IRC Logs

Time Nick Message
23:41 cait ohok
23:41 cait but actually... not a bad idea
23:41 jcamins Yes, and that was for you. :)
23:41 cait that was for magnuse :)
23:41 cait oh
23:41 jcamins Good night, cait.
23:40 huginn cait: The current temperature in Taegerwilen, Taegerwilen, Germany is 3.8ï¿½C (12:35 AM CET on January 08, 2012). Conditions: Light Rain. Humidity: 91%. Dew Point: 2.0ï¿½C. Windchill: 4.0ï¿½C. Pressure: 30.22 in 1023.2 hPa (Steady).
23:40 cait @wunder Konstanz
23:40 rangi sleep well
23:40 cait good night magnuse :)
23:39 * magnus_away goes to sleep
23:36 rangi hope that stops before tomorrow
23:36 huginn rangi: The current temperature in Wellington, New Zealand is 14.0ï¿½C (12:00 PM NZDT on January 08, 2012). Conditions: Rain. Humidity: 100%. Dew Point: 14.0ï¿½C. Pressure: 29.86 in 1011 hPa (Steady).
23:36 rangi @wunder nzwn
23:28 magnus_away yay!
23:27 rangi to have done that
23:27 rangi you a literally one in 100 million or so
23:27 rangi ohh a wild one
23:26 magnus_away sure!
23:26 * magnus_away saw a wild kiwi on kapiti
23:26 rangi heh you woke up to say that?
23:25 magnus_away rotorua was fun!
23:24 cait hehe :)
23:24 rangi heh
23:24 jcamins I guess it doesn't meet his standards of cleanliness.
23:21 jcamins I'm not sure exactly why.
23:20 jcamins When I put my face near his, he licks my nose. It's very cute.
23:20 * cait waves at myshkin
23:20 jcamins He is currently *loudly* washing his paws.
23:19 rangi ahh
23:19 jcamins A higher life form has preempted my chair.
23:19 rangi oh?
23:19 * jcamins has been evicted.
23:01 cait :)
23:01 rangi yep I think we will
23:00 cait that
22:59 cait yes hat
22:59 rangi hot pool
22:58 cait rangi: going to take a hot bath too?... forgot how they are called
22:48 cait only kiwi here
22:48 cait yep
22:48 rangi maybe germany too cait?
22:47 rangi I think the us drops the fruit bit off the name
22:47 jcamins I didn't know that.
22:47 jcamins Ah.
22:47 rangi small brown and hairy looking
22:46 rangi kiwi is the bird it takes it name from
22:46 rangi the fruit is kiwifruit
22:46 cait :)
22:45 jcamins Kiwis? I guess this is something other than the fruit?
22:45 rangi u can see them at zoodoo
22:45 rangi we have some here in the zoo
22:45 rangi yep
22:45 cait and they have kiwis
22:45 cait oh
22:44 rangi ill pass
22:44 rangi heh
22:44 cait maybe you should get some
22:44 cait I have heard they have 'mud for men' at the shop there :P
22:43 huginn New commit(s) needsignoff: [Bug 7284] Authority matching algorithm improvements <http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=7284>
22:43 cait and geysir
22:43 cait :)
22:43 rangi and paul, hdl and slef too
22:43 rangi cait and magnuse have been there
22:42 jcamins Mmm. Boiling mud.
22:42 rangi to see geysers and boiling mud etc
22:41 jcamins Cool!
22:41 rangi taking the family to rotorua
22:41 rangi on leave next week
22:41 jcamins Yeah, might be a good idea.
22:41 jcamins Hmm...
22:40 rangi from a csv file
22:40 rangi first I have to make the authorities
22:40 jcamins :)
22:40 jcamins Grr.
22:40 rangi but not for a week or so
22:40 jcamins Well, give it a whirl.
22:40 rangi I do need to do authority linking
22:40 rangi Hehe
22:40 jcamins So, if anyone was looking for something *incredibly absurdly totally and utterly* awesome, today is your lucky day.
22:38 cait very cool :)
22:38 rangi cool!
22:38 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=7284 enhancement, P2, ---, jcamins, ASSIGNED , Authority matching algorithm improvements
22:38 wahanui Bug 7284 is probably the first step to making that happen.
22:38 jcamins Bug 7284
22:38 jcamins Already posted the patches.
22:37 jcamins If I remember I'll try next time.
22:37 jcamins Hm.
22:37 rangi it might work :)
22:37 rangi and try that
22:37 rangi read up on git, how to specify a range
22:36 jcamins cait: that's what I do, too.
22:36 cait but not really elegant...
22:35 cait and then just delete one ^ after the other
22:35 cait and only -e for the first
22:35 cait jcamins: i do something like HEAD^^^
22:35 jcamins That's what I figured.
22:35 rangi not without hackery I think
22:34 jcamins Is there a way to get git-bz to attach multiple patches with one command?
22:26 magnus_away c ya cait!
22:26 magnus_away rangi: will do
22:26 cait bye magnus_away
22:25 rangi make sure u write it up :)
22:25 magnuse see ya tomorrow, folks!
22:25 rangi :)
22:25 magnuse well, that's a headache for another day, i'm off to watch the last half hour of "Lewis", season 5 episode 3!
22:25 rangi hmmm
22:23 magnuse yeah, weird
22:23 rangi looks like man page stuff
22:22 magnuse E: Failed autobuilding of package
22:22 magnuse dpkg-buildpackage: error: fakeroot debian/rules binary gave error exit status 2
22:22 magnuse make: *** [binary] Error 2
22:22 magnuse dh_install: koha-common missing files (debian/tmp_docbook/*.8), aborting
22:22 magnuse no luck...
22:21 cait crossing fingers
22:21 magnuse still no guarantee of ultimate success...
22:20 jcamins Hooray!
22:20 magnuse ooh, it's generating nb-NO templates from the translations...
22:19 rangi jcamins: there might be, but can't think of it
22:18 cait lol
22:18 jcamins rangi: I was hoping there was a nicer way.
