Message
00:55 Dyrcona joined #koha
00:58 Dyrcona joined #koha
04:03 Oak joined #koha
04:55 lds joined #koha
06:04 reiveune joined #koha
06:04 reiveune hello
06:34 alex_ joined #koha
06:34 alex_ Bonjour
06:34 wahanui hey, alex_
06:43 cait joined #koha
06:55 ashimema morning
07:01 cait1 joined #koha
07:06 cait2 joined #koha
08:34 ashimema any k-t-d users here?
08:35 I've generally used it for testing, but do my dev in a good old fashioned dev install running under cgi.. means I spot performance issues rather quick and live reload isn't requied..
08:36 anyone got live reload working in k-t-d?
08:36 i.e. a file watcher to auto restart plack
08:56 paulderscheid[m] Good idea, I always restart manually
08:58 ashimema I'm surprised we don't have an option in k-t-d to just start it with way by default or something
09:02 cait2 also restartig manually
09:26 angela joined #koha
09:43 Guest2735 Hi everyone! I want to be sure about the usage of two tables. Could you please give an idea. As I know, borrower_debarment table is used for holding the information of  borrowers, who are  blocked for some reason. Patrons in this table are still member of the library.   Discharges table is used for holding the information of discharged patrons. Patrons in this table are not a member of thr library anymore. Is this right?
09:56 mtj ashimema: re: ktd autorestart, modify /etc/koha/plack.psgi
09:57 76     enable "+Koha::Middleware::SetEnv";
09:57 77     enable "+Koha::Middleware::RealIP";
09:57 78     enable "Refresh", cooldown => 3;
09:57 the refresh seems to work ok
09:58 https://metacpan.org/pod/Plack[…]ddleware::Refresh
09:58 i think ktd should be using it by default
09:59 i think shotgun module has a problem with starman :/
09:59 https://metacpan.org/pod/Plack::Loader::Shotgun
09:59 ashimema yeah.. I reckon we should set it by default on k-t-d.. or at least give us a really simple option to enable/disable it in k-t-d without having to modify plack.psgi
10:03 so the question is.. should we add that to plack.psgi inside a block that detects dev mode or something?
10:03 and then set a dev mode env in k-t-d..
10:03 or is that too invasive to core koha code/packages
10:05 mtj enable "Refresh", cooldown => 3 if ($ENV{DEV_INSTALL} ;
10:05 seems to work ^
10:06 will make many people happy
10:07 ashimema indeed 🙂
10:10 mtj i can log a bug
10:10 ashimema awesome..
10:10 I can SO/QA 🙂
10:10 nice one mtj
10:22 cait1 joined #koha
10:39 oleonard joined #koha
10:42 oleonard Hi #koha
11:00 tundunf joined #koha
11:00 tundunf hi all
11:03 oleonard Hi tundunf
11:04 * cait1 waves
11:18 ashimema anyone here used 'here documents' for adding notice templaces to updatedatabase?
11:18 especially with TT syntax templates?
11:25 cait ?
11:25 paulderscheid[m] what are you aiming at ashimema ?
11:25 cait yes, what is the problem you are trying to solve :)
11:25 *get on her support hat*
11:25 ashimema as in.. it's a right pain adding in lots of \n\n\n and spacing..
11:26 her docs are much easier to read and write..
11:26 paulderscheid[m] I got a tip for you
11:26 ashimema but.. when I try to use them the % get interpretted
11:26 and my brain is failing to see the solution.. likely just not had enough coffee yet
11:26 fire away paulderscheid
11:26 cait maybe not the same tip paul has, but what I do: use export as INSERT from workbench...
11:27 then I get something properly formatted that I can use with SQL
11:27 ashimema see https://bugs.koha-community.or[…]attachment=141573 for an example of what I mean
11:27 but the update fails at the minute
11:27 cait maybe if you dump the table with your notice, that would work too
11:27 paulderscheid[m] my $doc <<~ "EOL"
11:27 template...
