IRC log for #koha, 2014-09-03

All times shown according to UTC.

Time S Nick Message
00:05 papa joined #koha
00:08 eternalsword next general meeting is in the past. didn't realize koha did time travel too :P
00:09 Topic for #koha is now Koha 3.16.3 is available! Next general meeting was 13 and 14 August 2014 at 22:00 and 15:00 UTC. Welcome to the IRC home of Koha Please use for pastes.
00:09 eythian there we go
00:10 * dcook seems to recall having read an article about time travel this morning
00:10 dcook Or maybe that is tomorrow morning. I have problems keeping my schedule straight when it comes to time travel.
00:10 dcook Or was that yesterday? Timezones...
00:11 dcook Yes, Oz and Nz are in the future...
00:11 dcook I forget these things sometimes.
00:30 eythian @wunder nzwn
00:30 huginn eythian: The current temperature in Wellington, New Zealand is 17.0°C (12:00 PM NZST on September 03, 2014). Conditions: Clear. Humidity: 68%. Dew Point: 11.0°C. Pressure: 29.86 in 1011 hPa (Steady).
00:31 eythian that is the warmest it's been here in a very long time
00:37 dcook @seen khall
00:37 huginn dcook: khall was last seen in #koha 5 days, 10 hours, 34 minutes, and 10 seconds ago: <khall> sounds like you are missing one of the qt5 dev packages most likely
00:37 dcook @seen khall_away
00:37 huginn dcook: khall_away was last seen in #koha 1 year, 15 weeks, 0 days, 2 hours, 34 minutes, and 44 seconds ago: <khall_away> happy bday rangi!
00:38 dcook @later tell khall Quick question. Do you have any plans to work on bug 7710 soon? I'm not sure what the best way to do it is, but your patch seems reasonable. Happy to test a rebased patch.
00:38 huginn dcook: The operation succeeded.
00:55 tcohen joined #koha
01:26 mtompset anyone know how to trigger the template?
01:30 eythian * olly is now itching to write Class::Tiny::Simple
01:30 cjh :D
01:31 pastebot "tcohen" at pasted "jcamins: getting closer, look at this output" (23 lines) at
01:31 eythian mtompset: by going to opac/ I expect
01:31 mtompset I was hoping to get there without cut and paste... you know, like a user.
01:32 tcohen mtompset: grep -R opac/
01:32 eythian koha-tmpl/opac-tmpl/prog/en/modules/
01:32 eythian 1236:        <p><a href="[% biblionumber %]">More details</a></p>
01:32 eythian just trace it backwards until you find somewhere you know
01:37 dcook If you have a record with more than 3 serials, there should probably be a link to that template which will show you all the issues for a record
01:37 dcook Yeah, what eythian said :p
01:39 mtompset staff client, home, serials, submit by the title text box, click one of the subscriptions, determine the biblionumber, opac, search for the biblio number and get to an opac details page, click on more details.
01:39 mtompset That's more of what I was looking for... instructions for test dummies. :)
01:40 dcook You could probably get away with half that but sure
01:40 dcook That works
01:54 mtompset @later tell oleonard Bug 12862 is a trivial one for you. :)
01:54 huginn mtompset: The operation succeeded.
02:16 jmsasse joined #koha
02:33 dcook "The caller can, of course,
02:33 dcook # break this assumption by playing silly buggers, but that's
02:33 dcook # harder to do than doing it properly"
02:34 dcook "Silly buggers"?
02:34 dcook See something different in Koha every day..
02:36 eythian how long have you been in Australia that you haven't heard that term?
02:36 dcook Oh, I think I've heard it before. Just surprised to see it in code comments :p
02:36 eythian ah :)
02:38 eythian I'm tempted to semantically fix these authority records to move a small step past marc stupidity.
02:39 eythian "Heading-use-main-or-added-entry":"|"
02:39 eythian that's just silly
02:39 dcook O_O
02:39 eythian because the MARC rules say it's in 008/15 and must be 'a', 'b', or '|'.
02:40 eythian in related news, I now have authorities going into elasticsearch
02:40 pianohacker eythian++
02:41 dcook Hmm, even more fun... people sometimes will use # or / instead of |
02:41 dcook also...
02:41 dcook eythian++
02:41 eythian see, I always thought that '#' was a synonym for ' '
02:41 dcook Ah wait
02:42 dcook Yeah, I'm totes wrong
02:42 dcook # and / are the same as ' '
02:42 dcook My bad :)
02:42 dcook Hunger interfering with brain
02:42 dcook brb
02:42 eythian I can totally see how they'd get mixed up though
02:42 eythian because # is used everywhere in the documentation
02:59 indradg joined #koha
02:59 indradg good morning #koha
03:03 indradg I just ran `make test` against master, ended up getting -> make: *** [test_dynamic] Error 137
03:04 indradg just before that line was -> t/Installer_PerlModules.t ........... 1/17 Killed
03:04 eythian probably out of memory or something
03:04 indradg was that what the test barf up on?