22:18 rangi making popcorn
22:18 * rangi attempts to burn down the house
22:17 rangi foreach loop?
22:14 jcamins Any idea how to sum up *all* the values in a hash?
22:11 jcamins +1709, -578, after squashing a bunch of small commits together.
22:11 jcamins Wow. I really didn't expect this to involve quite so many lines.
22:10 jcamins That sounds right.
22:10 rangi should do it
22:10 magnuse oops, without the "[^]" of course
22:10 magnuse sudo pbuilder create --othermirror 'deb http://debian.koha-community.org/koha [^] squeeze main'
22:10 magnuse yeah, i think eythian told me to do:
22:10 rangi ull just need that in ur sources list
22:09 rangi that module is at debian.kc
22:09 rangi yep
22:04 magnuse but i think i might know the solution, at least in theory
22:04 magnuse "debian/list-deps: unknown package for CGI::Session::Driver::memcached" does not look too promising, perhaps...
22:04 magnuse thanks
22:03 cait we believe in you :)
22:03 cait magnuse++
22:00 magnuse well, i did say *try to*!
22:00 jcamins magnuse++
22:00 jcamins Woohoo!
22:00 * magnuse has tried to rebase catalyst's l10n work on the 3.6.x branch and is now trying to build packages with translations off the result...
21:46 magnuse yay!
21:46 jcamins That worked.
21:44 magnuse yeah, a copy sounds like what you want
21:43 jcamins I can checkout a copy of my branch, then rebase the copy onto kc/master.
21:43 * magnuse is waiting
21:42 jcamins Wait, I have an idea.
21:40 jcamins I thought that might be the answer.
21:40 jcamins Yes, because on the second branch I'm going to squash commits, and I want to have a working branch if I screw stuff up.
21:40 magnuse maybe cherry-picking?
21:40 jcamins Yeah, that's what I'm trying not to do.
21:40 magnuse but would you want to have the commits in two places?
21:39 magnuse hm, i'm no expert, but i think it will
21:39 magnuse the branch that you want to get the commits from?
21:39 * magnuse was doing it the wrong way a minute ago
21:38 jcamins But won't that eliminate the original branch?
21:38 magnuse "just" rebase that branch onto the branch where you want the commits, then, i would guess?
21:38 jcamins Usually I'd use git format-patch and apply that, but I'd prefer to do it the right way.
21:38 jcamins Commits on a branch.
21:38 magnuse jcamins: are they patch files, or commits on a branch?
21:37 jcamins *do I
21:36 jcamins Okay, how to I rebase a whole bunch of patches onto a new branch?
21:33 cait wanted to it for testing links in the opac
21:33 cait so then the subjects should be correct, but it's quite small
21:33 cait waiting for the 'converted' version
21:33 jcamins Perfect!
21:32 cait quite small... but might be interesting for you, lots of links and things
21:32 cait I have created a small dataset for testing
21:31 cait hm k
21:30 jcamins cait: me too, but it'd take days to reindex, too.
21:30 cait but remind me to dig for some data on monday
21:30 cait :)
21:30 cait jcamins: I was interested in the results of that!
21:30 druthb kia ora, magnuse. :)
21:29 magnuse kia ora druthb!
21:29 * jcamins will beg you for a smaller dataset that you use for testing.
21:29 magnuse "If a table-full error occurs, it may be that the disk is full or that the table has reached its maximum size." http://dev.mysql.com/doc/refman/5.0/en/full-table.html
21:29 druthb hi cait
21:29 cait hi druthb
21:29 * magnuse shares some equinoxe from valrhona with the weekend/late night crew
21:29 cait oh :(
21:28 jcamins I think I'll just drop that database and pretend it didn't happen.
21:23 jcamins Agreed.
21:23 cait that's kinda.. scary
21:23 jcamins I have no idea.
21:22 cait what doest hat mean full?
21:22 * jcamins shrugs.
21:22 cait full??
21:22 cait oh
21:21 jcamins DBD::mysql::st execute failed: The table 'auth_header' is full
21:21 cait ?
21:21 jcamins The tablespace was filled.
21:21 cait what happened?
21:21 cait oh
21:20 jcamins Don't load Mannheim into a Koha database.
21:20 jcamins So...
21:20 jcamins Oh dear.
21:06 jcamins lol
21:06 cait jcamins: not you if you are worried about that :)
21:04 cait :)
21:04 magnuse guten abend cait
21:03 cait hi magnuse :)
21:03 * magnuse waves
20:59 jcamins lol
20:59 jcamins What are you testing?
20:59 cait if I did tell you everything... would not be a lot of fun
20:59 cait this is a test
20:59 cait hehe
20:59 cait always 9 digits
20:59 jcamins Something else you should've warned me about? :P
20:59 cait that' scorrect
20:59 cait ah
20:59 jcamins Yeah.
20:59 jcamins 001.
20:59 cait the number in 001?
20:59 cait what does it look at?
20:58 cait hm
20:58 jcamins Strangely it's starting at 16million.
20:58 jcamins Authorities are working fine.
20:58 cait that wouldn't be very nice... to ignore yo
20:57 jcamins Ignore me.
20:57 jcamins Oh, I just don't know how importing authorities works, apparently.
20:57 cait so now they show?
20:56 cait hmm
20:56 jcamins Actually, I'm just having issues figuring out how to access authority records.
20:55 jcamins I have 171k authority records in my database, but none show up in the staff client.
20:55 cait so it's a good time to make changes to the converter and improve things
20:55 cait the advantages of a central cataloging database :)
20:55 jcamins This is weird.
20:55 jcamins Well...
20:54 cait but also the bibliographic records
20:54 cait authorities especially
20:54 cait and we will reload all koha libraries with union catalog data
20:54 cait yep
20:54 jcamins Exciting times. :)
20:53 cait ...
20:53 cait training catalogers
20:53 cait all union catalogs closing, converting data to new format
20:53 cait it's not bad news, but it's really big
20:53 jcamins I suppose that's possible.
