Time Nick Message
00:29 dcook Ooo just learned about "corelist"
00:29 dcook That's neat
00:29 eythian what's that?
00:29 dcook It lets you know whether or not a Perl module is part of the core :)
00:30 eythian oh right
00:30 dcook I was wondering about that a bit...
00:30 dcook I mean... we don't include modules that in koha_perl_deps, which are part of the core
00:30 dcook But isn't CGI on its way out?
00:30 dcook I suppose we could just add that later of course..
00:30 eythian it's on its way to being non-core
00:30 dcook Yeah
00:31 eythian I can special case that though, by doing something like 'depends: perl (<5.22) | libcgi-perl'
00:31 dcook Only for packages though
00:32 dcook We other suckers would need it in koha_perl_deps.pl as well
00:32 dcook Which isn't a big deal of course
00:32 dcook I wonder if koha_perl_deps would benefit from listing whether or not the module is a core module...
00:33 * eythian only cares about dependencies in packages. Anything more becomes a larger headache.
00:33 dcook True dat
00:33 dcook Hmm
00:34 dcook C4::Sip::Sip is actually packaged just as Sip...
00:34 dcook C4/SIP/Sip I suppose..
00:35 eythian yeah
00:35 eythian it's its own little namespace
00:36 dcook I didn't know that we did that :S
00:36 dcook Looks like the ILSDI stuff might do that as well?
00:36 eythian see also bug 7904
00:36 dcook Hmm nope... must be something else
00:36 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=7904 enhancement, P5 - low, ---, colin.campbell, Passed QA , SIP modules handling of @INC is confused
00:36 eythian no, ilsdi I think crams a bunch of packages into one file
00:36 eythian something like that does that anyway
00:36 eythian it's very AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAa
00:37 dcook oai.pl
00:37 * dcook shudders
00:37 dcook But yay for bug 7904
00:37 eythian that's the one
00:37 dcook There is "use ILS" somewhere... not sure what that's for yet
00:38 dcook That's also for SIP it seems
00:38 dcook Ah, that's part of the bug too
00:38 dcook Sweet as
00:39 eythian oh, the ILS.pm in Sip looks interesting
00:39 * eythian ignores that for now and gets back to authority searching
00:41 eythian I had a realisation the other day that authority searching is a /lot/ more strict than biblio searching.
00:41 eythian this makes life easier
00:41 eythian no having to parse free-form search queries
00:42 dcook Oooh. That's nice :)
00:43 dcook Some day I'll look at query parsing again..
00:43 eythian https://i.imgur.com/5zuupyE.jpg
00:47 dcook hehe
00:56 dcook Hmm... what's support like for EDS?
00:56 dcook It looks like it uses "LWP"
00:57 eythian talk to rangi if you're working on that, he's going to be doing some work with it some time soon probably
00:57 eythian I can't remember the exact scope though
00:57 eythian but it'll involve genericising all that stuff
00:57 dcook Ah, I'm not working on it. I'm just looking at what modules are being loaded up by stuff
00:57 eythian ah, ok
00:57 dcook Double-checking deps and what not
00:58 dcook Looks like Try::Tiny is also used in EDS and a maintenance script
00:59 dcook I might just make a bug for this... or maybe email the listserv..
01:02 eythian it would be interesting to have something that scans what modules are used that aren't included in the dependency list
01:02 dcook Agreed
01:02 dcook I'm not sure how one would go about it
01:03 eythian although, if in general, they'll get picked up when I do builds
01:03 dcook At the moment, I'm just grepping the hell out of everything
01:03 eythian as the 00-load.t will fail
01:03 dcook Oo what's 00-load.t?
01:03 eythian (but it does deliberately skip some things)
01:03 dcook Looks like "English" gets used by a few scripts as well
01:03 eythian it's a test case to check that every module can be parsed.
01:03 eythian is English not core?
