IRC log for #koha, 2007-11-21

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

All times shown according to UTC.

Time Nick Message
12:27 kados hdl: you around?
12:27 hdl: I can send you the patches I've done
12:27 hdl: I've fixed simple searching, and boolean searching, truncation, field weighting, stemming and relevance
12:28 (and fuzzy operator handling)
12:28 hdl: still to do: limits, mc-*, and repairing facets
12:37 fbcit g'morning kados
12:40 kados morning fb
12:40 oops
12:40 fbcit:
12:40 :-)
12:42 fbcit kados: talked with chris last evening and it looks like he has the installer about wrapped up for now.
13:09 hdl kados : ok. send it to me.
13:12 kados hdl: I plan to work on the rest of the tasks this afternoon
13:12 (this morning I have a meeting :/))
14:55 Brooke howdy
14:58 owen Hi Brooke
14:59 Brooke hi owen :)
14:59 do you know when 3.0 is due out?
14:59 folks just asked me and I was kinda like, uh soon. :)
14:59 owen Not soon enough ;)
14:59 Brooke exactly!
14:59 owen A developer's release should be coming out soon, I think.
14:59 Brooke i did the whole I'm just guessing routine, but I've seen it with my own beady eyes
15:00 owen But I don't know about final release.
15:00 Brooke *nod*
15:00 do you know where they'll post the dev release?
15:00 owen Presumably on koha.org
15:00 Brooke *nod*
15:00 owen paul_ or hdl around?
15:01 Brooke paul's usually away when he _
15:01 hdl hdl yes
15:01 owen Hi hdl
15:01 Brooke bonjour henri
15:01 hdl hi
15:01 owen There's a check for "uppercasesurnames" in memberentrygen.tmpl, but I don't see that in system preferences
15:03 hdl Then it should be added to systempreferences.sql in install. Maybe in updatedatabase too... depending on Joshua's opinion.
15:05 owen Why is memberentry doing this: $data{'firstname'}=ucfirst(lc $data{'firstname'}); ?
15:05 Wouldn't that mess up an odd first name like "DuPage" ?
15:07 ryan the syspref could be a perlish expression for name handling:
15:07 uc(lastname) ucfirst(firstname)
15:07 or smthg
15:08 if extra parsing was worth it
15:08 but i say delete it.
15:08 javascript would be a better way maybe
15:09 JYL57 hello koha
15:10 could someone explain me the use and importance of table marc_word ?!
15:10 Brooke okay, back to trying to find a jorb.
15:11 hdl ryan : I hate when all is done in Javascript. This means not handy sustainable code.
15:12 JYL57 some mysql expert around ?! how long could a REPAIR TABLE normally go ?!
15:12 owen hdl: but more easily customized to the needs of a particular library
15:12 JYL57: Are you repairing marc_word?
15:13 ryan hdl: I maybe agree, but i also don't like two or more sytemprefs for handling name casing
15:13 hdl JYL57: it can be VERY LON
15:13 G
15:15 owen : if you have to modify one js each time for your customer... It will soon be a mess for you to manage devs
15:15 owen On the other hand, it's better than hard-coding text-transformation that isn't appropriate for all customers.
15:15 JYL57 yes, I'm trying to repair marc_word...
15:15 hdl marc_word is used in rel 2_2 in order to index data
15:16 JYL57 ok, I need advice hdl, would the option USE_FRM be quicker ?!
15:16 hdl JYL57: you had better drop the table and recreate it from a script which will use your database.
15:17 JYL57 hum, how to do that ? do you have sample code ?!
15:18 hdl build_marcword in misc is there for that.
15:19 JYL57 any prerequisite before using this script ?!
15:19 hdl may need customisation.
15:19 You should first read it.
15:19 Joshua did good job on inline documentation
15:21 JYL57 ok, thanks for advise, have to go now leaving my REPAIR working and will check it tonight...
15:22 have a nice day
15:23 atz it's funny to see people jump in w/ a problem and jump out on the same screen
15:25 hdl :)
15:26 And it is quite reassuring. Some questions are quickly answered.
15:27 atz that's true.  sometimes all you need is for somebody knowledgeable to say "that's normal."