11:27 EOL
11:27 wahanui somebody said template was <backend_directory>__PERL_MODULE_DIR__​/Koha/Illbackends</backend_directory>, so just to that dir, and everything below is seen as a backend
11:27 ashimema ah.. I've never used workbench
11:27 interesting `~`.. not see that one
11:28 paulderscheid[m] But I see that's not really your problem right now..
11:28 cait ashimema: just trying to say.. i let the dabase do the formatting for me, but was paul has looks promising
11:28 paulderscheid[m] So you can indent
11:28 ashimema hmm
11:28 the ~ made no difference..
11:29 oleonard wahanui: forget template
11:29 wahanui oleonard: I forgot template
11:29 cait maybe there are example sin the existing updates?
11:29 m_ruddy[m] joined #koha
11:29 m_ruddy[m] left #koha
11:29 cait quick meeting bbiab
11:29 alex_ joined #koha
11:29 paulderscheid[m] Here's an example for heredocs
11:29 ashimema hmm.. maybe it's the `.` char it's not liking
11:29 paulderscheid[m] https://github.com/LMSCloudPau[…]omReservations.pm
11:30 tundunf sorry to but in, and it's probably a long shot anyway lol, but is anyone aware of a bug that could be sending out Item Due notices even though the items aren't due? (e.g. no reservations, active accounts, unlimited automatic renewals enabled)
11:31 I've raised a ticket but just wondering if this is by any chance affecting other users in the same way ;_;
11:33 oleonard tundunf: Yes, we had to turn off item due notices because it didn't work with automatic renewals
11:34 tundunf I see, we turned off the PREDUE notices for this reason but we hadn't experienced any issues with the DUE notices until now
11:36 ashimema sorted my quoting.. I forgot the quote the whole here doc!..
11:37 I've worked on the autorenewal notices a fair bit recently... but I'm not recalling any solutions I'm afraid.
11:38 oleonard tundunf: Actually I'm wrong, it was the PREDUE notices we tried. I assume it's the same issue for both, though?
11:38 It was a disaster.
11:41 tundunf thanks both :')
11:41 ashimema Our support chaps might have some better ideas.. hopefully
11:42 tundunf our worry is that if we disable the Item Due notices there won't be a "buffer" before users receive overdue notices on reserved items or before they expire
11:42 ashimema It's an area I get pulled into and fix things in fairly frequently at the moment .. but not enough to know like the back of my hand right now
11:42 oleonard ashimema: I'd buy your support chaps a round if they did
11:43 tundunf thanks @ashimema!
11:43 ashimema Autorenewals is the challenge.. I think the other notice are generally ok these days.. but autorenewals adding into the mix, especially with the digest version, gets really confusing as its setup more as a summary than a call to action notice
11:48 tundunf That makes sense
12:01 cait hm initially that worked... because the library who did the auto-reneawls used the pre-dues... but that has been a long while ago
12:01 now I hope that won't expode for us on the nextu pdate
12:01 what's the issue exactly? shoudl the auto-renew not be sent before the due notices?
12:03 oleonard cait: The issue is that the DUE and PREDUE notices shouldn't go out if something is eligible for automatic renewal
12:03 I would think this could be solved by the timing of the notices, but I don't really understand how it works.
12:09 cait I see - maybe having a switch to exclude the auto-renew marked issues coudl be helpful then?
12:09 as a cronjob parameter
12:54 ashimema Or just more examples of how to time it all..
12:54 I have a feeling if the circ rules are carefully defined then it's not an issue
13:16 Dyrcona joined #koha
13:23 cait hm no new bugzilla emails?
13:30 paulderscheid[m] Anyone seen this before?
13:30 [pauld@rocky-4gb-fsn1-1 ~]$ git bz
13:30 fatal: 'bz' appears to be a git command, but we were not
13:30 able to execute it. Maybe git-bz is broken?
13:40 jajm paulderscheid[m]: you need python2
13:41 ashimema ah yes.. I got hit hard by that on my new laptop..