03:04 mtompset free -m
03:04 mtompset check eythian's hypothesis.
03:05 eythian err 137 means SIGKILL and coredump
03:05 eythian which is probably the OOM killer
03:05 eythian syslog will tell you more
03:06 mtompset Does syslog tell you that? -- perhaps it was the poor VM environment I was in before where it didn't.
03:06 indradg devtest1@vps:~/kohaclone$ free -m
03:06 indradg total       used       free     shared    buffers     cached
03:06 indradg Mem:          4096       3706        389          0          0       3587
03:06 indradg -/+ buffers/cache:        118       3977
03:06 indradg Swap:         4096         15       4080
03:06 pastebot "indradg" at pasted "devtest1@vps:~/kohaclone$ free" (5 lines) at
03:06 mtompset What else are you running to hog so much memory?
03:07 mtompset I wouldn't expect a barf with 389MB free.
03:07 eythian you're not reading it right, mtompset
03:07 eythian there's 3.9GB free
03:07 eythian which is even more perplexing
03:07 eythian unless there's a bug that causes it to swell up and die
03:08 eythian I'd look in syslog to see what happened
03:08 mtompset oh my... that is perplexing. You're right, I misread.
03:08 mtompset no... wait... 389MB... 4096 = 4GB RAM.
03:08 eythian you're reading the wrong line
03:08 mtompset with 4GB Swap.
03:08 eythian although, with 3.5GB cached, it doesn't look like it's had recent memory exhaustion
03:09 eythian so it may have been something else causing the sigkill.
03:09 pianohacker not selinux or apparmor, right?
03:09 indradg mtompset: nope! syslog clean as a slate
03:09 mtompset 4GB total, ~3.7GB used, 389MB free. -- with ~3.6GB cached.
03:11 eythian right, which means that it's effectively 3977MB free
03:11 mtompset because you can flush cache... right.
03:11 eythian which is why you read the -/+ buffers/cache line, as anything else won't give you the effective values
03:11 eythian well, the system does it
03:11 mtompset right.
03:12 mtompset Are you running a zebra full reindex when you tried to make test? ;)
03:13 eythian if there's no info in syslog, try dmesg
03:14 pastebot "indradg" at pasted "ran `make test` again with different result this time, but unlike it did not barf at t/Installer_PerlModules.t ........... 1/17 Killed" (154 lines) at
03:14 eythian
03:14 eythian something is wrong with your system
03:15 eythian it's killing things for some reason.
03:15 eythian does dmesg tell you?
03:18 jcamins @later tell tcohen I would like to add my vote to rangi's proposal for keeping an un-UTF8-ed branch.
03:18 huginn jcamins: The operation succeeded.
03:19 eythian an "un-UTF8-ed branch" ?
03:19 jcamins eythian: as opposed to a UTF8-ed branch, yeah.
03:19 pianohacker in case odd encoding bugs pop up, or what?
03:19 eythian that was the least helpful explaination ever :)
03:21 jcamins pianohacker: in case this attempt to make Koha Unicode-safe is as disastrous as previous attempts, it provides a safety valve.
03:21 eythian ah
03:42 indradg eythian, mtompset the kills are happening at random on make test. something funky here. prolly some sort of admin override by the VPS provider.
03:42 eythian could be. It's likely dmesg will tell you.
03:43 mtompset indradg: Do you know if your VPS is using vserver? :)
03:43 indradg dmesg and syslog says nothing, no mails in /var/mail/root :(
03:43 indradg mtompset: OpenVZ
03:44 mtompset No idea. Because that sort of behaviour only happened to me on a hosting provider which used vserver.
03:44 pianohacker indradg: what vps provider?
03:45 indradg pianohacker:
03:47 mtompset --- my guess is they've loaded more people than the server can handle, and other VMs are causing yours to randomly sigkill.
03:48 eythian oh, openvz
03:48 eythian it could be anything then
03:49 pianohacker yeah... there's a reason openvz vms are often cheaper than xen ones
03:49 eythian your processes are visible to the host, so if it thinks you're using too much cpu or whatever, it can kill the process.
03:49 eythian they're easier to safely oversubscribe.
03:49 eythian you could try doing it with 'nice' to make it a lower priority.
03:59 wnickc joined #koha
04:00 indradg nah.... even nice didn't cut it. filed a support ticket. thankfully I paid only for 1 month :-)
04:02 * dcook thinks it's been a while since anyone has touched ""...
04:04 eythian it's got the old copyright notice
04:04 eythian apparently I was the last person to touch it, 2 weeks ago :)
04:04 dcook hehe
04:05 dcook I've been thinking more about frameworks.
04:06 dcook I was working on a tool to update frameworks using the XML that bgkriegel put together, but it's cumbersome.
