IRC log for #koha, 2006-08-05

All times shown according to UTC.

Time S Nick Message
12:08 paul bye bye world. see you later, alligator
12:30 owen kados: ping me when you have time to talk about bugs
12:40 kados owen: ok, now's a good time
12:40 owen Hi
12:41 owen I'm looking at this one: http://bugs.koha.org/cgi-bin/b[…]w_bug.cgi?id=1129
12:41 kados yea
12:41 owen The NPL and default templates are only cosmetically different
12:41 kados strange
12:41 kados I can give you access to smfpl
12:42 owen Are you sure smfpl has an up to date copy of the template?
12:42 kados yep ...
12:42 kados I'll try again right now to double check
12:44 kados ok ... just tried again
12:44 kados same deal
12:44 kados I search for an authority
12:44 kados view the full heading
12:44 owen I do see where that missing image is. I missed that
12:44 kados click on 'edit'
12:44 kados add a field
12:44 kados click save
12:44 kados it acts like it's saving
12:45 kados but when I look at the marc, the field doesn't show up
12:47 owen Oh hang on... I see a problem.
12:47 owen NPL's got some javascript missing
12:47 kados ahh
12:56 owen kados: when you try to save, does it forward you to a blank search results screen?
12:56 kados yes
12:59 owen Shoot... I need to get some lunch and I'll tackle this after that.
12:59 owen Too...hungry...to...concentrate!
12:59 kados k :-)
13:09 thd kados: have you seen http://wiki.koha.org/doku.php?[…]er_meta_record_db ?
13:10 kados thd: looking now
13:11 thd kados: note at the end I became too tired to start a DTD
13:11 thd or link to supported relationships
13:13 kados thd: nice summary
13:13 thd kados: tumer has a single minded focus
13:13 kados well ... he _has_ to get something working
13:13 kados nothing like that for motivation :-)
13:13 thd kados: he is liable to implement something soon
13:13 kados yep
13:15 thd kados: therefore I propose that we invite wide suggestion on how to construct a relationship DTD after or maybe even before my initial draft of one
13:16 thd kados: I might have some good thought but this should have many thoughtful people looking so that we can get something for tumer too work with in a couple of weeks or so if that is not too late for him
13:17 thd kados: therefore, i would propose to advertise that wiki page widely
13:18 thd kados: to invite comment on the best XML meta-record design
13:19 thd kados: any comment?
13:19 kados not a bad idea
13:19 kados but I think we should use tumer's dtd as a starting place
13:19 thd kados: yes of course
13:20 thd kados: and you started to give a suggestion
13:20 thd kados: Can Zebra index <whatever><bibliographic><syntax name="some_syntax"><100>  differently from <whatever><bibliographic><syntax name="other_syntax"><100> ?
13:21 kados I don't think it can do that yet
13:21 kados but it would be a good question to ask the koha-zebra list
13:21 thd kados: really
13:22 thd kados: my concept presumed that could be differently indexed
13:23 thd or at least the multi-MARC multi-authority file part of the concept
13:27 thd kados: well we could always do <100_unimarc> , <100_marc21> , <100_ibermarc> but that would be icky
13:33 thd kados: we need to do <100_bib> , <100_hold> , <100_auth>, <100_class> to distinguish even within MARC 21
13:33 thd if paths cannot be indexed separately
13:56 kados owen: quick quetion
13:57 kados /cgi-bin/koha/bull/statecollection.pl?
13:57 kados the callnumber and barcode fields have sizes
13:57 kados any reason I can't take those out?
13:57 kados NBBC has longer values than are fitting
13:58 kados maxlength=15
13:58 owen No idea why those are there.
13:58 kados k ... I'm taking them out
13:58 owen A maxlength would only be /necessary/ if there was a limit to the field size in the database
13:59 kados I think it's a text field
13:59 kados also ...
13:59 kados on that same page
13:59 kados we need to distinguish between 'current expected' and 'recently checked in'
14:05 kados hmmm ... looks like koha doesn't distingish between them
14:05 kados there's just one loop ... serialslist
14:06 kados i think?
14:09 kados owen: i updated smfpl with your recent commit
14:10 owen Kados, on the authorities page, in the default template, there are little left-arrow graphics next to the subfield. What are those supposed to do?
14:10 kados I wonder if those are reordering arrows
14:10 kados in the bib editor you can reorder subfields
14:10 kados are they attached to a cloneSubfield() function?