15:29 owen hdl: it looks like we lost "Alternate Contact" from member entry somewhere along the way?
15:47 hdl owen : I have not touched members pages for some time
15:56 owen If IndependentBranches is turned on, a librarian should not be able to log in to anything but their home branch (as specified in their patron record)?
16:00 hdl yes
16:02 owen So if IndependentBranches is turned on, we should hide all interface elements related to switching branches
16:15 hdl aye you're right owen. (Unless user is superlibrarian... he should not be able to switch branches)
16:16 owen Easy enough to do a check for AutoLocation and IndependentBranches, but how to tell if the user is a superlibrarian?
16:17 Do we need to add $template->param( IS_superlibrairan       => 1 ); to Auth.pm?
16:18 hdl user flag is >0
16:19 C4::Context->userenv->{flag} >0... After a getuserandtemplate.
16:20 owen Sorry hdl, I don't understand
16:22 hdl see neworderempty.pl
16:23 $onlymine is set when
16:23 IndependantBranches
16:23 And user have a branch and is not superlibrarian.
16:23 line 172
16:33 kados : search adds implicit left truncation. This is not considered as a feature.
16:35 For us when you query Jean, unless you query Jean*, you should not get any author like jeanmaire Jeanine Jeaneune
16:35 Is this a requirement for you ?
16:36 atz hdl: you mean on an author query, i.e. not by keyword?
16:37 hdl yes
16:37 even keyword should not have implicit truncation.
16:37 We have a simple facility to add it : *
16:38 atz i'm not sure about that.  If you have adequate relevance ranking, then it does not hurt to use implicit truncation.
16:38 exact matches will be recognized and returned w/ priority
16:39 chris could easily be a syspref though eh?
16:39 implicit truncation on or off
16:40 atz and you could force it off with search by "jean"...
16:40 (quoted string)
16:40 chris then the library could chose the default behaviour it wants
16:41 atz hdl: i know that K-12 school libraries will appreciate a more robust return set
16:42 kids frequently can't correctly spell or remember all of authors' names
16:42 but I can see why precision would be important in a research environment
16:43 hdl atz, chris : we have research institutions as clients and they require that asking for an author would not bring up 10s of others.
16:43 chris right
16:44 i think a system preference is the way to go then
16:44 on by default, and we can switch it off for libraries that want precision
16:45 hdl Can't you tell me what to do to disable wider scope of results ?
16:46 atz i haven't looked at the searching code
16:46 hdl here is my query :au,phr: "Jean" and (mt:OUVR ) and ( homebranch:CPPM)
16:47 I got :
16:47 Jeandeau and Jean as results.
16:48 chris i dont know how to either, but i think we should get joshua to do a patch so that its controlled by a syspref then you can just switch that off
16:57 owen hdl: should a librarian be able to choose the patron's home branch if IndependentBranches is turned on?
17:03 paul owen : mail in your mailbox, with some more investigations about detail.tmpl
17:10 owen paul, it must be coming by steam ship
17:11 paul steam ship ?
17:11 ah, ok
17:24 hdl owen : when IndependantBranches is on, patrons homebranch is the same as librarians' one.
17:24 branch list is a single item list.
17:25 owen I'm seeing the whole list
17:25 hdl Are you superlibrarian ?
17:26 owen Yes
17:29 hdl owen : Then it is OK.
17:29 superlibrarian has the right to choose any branch.
17:29 But not others.
17:57 owen Logged in as a non-superlibrarian, I'm not able to search patrons
18:10 ryan owen: you can only see patrons registered at your branch
18:11 owen ah, of course
18:14 ...but since I'm not getting logged in to the correct branch, I can't even find patrons registered at my own branch :(
18:16 ryan huh
18:16 was working for me a couple days ago
18:57 martinmorris hi i was asking last night about getting hold of ZOOM 1.16 (i'm playing with koha 3.0)
18:58 had installed Net-Z3950-ZOOM-1.16 (and the guy who replied confirmed that) but when i run perl Makefile.PL it still complains i haven't installed it
18:59 if anyone can advise i'd be grateful ta :)
19:01 ryan martinmorris: for now I would edit Makefile.PL
19:01 and remove that check.