20:53 cait nice data? :P
20:50 jcamins That worries me.
20:50 jcamins This is blazing fast.
20:50 jcamins 140k imported.
20:50 jcamins Uh-oh. Big projects are always bad news. ;)
20:50 cait it's a real big project
20:49 jcamins Why?
20:49 jcamins In May?
20:49 cait for the authorities
20:49 cait because they change the fomrat
20:49 cait and it would be cool to figure that out until may
20:49 jcamins Yes, but maybe you can find a use for it. :)
20:49 cait but use koha internal numbers for linking
20:49 cait but it would be nice to not rely so much on our authority number
20:49 cait we import them linked
20:49 cait them
20:49 cait at the moment we don't link then
20:48 jcamins :)
20:47 jcamins Perhaps Uli will be interested in testing the linking, to see if it can simplify things for you at all.
20:47 jcamins That's very diplomatic.
20:46 cait the structure of the 689 is ... not so nice I think
20:46 jcamins 40GB each better be enough, 'cause I don't have infinite disk space on this server.
20:45 jcamins Better increase the size limits on the Zebra directories...
20:45 jcamins :D
20:45 jcamins Eh. It's running on a server. We'll see what happens.
20:45 cait topical, geographic
20:45 cait no meeting
20:45 cait personal names
20:44 cait corporate authors
20:44 cait there will be
20:44 jcamins Well, I have 90k authorities from BSZ-BW imported now.
20:44 cait about that part... the other things should not make a big difference
20:44 jcamins That's okay.
20:44 cait didn't think about it
20:44 cait sorry
20:43 jcamins Yeah...
20:43 cait I should have told you that...
20:43 jcamins That won't help me test your subject headings. :P
20:43 cait oh right
20:43 cait there might be some from loc
20:43 cait hm
20:42 jcamins Uh-oh. So there aren't subjects in 650s?
20:42 cait look at that
20:42 cait 689
20:42 cait breaking them up into single subjects and move them to the proper marc21 fields
20:42 jcamins Moving the subject chains?
20:42 cait like moving the subject chains
20:42 jcamins Well, I'll bug you if it doesn't work. :P
20:42 cait and some other evil things
20:42 cait unicode normalization
20:42 jcamins Hm.
20:42 jcamins Ah.
20:42 jcamins What does your Koha converter do?
20:42 cait so you will have wrong normalization perhaps
20:41 cait and it's not converted... with our koha converter
20:41 cait oh
20:41 cait it's data from the union catalog
20:41 jcamins I guess it's not, then.
20:41 jcamins Right.
20:41 jcamins Oh.
20:41 jcamins It seems to me that this must be one of the larger installations.
20:41 cait nope
20:41 jcamins cait: you have this in a Koha instance?
20:39 jcamins 50k imported...
20:35 jcamins 20k imported...
20:34 jcamins Yup.
20:34 cait are there 2 auth files?
20:34 jcamins Hopefully that will be enough.
20:34 * jcamins will try just importing _1.
20:34 jcamins Eek.
20:34 jcamins 1.4 million titles?
20:33 jcamins Importing Mannheim's data.
20:31 jcamins In this case, it means "more than one authority matched."
20:31 cait sounds right
20:30 cait yep
20:30 jcamins cait: oh, in answer to your question earlier, "fuzzy" is whatever the linker defines it as
20:30 jcamins cait: aut = authorities, tit = biblios?
20:28 jcamins Ah. Maybe.
20:27 cait http://www.belwue.de/ - perhaps that's the reason?
20:26 jcamins Oh, maybe they have Internet2.
20:26 cait our servers are at the data center of university of konstanz
20:26 cait and you are probably downloading at a good time
20:26 jcamins It's nice and fast, whoever hosts it. :)
20:25 cait perhaps has something to do with belwue?
20:25 jcamins :)
20:25 cait oh
20:25 jcamins I've never gotten 5MB/sec download before.
20:25 jcamins BSZ-BW's data center.
20:24 cait my what?
20:24 jcamins Pretty please? :)
20:24 jcamins cait: can I please use your data center?
20:24 jcamins Wow!
20:17 * cait offers dried cherries
20:17 cait i don't know
20:17 cait don't look at me
20:16 * jcamins refuses to ask the follow up question: "Should I?"
20:16 jcamins Answer: yes.
20:14 jcamins Can I softlink to a file which is under a softlinked directory?
20:11 cait jcamins: will be even happier when you help me doing it :)
20:10 wizzyrea bbiab
20:08 jcamins :)
20:08 jcamins cait: you'll be glad to know that writing your linker module will be super-easy.
20:06 cait says 2 parts
20:06 cait https://wiki.bsz-bw.de/doku.php?id=v-team:daten:openaccess:maub
20:06 cait yep
20:06 cait sliced up a bit
20:05 cait I think so
20:05 jcamins Just different parts of the data?
20:05 jcamins What's the difference between _1 and _2?
20:05 jcamins No, I have it.
20:05 cait want me to find the link?
20:05 jcamins But I want to test it on a server, and I have a lousy home connection.
20:05 jcamins I did.
20:04 cait thought you had downloaded it
20:04 jcamins Wait, found it.
20:04 cait ah
20:04 jcamins cait: remind me where to find your data?
20:03 jcamins (but, when you don't have an authority record, you're kind of SOL on the whole "automatic linking" thing.
20:02 jcamins The cataloger is going to be *pissed*.
20:02 jcamins No auth record.
20:02 jcamins Figured out why these aren't linking.
20:01 jcamins Ohh.
20:01 rangi [off] yeah it makes it harder for people to help you
20:01 jcamins [off] Especially not when I'm grouchy.
20:00 jcamins [off] This is why you shouldn't be a jerk.
20:00 rangi :)
20:00 cait jcamins: I should have seen that coming...
20:00 * jcamins will consider.
19:59 jcamins True.
19:59 jcamins Not that it's likely they'd refuse to give it to me.