01:03 dcook Ah, but English is core
01:04 dcook I just tried one of these "sh find-perl-modules.sh | xargs corelist"
01:04 dcook Kinda handy
01:04 dcook I'm sure I'm probably missing something somewhere though
01:04 dcook And just ignoring Koha|C4|Sip|ILS atm
01:04 dcook Lots of egrep followed by an awk and a sort -u
01:05 dcook Hmm, I misised the PDF modules somehow..
01:05 dcook missed*
01:07 dcook Ah, I see..
01:14 dcook Hmm
01:15 dcook eythian: I wonder if 00-load.tt just gets lucky sometimes?
01:16 eythian whaddaya mean?
01:16 dcook I don't see "URI" in koha_perl_deps.pl
01:16 dcook And it's not core
01:16 dcook But it's used by PazPar2
01:16 dcook Which is loaded by 00-load.tt
01:16 dcook But we do have URI in our system for some reason
01:17 eythian could be it's pulled in as part of something else
01:18 dcook I figure that must be it
01:18 eythian my package builds are done in a cleanroom type environment, so if it doesn't end up in there, it'll fail.
01:18 dcook Ahh, that's good :)
01:19 dcook Then when something is missing you add it to koha_perl_deps.pl and make a deb package for it?
01:19 eythian well
01:19 eythian it depends
01:20 eythian debian/control is generated from C4::Installer::PerlDependencies
01:20 eythian if it doesn't end up in there, then it'll fail.
01:20 eythian typically I just report a bug on it.
01:20 eythian sometimes fix it myself.
01:20 eythian I try to encourage people from just assuming I can make a package, as it's happened before that it's not always possible
01:21 eythian sometimes it requires updating half the universe, sometimes they're not copyleft, etc.
01:21 eythian s/encourage/discourage/
01:21 dcook Mmm, yeah, I think you've said that before
01:22 dcook Cool beans
01:22 dcook Well, I figure any issues with Koha will find their way out individually rather than a big sweep..
01:22 dcook So perhaps I'll just focus on making sure we track any extra modules we add locally...
01:24 eythian yeah, it's part of periodically building a master package, they can get spotted quickly.
01:24 eythian Though the biggest issue I have is people making test cases that require a database to be present.
01:25 eythian if you add the modules to C4::..PerlDeps, then it'll just follow it around happily
01:25 dcook Yeah the database dependent ones seem suboptimal
01:25 dcook I'll keep that in mind :)
01:25 dcook The C4::..PerlDeps thing that is
01:25 dcook Well both :p
01:28 dcook Ah, it turns out that CGI "is" listed in C4...PerlDeps
01:28 dcook That's cool
01:31 dcook Hmm, we list a few core modules
01:31 dcook Data::Dumper too
01:31 eythian innit core too?
01:31 eythian or just everywhere...
01:31 eythian oh, I see
01:31 eythian there's no harm in listing them
01:31 dcook Yeah, it's core
01:31 dcook Nope, no harm at all
01:32 dcook Yeah, I'd almost prefer everything were listed
01:32 eythian though, it makes my update-dependecies script run slower, but I probably shouldn't run it in a RAM constrained VM anyway.
01:33 dcook update-dependencies?
01:34 eythian update-control I meant
01:34 dcook Mmm
01:34 dcook If it's in core, would it also have a debian package?
01:34 dcook Or sometimes yes, sometimes no?
01:34 eythian generally no
01:35 dcook So that could get dicey
01:35 dcook Hmm, I always figured DBI would be core
01:35 eythian $ dpkg -S Data/Dumper.pm
01:35 eythian perl: /usr/lib/x86_64-linux-gnu/perl/5.20.1/Data/Dumper.pm
01:35 eythian so, it's in perl
01:35 eythian no, my update script understands that things might be core
01:35 eythian (well, it's mostly liw's update script, I've just change it a bit.)