14:10 owen The javascript links to "upSubfield"
14:11 kados yea, that's it
14:11 kados I would be surprised if they work though
14:11 kados they work well in the biblio editor
14:12 kados but i think the auth editor is a few generations behind
14:15 owen Okay, so I've committed an updated authorities-topmenu.inc and authorities.tmpl
14:15 owen Editing should be working now
14:15 kados w00t
14:16 owen Uh oh...looks like paul made some changes to authorities.tmpl in default today when I wasn't looking
14:16 kados hmmm
14:16 kados I'm getting the same prob
14:16 kados as before
14:19 owen Maybe not.
14:19 owen Same problem, kados?
14:19 owen It's forwarding you to the search page still?
14:19 kados yea
14:21 owen Are you sure you've got the most current authorities-topmenu.inc?
14:21 owen It doesn't look like it.
14:22 kados oops
14:22 kados updated the wrong cvs repo :-)
14:22 kados ok ... testing again
14:23 kados owen: worked!
14:23 kados owen: thanks ... feel free to close out that bug :-)
14:24 owen Yay.
14:25 owen What's up with 1128?
14:26 kados I'll test it
14:26 kados now
14:26 kados still a bug in npl
14:27 kados interestingly, clicking once gives a js bug, twice a 500 error
14:27 kados [Thu Aug 03 09:21:32 2006] [error] [client 69.58.13.236] Tag "12" is not a valid tag. at /home/koha/intranet/cgi-bin​/authorities/authorities.pl line 371, referer: http://koha.smfpl.org/cgi-bin/[…]l?authid=12080061
14:27 kados huh ... looks like maybe a leader problem
14:27 owen I was trying it on smfpl.org and was just getting a 500 error.
14:27 kados encoding not set properly or something
14:27 kados ahh ... the framework prolly doesn't have a leader at all
14:29 kados huh ... can't seem to edit the subfields in the thesaurus editor
14:29 kados the thesaurus frameworks
14:30 kados works in default
14:30 kados maybe auth_subfields_structure not up to date in npl?
14:31 owen I'll check it.
14:32 kados also, no plugins available for either default or npl
14:32 owen ?
14:33 kados there are authorized values and plugins
14:33 kados when editing a subfield, in the biblio framework, you can specify both
14:33 kados in authorities you only have authorized values
14:35 owen You were talking about /parameters/auth_subfields_structure.tmpl, right?
14:35 kados yea
14:38 kados ok, paul didn't fix 1128, I get 505 in default too
14:40 kados :q
14:41 kados re-reassigning to paul
14:41 owen Bug tennis.
14:41 kados hehe
14:41 kados i discovered another one
14:42 kados # of bib records in opac only shows up with authors :/
14:42 kados weird
14:49 owen kados: I updated auth_subfields_structure.tmpl, but there was only one minor difference.
14:49 kados k, thx
14:50 kados auth_finder.tmpl isn't working for some reason
14:50 kados well ...
14:50 kados i thought it was working yesterday
14:51 kados the Main entry $a only doesn't work
14:51 kados the other seem to work
14:53 kados and no idea what that 'clear' link is supposd to do, but it doesn't work :(
14:55 kados weird
14:55 kados it works on smfpl for pname
14:55 kados but not on tterm on another site
15:09 owen kados: what can I look at for you?
15:09 kados I think I figured it out
15:09 kados it was a liblime thing :-)
15:57 kados owen: somehow ... in auth_finder and authoritis-home templates 'mainentry' turned into $a
15:57 owen Hunh?
15:59 kados well ... remember I said the search stopped working?
15:59 kados it did :-)
15:59 kados there's this line:
16:00 kados <input type="hidden" name="marclist" value="$a" />
16:00 kados which used to be:
16:00 kados <input type="hidden" name="marclist" value="mainentry" />
16:00 kados $a doesn't work
16:00 kados so we need to change it back
16:00 kados it's been changed in the opac and the intranet strangely
16:02 owen That was changed by paul on 7/31: "adding third option to authority search (search on $a of on all subfields of main entry)
16:03 kados and it looks like the 'used' tmpl_var has been taken out too
16:03 kados which is the count bit
16:03 kados note that the NPL display is a bit different in the OPAC than paul's
16:04 owen On which page?
16:04 kados on the results page
16:05 kados well i don't understand the $a thing
16:05 kados things aren't stored in the db with a 4
16:05 kados s/4/$/
16:05 kados so why would he query on $?
16:07 kados if (@$tags[$i] eq '\'$a\'') {
16:08 kados in AuthoritiesMarc.pm
16:08 kados grrr
16:09 kados if (@$tags[$i] eq '\'$a\'') {
16:09 kados so he does use it
02:18 hdl hi
02:23 hdl kados there?
02:42 hdl paul_away: ?
04:10 hdl hey
04:21 toins ho
07:53 hdl tumer around ?