19:59 rangi i just thought might be a nice PR
19:59 rangi yep
19:59 * jcamins would probably prefer to beg MDAH for data, though.
19:58 jcamins I'm sure he would.
19:58 jcamins cait: you can write your own linker for that, you know. Very easy with my new architecture.
19:58 rangi i wonder if that paul guy would give you data
19:58 rangi hmm
19:57 cait :P
19:57 jcamins Right.
19:57 cait and you know that our data could be perfectly linked... if the script understood the $0
19:57 jcamins I didn't try your data yet.
19:57 cait how long did it take for those now?
19:57 * jcamins will be trying with cait's data. :)
19:57 cait I gave him data!
19:57 * cait hides
19:57 jcamins :)
19:57 rangi hehe
19:57 rangi like cait
19:57 rangi and worth getting others to test
19:56 cait what does fuzzy link mean?
19:56 jcamins Yeah, certainly.
19:56 rangi well i think thats a good improvement from the current one
19:55 jcamins Now, how to improve it?
19:54 rangi oh thats pretty darn good
19:54 jcamins 340k unlinked; 495k linked; 5877 fuzzily linked.
19:49 jcamins 12.5MB for the report only.
19:48 cait jcamins++ :)
19:47 jcamins First time running it on the entire database.
19:47 jcamins Exactly.
19:47 rangi that'd do it, well too much info beats not enough
19:47 jcamins Right... guess I need to write a script that will remove that list.
19:46 rangi ahhh
19:46 jcamins It includes a list of every single bib it processed.
19:46 rangi whoa
19:45 jcamins Wow. It's a 21MB file.
19:45 jcamins Heh. Producing the linking report takes a long time.
19:45 jcamins Yup.
19:44 rangi and because it doesnt need to rebuild everything in one go thats totally good enough
19:44 rangi nice
19:44 jcamins ^^ do you see that? In just a few seconds it did over a hundred!
19:43 jcamins 167313
19:43 jcamins Bib 167198.
19:42 rangi it was at the market, was quite funny
19:42 rangi i suspect he has been watching his dad play games
19:42 jcamins lol
19:42 rangi "where's my sword, brigands are attacking"
19:41 rangi my nephew (who is 4) was running around yesterday yelling
19:41 cait :)
19:39 rangi i played 1 - 5
19:39 rangi it was fun
19:39 rangi yeah
19:39 rangi http://en.wikipedia.org/wiki/Ultima_(series)
19:39 jcamins I know about it, though. A bit of a cult classic, I think.
19:38 jcamins Not me.
19:38 jcamins Ooh... almost done with the 167k linking.
19:38 rangi anyone play ultima ?
19:38 rangi hehe
19:38 cait and when I looked at it years later... it seemed quite silly
19:38 cait i got soo scared in those dungeions
19:35 rangi ah yeah i played that one
19:35 cait http://en.wikipedia.org/wiki/Might_%26_Magic
19:35 rangi so actually i spent years listening to a tape squeal
19:35 * jcamins didn't have one of those.
19:35 rangi on my zx spectrum
19:34 cait only might and magic
19:34 cait hm not heroes
19:34 rangi course it took about 17 minutes to load each time
19:34 cait heroes of might and magic - I drew maps for that
19:34 rangi drew maps on graph paper
19:34 rangi i played that for years
19:34 rangi i was 11 when we got that
19:34 rangi http://en.wikipedia.org/wiki/The_Lords_of_Midnight
19:34 jcamins Never played either of those.
19:33 cait but was young too :)
19:33 * cait loved that
19:33 cait has anyone played day of the tentacle?
19:33 rangi lords of midnight
19:33 jcamins Apparently Myst came out almost 20 years ago.
19:32 jcamins Okay, that might be because I was... errr... not very old... when it came out.
19:31 jcamins Myst took *years*
19:31 jcamins It took a week to win. :(
19:31 * jcamins wasn't so impressed with Myst III.
19:30 * wizzyrea finds this funny
19:30 rangi i only play games at my parents house
19:30 rangi heh
19:30 jcamins Wow... there have been two more Mysts.
19:29 jcamins Morrowind is the only game I've played since Myst 3.
19:29 wizzyrea but usually 6-8 months and they'll have a tweak for it
19:29 wizzyrea of course they keep thinking of new, novel ways to break it
19:29 wizzyrea it really is, quite impressive
19:29 rangi wine is pretty sweet these days
19:28 wizzyrea and WoW
19:28 wizzyrea I got Rift to play in Wine
19:28 wizzyrea *I think*
19:28 jcamins That's incredible!
19:28 wizzyrea would be better to run it in windows
19:28 jcamins Wow!
19:28 jcamins Skyrim runs under Wine?
19:28 wizzyrea I think skyrim will, but I don't have my super good video card in linux
19:28 jcamins Wait...
19:28 * rangi hasnt played it, my dad has it tho
19:27 rangi yup
19:27 jcamins rangi: I know, but Morrowind is old enough it might work in wine.
19:27 rangi the latest one
19:27 wizzyrea ^^
19:27 rangi skyrim is elder scrolls jcamins
19:27 jcamins Morrowind?
19:27 rangi http://soundcloud.com/dubstep/skyrim-in-the-knee-by-rameses-b
19:26 wizzyrea but i've really only played SC2
19:26 wizzyrea hehe, I love starcrack
19:26 cait starcraft
19:26 cait oh starcract
19:26 jcamins wizzyrea: you could play Starcraft!
19:25 jcamins wizzyrea: excellent!
19:25 cait morning rangi, hi wizzyrea and jcamins :)
19:25 wizzyrea wine*
19:25 wizzyrea i am installing ancient windows games in wind
19:25 rangi hey cait!
19:25 jcamins Hi bag and cait.
19:25 jcamins Ah. Yeah, I won't get those.
19:25 wizzyrea wow!
19:25 bag yo cait
19:24 cait heya #koha
19:24 wizzyrea heya bag
19:24 wizzyrea see: skyrim memes
19:24 bag hi there wizzyrea
19:24 * jcamins didn't get the joke.