13:41 I even stuck it on the dev meeting agenda.. I think we should perhaps be considering alternatives at this point.. it's getting harder to install Python2 now
13:42 my arch really didn't want to compile it.
13:44 paulderscheid[m] Yeah, support's getting dropped..
13:51 jajm it shouldn't be that hard to rewrite git-bz for python 3, right ?... :)
13:52 ashimema haha.. I've looked.. but I'm not python coder.. I just dabble occasionally..
13:52 I was more wondering if this might give us the push we need to consider alternative workflows like git pull/merge requests rather than patches
13:53 there are some lovely cli wrappers for github and gitlab out there 😜
13:53 bet it wouldn't be horrific to adapt one for gittea either..
13:57 paulderscheid[m] ashimema: I'll drink to that
14:03 ashimema it would be a big change.. we should revisit the whole bugzilla vs the rest issue trackers some time for sure...
14:04 I've become accustomed to bugzilla over the years.. but it was already an old fashioned way of doing things when i first came into Koha land 😜
14:18 ashtart joined #koha
14:32 reiveune left #koha
14:48 oleonard-away I can't remember... Is a move away from Bugzilla partly stalled by the difficulty in migrating old data?
14:48 ashimema yup
14:49 and dashboard.koha-commnuity.org uses the bugzilla db directly
14:49 and splitter.koha-community.org also uses the bugzilla db
14:49 so those would both need accounting for too
14:52 * oleonard sighs in MediaWiki
15:03 tundunf left #koha
15:10 tuxayo ashimema: «my arch really didn't want to compile it.» Lol me too, some post compile automated test fail.
15:10 ashtart left #koha
15:12 tuxayo «Is a move away from Bugzilla partly stalled by the difficulty in migrating old data»
15:12 Do alternatives have a way to keep being able to link tickets hierarchically (non-libre feature for gitlab) and be able to seamlessly add or rebase patches on any submission?
15:12 Easily reviving old patches and adding followups is strong point of our current workflow.
15:13 ashimema the CLI one
15:13 I ended up setting it to ignore tests
15:13 as only one was failing and it was for a feature we're not using
15:14 seamlessly rebase.. gitlab does that way better than we do with BZ..
15:14 most of the time..
15:14 wahanui i think most of the time is doing the test plan. And some is conflict resolution.
15:14 tuxayo I wonder if it would work to open a new merge request and post it in the ticket. But the old one would be open, it would be a mess.
15:14 ashimema there are certainly caveats.. a move to anything other than bugzilla will be hard
15:15 tuxayo > gitlab does that way better than we do with BZ..
15:15 It would avoid the sha1 missing error. But still the same conflicts to solve.
15:15 ashimema gitlab has an automated rebase option for open merge requests that are seen to conflict
15:15 and.. you can just push over the branch in the merge request for other cases..
15:15 tuxayo And one can't push to the same merge request the result I think
15:15 ashimema rebases are significantly easier in gitlab and github that they are for bz
15:15 bz is insanely noisy with patches.
15:16 tuxayo > gitlab has an automated rebase option for open merge requests that are seen to conflict
15:16 How does it deal with conflict. IIUC, it can do it when there is no conflict.
15:17 > you can just push over the branch in the merge request for other cases..
15:17 How? Maybe by adding everyone with developer roles in the prohect?
15:17 ashimema huge projects like gnome have moved away from bz
15:17 tuxayo By default project devs can push in the merge request from others.
15:17 ashimema we'd need to investigate it more..
15:18 tuxayo Indeed, there are certainly big gains in moving. And hopefully ways to keep the current advantage's
15:18 ashimema I'm not suggesting we go jump to anything in particular.. more suggesting that our bugzilla workflow is really putting off new developers
15:18 tuxayo True
15:18 ashimema including that fact that you can't easily install git-bz now on any modern distro.. you have to jump through hoops to install python 2 which is well and truly end of life
15:18 tuxayo Installing and learning git bz alone is filtering out some devs.