04:06 dcook It seems to me that it would make sense to make an "untouchable" framework which end-users never see.
04:06 dcook But that we can update during releases
04:07 dcook If there's an untouchable framework, it gets easier to do comparisons via SQL
04:07 dcook Then you can update to your heart's content, put in news warnings, etc.
04:08 dcook Oooh...
04:08 eythian hmm yeah
04:08 dcook Maybe the way to do that is just to add a framework and not reference it in biblio_framework.
04:08 dcook thoughts, eythian?
04:09 eythian it's probably not a bad idea to do something like that
04:09 eythian I don't know the area well enough to have any strong opinions :)
04:09 dcook Fair enough :)
04:09 dcook I appreciate you saying it's not insane at least ;)
04:12 dcook Hmm..
04:13 dcook You know...I may have been a fool all this time
04:13 dcook /cgi-bin/koha/admin/
04:13 dcook We already have a script that is supposed to be checking the MARC bibliographic frameworks
04:13 dcook Hmm, maybe not a fool..
04:14 eythian it's not super
04:14 dcook No, I've always avoided it as I didn't see much good with using it
04:14 eythian I'm not sure it's feasible to validate the whole thing reliably, really
04:14 eythian well, it's so when you change a framework, you can run that and see that you haven't done anything too dumb
04:14 dcook Mmm, fair enough
04:15 dcook I was thinking of adding a button/link on cgi-bin/koha/admin/
04:15 dcook Something about updating or checking against the official MARC
04:15 dcook Which would be represented by the hidden framework
04:16 mtompset Have a great day, #koha dcook eythian.
04:23 dcook Hmm, the hidden framework idea seems to work so far
04:23 dcook Hmm... unless someone tries to make a framework with the same frameworkcode...
04:29 dcook C4::Koha::getframeworks() sure has some awful variable names..
04:29 dcook return ( \%itemtypes );
04:29 dcook kthx
04:31 dcook Hmm, maybe I'm thinking too small..
04:49 dcook Hmm, maybe it's not that slow..
04:49 dcook Much much slower using DBIC
04:50 eythian yeah, we need to find a way to make DBIC's initialisation be faster.
04:53 dcook Yeah, I'd love to know where I'm losing my time. I want to say I already tried using NYTProf and it didn't help me out too much
04:54 pianohacker oh good, I'm not crazy, saw a huge delay in schema loading...
04:58 francesca joined #koha
04:59 dcook I despise marc frameworks...
04:59 dcook Maybe MARC in general too :P
05:01 dcook Mostly just how incoming records will lose data if the corresponding tag isn't in the framework...
05:01 dcook Although that also makes some sense
05:03 dcook But fixing that would require re-writing the cataloguing editor and that's not happening..
05:05 indradg joined #koha
05:05 eythian talk to pianohacker about that :)
05:05 dcook hehe. already have a while back.
05:05 dcook I think I've finally reached my tipping point though.
05:08 dcook It makes sense to have frameworks where you can customize the labels and the visibility (as well as all the other stuff... authorities, plugins, etc.)
05:08 dcook But since it's MARC... it should always at least have all the fields/subfields from the spec available...
05:09 cait joined #koha
05:09 eythian hi cait
05:09 dcook hey cait
05:09 cait :)
05:12 pianohacker dcook: don't worry, the new editor doesn't throw away tags
05:12 dcook That's good to hear :)
05:13 dcook Although I imagine none of my libraries will use the new editor :/
05:13 dcook Well, maybe a few
05:32 indradg hi cait
05:32 cait hi indradg
05:33 indradg eythian: there you go! Vimal wants new packages :D
05:34 eythian he sure does
05:34 * eythian is off now, later all
05:38 pianohacker Out, bye
05:43 dcook hehe
05:43 dcook laters eythian
05:48 pianohacker joined #koha
06:33 reiveune joined #koha
06:33 reiveune hello
06:39 cait left #koha
07:04 ashimema joined #koha
07:15 indradg_ joined #koha
07:15 indradg_ joined #koha
07:17 paul_p joined #koha
07:21 Joubu hello #koha
07:22 indradg joined #koha
07:24 ashimema meeting was super quick last night by the looks
07:24 ashimema morning #koha, Joubu
07:25 drojf joined #koha
07:26 drojf good morning #koha
07:28 indradg morning ashimema drojf
07:29 drojf hi indradg
07:47 kivilahtio joined #koha
07:48 kivilahtio Hello! Is anyone using a log monitoring solution to get automatic alerts from failed cronjobs?
07:49 kivilahtio I am facing such issues and made the news again with failing to create predue notifications for 3 days.
07:49 Joubu kivilahtio: what is the error ?