19:24 wizzyrea naw, not a fail - still pretty funny.
19:23 * wizzyrea laughs
19:23 rangi joke fail
19:23 rangi knee
19:23 * rangi use to wave then he took an arrow to the kneww
19:23 wizzyrea whassup
19:23 * jcamins waves back.
19:23 * wizzyrea waves
19:23 rangi oh look its wizzyrea
19:14 jcamins :)
19:13 rangi thats as far as ive thought :)
19:13 rangi $biblio->biblionumber();
19:13 rangi so i can do
19:13 rangi using Class::Accessor as base
19:12 rangi if i were redoing biblio, i think its totally a candidate for OO
19:11 jcamins Right.
19:10 rangi :)
19:10 rangi in theory
19:10 rangi the nice thing with a new namespace is you can do it carefully without busting anything already existing :)
19:09 rangi and redo one subroutine to start
19:08 rangi but start with Koha::Biblio for example
19:08 rangi unless that module is already clean
19:08 rangi is one huge patch that shifts a whole module over
19:08 jcamins Agreed.
19:08 jcamins Right.
19:08 rangi what i dont want to see
19:08 jcamins Ah.
19:08 rangi no in general
19:08 jcamins No, just the part that does searching. :P
19:07 rangi and i dont think we need to create the whole module
19:07 jcamins (AuthoritiesMarc<->Biblio, not cycles in general)
19:07 jcamins If we had Koha::Search::Engine already, I could've mostly resolved that problem.
19:07 rangi but not to try to do it all in one go
19:07 rangi is make new modules there, more carefully and transition to them
19:06 rangi thats the main idea behind Koha:: in my mind
19:06 rangi yes
19:04 jcamins Also, the AuthoritiesMarc<->Biblio dependency cycle is very vexing.
19:02 jcamins Just thought I'd point that out.
19:02 jcamins rangi: incidentally, having ModZebra in C4::Biblio is horrifying.
18:59 rangi awesome
18:59 rangi :)
18:59 jcamins Several orders of magnitude faster.
18:59 jcamins 140k records matched today.
18:58 jcamins fredericd: UNIMARC pushed without typo, now.
18:57 jcamins I just pushed a typo.
18:57 jcamins Argh!
18:56 jcamins fredericd: UNIMARC pushed.
18:54 jcamins Good news! I remembered my alphabet after a few tries!
18:51 jcamins I was wrong about that last bit.
18:51 jcamins I was only doing it 'cause fredericd asked and it looked easy.
18:51 jcamins Right.
18:50 rangi kaizen :)
18:50 rangi but i wouldnt hold up your matching work trying to do it
18:50 rangi it can be a separate patch
18:49 jcamins Sorry, fredericd.
18:46 rangi yup
18:46 jcamins "No regressions" = good enough for me.
18:46 jcamins Right, that's the plan.
18:45 rangi id probably do the dynamic bit as a separate thing then
18:44 rangi ah
18:44 jcamins Well, it was up until I decided fifteen minutes ago not to hardcode it.
18:44 rangi so caching it will not be the same
18:44 jcamins My plan C is "put it back to hardcoded because this really isn't the issue that I'm trying to deal with."
18:44 rangi yep, but now its not
18:44 jcamins That has not bothered anyone in 3 years.
18:44 rangi yep
18:44 jcamins Right, it used to be hardcoded.
18:44 rangi either way, if you cache something, you have to clear it when you change the value
18:42 jcamins Okay... what's plan B?
18:41 jcamins Hm.
18:40 rangi i think preferences in Context does it
18:39 rangi in mod_perl .. it will last forever
18:39 rangi for cgi it doesnt matter so much, the variable dies when the script does
18:37 rangi right so that will now need to call a sub in the other one called something like clear_cached_heading or something
18:36 jcamins That's in a different file.
18:36 jcamins Hm.
18:36 rangi find the routine that sets the value in the db, and make that clear the variable
18:36 jcamins rangi: how do I do that?
18:33 rangi and make sure you clear/reset that variable whenever the db changes, or it will cause a big mess when running in a persistent environment like mod_perl or plack
18:30 jcamins Yup.
18:29 jcamins our?
18:29 jcamins Okay... it doesn't like variables declared with my.
18:12 jcamins Thanks.
18:12 jcamins Cool!
18:12 jcamins I have not.
18:11 fredericd Have you checked that: http://perldoc.perl.org/perlmod.html#BEGIN,-UNITCHECK,-CHECK,-INIT-and-END
18:11 jcamins Drat.
18:11 jcamins Wait, maybe it does.
18:10 jcamins That doesn't have to be specific to the MARC flavor.
18:10 jcamins Well, just in C4::Heading.
18:10 fredericd In C4::Heading::MARC21 ?
18:08 jcamins Maybe I'll try memoize.
18:08 jcamins Yeah, my difficulty is that I really need the equivalent of the C++ "static".
18:07 fredericd jcamins: I suppose you have during initialization and populate the appropriate data structure
18:05 jcamins I do not want to query MySQL *every* time I am processing a heading.
18:05 jcamins So I need to figure out whether a given field is authority-controlled.
18:04 jcamins We discussed looking up in the framework which fields were authority-controlled, rather than hardcoding it.
18:03 jcamins Un-never mind.
18:03 jcamins Oh, no, I didn't.
18:03 fredericd good for you
18:03 jcamins As I wrote out my question in more detail I found the answer.
18:03 jcamins Wait, never mind.
18:02 fredericd I don't understand
18:01 jcamins fredericd: which fields are associated with authority records.
18:01 fredericd jcamins: how to cache what?
16:50 ktk does it sort out issue with any latest version of koha? or is there any way to make changes in the database directly and get this error sorted out.
16:49 ktk is there any patch available to this bug then "to get rid of an error for patrons with message getting an error as patron has had overdue items and is blocked for x days" we are getting this error for several of our borrowers and we cant do any transactions
16:31 jcamins (I mean, very recently someone mentioned it)
16:31 jcamins ktk: I think I recall someone mentioning a bug that caused this, but I don't recall what it was.
