IRC log for #koha, 2006-11-26

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

All times shown according to UTC.

Time Nick Message
11:19 hdl hi owen.
11:19 seen kados ?
11:19 dewey kados was last seen on #koha 18 hours, 56 minutes and 51 seconds ago, saying: heads out [Fri Nov 24 08:23:08 2006]
11:20 owen Hi hdl. No, I just arrived.
11:20 hdl how are you ?
11:20 owen I am well, thank you. Had a nice Thanksgiving with my family yesterday. I'm back at work while the rest of America is out shopping :)
11:23 hdl Oh... Sure. It is a long weekend for you...
11:23 Happy thanksGiving
11:24 owen Thanks. What's new with you hdl?
11:27 hdl My wife is fantastic... little 2 and a half kid at school and doing fine
11:27 :D
11:33 kados hi hdl
11:34 hdl hi kados
11:34 Happy thanks giving...
11:35 I had a question about the use of chr files and where to call them.
11:35 (Can't find it in Zebra doc)
11:35 kados sure
11:35 it all depends on what you have in your idx file
11:36 you can define different mappings for each index type
11:40 hdl Thx. I saw it.
11:40 If I change word-string.chr, do I have to reindex data ?
11:41 (seems yes)
11:44 kados: it seems there is no equivalent directive in *.chr files any longer... Is there ?
11:45 And position is not a defined keyword. But it is used in our default.idx
11:45 Can you tell me more ?
11:59 kados hdl: yes, you will need to re-index
12:00 equivalent directive? do you mean 'map'?
12:00 hdl # equivalent ��(ae)
12:00 # equivalent øö(oe)
12:00 # equivalent å(aa)
12:00 # equivalent uü
12:00 kados hmmm ... I think it's called 'map' now
12:01 ahh, right
12:01 equivilent is for searching
12:02 so ...
12:02 you can define collations in two ways:
12:02 make characters equivilent to each other so that searching on one finds the other, etc.
12:04 hdl can you give me the link you search in ?
12:04 kados I'm just looking at the .chr file
12:05 hdl I couldnot see equivalent in Indexdata documentation.
12:05 I also read *.chr file.
12:05 But am unsure though.
12:06 kados what are you trying to acomplish?
12:19 hdl: ?
12:19 hdl kados : I am trying to make search for é è ë equivalent to search e.
12:20 If a user type determinisme, it could return déterminisme.
12:20 kados is that UTF-8?
12:20 (on my computer it's not for some reason)
12:21 é is the character?
12:21 you want:
12:22 equivalent ée
12:22 in the appropriate .chr file
12:22 hdl yes.
12:23 (I used ISO8859-15 because it seems to be the encoding of this channel)
12:23 kados ahh, ok
12:24 hdl ���� equivalent to e
12:24 (Can you read that ?
12:24 kados my irssi is set to use utf8
12:24 hdl )
12:24 kados no, I can't read it
12:24 strange
12:24 it looks like ????
12:24 (four question marks)
12:24 is it the multi-byte version or the single-byte version?
12:25 é << can you read that?
12:25 hdl No.
12:25 kados hehe
12:25 hdl looks like A~(c)
12:25 kados hehe
12:25 so our channel is poorly encoded for communication about encodings ;-)
12:26 I wonder if freenode is any better
12:26 could you join #koha on freenode for testing?
12:27 (irc.freenode.net)
12:29 it seems I could issue /charset utf-8 but I don't have oper privs here
12:32 UTF-8 is hard :/
12:36 hdl What I wanted to say was :
12:36 kados hdl: how about now: ééé
12:36 hdl yes.
12:36 So.
12:36 I wanted éëêè to be equivalent for research to e
12:36 (French configuration)
12:37 kados right
12:37 I believe you can try:
12:38 equivalent éëêèe
12:38 hdl Taht is exactly what I am trying.
12:38 kados (I can't promise more than two work)
12:38 cool
12:38 let me know if it works
12:39 hdl But when indexing, it finds new bugs in mrc files.
12:39 kados ahh, so it must not be correct then
12:39 try just one:
12:39 equivilent ée
12:39 equivalent ée
12:39 (wrong spelling ;-))
12:43 hdl it will be a pain for multiligual catalogue.
12:44 What is the difference with map ?
12:44 kados I agree
12:44 maybe we should ask ID
12:46 hdl the syntax you proposed worked.
12:46 kados great!
12:46 hdl one HAVE TO tell equivalents one char at a time!!!
12:46 kados I bet ID would fix it if we complained
12:47 hdl If you get spanish, german, french and dutch ....
12:47 you go crazy.
12:47 :D
12:47 kados yep
12:48 hdl (And some catalogues some customers have contain ukrainian, rumainian, and so on...)