01:36 dcook Speaking of liw, I was thinking of you last night, eythian, as I was deciding which Ubuntu release to use
01:36 dcook I like 14.04 so far, but 14.10 has the latest Obnam...
01:36 dcook Which reminds me I forgot to bring in my backup drive today..
01:37 eythian I tend to run the latest ubuntu for desktops, but I tend to recommend the latest LTS to other people, paradoxically/hyocritically.
01:38 dcook hehe
01:38 dcook Makes sense
01:38 dcook I suppose the average person doesn't want to upgrade every year
01:39 eythian 6 months, but yeah
01:39 dcook I thought it was 9 months of support?
01:39 eythian it is, but there's a new release every 6 months
01:39 dcook I guess you wouldn't necessarily want to be at the start or end..
01:39 dcook Ah, righto
01:46 eythian ok, ugly old authorities query now turned into a nicer hash form, now to turn that into an actual ES query.
01:54 dcook Oh wow... just received a bundle of Christmas presents from my grandmother
01:54 * dcook thinks it took the boat a while to get to Sydney
01:55 eythian heh
01:56 dcook Apparently it only takes 2 days to get something shipped from Amazon though :p
01:57 eythian it takes forever to get amazon things to here.
01:57 eythian well, last time I did it
01:58 eythian usually the shipping cost is ridiculous at the same time
02:00 dcook Yeah, I haven't bought anything from Amazon yet, but the time/shipping cost are the factors that would get me :/
02:04 dcook Hehe. I don't know what "GD.pm" does, but I like its name
02:04 eythian graphicsy things, iirc
02:04 dcook Ah yeah. Something with barcodes it seems.
02:04 dcook That's not as amusing as I was thinking
02:04 eythian yeah, lets you use drawing primitives to output a png
02:07 * dcook thumbs up
02:32 eythian dcook: when you do an authority search,
02:32 eythian oh, never mind
02:32 wahanui Good, I'm glad you figured it out. I didn't understand, and probably never will, being a bot.
02:32 eythian thanks for helping
02:33 dcook hehe
02:33 dcook :)
02:42 eythian I wonder why the authorities code makes one name to another:
02:42 eythian elsif ( @$tags[$i] eq "mainentry" ) {
02:42 eythian $attr = " \@attr 1=Heading ";
02:42 eythian }
02:42 eythian there's a few options like that
02:42 eythian maybe it's an old name to a new name
02:48 eythian beer o'clock in 12 minutes, followed by whisk(e)y o'clock
02:49 eythian my Japanese whiskey arrived today, which was handy too.
02:53 dcook Mmm
02:53 * dcook is jealous of the Catalyst folk and their booze
02:54 eythian wellington is nice this time of year, just sayin'...
02:57 dcook hehe
02:57 dcook I need to look at my vacation leave again...
02:57 dcook And possibly wait until we get some money back from these vet bills
02:57 dcook Turns out giving your cat radiation therapy is about $2000
02:58 dcook Well, a radioactive capsule
02:58 dcook Hopefully insurance will cover most of that though
02:58 eythian ouch, not cheap.
02:58 dcook Nopes
02:58 eythian can't you just feed them a lot of bananas?
02:58 dcook Plus he'll be radioactive for a while so we're putting him up in boarding for a while
02:58 dcook hehe
02:58 dcook I forgot about the banana thing
02:59 dcook Well, it's quite interesting. He has hyperthyroidism.
02:59 dcook Which apparently is caused by a cancerous growth in the thyroid gland.
02:59 dcook Same thing happens with humans.
02:59 eythian ah yep
02:59 dcook So they give you some radioactive iodine, which gets absorbed into the gland, and apparently the growth shrinks/is destroyed
03:00 dcook One day I will visit Wellington though, especially on a Friday ;)
03:00 eythian that's the inverse of how you plan for nuclear fallout/exploding reactors.
03:00 dcook As for the authority stuff...
03:00 dcook I wonder that too
03:01 dcook Hmmm?