15:19 ashimema yup
15:19 Amit_Gupta joined #koha
15:24 tuxayo Maybe there are external tools that plug into GitLab or Gitea that can add ticket dependency and graph for it.
15:24 There are some for GitHub that add other project management feature. So maybe.
15:25 Or maybe now that I'm getting some income, I could fund a bounty for having deps and graphs on gitea: https://github.com/go-gitea/gitea/issues/8794
15:28 cait git bz is set up if you use ktd
15:29 i am not sure using any other dev environment makes any sense for new devs
15:29 too many moving bits and pieces
15:29 but bugzilla is kinda dead... so at some point migration is probably a topic we cannot avoid
15:29 tuxayo Yes
15:30 ashimema I thought the recommended approach was still to do your git work outside of k-t-d
15:30 tuxayo > i am not sure using any other dev environment makes any sense for new devs
15:30 I forgot if I have actual reasons to use bit-bz outside ktd
15:30 cait ashimema: i do mine inside
15:30 ashimema else you get permissions issues and things
15:30 cait i get the permission issues if i don't
15:30 tuxayo Ah, everything own by root
15:30 ashimema interesting
15:30 I do all git outside
15:31 cait do you also run qa tools outside?
15:31 tuxayo Well if you do everything in ktd, absolutely everything, the permissions might be a mess but it will work.
15:31 cait to me it semes like as soon as I run those my permissions get borked
15:31 ashimema yup
15:31 tuxayo > do you also run qa tools outside?
15:31 Nope, too many dependencies
15:31 So in ktd
15:31 cait that's the only thing I really don't like about ktd
15:31 tuxayo To avoid installing too much unofficial perl packages in my distribution.
15:32 cait all the times I write up a prety nice comit message... and then it doesn't save because permission error
15:32 and then i start again... and I can guarantee you, the second is not as nice
15:32 or the many error from my editor because it can not save my changes...
15:32 having a workign solution for that would be really nice
15:32 tuxayo I still can do permission issue even if you do everything from the root user of ktd?
15:33 cait i am not even sure what that sentence means :)
15:34 cait1 joined #koha
15:34 tuxayo > but bugzilla is kinda dead... so at some point migration is probably a topic we cannot avoid
15:34 At least bugzilla is in perl, not python 2. I hope it will continue to work with new perls for the time being ^^"
15:34 > i am not even sure what that sentence means :)
15:34 indeed! Sorry
15:34 It still can cause permission issues even if you do everything from the root user of ktd?
15:35 * tuxayo need more proofread
15:36 tuxayo *needs ><
15:38 ashimema I'm on the bugzilla irc channels.. there's just no one with any time to look at it and with the mass exodus to other platforms I think there's very little impetus to keep supporting it
15:39 it'll keep functioning but new features are constantly blocked.. I've tried to submit things to it a few times.. and the answer is always the same.. 'wait for the next version'.. which has be 'on the verge of release' for like 5 years at least.
15:39 cait1 joined #koha
15:40 cait i think my issue is: what is the root user of tkd?
15:47 cait1 joined #koha
15:58 tuxayo cait: when from your main system, you run `kshell` you end up in ktd, and the username is root, it's the name of the admin user in Linux.
15:59 cait I use a graphic editor for editing, so that has to happen outside ktd...
15:59 i run all other things inside
16:00 tuxayo And every file write including those done by git and the qa tools, will make the files belong to root. And from your system by default you have a normal user. Not admin one. And when running git commands or qa tools from the outside of ktd, it will try to change or delete file from the admin user. And won't have the permissions.
16:00 > I use a graphic editor for editing, so that has to happen outside ktd...
16:00 Same.
16:00 ashimema which is why I do everything outside of k-t-d
16:01 bar running tests
16:01 how do you run qa tools inside k-t-d and not screw up permissions then tuxayo
16:01 tuxayo > i run all other things inside
16:01 Weird that you have permission issues. The one expectable issues are with your text editor.
16:01 ashimema you've explained the issue well..