07:54 hdl kados ?
07:54 dewey kados is becoming a true Perl Monger...
08:12 hdl hi owen
08:18 owen Hi hdl
08:55 toins kados around ?
08:56 hdl toins: seems not.
08:56 toins :-/
10:07 owen kados: I have a question for you if you get a moment
10:47 kados owen: k ... shoot
10:48 paul hi kados.
10:48 kados hi paul
10:48 paul i've done my 1st on dev_week
10:48 paul (& look at it, as it will break something if you don't update your config file, for sure ;-) )
10:49 paul i still have some problems with encoding...
10:49 paul http://o19.bureau.paulpoulain.[…]RCdetail.pl?bib=5
10:49 paul my biblioitems.marcxml is utf8
10:49 paul if I save it in a file, I can confirm it's utf8
10:49 owen I'm trying to figure out whether I've got full-serial-issues.tmpl working in NPL
10:49 paul but when displaying, it's no more utf8...
10:50 kados cool!
10:50 kados paul: i agree
10:50 kados paul: nice ida
10:50 kados idea even
10:50 kados yep
10:50 kados hmmm
10:50 kados paul: have you followed guidelines on the wiki?
10:50 kados you must have mysql encoding correct at every level
10:50 kados owen: I can look
10:50 paul kados : wiki is down (from france)
10:50 kados paul: try this:
10:50 kados paul: it works for me
10:50 kados paul: try this:
10:50 kados mysql -ukohaadmin -ppassword
10:51 kados show variables;
10:51 kados look for some lines like:
10:51 kados character_set_client    
10:51 paul | character_set_client            | latin1                                                 |
10:51 paul | character_set_connection        | latin1                                                 |
10:51 paul | character_set_database          | latin1                                                 |
10:51 paul | character_set_results           | latin1                                                 |
10:51 paul | character_set_server            | latin1                                                 |
10:51 paul | character_set_system            | utf8                                                   |
10:51 kados ok ... so that's your 1st prob
10:52 paul ok, i go to modify this ;-)
10:52 paul (is it possible to modify it only for 1 DB ?)
10:52 kados in my.cnf do this:
10:52 kados init-connect = 'SET NAMES utf8'
10:52 kados character-set-server=utf8
10:52 kados collation-server=utf8_general_ci
10:52 kados yes
10:52 owen kados: the serials data on 101 is so bare-bones. I wondered if you've got some good data somewhere.
10:52 kados paul: to do this do:
10:52 kados ALTER DATABASE kohawipofinal DEFAULT CHARACTER SET utf8 DEFAULT COLLATE utf8_general_ci;
10:52 kados paul: then you must alter all tables too :-)
10:53 kados paul: for 'text' you must first conver to blob
10:53 paul I already did this for all tables
10:53 kados paul: then tu utf8
10:53 kados like this:
10:53 kados alter table biblio modify title BLOB;
10:53 kados alter table biblio modify title TEXT CHARACTER SET utf8;
10:53 kados for char and varchar you must go:
10:53 kados alter table marc_subfield_table modify subfieldvalue BINARY(255);
10:53 kados alter table marc_subfield_table modify subfieldvalue VARCHAR(255) CHARACTER SET utf8;
10:54 kados if you don't convert to blob/binary first, mysql will double-encode your strings and you will have to start from scratch :-)
10:54 kados owen: yea, I do
10:54 kados owen: sec
10:55 paul kados : biblio.title is already utf_8_general_ci, collation too, and biblio.title show correct utf8 afaics
10:56 kados paul: yes, but did you check biblioitems.marc/biblioitems.marcxml?
10:56 kados paul: that is how dev_week creates MARCdetail.pl
10:59 paul i'm 99% sure it's correct utf8
11:00 kados paul: mysql can store utf8 data in a latin1 database/table/column
11:00 kados paul: also, do you have firefox developer toolbar?
11:00 paul yes I know. but : it's utf8 inside, the collection is marked utf8
11:00 kados paul: check the 'response headers' and see what the encoding is
11:00 paul yes
11:01 paul Date: Fri, 04 Aug 2006 16:01:16 GMT
11:01 paul Server: Apache/2.0.54 (Mandriva Linux/PREFORK-13.2.20060mdk)
11:01 paul Set-Cookie: sessionID=46821-1154704457; path=/
11:01 paul Keep-Alive: timeout=15, max=98
11:01 paul Connection: Keep-Alive
11:01 paul Transfer-Encoding: chunked
11:01 paul Content-Type: text/html; charset=utf-8
11:01 paul 200 OK
11:02 kados good, so apache is correct
11:02 kados paul: is the opac online ... can I look?