19:01 martinmorris fair enough, thanks ryan
19:01 ryan install the latest (1.20) zoom
19:01 and it should be fine.
19:01 martinmorris will do, cheers
19:01 ryan np
19:06 martinmorris and i'm assumign it's the standard perl Makefile.PL in the kohaclone directory?
19:10 ryan martinmorris: the installer is currently being revised...
19:10 i know chris has done some work on it and will know its state better
19:10 he should be back shortly
19:10 martinmorris ah OK, and i take it there's no notes or anything like that at this stage?
19:11 ryan no - we should have a developer's release out in the next couple weeks
19:11 it'll be cleaned up by then
19:11 martinmorris so i'm  not going to be able to install it that easily then?
19:12 chris yeah it works now
19:12 martinmorris i checked out my copy last night, do i need to check it out again?
19:12 chris nope none of my patches have made it through yet
19:13 martinmorris fair enough
19:13 perl Makefile.PL -> make -> make install?
19:13 chris there is a patch from slef waiting for the RM to check which fixes the check for the
19:13 zoom
19:14 what i have been working on is making it work better
19:14 :)
19:14 martinmorris :)
19:14 chris perl Makefile.pl PREFIX=/somewhere
19:14 if you want
19:14 currently the make will try to copy things
19:14 what i have been doing, is making it so nothing is copied until make install
19:14 ie more standard
19:15 ryan chris: so PREFIX is working now ?
19:15 chris and making it easier to get the right record.abs file for zebra (unimarc or marc21)
19:15 prefix has worked for a while
19:15 what i have done is change it to INSTALL_BASE
19:16 which works the way people expect PREFIX to do :)
19:16 ryan but it still installs modules to /usr/local/share... ?
19:16 fbcit PREFIX/share....
19:16 chris in its current version it might
19:17 i think its the cgi-bin its doing wrong
19:17 martinmorris mine was trying to copy stuff to /usr/local/lib/cgi-bin/koha
19:17 chris yeah thats the one it does wrong
19:18 cgi-bin
19:18 on the make
19:18 martinmorris yep
19:18 chris that should be fixed soon
19:18 ive just sent all my patches to galen who is gonna be taking over fixing the installer
19:18 martinmorris ok
19:18 will there be a z39.50 server included?
19:19 chris yep, if you are running zebra
19:19 martinmorris marvellous
19:19 chris then its a z3950 server
19:19 martinmorris OK
19:19 need to learn more about that
19:23 but in the meantime i'm not going to be able to install this easily without asking lots of questions i take it? :)
19:27 chris lemme just check the wiki
19:28 nope no good page there
19:29 basically you perl Makefile.PL
19:29 make
19:29 make install
19:29 set up your db, and a user and password
19:29 edit koha-httpd.conf
19:29 and koha-conf.xml
19:29 hit the staff site in your browser (after restarting apache)
19:30 and the web installer does the rest
19:31 then its just setting up zebra after that
19:31 fbcit martinmorris: take a look @ README.txt
19:32 martinmorris what about the cgi-bin copy problems at the make install stage?
19:35 fbcit actually, as chris mentioned, the installer does "work" in its present state.
19:36 martinmorris including the cgi-bin bit?
19:37 fbcit right
19:37 martinmorris he just said it was the cgi-bin bit it was doing wrong, am I being a bit blonde?
19:37 :)
19:38 chris its wrong, but it works :)
19:38 fbcit what it is doing wrong does not equate to "doesn't work"... :-)
19:38 martinmorris ah, got you :)
19:38 fbcit though one might think the two are eq...
19:38 martinmorris thanks for your patience
19:39 one last question, i assume i don't play with this on a machine which already runs koha 2.2 happily?
19:41 chris i wouldnt :)
19:41 martinmorris cheers :)
19:41 chris but if you wanted to, you could take a copy of your 2.2 database
19:42 and copy it somewhere and then after doing the make install
19:42 http://wiki.koha.org/doku.php?id=22_to_30
19:42 but its probably best to play with a fresh 3.0 first
19:42 martinmorris i think so
21:20 owen Hmmm... can't log out of the OPAC
21:26 Anyone here know how anonymous suggestions work? The description says "Set to anonymous borrowernumber to enable Anonymous suggestions"
21:26 What is an anonymous borrowernumber?