07:50 kivilahtio Looking into logging best practices and looks like logwarn is a recommended way to go and planning the nagios plugins and configuraitions
07:50 Joubu kivilahtio: I think we just set an email address in the crontab
07:50 kivilahtio Joubu: I would like to know if we get errors, not if get successfull runs. For ex. runs every minute
07:51 kivilahtio Joubu: I didin't get any errors, the cronjobs just didin't run
07:51 Joubu kivilahtio: you can monitor the mailbox :p
07:51 kivilahtio Joubu: It could be my custom wrapper which times their execution and collects logs to respective files
07:51 Joubu ha...
07:51 kivilahtio it has a locking mechanism top prevent cronjobs from restarting before they close
07:51 kivilahtio and a single lock file to prevent running all cronjobs
07:52 kivilahtio but i dont understand how the disable flag was set?
07:52 indradg @seen eythian
07:52 huginn indradg: eythian was last seen in #koha 2 hours, 18 minutes, and 18 seconds ago: * eythian is off now, later all
07:52 kivilahtio anyway, to make sure that never happens and save me some sleep, I am looking into a more intelligent solutino
07:52 indradg @seen mtompset
07:52 huginn indradg: mtompset was last seen in #koha 3 hours, 36 minutes, and 13 seconds ago: <mtompset> Have a great day, #koha dcook eythian.
07:53 kivilahtio Joubu: Also we noticed it like a week later and all the syslogs had already rotated out of reach :(
07:53 Joubu arf
07:53 kivilahtio well that wont happen again I hope :)
07:53 Joubu kivilahtio: so the advance_notices cronjob did not run because of what?
07:53 kivilahtio Joubu: I don't know
07:54 kivilahtio Joubu: I can only guess, it was not the only cronjobs, some of the did run, some of them not
07:54 kivilahtio but no way to find out, I'll set the monitoring server to check the logs periodically so I can catch future issues
07:54 Joubu kivilahtio: did you add these cronjob at the same time? Are they working now?
07:54 kivilahtio Joubu: yes, no panic
07:55 kivilahtio Joubu: just lots of bad PR from pissed off customers getting rather large fines
07:55 Joubu kivilahtio: Last week, we had the problem on 1 machine. crontab -e; save+quit solves the problem...
07:56 kivilahtio I'll share this script set as soon as I can, maybe you can find it helpful as well.
07:56 Joubu kivilahtio: maybe your custom script does not log the right things :)
07:56 kivilahtio Joubu: how are you doing? We just moved to our own terrace apartment
07:56 kivilahtio 2&>> ?
07:57 kivilahtio its pretty neat
07:57 Joubu yes, and catch errors if the script crashes
07:57 kivilahtio there's not much to crash :)
07:58 Joubu kivilahtio: I am fine :) excited to go in Argentina next month!
07:58 kivilahtio
07:59 kivilahtio Joubu: wow, happy to hear that!
07:59 indradg @later tell eythian `make test` finally worked, after supported moved over a bunch of us to a new node. Only t/00-checkdatabase-version.t FAILED. :)
07:59 huginn indradg: The operation succeeded.
07:59 kivilahtio Joubu: It must be so cool! I have never been to South America
07:59 Joubu me neither!
07:59 indradg @later tell eythian s/supported/sysadmin
07:59 huginn indradg: The operation succeeded.
07:59 Joubu Shouldn't you redirect stdout to stderr?
08:00 kivilahtio Joubu: Just don't joke about soccer
08:00 kivilahtio I am?
08:00 kivilahtio $croncommand 2&>> $logpath
08:01 Joubu I never used this syntax, I always use 2>&1
08:01 kivilahtio I need appending, not truncation
08:01 kivilahtio this syntax is regardless pretty whacky :)
08:11 Joubu kivilahtio: did you see you use $logfile and $logpath?
08:12 kivilahtio $logfile is actually cronjob/
08:12 Joubu echo "Lockfile present" >> $logfile
08:12 kivilahtio it has the directory embedded due to cronjobs being in two separate folders, migration_tools and cronjobs
08:12 kivilahtio I made those lockfile additions after the accident
08:13 Joubu yes but you write into $logfile
08:13 kivilahtio yeah there is a bug :)
08:13 kivilahtio thanks for spotting it
08:13 Joubu but you are modifying the perl script!?
08:14 kivilahtio which perl script?
08:14 Joubu cronjob/
08:14 kivilahtio i am not modifying any cronjobs scripts
08:14 Joubu if $logfile = cronjob/
08:14 kivilahtio this is just a wrapper to make sure those cronjobs dont run over each others and time their execution and log it
08:15 Joubu yes, I understood that :) Just wondering the impact of writing into $logfile
08:15 kivilahtio Joubu: well it writes to somewhere where it shouldnt
08:15 kivilahtio depending on what crontab thinks of it's cwd()
08:16 kivilahtio apparently there is no such thing as a simple fix :)
08:16 Joubu if cwd == $KOHA_PATH, it could explain why your scripts fail
08:16 kivilahtio i made that modification after the accident
08:16 Joubu ok
08:16 kivilahtio i dont think that branch has been triggered after that
08:17 kivilahtio but thanks for the tip. I've got one for you!      "Don't cook it, bake it naked!"
