IRC log for #koha, 2008-04-18

Today | Next day → | Search | Index

All times shown according to UTC.

Time Nick Message
15:07 fbcit I used to have a button to add a new MARC record
15:08 arsenic fbcit: did you change something in the YUI config?
15:08 fbcit no, but yui has been upgraded since yesterday
15:09 arsenic ok.. that is probaly wuy.  who changed that?
15:09 wuy = why
15:09 fbcit it was upgraded to the latest ver
15:10 arsenic the latest version of YUI doesn't support button.beta
15:10 it's now just button
15:16 hdl owen around ?
15:44 fbcit anyone else rebased up and missing the tool bars from either addbooks.pl or members-home.pl?
15:44 paul_ yes, i've rebased
15:44 and see that the nice buttons have disappeared :-(
15:44 replaced by a standard link
15:44 (hi fbcit)
15:45 fbcit paul_: I'm missing the link too
15:45 hello paul
15:45 hdl think it is because of change in yui version.
15:45 atz most likel
15:45 *likely
15:45 paul yep
15:45 hdl patch it up with frederic patch
15:46 kados paul: can you confirm bug 2020?
15:46 have you guys tried shift-refresh?
15:48 fbcit tnx hdl
15:48 paul kados : nope, I can't confirm it. I don't have the problem.
15:48 hdl mice and men is searching OK.
15:48 paul http://o18.bureau.paulpoulain.[…]=harry+and+potter
15:49 hdl and returning no biblio
15:49 paul (same for i18, but no public login)
15:49 MatthewMetzger hdl: really? I can help isolate it, if needed
15:49 kados :(
15:49 paul (and = et in french. I suspected an empty word problem)
15:49 kados paul: it sounds like a load problem
15:50 paul: how big is your database?
15:50 paul around 15 000 biblios
15:51 no problem, even when i've added AND as a stopword.
15:51 kados I will ask some additional questions on bugzilla
15:51 so it could be a MARC21 prob
15:52 or it could be too many records
15:52 or maybe not right hardware for him
15:52 I'll ask
15:52 MatthewMetzger kados: bugzilla is good. I'm the reporter of the bug, so if you need fast answers, I'm here.
15:52 kados oh, hi MatthewMetzger
15:52 MatthewMetzger: so ... how many bib and items?
15:53 MatthewMetzger I believe just under 6000 items. Is there a fast way to check from the admin interface?
15:53 kados ok, that's small
15:53 very strange
15:53 MatthewMetzger we have a very small library
15:53 kados MatthewMetzger: I may have some time tonight to test marc21 searching with nozebra on
15:53 MatthewMetzger: otherwise, maybe paul or hdl can help right now
15:54 I've got to head to a presentation
15:54 ttyl
15:54 MatthewMetzger okay. I'll answer the questions on bugzilla, so it's documented.
15:54 kados: good luck on your presentation
15:54 paul MatthewMetzger: do you have the problem with a search containing "OR" ?
15:54 "NOT"
15:54 MatthewMetzger paul: I'll check right now...
15:54 paul like "mice or men"
15:55 q=mice and men should be interpreted as "search mice" and do a AND with "search men"
15:57 MatthewMetzger mice or men - same behavior, ramps up cpu and memory usage. I have to kill the process.
15:57 paul a search on "mice" works fine ?
15:57 idem for a search on "men"
15:57 ?
15:59 MatthewMetzger mice works fine.
16:00 individual words and author's names return searches fine.
16:00 paul could you set debug ON and tell me what appears on the log ?
16:00 MatthewMetzger It seems that "and" and "or" in the search cause the "crash"
16:01 paul: could you remind me where that setting is?
16:01 It may already be on.
16:01 paul in the virtual host, just add : SetEnv DEBUG 1
16:01 MatthewMetzger what log file am I looking for
16:01 paul in the virtual host, just add : SetEnv DEBUG=1
16:01 the one you see in the apache VHost
16:01 MatthewMetzger okay, just a minute
16:02 paul no, it's SetEnv DEBUG 1
16:02 definetly
16:02 i have to leave.
16:03 you'll see things like :
16:03    warn "query =$query" if $DEBUG;
16:03 query = ...
16:03 string : ....
16:03 no parenthesis. left : ... operator: ... right: ...
16:04 node : .../.../...
16:04 leftresult : ...
16:04 rightresult : ...
16:04 paste everything in bugzilla pls
16:04 you may also get : "NZAND"
16:23 hdl gmcharlt: around ?
16:24 gmcharlt hdl: yes
16:25 hdl I am experiencing some differences between "box search" and advanced kw search
16:25 is this expected ?
16:26 http://o17.hdlaurent.paulpoula[…]ha/opac-search.pl
16:26 24.082
16:26 you have 2 results
16:27 with advanced search, you get 29
16:31 in fact, queries generated are not the same :
16:32 24.082 for the simple one
16:32 @attr 1=1016 @attr 4=6 24.082 for advanced one.
16:34 gmcharlt hdl: yeah, looks like one is interpreting as a phrase search, and one is interpreting as a keyword search with like "24 and 082"
16:35 hdl well... This is quite embarassing... Because depending on the place you type your query, results are not the same.
16:35 Worse,
16:35 gmcharlt probably should be made consistent, but we should think carefully about when and where to add qualifiers to a default search
16:36 hdl results with advanced search are less accurate than with simple box.
16:36 I agree.
16:37 Would there be a place where we may think together and define some test cases ?
16:37 Maybe a google document would be fine ?
16:38 gmcharlt that, or use the dev wiki
17:11 tim When I click on a call number in the search results of the opac or intranet interface I get the error No results match your search for “callnum:AF KOO" using version3 beta2.
17:17 There are no errors in the logs for this, but getting to the search results screen gives me an error 3 times.
17:17 File does not exist: /usr/share/koha/intranet/htdocs/opac-tmpl, referer: http://koha2.wlpl.org:8080/cgi[…]sort_by=relevance
17:19 That was from Koha-error_log.  I don't get any errors on the OPAC.
17:20 Or at least not in the logs.  Just in the browser.
17:21 hdl you could have errors in zebra logs.
17:25 tim That's wierd.  Zebra hasn't logged anything.
17:49 I got zebra logging.  Nothing in the error log, but there was an error in the output log.
17:49 Search biblios ERROR 114 1 1+0 RPN @attrset Bib-1 @attr 1=20 "AF KOO"
17:50 That was the koha-zebradaemon-output.log
17:50 gmcharlt tim: looks like a missing indexing definition (or a mismatched one)
17:51 enough for me to go on
17:54 tim I was thinking it would probably be something like that.
17:55 I was also thinking I don't know what to do about it.  Would that be in record.abs?
18:13 gmcharlt tim: yes, looking in record.abs to verify that appropriate MARC fields are mapped to callnumber index, bib1.att to confirm attribute (1=20) and ccl.properties to verifying that callnumber is mapped to the 1=20 attribute
19:32 ryan tim: also, be sure you actually have data in the itemcallnumber field.
19:33 tim: note the default frameworks have changed quite a bit from 2.2 -> 3.0 , so you'll have to deal with mapping
20:26 MatthewMetzger_ Regarding bug 2020: I noticed a discrepancy with with my perl install. mod_perl is 5.8.8 while koha reports it as 5.010000. (5.10.0). Could this be a problem?
22:28 tim thanks gmcharlt.  I finally got the time to try what you suggested (a long time ago now) and it worked.
22:43 gmcharlt tim: great!

Today | Next day → | Search | Index

koha1