16:01 but not a solution
16:02 tuxayo > which is why I do everything outside of k-t-d
16:02 And then the issue of installing git bz comes up.
16:02 > how do you run qa tools inside k-t-d and not screw up permissions then tuxayo
16:02 I don't know why I never had issues!
16:02 ashimema yup
16:03 I think I will always code outside of k-t-d.. there's too much personal config to add into k-t-d
16:03 tuxayo > you've explained the issue well..
16:03 Though incomplete since I realize my setup should be messed up when running qa tools but it's not ^^"
16:03 ashimema git config, vim config, signing keys etc
16:03 I think I run the QA tools as the kshell user not root on the odd occasion I do run them in k-t-d
16:05 isn't `echo 'export LOCAL_USER_ID=$(id -u)' >> ~/.bashrc`
16:05 tuxayo > I think I will always code outside of k-t-d..
16:05 For personal config and also because a lot of people are using a graphical editor, so it should be well supported. I don't know if there is config for docker to use the same user id for root than for the regular user outside ktd.
16:05 ashimema to do with getting permissions right?
16:05 I've never totally wrapped my head around it.
16:06 cait i really want do deal with installing qa test tools and git bz manually again...
16:07 i do have certificate sfor pushing to remote externally too
16:08 pastebot "tuxayo" at 127.0.0.1 pasted "Here is what seems to be the finallity of LOCAL_USER_ID" (11 lines) at https://paste.koha-community.org/16188
16:09 ashimema https://blog.gougousis.net/fil[…]l-side-of-docker/ is an interesting read
16:10 tuxayo ashimema: So LOCAL_USER_ID doesn't seem to do anything about the .git and code directory permissions.
16:10 ashimema: thanks for the find
16:13 Maybe LOCAL_USER_ID should be used to create a user in ktd and kshell should lead to that user instead of root.
16:13 ashimema when tcohen is back we can maybe talk about this stuff with more knowledge
16:14 sounds like it's plausible to me..
16:15 tuxayo the start of ktd doesn't a bunch of stuff so creating a user should be doable.
16:15 «the default user under which a container is executed is root»
16:15 Hopefully that means it's possible to change that
16:16 Last resort we can have some ugly stuff in root's bash_profile like `su - my_user`
16:16 See you latter, gotta go
16:19 ashimema I commented on your issue on gitlab tuxayo https://gitlab.com/koha-commun[…]cker/-/issues/270
16:19 see ya
16:39 ashtart joined #koha
16:50 mtj ashimema: bug 31729, sorry its a bit late...
16:50 huginn` Bug https://bugs.koha-community.or[…]_bug.cgi?id=31729 enhancement, P5 - low, ---, mtj, Needs Signoff , Enable automatic filesystem refresh in Plack
16:59 ashimema Nice one mtj
17:02 Hmm, on my phone right now, but will test later
17:02 Quick look.. isn't there a closing brace missing in the if condition?
17:03 Or maybe I'm just getting my psgi syntax wrong
17:32 mtj ah, yes :)
17:56 oleonard There's not really a way to check a user's permissions when working with system preferences is there...
18:21 cait left #koha
18:47 cait1 oleonard: what are you trying to do?
18:48 oleonard Bug 31730
18:48 huginn` Bug https://bugs.koha-community.or[…]_bug.cgi?id=31730 enhancement, P5 - low, ---, oleonard, ASSIGNED , Link to authorized value interface when an authval is mentioned in preferences
18:48 oleonard Theoretically a person could have permission to edit preferences but not authorized values
18:48 cait1 hmm
18:49 oleonard But I think I'll just be fine with those people hitting an "access denied" page
18:49 cait1 yeah, i tend to agree
19:06 paulderscheid[m] For the record: I compiled python2.7 and everything’s fine
19:07 But seems like I started something ^^
19:09 oleonard We can hope :)
19:11 paulderscheid[m] o/
19:20 cait1 oleonard: isn't today a holiday for you?
21:52 ashtart joined #koha

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

koha1