08:19 Joubu thanks for this tip, very helpful! :p
08:20 kivilahtio Ask Brendan more about it if you meet him
08:20 kivilahtio Joubu: .
09:10 cait joined #koha
09:33 ashimema morning cait
09:34 cait morning ashimema
10:02 ztajoli joined #koha
10:14 jesse joined #koha
10:14 jesse hello
10:53 jcamins joined #koha
11:37 mveron joined #koha
11:37 mveron Hi #koha
11:38 mveron @wunder basel
11:38 huginn mveron: The current temperature in Wetter Allschwil, Allschwil, Switzerland is 18.7°C (1:38 PM CEST on September 03, 2014). Conditions: Mostly Cloudy. Humidity: 77%. Dew Point: 15.0°C. Pressure: 30.04 in 1017 hPa (Steady).
11:38 ashimema khall around?
11:38 khall yo
11:39 ashimema Think I've found a trivial typo in bug 11703
11:39 huginn Bug[…]_bug.cgi?id=11703 enhancement, P5 - low, ---, kyle, Pushed to Master , Convert checkouts table to ajax datatable
11:40 ashimema holds.js refers to /svc/ instead of /svc/holds around line 113..
11:40 ashimema 'means that the holds table never loads on some pages..
11:40 ashimema any chance you could varify the typo..
11:41 ashimema I have no idea how we missed this during QA..
11:41 khall will do! good catch!
11:41 ashimema Cheers Khall.
11:42 cait i think we had fixed it
11:42 cait it must have crept back in
11:42 khall yep, looks fixed in master to me
11:42 cait same explanation amrtin just gave
11:42 ashimema I think it must have crept back in somewhere... it's in 3.16.3 too
11:42 ashimema :(
11:43 ashimema Want me to post the tiny fix to a bug..
11:43 ashimema I've fixed it locally here.. drove me nuts trying to work out what was wrong.
11:43 cait i am pretty sure i found it once and kyle fixed it
11:43 khall Tomas fixed it with a followup
11:43 khall commit 9657a2c6dc92516963ffa960b7d561cdcc2dc5e8
11:45 ashimema cool..
11:45 ashimema so it's in master.. but not in 3.16.3
11:46 khall is gmcharlt around?
11:46 phasefx_ joined #koha
11:46 ashimema snap
11:47 khall Also, has anyone else had a problem with Koha 3.16 not sending notices if both email and digest are selected?
11:47 khall I was wondering if there was already a bug filed for that
11:47 cait hm
11:47 cait isn't that the normal thing, to have digest and email?
11:47 cait in my tests i did with master for th reply to it semed ok
11:47 cait what kind is the problem?
11:48 khall our experience at least for some libraries, is that if both are checked, no email is sent at all. You must either check only email, or only digest to recieve any email notifications ( for notifications with a digest option )
11:52 cait hm that sounds weird
11:52 khall yeah, I haven't yet tracked down exactly what's going on. That's on the todo list. We've had a few hiccups from the new message transport system.
11:53 cait hm
11:53 khall That being said, I still think it's a great improvement
11:53 cait i didn't see htat so far
11:54 jcamins_ joined #koha
11:54 cait khall: testing on my system now
11:55 meliss joined #koha
11:55 oleonard joined #koha
11:55 khall It could be something specific to our configuration. I was just wondering if anyone else had seen the behavior in the wild.
11:55 cait we are sendind our first library live with 3.16 on monday
11:56 cait i am done with bad news about bugs for a day :)
11:56 oleonard Already? But I just got here!
11:58 cait i can't confirm the bug on my master intallation
12:06 cait khall: in master i don't see that - i tested with an advance notice - digest, email checked, email went out
12:06 khall ok, thanks for checking! I'll have to do some more digging to figure out what's going on.
12:06 cait testing 3.16.2 now, don't have a .3
12:09 drojf joined #koha
12:09 Viktor joined #koha
12:09 cait intersting
12:09 cait i can see it in 3.16
12:11 cait ah no, was me
12:11 cait it works on 3.16.2 as well for me
12:12 cait khall - works on master and 3.16.2 for me - not sure now if that is good or bad
12:13 khall thanks for testing!
12:13 cait thx for scaring me :P
12:13 cait drojf: you are alive?
12:18 nengard joined #koha
12:18 oleonard I thought someone fixed the bug causing the error "Open of share file /tmp/sharefile-koha-koha failed: Permission denied at /usr/lib/perl5/Cache/ line 640."
12:18 oleonard ...but I just got it.
12:19 cait you have to uninstall the perl module
12:19 cait i think it just fixed the tests :(
12:19 cait for me at least - i still get it too
12:19 oleonard So I have a Perl module that I shouldn't have?