03:01 dcook So what you're saying is... if I would've just taken him near a exploding reactor, I wouldn't have to pay $2000... ;)
03:01 eythian stock up on normal iodine, so that the radioactive stuff doesn't get absorbed into your thyroid causing cancer.
03:02 eythian I think Finland did it around Chernobyl time
03:02 dcook Huh... interesting!
03:02 wahanui i think interesting is sometimes good and sometimes bad
03:03 dcook Apparently it's used as a preventative for breast cancer
03:03 dcook "the FDA recommends 165 mg of iodine to prevent thyroid cancer in the event of a radiation emergency"
03:04 dcook Learn something new everyday
03:08 eythian right, things pushed, now it's beer o'clock.
03:08 eythian bye!
03:11 dcook laters eythian
03:22 dcook @later tell eythian When you use EDS for your clients, do you turn LWP into a debian package? Or is it provided by a different one?
03:22 huginn dcook: The operation succeeded.
03:22 dcook Don't think it's provided by anything else..
04:04 dcook Huh... LWP is there...
04:04 dcook I wonder if it's just part of LWP::UserAgent
04:06 dcook Or it's part of libwww-perl...
04:08 dcook @later tell eythian Nevermind. I see that LWP.pm is included in libwww-perl, which you already have in koha-perldeps :)
04:08 huginn dcook: The operation succeeded.
07:38 reiveune hello
07:38 * magnuse waves
07:45 malik Hello to All : I am working on Active Directory integration with koha. Can anyone have an idea about this that how to install AD on ubuntu because i have my koha on ubuntu and i don't have window server
07:58 alex_a bonjour
08:05 fridolin hie all
08:17 gaetan_B hello
08:29 cait morning
08:29 cait Joubu: around?
08:31 Joubu cait: yep
08:35 cait ok
08:35 cait going to test again here in our 3.18.3
08:35 cait give me a second
08:35 cait (serial expiration)
08:36 cait ok, i am on the serial subscription page, using the filters to the left - the expiration date of the subscription is 1.3.2015
08:36 Joubu cait: I tried on master
08:36 cait using expires before 01.01.2016 finds it
08:36 cait Joubu: i did this mornign - but don't have a master here
08:36 cait i tried with and without the patch
08:37 cait interesting, you are right, it works now here
08:37 Joubu cait: Ho! I wasn't! It was a 3.18.x too
08:37 cait ok - so we retest on master?
08:37 Joubu ok, git checkout master and I retry :)
08:37 cait thx :)
08:37 cait wonder if it's some weird logic and only some dates fail (you never know :( )
08:38 cait if you have to create a subscription, maybe you could try with the dates on the bug
08:38 Joubu same, it works with the same example
08:38 cait weird
08:38 cait i tried more than one time
08:38 cait hmpf.
08:39 cait hmm
08:39 cait could you check something?
08:39 cait do the filters go on the same field?
08:41 Joubu erk , does not work with 10/09/2014
08:41 Joubu funny...
08:45 Joubu cait: actually it is in C4::Serials::GetExpirationDate, there are some check on the frequency, the number of issues, etc.
08:45 Joubu the enddate is not checked in these cases
08:47 cait ok, 2 filters, labelled the same, doing different htings?
08:48 cait Joubu: can we do something to make it more clear? not sure i understand it yet myself
08:49 Joubu cait: in the sub search, the check is done on subscription.end_date
08:50 Joubu actually the algorithm in "check expiration" is really awful: retrieve all subscriptions from the DB (with title and isbn match, but if don't provide them, retrieve all sub...)
08:51 Joubu and do some funny checks on length, frequency, etc.
08:51 Joubu do know if it is expired or not
08:51 Joubu s/do/to
08:51 cait sounds too complicated to work nicely
08:52 cait do we update the expiration date when we renew a subscription?
08:52 Joubu maybe it sounds complicated because of my English and I did not manage to explain correctly/easily...