16:28 ktk hw to get rid of an error for patrons with message getting an error as patron has had overdue items and is blocked for 833 days We are getting this error for many patrons
16:19 jcamins It's too early for NZ still.
16:19 jcamins fredericd: any thoughts?
16:19 jcamins fredericd: I'm still having trouble figuring out how to cache the results of retrieving a list of tags linked to authorities.
15:46 jcamins druthb: lol
15:44 * jcamins grumbles about circular dependencies.
15:44 druthb "Thick as mud", by D Ruth Bavousett. "Living with a Crazy Transsexual", by Lindsey Bavousett. "I'm a Teenager, Therefore My Life Sucks," by Libby Bavousett. "Meh, Whatever." by Captain Midnight Bavousett
15:44 jcamins Excellent, this works.
15:43 druthb lol
15:42 * jcamins moves on to cataloging imaginary books by members of druthb's family.
15:42 jcamins Having exhausted members of his own family...
15:41 jcamins :)
15:38 * druthb is a little slow sometimes, but she usually gets it, eventually.
15:38 jcamins Exactly.
15:37 druthb probably oughta put that in C4::Search::Zebra or some such, then, in the long run.
15:37 druthb ahhh...
15:37 jcamins Providing a "biblionumber" that looks suspiciously like the authid.
15:37 jcamins In C4::Biblio.
15:37 jcamins In order to modify an authority record, you must call ModZebra.
15:36 jcamins Ah, no, you misunderstand me.
15:36 druthb Properly, that probably ought to be two separate functions. If you check out an item, you need to update the index---you *don't* need to tinker with its' attached authorities at that point.
15:35 jcamins Yeah, I know.
15:35 druthb it's also used to trigger reindexing of a biblio.
15:34 jcamins authorities.
15:34 jcamins The function used to modify...
15:34 wahanui it has been said that modzebra is part of C4::Biblio.
15:34 jcamins ModZebra.
15:34 druthb yes....?
15:34 jcamins druthb: ModZebra is part of C4::Biblio.
15:34 * druthb occasionally worries about jcamins, you know.
15:26 jcamins We do the stupidest things!
15:26 jcamins AAAAAAAHHHHH!!!!!!!
14:41 cait I have to run to catch my bus - bye all
14:41 cait hm bug not patch
14:40 cait perhaps link that to the diacritics patch?
14:37 cait worth a test :)
14:35 fredericd Yes. But it may have been fixed in Zebra/YAZ itself
14:33 cait we never really figured it out, but there is some weirdness about them and icu
14:33 cait I think it was stemming, fuzzy or/and field weighting
14:33 cait there are some system preferences that don't work well with icu
14:33 cait and another for marc21
14:32 cait I think a sign off from someone knowing unimarc would be good then
14:32 cait oh!
14:32 fredericd cait: This patch is essential for UNIMARC libraries. Without it, authors facets don't include firstname
14:28 cait hope to find time to look at it... but probably not for a while :(
14:28 cait sounds really good
14:24 fredericd and you can choose during installation
14:24 fredericd general
14:23 cait is this only icu for facets or will it use it in general then?
14:23 cait we use icu for one library with hebrew
14:23 cait and hi fredericd :)
14:22 fredericd yes
14:22 cait will it work for marc21 too?
14:22 cait 3216 looks really interesting
14:19 jcamins Yeah, I was just thinking of doing that.
14:19 fredericd so, code using that should be removed before
14:18 jcamins Yes, but that doesn't serve any useful purpose.
14:18 fredericd it doesn't exist in auth_tag_structure and auth_subfield_structure tables...
14:18 fredericd jcamins: C4::Heading::MARC21::$bib_heading_fields has something to distinguish subject/series/main_entry
14:14 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=3216 normal, PATCH-Sent, ---, frederic, NEW , UNIMARC author facets
14:14 fredericd I have a pending patch for improving Zebra indexing, and facets building: bug 3216
14:14 jcamins I'm getting paid to fix linking for MARC21, and I can add a UNIMARC class easily enough.
14:13 * jcamins is not writing UNIMARC indexing.
14:13 fredericd we don't use Zebra as we should/could
14:12 jcamins GRS-1.
14:12 jcamins Actually, bigger problem: UNIMARC uses different indexing.
14:11 jcamins The form subdivision in UNIMARC is $j.
14:11 fredericd was replying to your subdiv question
14:10 fredericd Yes, but they could stay hardcoded. Not absolutely sure, but I think that MARC21 and UNIMARC share the same letters to identify subdivisons
14:10 jcamins fredericd: do headings currently link with authorities in UNIMARC at all?
14:09 jcamins Oh, there's the problem. Subdivisions.
14:08 jcamins memoize valid_bib_heading_tag, and have it take $tag and $framework?
14:07 jcamins Except for one thing. How do I cache the results of the database access?
14:07 jcamins Yes, I like that idea.
14:05 fredericd This way, we even could merge MARC21 and UNIMARC classes...
14:03 jcamins Wait, no I don't.
14:01 jcamins Ah, got it.
13:59 jcamins So, the question is, how do I cache the results of SELECT authtypecode,tagfield FROM ...
13:57 fredericd I have libraries who use specific authorities types located in the 69X biblio tags. For them, C4::Heading::MARC should have to be modified by hand ($bib_heading_fields)
13:53 jcamins Since someone might not want to link 655.
13:53 jcamins Yeah, and I've been thinking that I'd prefer to use that.
13:53 fredericd The framework must contain info required to build $bib_heading_fields data structure
13:51 jcamins Thoughts?
13:50 jcamins I've been thinking of making it use the framework, though.
13:50 jcamins It's certainly not ideal, but it's been in Koha for years, so it can stay a bit longer.
13:48 fredericd and I don't like that much how C4::Heading::MARC21 hard code authorities definitions...
13:48 jcamins I will add one. :)
13:48 fredericd I can't test your patch without having a C4::Heading::UNIMARC class!