21:28 kados it's the borrower that you want the anonoymous suggestions set to
21:28 ie, they are still linked to an account
21:29 owen Hi kados. So you create a dummy account or something?
21:29 kados yea, exactly
21:29 can't log out of the opac sounds like a bug
21:31 owen I thought I saw a commit related to logouts recently, but now I can't put my finger on it
21:32 Oh, there it is http://git.koha.org/cgi-bin/gi[…]541eb04e00d5ac45d
21:35 I was pointing to opac-logout.pl instead of opac-main.pl. Silly me :)
21:37 chris heh
21:54 owen I log in as one user to add a purchase suggestion. When I submit the form the page reloads and I'm told there are no purchase suggestions (instead of showing me mine)
21:54 When I log in to the OPAC as another user (a superlibrarian, by the way), I see my suggestions and the one I submitted as the other user
21:56 kados hmmm
21:56 sounds like a bug
23:00 [K] *** join #koha@FreeNode: rangi n=chris@203-118-134-114.netspace.net.nz
23:01 *** join #koha@FreeNode: ru55el n=russel@203-118-134-114.netspace.net.nz
23:06 *** join #koha@FreeNode: rangi n=chris@203-118-134-114.netspace.net.nz
08:08 JYL57 hi koha world, paul, hdl around ?!
08:08 js hi JYL57
08:09 hdl hi$
08:23 chris hi hdl
08:23 some patches from joshua should have come through now
08:24 so hopefully the search is working better and did you see that you can turn the stemming/left truncation off in system preferences?
08:25 and hi paul and jyl57 :)
08:25 paul hello chris.
08:25 hdl IT was not stemming.
08:26 It was an unfortunate copy/paste on ccl.properties which added 5=1 to author
08:26 chris ahhh
08:26 cool paul
08:29 paul: you use mandriva?
08:29 paul yep
08:29 2008
08:29 chris cool
08:29 hdl ccl.properties in /etc contains 5=1 for some attributes
08:30 chris at least i know it can be done :)
08:30 ahh, 5=1 is the truncation?
08:31 hdl yes
08:36 chris paul: whats the biggest library you have running nozebra?
08:36 paul live ?
08:36 chris oh have tested
08:36 paul i have only 1 and it's 30 000 items
08:36 chris oh = or (cant type)
08:36 paul same answer
08:36 chris :)
08:37 paul but ENSMP, that is more than 100 000 items wide will move to 3.0 NZ soon
08:37 as they can't install zebra (their techie guys refuses if I have understood well)
08:37 chris i plan to do some testing with one that is 148472 items
08:38 so i will let you know what performance i get
08:38 paul the reindexing will be quite long I bet...
08:38 chris yep
08:38 paul i'm sure it will be OK (and much faster than for 2.2)
08:38 chris but i think it will be much faster than 2.2
08:38 heh
08:38 great minds think alike :)
08:39 i really like how you can define the indexes, thats really neat, you guys did a great job there
08:40 paul so we just lack a release now :-\
08:41 chris yep i think once this installer is fixed, very close now, and search is fixed then release time
08:41 i sent all my patches i did on the installer to galen who is going to test and fix
08:42 and kados is working hard on fixing the search, so im hopeful we can release soon
08:47 JYL57 it would be a nice christmas gift ! ;-)
08:51 [K] *** join #koha@FreeNode: phasefx n=phasefx@68-190-35-14.dhcp.gwnt.ga.charter.com
10:30 hdl kados, Joshua : hi.
10:31 You work on searches and it is obvious it works better.
10:31 Some questions though.
10:32 When searching kw, things are queried as if it was phrase.
10:32 For instance, I searched particules test
10:33 It searched not a wordlist "particules" "test".
10:33 But the expression : "particules test"
10:33 Is this the expected behaviour ?
10:33 Is there a way to search on wordlist by default ?
10:34 You told me it would be so. But it seems it is not.
10:34 Is there a ccl atribute or a pqf attribute I should remove ?
10:37 I tried with s=pw but it would not answer the way I want
10:39 looking at zebra query, @attr 4=1 appears when I didnot ask for it.
10:39 And did not add it to ccl.properties

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

koha1