12:20 cait ye
12:20 cait s
12:21 cait if you can find the bug, i think i listed it there
12:21 BobB joined #koha
12:23 Dyrcona joined #koha
12:39 tcohen joined #koha
12:41 tcohen hi
12:46 oleonard Hi tcohen
12:46 tcohen hi oleonard
12:50 cma joined #koha
12:53 tcohen gmcharlt: wake up
12:53 ashimema hi tcohen
12:53 tcohen hi ashimema
12:53 ashimema hi oleonard
12:54 tcohen i've been extending the xml + xsl > xsl DOM toolchain
12:54 tcohen and thought i was doing it wrong
12:54 tcohen but it seems it does wrong without my patches... :(
12:55 tcohen_ joined #koha
12:55 ashimema oh
12:55 ashimema not so good.
12:56 edveal joined #koha
12:58 ashimema tcohen.. were you about when barton was expressing some thoughts on that toolchain?
12:59 tcohen nope
13:02 tcohen logs?
13:02 wahanui i guess logs is
13:02 tcohen today logs?
13:02 ashimema was a while back..
13:02 barton tcohen no, it was a few weeks ago.
13:02 tcohen wahanui: today logs is
13:02 ashimema I'm just trying to fidn the relevant ...
13:02 wahanui OK, tcohen.
13:02 ashimema barton is here.. he can talk ;)
13:03 barton I put a link on the wiki, lemme see if I can't find it...
13:03 ashimema where the RFC barton.. our wiki is crap.. I can't find it anymore.
13:03 tcohen i see no necesary problem with the toolchain itself
13:03 barton[…].50_attribute_rfc
13:04 barton tcohen -- my problem with the toolchain is that there is *way* too much duplication of data.
13:04 ashimema[…]4-08-12#i_1551164
13:04 ashimema and the irc conversation to go with it ;)
13:05 tcohen how would you cope with marc flavour specifics?
13:05 barton good question...
13:06 tcohen i don't have the answer of course
13:06 tcohen heh
13:06 ashimema anywho.. thought it worth getting you two talking ;)
13:07 tcohen what i've seen is that the current xsl build weird things
13:07 tcohen barton: any ideas are welcome
13:07 NateC joined #koha
13:08 kmlussier joined #koha
13:09 ashimema joined #koha
13:09 barton I actually wrote a script last night that takes yaml input and spits out chunks of text that you can paste into, bib1.att, biblio-koha-indexdefs.xml, and record.abs ... just that much makes creating patches much easier...
13:13 ashimema I think the problem is basically endemic of how zebra expects config files to be spread all over the place
13:17 barton If we used a yaml file to store the data, we could also add a 'marc-flavour' and perhaps also a 'search-type' (biblio vs authorities) field, and store all marc data in a single file, then generate {marc21,normarc,unimarc}/{biblios,​authorities}/*-koha-indexdefs.xml files from that.
13:19 barton ashimema, I don't disagree, but there's no reason to bang our heads against zebra's mess if we can centralize it.
13:20 tcohen we shouldn't blame zebra
13:20 tcohen for our decisions
13:20 ashimema I'm pretty sure there's a reason zebra does it the way it does..
13:21 ashimema I like the idea of abstracting into a single point of truth.. but I also don't know zebra well enough to know whether we can cover all eventualities with such a file..
13:21 cait i have my doubts
13:22 cait especially with some of the recent efforts - we'd have to change the structure of the yaml file additionally to other thigns
13:22 ashimema I know we can 'get more' out of using DOM configs.. but don't know what 'more' is yet... I wouldn't want us to paint ourselves into a new cornder 'tis all..
13:22 cait ashimema: i think there is a patch from mathieu for a little more - indexing over multiple subfields into a phrase index
13:22 cait the one they talked about on the mailing list
13:22 ashimema cool
13:23 ashimema we need to go DOM only as soon as possible in my mind..
13:23 ashimema at which point I'll learn how the DOM config fits together too ;)
13:23 barton cait, do you have a link to that thread in the mailing list?
13:23 cait sory not right now
13:24 cait pretty busy here
13:24 barton np, cait.
13:24 cait i tried finding th bug, but didn't spot the right one immediately
13:24 cait was in the last few days
13:24 barton oh, that *never* happens ;-_
13:24 barton er :-)
13:45 tcohen ashimema: in DOM indexing zebra just expects you point it to an xsl that transforms the marcxml into an xml that tells zebra what to put on which index, like this
13:45 pastebot "tcohen" at pasted "sample xsl output for zebra indexing" (41 lines) at
13:45 tcohen it is pretty simple, and clean
13:46 tcohen that xsl can be whatever we want it to
13:46 tcohen maintaing the xsl is way harder than using the XML structure galen et all defined
13:46 tcohen s/et all/et al/
13:47 ashimema I see..