11:03 paul (yes, link 10mn ago)
11:03 paul http://o19.bureau.paulpoulain.[…]RCdetail.pl?bib=5
11:03 paul for example
11:03 paul 200$a :  a titre propre  La Philosophie m?edi?evale
11:03 paul it should be médiévale
11:03 paul note that the 100 header : 100 -Données générales de traitement
11:04 paul is correct, and it comes from mySQL too
11:04 kados link doesn't work for me
11:05 kados Server not found
11:05 paul I think there is a problem between europe & US
11:05 paul because I can't reach wiki.koha.org
11:06 kados right :-)
11:06 paul (fortunatly, irc is up. thx new zealand :-D )
11:06 kados :-)
11:06 kados paul: have you done:
11:07 kados ALTER TABLE bibliosubtitle MODIFY marcxml BLOB;
11:07 kados oops
11:07 kados biblioitem I mean
11:07 paul nope
11:08 kados this must be done
11:08 kados then back to utf8
11:08 paul BEFORE the alter table collate utf8 ?
11:08 kados ALTER TABLE biblioitems MODIFY marcxml BLOB;
11:08 kados ALTER TABLE biblioitems MODIFY marcxml TEXT CHARACTER SET utf8;
11:09 kados collate is just to relate one character to another
11:09 kados character set is to define what encoding is used
11:10 kados so right now you have a mysql system with latin1 encoding, your db is utf8 encoding, your table is latin1 encoded, apache is utf8 encoded
11:10 paul it's done, but changes nothing
11:10 kados and your column marcxml is latin1 as well
11:11 kados I don't know, maybe mysql latin1 is a prob
11:11 kados in my tests, everything is utf8
11:11 kados and it works ok
11:11 kados also, maybe your data is not utf8 :-)
11:11 kados and ... third possibility
11:12 kados maybe your browser can't show utf8 ... or maybe the font can't display utf8 in your OS
11:12 kados the examples you posted above are 'combining' characters in unicode
11:13 paul I have correct combining characters, as 100$a description is correctly shown
11:14 paul my marcxml column is declared as utf8, and, if I mysqldump the table, I get utf8 well formed in the dump
11:14 kados 100a wasn't combining
11:14 paul why ?
11:15 paul it's a "données générales de traitement"
11:15 paul so, 2 é
11:15 paul (i speak of the description, from marc_subfield_structure, not the content)
11:16 paul tr[root@i1 paul]# traceroute wiki.koha.org
11:16 paul 16  vox.spry.com (209.59.197.6)  239.058 ms  236.019 ms  235.756 ms
11:16 paul 17  pippin.metavore.com (66.249.19.54)  237.055 ms  236.931 ms  236.786 ms
11:16 kados so dns works :-)
11:16 paul seems I can reach the server but not http traffic
11:16 kados strange
11:16 kados maybe http is blocked by your firewall?
11:16 paul nope
11:17 kados maybe your ISP?
11:17 paul (I can surf everywhere else)
11:17 kados strange
11:26 paul kados still here or i've been disconnected ?
11:26 paul anyway :
11:26 kados paul: I"m here
11:27 paul i've mysqldump biblioitems : I see "   Dogmatique chrA@tienne"
11:27 paul which is utf8 data
11:27 paul (it's the é)
11:28 paul is it enough to be sure the datas are really utf8 & identified as such by mySQL ?
11:28 paul I think yes, because, the table definition says :
11:28 kados mysql defines the following places for character sets:
11:28 paul create table ... ENGINE type=myISAM DEFAULT CHARSET=utf8
11:28 kados client, connection, database, server, system, table, column
11:29 paul phpmyadmin says "column is collate utf8_general_ci"
11:29 kados good
11:29 kados so mysql looks ok then
11:30 kados and apache is ok
11:30 paul the dump says the table is utf8
11:30 kados what about your browser?
11:30 paul it's firefox (1.5)
11:30 kados ok ...
11:30 kados what about your font?
11:30 kados some fonts don't have combining characters
11:30 paul view > encoding > utf8
11:31 kados I wrote a mail to koha-devel about this this morning
11:31 paul Remember that on the same page (http://i19.bureau.paulpoulain.[…]Cdetail.pl?bib=1)
11:31 paul I have valid combining
11:31 paul (for what comes from marc_subfield_structure)
11:31 kados I don't think 100 is combining
11:31 paul ???
11:31 kados there are many ways to write the same char in unicode
11:32 paul ah, ok
11:32 kados combining characters are when you do something like:
11:32 kados e/
11:32 kados for É
11:33 paul in the dump for marc_subfield_structure, I have A@ (same thing than for biblioitems) for é
11:33 paul does it means it's a combining char ?