08:53 Joubu yes, using GetExpirationDate :p
08:53 cait still confused i think :)
09:02 cait Joubu: so does it work... or does it not?
09:04 Joubu it does like the code wants it to work...
09:12 cait Joubu: i thik we display the subscription.enddate in the search result - but it is not used for the searhc
09:12 cait that's the confusing bit
09:12 cait it would be interesting to show the date the system calculated
09:14 Joubu cait: it's what it's done, I think. On checkexpiration.pl, I see:
09:14 Joubu 78 my $expirationdate = GetExpirationDate($subscriptionid);
09:14 Joubu 80 $subscription->{expirationdate} = $expirationdate;
09:14 cait hm ok
09:14 cait something for another bug maybe
09:15 cait to make it visually clear when we show a calculted date
09:15 cait calculated
09:16 Joubu What it certainly happen: the expiration date is not the same, depending on the page you use
09:16 cait yeah
09:16 cait not very clean
10:01 cait hm someone good with javacript regex?
10:23 cait hm found a workaround
10:48 paxed cait: window.location.hash?
10:50 cait i figured out a workaround - the probem is another script that comes as last and on click
12:46 Joubu khall: hi! Jacek have opened an interesting bug report (bug 13645). Do you have any idea on how could we do that in an elegant way?
12:46 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=13645 normal, P5 - low, ---, gmcharlt, NEW , Koha::Acquisition::Bookseller->search() performance / efficiency problems
13:20 drojf hi #koha
13:32 oleonard Hi #koha
13:48 cait hi oleonard
14:08 magnuse hm, on a completely fresh install of 3.18.3, when i enter a patron with "norwegian wovels", save and view the patron, the they show as garbled text. anyone else seen that?
14:08 cait not yet
14:08 cait but let me try
14:09 cait hm looking ok here
14:09 cait which page are you looking at? i used umlauts in the name
14:11 magnuse the member detail page /cgi-bin/koha/members/moremember.pl?borrowernumber=1
14:11 magnuse and the check out page
14:12 cait hm ok for umlauts here
14:12 cait äüö
14:12 cait can you give me your name? :)
14:13 magnuse HÃ¥vard
14:13 magnuse or did you mean Magnus? ;-)
14:13 magnuse looks ok in the db
14:14 magnuse umlauts are garbled too
14:17 magnuse works ok in one of my old, upgraded instances also running 3.18.3
14:17 cait hm i tested in a new one
14:18 cait but updated from 3.18.2 i think
14:18 cait might be a new 3.18.3 not sure when we set it up
14:18 magnuse weird
14:22 drojf hi cait and magnuse
14:28 Joubu khall: around ?
14:28 khall what's up?
14:29 Joubu khall: I have retested bug 13021, and it looks ok here
14:29 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=13021 enhancement, P5 - low, ---, jonathan.druart, Failed QA , DataTables server-side processing - guarantor search
14:29 Joubu could you confirm you correctly clean the cache?
14:29 khall ok, I'll retest and figure out what the heck is going on
14:29 Joubu Otherwise, it cans come from a bad formated json
14:30 khall I cleared the cache multiple times
14:30 Joubu caused by non escaped char
14:30 khall Joubu: I'm using the example patrons fwiw
14:31 khall no errors in the js console or the error log either
14:32 Joubu what do you search for?
14:35 Joubu ho wait, I got it
14:38 Joubu khall: it comes from b8989510ec8b6b58249707bf2ddc6cf80fbfef4d, bug 8096
14:38 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=8096 normal, P5 - low, ---, veron, Pushed to Stable , Card number twice on patron search results
14:38 Joubu it added new breaklins
14:40 Joubu khall: ok I have a patch
14:40 khall excellent!
14:42 Joubu khall: ok sent
14:42 Joubu khall: did you see my previous ping?