13:47 fredericd you can do a git merge
13:47 jcamins I guess I should figure out why and fix that at some point.
13:46 jcamins GitHub won't let me push to it anymore.
13:46 jcamins The "master" branch on there is spectacularly out of date.
13:46 jcamins Ah. Cool! I didn't know that.
13:46 fredericd I can with git diff jcamins/master, where jcamins is the name of your remote repo
13:45 jcamins fredericd: does it make sense?
13:45 fredericd I was trying to figure out you modification
13:45 fredericd But I don't need it anymore
13:44 fredericd Yes. You could have a branch bug_7284_v2 for example
13:43 jcamins That should work... right?
13:43 jcamins fredericd: what if I used a different branch?
13:40 jcamins (I guess that's the disadvantage of frequent upstream changes... it makes sharing repositories difficult)
13:38 jcamins I can't push branches after rebasing.
13:37 jcamins fredericd: it is based on head as of two days ago.
13:37 jcamins fredericd: I'll rebase when I'm done.
13:34 fredericd jcamins: I got your branch. Could you rebase it to HEAD? It would help tracking your modifications.
13:31 jcamins Seems to.
13:31 jcamins right?
13:31 jcamins It returns a MARC::Record object...
13:31 jcamins What does TransformHtmlToMarc do?
13:27 fredericd I don't say no :-)
13:27 jcamins :P
13:27 jcamins fredericd++ # for volunteering to test linking with UNIMARC.
13:27 jcamins fredericd: but that means someone will have to test it. :P
13:26 jcamins fredericd: I'll add a simple one, though.
13:26 jcamins fredericd: we've never had one in the past.
13:26 fredericd jcamins: A C4::Heading::UNIMARC class is required...
13:24 * jcamins may add an additional sign-off on that.
13:24 jcamins fredericd++
13:23 jcamins That's exactly what I was just thinking we needed!
13:23 jcamins Cool!
13:23 jcamins fredericd: https://github.com/jcamins/koha/commits/bug_7284
13:23 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=6990 enhancement, PATCH-Sent, ---, frederic, NEW , TransformKohaToMarc enhancement
13:23 fredericd patch for bug 6990 could help
13:22 fredericd jcamins: is your code somewhere?
13:21 jcamins I mean, the fact that updating a record is hugely inefficient is bad, but it's progress.
13:20 jcamins cait: yes, that's good.
13:20 cait um yay? :)
13:09 jcamins Woohoo! MySQL now takes more time than Zebra!
13:01 jcamins Namaste, kmkale.
13:01 kmkale Namaste jcamins :)
13:00 jcamins I should make BiblioAddsAuthorities use the new working linking.
13:00 druthb Indeed. It's been a briskly busy fall for me, so I haven't been on here much.
12:59 kmkale LTNC
12:59 kmkale Namaste druthb :-D
12:57 druthb Namaskar, kmkale!
12:52 magnuse bbl
12:47 jcamins I'll profile 10k, then try my modification.
12:46 cait go jcamins go
12:46 cait heh
12:45 * jcamins has another idea for optimization.
12:44 wahanui jcamins: excuse me?
12:44 jcamins wahanui: interesting things? Yes indeed.
12:43 huginn 04Bug http://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=7284 enhancement, P5 - low, ---, jcamins, ASSIGNED , Authority matching algorithm improvements
12:43 wahanui Bug 7284 is the first step to making that happen.
12:43 jcamins Bug 7284
12:43 fredericd intersting...
12:42 jcamins fredericd: more generally, I completely rewrote all the linking code.
12:42 jcamins fredericd: for the performance change I just uncommented the oAResultSet->destroy().
12:41 fredericd what have you modified?
12:41 jcamins Yes.
12:41 jcamins No typo there.
12:41 cait so it should be faster now :)
12:41 cait oh
12:41 jcamins 33238s.
12:40 jcamins cait: previously 5k records took...
12:40 jcamins cait: well, 3k records just took... 308s.
12:40 jcamins Being impatient, that last is problematic.
12:40 cait I think you began with an hour?
12:40 cait how long does it take now?
12:40 cait hehe
12:40 jcamins Without dying of boredom.
12:40 jcamins fredericd: run link_bibs_to_authorities.pl on 167k records.
12:39 * druthb finishes loading 400K+ bibs and the quarter-million patrons, and runs some touch-up scripts.
12:39 fredericd jcamins: what are you trying to do exactly?
12:38 jcamins No, thank goodness.
12:38 cait so zebra is not so bad? ;)
12:36 jcamins Whoops.
12:36 jcamins But it was commented out.
12:36 jcamins I didn't recognize that was the issue at first because we already had the destroy command.
12:36 * magnuse hums "anarchy in the uk"
12:36 cait sounds radical :)
12:36 cait destroy!
12:35 jcamins The problem, in case you were wondering, is that you have to destroy your Zebra recordSets when you're done with them.
12:35 druthb At this hour on a Saturday morning...he's trying to set an example for you.
12:35 jcamins He doesn't like Zebras.
12:35 jcamins Actually, Myshkin has abandoned me to cuddle with Shari.
12:34 druthb "we." You and Myshkin?
12:34 cait woohoo
12:34 magnuse woohoo!
12:34 druthb yes, yes it is.
12:34 jcamins Ladies and germs, we have solved the problem!
12:33 jcamins 14.6s < 71.1s.
12:33 * jcamins does some math.
12:33 jcamins Wait...
12:33 magnuse yay
12:33 cait :)
12:33 cait heh
12:33 druthb lol
12:33 jcamins This is good.
12:33 jcamins Hrm. Microseconds are smaller than milliseconds.
12:32 * druthb enjoys watching jcamins sometimes.
12:32 * magnuse knows the feeling ;-)
12:32 jcamins Drat.
12:31 jcamins Wait, maybe I didn't solve it.
12:31 jcamins Woohoo!
12:29 jcamins To be clear, the problem is not Zebra, it's the Perl interface to it.