11:33 kados if it is a combining character, it is two characters to build one character
11:33 paul you mean 2Bytes for 1 char isn't it ?
11:34 kados yes
11:34 paul mmm... one minut... I've an idea ...
11:34 kados you can check the length in mysql
11:34 paul mmm... maybe i've something :
11:34 paul i'm UNIMARC.
11:34 paul and the encoding is in 100$a
11:34 paul and 100$a says "03" (in position 26)
11:34 paul that's wrong !
11:34 paul because utf8 is 50 :
11:35 paul !
11:35 kados we might also need to look at MARC::Record
11:35 kados MARC::File::XML and MARC::Charset
11:36 paul no change :-(
11:36 paul but I will follow this path.
11:36 paul do you plan to work tomorrow ?
11:36 paul (because I will)
11:36 kados yes
11:36 kados before you follow that path
11:36 kados check the font
11:36 paul so we will be at least 2 ;-)
11:36 kados because I spent many many hours
11:36 kados and it was a font problem
11:36 paul how can I check my font ?
11:36 kados in your css
11:37 kados font-family
11:38 paul font-family: arial, geneva, sans-serif;
11:38 paul it's the default dev_week/npl one
11:38 kados yep ...
11:38 kados try this:
11:38 kados font-family: Lucida Grande, Tahoma, Arial, sans-serif;
11:38 paul no change
11:38 kados :(
11:38 kados try delete the line
11:39 kados so browser will use defaults
11:41 kados maybe not
11:42 paul mmm... strange... i've modifier h2 to add a background red, and nothing visible.
11:43 kados which css are you changing?
11:43 paul intranet.css
11:44 kados you can override with a inline css
11:44 kados add style="font-family: Lucida Grande, Tahoma, Arial, sans-serif;"
11:44 kados to the body tag
11:44 kados maybe?
11:46 paul done, but no change
11:46 paul (& I can see the style="...")
11:47 kados really strange ...
11:48 kados are you sure it's the css used in intranet?
11:48 kados (I mean intranet.css)
11:48 kados I would suspect font first if other utf8 characters look ok but combining dont
11:48 kados paul: http://wipoopac.liblime.com
11:49 kados paul: do some searches there ... do all chars look ok?
11:49 kados (it's a fully utf8 catalog)
11:50 paul any idea of a search that will provide combined chars ?
11:50 kados i think patents
11:51 kados title starting 1983 ars
11:51 kados has a a with umlaut that is combining
11:52 kados also, search for french word for information
11:52 paul the 2nd a, the one 1st letter of andringar ?
11:52 paul becaure I see a and AFTER "
11:52 paul (umlaut)
11:52 kados ahh ... so that means your system has trouble with combining characters already
11:52 kados on my system umlaut is above the a
11:53 kados (but only with lucindia grande font)
11:53 kados (with arial font, i get a box instead of umlaut)
11:53 kados now I will switch to dev_weeks default font
11:53 paul with konqueror I just see a, without umlaut
11:53 kados and you can reload the page
11:53 kados one second
11:53 paul and with opera 8.5, same as firefox
11:54 paul (a and " on the right)
11:54 kados ok, shift-refresh the page
11:54 kados see if there are any changes
11:54 paul no
11:55 kados hmmm
11:55 kados now it is font-family : Arial, Geneva, Helvetica, sans-serif;
11:55 kados same as dev_week?
11:56 paul dev_week is :font-family: arial, geneva, sans-serif;
11:57 paul si almost the same
11:57 paul must my linux box be utf8 ?
11:57 paul (iirc there is something about this on your wiki page)
11:57 paul because I don't think i'm
11:57 kados type:
11:57 kados locale
11:57 kados on your local box
11:58 paul i'm not utf8
11:58 paul LANG=fr_FR
11:58 kados (I don't know if it's necessary as I'm alwasy utf8 on OSX)
11:58 paul LC_CTYPE=fr_FR
11:58 paul LC_NUMERIC=fr_FR
11:58 paul LC_TIME=fr_FR
11:58 paul LC_COLLATE=fr_FR
11:58 kados that may be the prob ... but it seems you can see combining chars ok on wipoopac
11:58 kados just not combined
11:58 kados but at least not a ?
11:58 paul combining / combined ?
11:58 kados combining I think
11:59 paul I don't see things correctly, right ?
11:59 kados well ...
11:59 paul because the umlaut is on the left of th a
11:59 kados on wipoopac you see combining characters not combined
11:59 kados but ...
11:59 kados on your system, you don't even see them :-)

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