13:47 ashimema that makes allot of sense
13:48 pastebot "tcohen" at pasted "defining indexes for 100$a" (9 lines) at
13:48 tcohen it does make a lot of sense
13:49 tcohen i'm adding a <facet>index_name</facet>
13:50 tcohen too
13:50 tcohen the difficult part
13:50 wahanui i guess the difficult part is keeping our data intact I guess
13:50 tcohen is that if you need to some some magic trick
13:50 tcohen like taking stuff from several places for putting together
13:50 tcohen you don't write xslt code that does that
13:51 tcohen you need to write xsl code, that generates code that generates what you want
13:51 tcohen so, think about having more abstraction :-D
13:52 tcohen but, overall, that's our own problem
13:52 tcohen and not zebra-related
13:52 tcohen besides all cr*p we can find about zebra
13:53 * tcohen feels he wrote too much, shuts for a while
13:53 ashimema was just throwing that around my head for a bit..
13:54 ashimema I 'think' it still makes sense to me..
13:56 tcohen ashimema: try xsltproc etc/zebradb/marc_defs/marc21/bib​lios/biblio-zebra-indexdefs.xsl marc_record.xml
13:57 tcohen on any marcxml record you download from a koha
13:57 tcohen instance
13:57 wahanui instance is, like, called koha_test
13:57 tcohen that will help you see
14:03 ashimema lightbulb just went on..
14:03 ashimema thanks tcohen.. that makes allot more sense now
14:04 tcohen see? that's all that is happening behind the scenes :-D
14:10 ashimema So.. to edit indexes on a package install you'de need to edit /etc/koha/zebra/marc_defs/marc21/​biblios/biblio-koha-indexdefs.xml and then run it through xslproc to generate a new biblio-zebra-indexdefs.xsl in the same place..
14:10 ashimema and 'all the really clever stuff' happens in that conversion pass
14:12 rocio joined #koha
14:18 tcohen it just translates those index definitions into xsl that transforms that into zebra's format
14:18 tcohen in a convenient way
14:19 tcohen[…]iblionumber=19433
14:19 tcohen can anyone try the on-line link?
14:19 ashimema get 404 on it.
14:20 tcohen it points you to[…]ook/9789500604987
14:20 tcohen right?
14:20 tcohen the href points to http://www.medicapanamericana.[…]ook/9789500604987
14:20 tcohen i cannot understand what's going on
14:23 ashimema can varify what you just said..
14:23 ashimema points to medicpana... not fcb.biblio...
14:24 tcohen oh
14:25 tcohen the cataloguer put only one / on the URL
14:27 ashimema l
14:47 tcohen jcamins++
14:58 tgoat joined #koha
15:16 reiveune bye
15:16 reiveune left #koha
15:45 ztajoli left #koha
16:01 Joubu good bye #koha
16:10 cait left #koha
16:14 pianohacker joined #koha
16:36 mveron joined #koha
16:38 carmenh joined #koha
16:44 cait joined #koha
16:45 wnickc joined #koha
16:50 cait left #koha
16:50 cait joined #koha
17:00 ashimema tcohen: fancy making one of my colleagues day?  She's just posted her first patch bug 12866 :)
17:00 huginn Bug[…]_bug.cgi?id=12866 minor, P4, ---, janet.mcgowan, Passed QA , Incorrect display of Email Address in Alternate Address Patron Details
17:01 cait and she is a very lovely colleague of ashimema
17:01 ashimema hehe..
17:01 cait :)
17:15 mtompset joined #koha
17:15 mveron ...assigned 16:01:20, QAed 16:58:05...  Would be great if every patch would be  so fastgoing...  :-)
17:15 mtompset Greetings, #koha.
17:15 mtompset Greetings, mveron. :)
17:16 mveron Hi mtompset :-)   And good daytime to everybody... :-)
17:16 mtompset oleonard: Did you get my message? :)
17:16 tcohen ashimema: don't think so
17:17 * tcohen is just kidding
17:17 ashimema haha..
17:17 akafred joined #koha
17:17 ashimema I was impresseed by it's speedy signoff..
17:18 ashimema I couldn't resist ;)
17:18 mtompset I notice that bug 11592 has been sitting on the Needs Sign Off List of 10 oldest bugs. :)
17:18 huginn Bug[…]_bug.cgi?id=11592 normal, P5 - low, ---, mtompset, Needs Signoff , opac detail scripts do not respect MARC tag visibility
17:20 mtompset but how international was it's transit? Reminds me of the security patch rangi posted a while back.