12:26 jcamins It's unfortunate that I can't check whether my changes help quickly.
12:24 jcamins With mod_perl, not so much.
12:24 jcamins With CGI, Zebra connections are short-lived.
12:24 * magnuse is confused
12:24 magnuse ah, so where does mod_perl come into it?
12:23 jcamins Nope, running it from the command line.
12:23 magnuse ah, and you are running that with mod_perl?
12:20 jcamins link_bibs_to_authorities.pl.
12:20 jcamins I am currently telling myself "surely there's something I can do to make it happy."
12:19 magnuse what kind of operation are you profiling there?
12:19 jcamins It is *100* times as long per call.
12:19 jcamins It is.
12:19 magnuse 32735s is quite a long time...
12:17 jcamins For 5k routines, Net::Z3950::ZOOM::event took 32735 seconds to run 95441 times. For 1k records, 71.1s to run 19181 times.
12:16 jcamins The important bits are columns 1 and 4.
12:15 pastebot "jcamins" at 98.14.120.176 pasted "Yikes! This is NOT O(n)" (31 lines) at http://paste.koha-community.org/192
12:15 magnuse would it be the same for plack? i don't know either of them...
12:14 magnuse ouch?
12:13 jcamins Alas, this is Zebra.
12:13 cait and hi jared
12:13 cait but how we use it?
12:13 cait perhaps it's not zebra
12:12 jcamins Welp, looks like using mod_perl with Zebra Might Not Be A Good Idea.
12:10 magnuse yay
12:09 cait ground
12:09 cait it melts when hitting the grond
12:09 magnuse +2 and light snow = slush ?
12:09 druthb guten tag, cait. :) *hug*
12:09 huginn cait: The current temperature in Konstanz, Germany is 2.0ï¿½C (1:00 PM CET on January 07, 2012). Conditions: Light Snow. Humidity: 77%. Dew Point: 0.0ï¿½C. Pressure: 30.24 in 1024 hPa (Falling).
12:09 cait @wunder Konstanz
12:09 magnuse guten tag cait
12:08 huginn druthb: The current temperature in Channel 6 Downtown, Lawrence, Kansas is 0.9ï¿½C (6:08 AM CST on January 07, 2012). Conditions: Clear. Humidity: 63%. Dew Point: -5.0ï¿½C. Windchill: 1.0ï¿½C. Pressure: 30.12 in 1019.9 hPa (Steady).
12:08 druthb @wunder lawrence, ks
12:08 cait heya
12:08 wahanui That's Ms. Cait to you! or in a very good mood today :)
12:08 druthb cait?
12:08 huginn magnuse: The current temperature in Bodo Vi, Norway is -7.0ï¿½C (1:00 PM CET on January 07, 2012). Conditions: Partly Cloudy. Humidity: 67%. Dew Point: -10.0ï¿½C. Pressure: 29.45 in 997 hPa (Falling).
12:08 magnuse @wunder boo
12:08 magnuse hehe
12:08 wahanui I'M TAKING THE PISS OUT OF 'EM. or a ratbag.
12:08 druthb wahanui: Brooke?
12:08 * druthb giggles.
12:08 wahanui magnuse is afraid that we added another 10000 bugs while he was eating pizza.
12:08 druthb wahanui: magnuse?
12:08 druthb hm.
12:04 magnuse oops?
12:04 * druthb nods.
12:04 wahanui She finally snapped, like we all knew she would.
12:04 druthb wahanui: druthb?
12:03 magnuse it has been rumoured...
12:02 jcamins Well, eww. Zebra has some major issues, doesn't it?
12:02 druthb :)
12:02 magnuse lucky quarter-million!
12:01 * magnuse is wowed
12:01 * druthb is bringing another quarter-million library patrons into the world of being served by a Koha-powered library. :)
12:00 jcamins_away o/
12:00 magnuse o/
11:57 druthb \o/
11:57 druthb \o
11:57 druthb o/
09:38 magnuse hiya cait and rangi
08:48 rangi hi magnuse
08:47 cait hi magnuse :)
08:42 magnuse nick magnus_afk
08:42 magnuse bbl
08:41 magnuse kia ora #koha
08:06 rangi sleep well
08:03 bag :D
08:03 bag alright I'm off to bed
08:03 cait heya bag
08:03 bag hey there rangi and cait
07:56 cait wb rangi
07:55 rangi back
07:26 * rangi will bbiab
07:26 rangi :)
07:26 cait don't forget the cookies :)
07:24 * rangi has to head off in aminute to buy some groceries
07:08 trea yea, just noodlin around
07:08 rangi ur up late trea
07:08 trea hi cait
07:08 cait hi rangi :) hi trea
07:08 trea hey rangi
07:08 rangi hey trea and cait
05:54 huginn jcamins_away: The operation succeeded.
05:54 jcamins_away @later tell rangi In case you were wondering, ZOOM doesn't store IDs.
03:01 druthb o/
00:46 rangi xs means its a wrapper for compiled c binaries
00:45 jcamins_away Oh, right. They're different. Good point.
00:44 rangi expat_xs maybe
00:44 jcamins_away According to cfouts, expat is faster anyway.
00:43 rangi I suspect you can tell it what to use somehow
00:43 rangi heh
00:42 jcamins_away Wow. We use every single parser module.
00:41 jcamins_away I was just hoping you knew off the top of your head. :)
00:41 jcamins_away I'll poke around.
00:41 * rangi is on phone so can't look
00:40 rangi its marc::file::xml eh?
00:40 jcamins_away The configured parser is LibXML, and it's using that, but it also uses Expat, which is incredibly slow in comparison.
00:39 rangi maybe you can tell it what to use
00:39 jcamins_away That's what I would've thought, but it looks like we're wrong.
00:39 rangi try libxml::xml instead or something
00:39 rangi you could try messing with parserdetails.ini
00:38 rangi hmm I think it uses whatever the xml parser is
00:36 jcamins_away rangi: any idea why MARC::Record uses expat?