14:42 Joubu It was "khall: hi! Jacek have opened an interesting bug report (bug 13645). Do you have any idea on how could we do that in an elegant way?"
14:42 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=13645 normal, P5 - low, ---, gmcharlt, NEW , Koha::Acquisition::Bookseller->search() performance / efficiency problems
14:42 khall I'll take a look at that
14:43 oleonard I don't want to throw fuel on an email thread by that "other paul," but is there something to his contention that searching in 3.18 is slower?
14:44 oleonard My library's experience of 3.16 over 3.14 is that searching is significantly slower.
14:44 jcamins oleonard: I suggest rewriting Zebra with pthreads.
14:45 jcamins It'll be awesome!
14:45 jcamins ... as long as someone who isn't me does it.
14:45 * jcamins goes to get some yogurt for breakfast.
14:45 oleonard jcamins: Like I said, I'm not getting into that, but I would like serious discussion if anyone has any.
14:46 Joubu could be related to 13645 too
14:48 khall Joubu: I suppose we could make the schema and resultsets package level vars and make it a singleton perhaps?
14:48 khall and perhaps implement Koha::Cache for it
14:48 Joubu khall: using the same way as we do for C4::Context?
14:49 khall essentially, yes
14:50 magnuse oleonard: my server has seen higher load and slower everything after the upgrade from 3.16.7 to 3.18.3. but apache and mysql were updated at the same time, so there are a few "moving parts"
14:50 magnuse i suspect there is something amiss with my mysql config
14:51 khall I think as a general rule Koha adds additional complexity over time which leads to more processing which leads to less speed overall.
14:52 khall I think a solid REST api which can be leveraged by Koha itself would help a lot for overall speed
14:52 oleonard If that's true khall then Koha will inevitably get slower and slower until it is unusable.
14:53 khall oleonard: to some extent that is true, I am of course over-generalizing
14:53 khall the solution is to make a page do less ( at least on initial load )
14:54 magnuse one one my munin grpahs called "http loadtime of a page" shows the loadtime ~doubling after the upgrade
14:54 khall however, some complexity just can't be avoided ( checkins and checkouts behavior for example )
14:55 magnuse uptimerobot shows a less dramatic increase, but still a visible difference
14:55 magnuse plackify all the things!
14:56 khall here here!
14:56 oleonard How can we make quantitative judgements about how updates to Koha change various primary functions like check in checkout and searching?
14:57 khall relative subroutine call timings via unit tests?
14:57 khall can Jenkins tell us anything useful?
14:57 magnuse i took a shot at that with https://github.com/digibib/kohaprofile
14:58 khall neat! magnuse++
14:58 magnuse maybe i should try running it again, to see if there is a difference for 3.18
14:59 magnuse it needs some more meaningful tests, i think
14:59 magnuse patches welcome ;-)
15:00 drojf magnuse++
15:05 cait magnuse++
15:08 Joubu thanks khall!
15:08 khall np!
15:10 drojf @wunder berlin, germany
15:10 huginn drojf: The current temperature in Berlin Tegel, Germany is 2.0°C (3:50 PM CET on January 30, 2015). Conditions: Light Snow. Humidity: 100%. Dew Point: 2.0°C. Windchill: -2.0°C. Pressure: 28.74 in 973 hPa (Steady).
15:10 drojf :/
15:13 drojf /o/ into the cold… later #koha
16:03 reiveune bye
17:25 gaetan_B bye
19:01 mtompset Out of curiosity, the crowdfunding page that Bywater put up recently... those funds go where and how?
19:37 bag mtompset: the fund come to bywater - but it doesn't totally mean that bywater is doing the development - things could be shared to other developers etc. But bywater would be the fund manager and project manager from that bywater hosted site
19:40 mtompset So, if I go, "Oooo... I like that project," and as a develop report the bug, take the bug, and code the bug. I could ask for the funds raised?
19:40 mtompset ^develop^developer^
22:41 mtompset Have a great weekend, #koha.