12:48 kados right
12:48 it can get quite complicated
12:49 and we still have some problems with mysql
12:49 hdl Yes.
12:49 kados for instance, multi-byte characters cause problems
12:49 because unlike zebra, we can't control easily the collations in mysql
12:50 we are stuck with what mysql provides
12:50 (unless you want to write some C ;-))
12:50 hdl + DBD::MYSQL. and the fact that it is not utf8 aware in perl.
12:50 kados right
12:51 hdl Is there a better Database utf8 supported system in PERL ?
12:51 kados postgres I think
12:51 hdl I hope too.
12:51 But didnot inquire deeply enough.
12:51 kados if I ever get sponsorship for it, I would like to write a driver for koha and postgres
12:52 it could be much better (though also harder to configure)
12:52 hdl And postgress would certainly require to dig into DB queries...
12:52 kados yep
12:53 hdl Should we do this for rel_3_0 ?
12:53 Could we have time ?
12:53 kados but it would be ideal
12:53 hdl I don't think it would be realistic.
12:54 But we certainly have to consider it.
12:54 ok.
12:54 kados I think it will depend on finding a client who needs it
12:55 hdl (+ to consider very well how much it impacts on Koha and time needed for developpement)
12:55 (Which certainly wouldnot be the easiest part)
12:55 :D
12:55 kados hehe
12:56 hdl Many THX...
12:56 kados np
13:01 hdl kados :
13:01 equivalent éèëêe
13:02 works for e
13:02 kados interesting
13:02 hdl one merely have to set the latest equivalence.
13:03 And serching dêterminisme finds déterminisme
13:03 Really COOL
13:04 (I wonder what went wrong when I told you it didnot work.
13:04 )
13:04 kados hmmm
13:04 that's great news
13:07 hdl Maybe we should inquire some "common languages" to see what default we should set for this list...
13:07 kados good idea
16:41 paul or hdl around?
17:55 hdl yes
17:56 kados: yes
17:57 kados hi hdl
17:57 I'm having trouble with the rel_3_0 Context.pm
17:57 it doesn't read the koha.xml file the same way as dev_week does
17:58 hdl what disturbs you ?
17:58 kados well, ccl2rpn isn't being read at all
17:58 do you hard code that in your install?
17:58 or does it properly read it from the serverinfo element in koha.xml?
17:59 also, I fear we have several different ways of doing things
17:59 for instance, the user/password elements
17:59 but most distirbing
17:59 easiest to past in an example
18:00 if I use Data::Dumper to warn Dumper($context) in dev_week the 'server' element gives me:
18:00 'server' => {
18:00 'authorityserver' => {
18:00  'listenref' => 'authorityserver',
18:00   'config' => '/koha/etc/zebra-authorities.cfg',
18:00   'directory' => '/koha/zebradb/authorities'
18:00 },
18:01 'biblioserver' => {
18:01 etc
18:01 but in rel_3_0 I get:
18:01 'server' => {
18:01  'cql2rpn' => '/home/jmf/kohademo/etc/pqf.properties',
18:01   'biblioserver' => {},
18:01 so it's not broken up into servers
18:02 in Context.pm in Zconn, $o->option(cclfile=> $context->{"serverinfo"}->{$server}->{"ccl2rpn"}); does not work
18:04 hdl on pastebin.ca :my koha.xml http://pastebin.ca/256957
18:04 kados hmmm, it looks the same
18:08 hdl: I found the problem
18:08 hdl: XML::Simple does not always handle comments in XMl correctly
18:08 hdl WTF !!!
18:09 Sorry.
18:09 Congrat !
18:14 One must have a nightmare once to enjoy when dreams happens to be lighter :D
18:14 kados hehe
18:15 hdl: there is a restart facility for koha managing the zebra server if a connection is lost, but IMO it shouldn't be used
18:16 hdl: are you OK if I comment it out and commit?
18:16 wait, nevermind, it's already wrapped in a non-existant syspref
18:16 ;-)
18:16 hdl why should it not be used.
18:16 kados do you allow koha to start zebra?
18:16 or do you start it by command line?
18:17 hdl always cmd line.
18:17 kados so it shoudn't be used then I think
18:17 esp since I don't think the process that it gets stuck in will ever due
18:17 die even
18:17 it doesn't background the process or anything
18:18 hdl but ppl will like to have it daemonized.
18:18 kados but maybe that's not best handled in Context.pm ...
18:18 we should create an init script for daemonization
18:19 IMO
18:19 (or actually, ID should ;-))
18:34 hdl :D
18:59 see ya

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

koha1