17:23 huginn New commit(s) kohagit: Bug 12866: Fixes Display of Email in Alternate Address <[…]ad75550b849c809c5>
17:24 cait woohoo :)
17:24 cait uk - switzerland - germany - argentina
17:24 cait not too bad on the international side of things :)
17:26 ashimema :)
17:28 ashimema mveron++, cait++, tcohen++ #fantastic fast response there :)
17:28 bag :)
17:29 mveron Wow... :-)
17:29 tcohen hey, we are a friendly community
17:29 tcohen this is expected, actuallly
17:31 mveron Assigned 16:01:20 - Pushed 17:19:22  - In less than 80 minutes around the world :-)
17:32 cait :)
17:36 * mveron Is playing around with NVDA and for testing accessibility in OPAC
17:38 cait :)
17:38 cait no... now who writes a patch for :)
17:40 tcohen that'd be rangi?
17:42 mveron cait: Do you mean Bug 12790 ?
17:42 huginn Bug[…]_bug.cgi?id=12790 enhancement, P5 - low, ---, veron, Needs Signoff , Perl modules: Display required version numbers on about page
17:43 cait mveron: :)
17:43 cait i meand adding janet, but that's also a good one
17:54 cbrannon joined #koha
18:28 WNickC joined #koha
18:53 tcohen unimarc gutus?
18:54 tcohen gurus?
18:57 mtompset Hmmm... networking issues with my VM... Have a great day, #koha.
19:11 mtompset joined #koha
19:17 mtompset Greetings, #koha.
19:18 mtompset Okay. This is rather annoying. If a perl script has use Foo; in it, then should I not be able to my $blah = Bar(); without having to Foo::Bar();?!
19:19 tcohen mtompset: if you're doing a static call, you should use Foo qw/Bar/;
19:19 mtompset i'm trying to run prove -v t/db_dependent/Circulation.t on a 3.14.x branch, it is barfing on GetItem.
19:23 mtompset ARG... it must be some change of mine SOMEWHERE....
19:23 tcohen bye #koha
19:28 jmsasse joined #koha
19:51 rangi   francesca has been tidying this up
20:14 nengard left #koha
20:16 BigRig joined #koha
21:04 kmlussier left #koha
21:09 sophie_m joined #koha
21:26 * cait waves
21:26 cait francesca++
21:28 jmsasse Hi cait. Wie geht's?
21:28 cait hi jmsasse :)
21:28 cait good, but sleepy
21:28 cait going to bed now :) night!
21:28 cait have a good day
21:28 cait left #koha
21:34 JasonBurds joined #koha
21:36 rambutan joined #koha
22:00 rambutan Has any thought been given to changing the channel topic?
22:01 pianohacker joined #koha
22:10 mtompset OH MY! Has no one seen the bug in GetMemberDetails before?!
22:14 rambutan left #koha
22:16 eythian I find your lack of specificity disturbing, mtompset
22:17 mtompset imagine possible values of $borrowernumber and $cardnumber... like undef and something valid.
22:18 eythian Now assume I don't have the function memorised.
22:18 mtompset GetMemberAccountRecords is called with undef, even though $borrower is a hash to something with a valid borrower number!
22:19 mtompset Now to figure out how to write a test to test the fix.
22:20 mtompset It uses the undef $borrowernumber, rather than the defined $borrower->{borrowernumber}
23:09 tcohen joined #koha
23:10 tcohen hi
23:10 wahanui bonjour, tcohen
23:11 eythian tcohen: I don't have access to your redmine
23:13 tcohen solved
23:22 mtompset redmine?
23:22 mtompset Ah, interesting.
23:25 tcohen mtompset: not that interesting
23:25 tcohen even though is quite a flexible tool for project management
23:25 mtompset well, i never knew of it.
23:26 tcohen not just bug tickets
23:26 mtompset that's why it is mildly interesting.
23:26 tcohen it is interesting, it is just that, it is... ruby
23:26 tcohen :-P
23:26 mtompset What can you do? Lord help us all. ;)
23:31 papa joined #koha
23:33 dcook mtompset: I think GetMemberDetails will also return a partially filled hash, rather than undef, if you provide undef or invalid arguments for both parameters.
23:33 dcook Or maybe it's that other Member sub
23:33 dcook I don't like either of them in any case :p
23:33 mtompset dcook: bug 12868 ;)
23:33 huginn Bug[…]_bug.cgi?id=12868 enhancement, P5 - low, ---, koha-bugs, Needs Signoff , GetMemberDetails calls GetMemberAccountRecords incorrectly
23:34 chrisvella94 joined #koha
23:34 dcook Cool. Alas, I'm buildin' a DSpace and fixin' up a OPAC atm.
23:39 mtompset come on. :P --- git bz apply interactive just the first, prove -v t/db_dependent/Member.t, failed GREAT!, koha-qa -v 2 -c 1, git bz apply interactive just the second, prove -v t/db_dependent/Member.t, passes GREAT!, koha-qa -v 2 -c 1... sign off.
23:39 NateC joined #koha
23:49 dcook ...might want to verify it myself as well rather than just banking on the tests :p

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