IRC log for #koha, 2006-05-31

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

All times shown according to UTC.

Time Nick Message
12:46 tumer hi kados?
12:47 kados tumer: hey
12:47 tumer: just committed some stuff
12:47 tumer I will not be able to attend tonight
12:47 kados tumer: you might find useful
12:48 tumer: wow, no tumer, no paul, no pierrick
12:48 tumer about?
12:48 kados tumer: Record.pm with a testing suite
12:48 tumer I've seen it did not have time to play
12:48 kados tumer: you can use it to test your system's capabilities to use the new API I'm proposing for records (which relies on MARC::* on the backend)
12:49 it's got some 'not fully implemented' features as well
12:49 tumer I'll do it soon but rather busy now have to attend a 4 day conf in Turkey this week
12:49 kados ahh, cool
12:49 I'll try to get the stuff you committed yesterday working
12:50 tumer: did you commit all the authorites management stuff?
12:50 tumer I suppose so
12:50 kados tumer: is dev_week == tumer's production koha?
12:51 tumer but I think there are bits that I have committed to head such as script to get marc into authorities table (i.e update from 2-2 to this ver)
12:51 I have to check these
12:52 kados k
12:53 tumer: anything else to discuss?
12:53 tumer I have not committed small bits. everything (with npl templates) regarding circulation, authorities marc editing and search
12:53 ok bye for now
12:53 kados small bits would be useful :-)
12:54 i assume nothing works without the small bits, right? :-)
12:54 tumer will get things complicated with hard coded NEU stuff
12:54 kados ahh, I see
12:54 tumer No everything I committed works
12:54 kados cool
12:55 I'm gonna have owen work on the npl templates this week for dev-week
12:55 tumer how did your system go
12:55 kados did you also commit the new z39.50 stuff?
12:55 tumer being able to edit? witzh zebra
12:56 kados I haven't had a chance to update with the latest commits
12:56 tumer yes all new z39.50 staff
12:56 have to go se you
12:57 kados I'm planning to do that this afternoon
14:18 hey Nick
14:19 Nick heyoh.
14:20 whats up?
14:20 hdl helo
14:27 kados hdl: you're definitely right that we need to start using XSLT
14:28 hdl Maybe. But it wil be a hard time to get all stuff working.
14:28 But that would be REALLY helpful.
14:29 are you using some in MARC2MODS ?
14:29 kados yea
14:30 LOC maintains a bunch of xslt files for mapping various records from/to each other
14:30 we've got about 30 minutes before the meeting, right?
14:30 hdl yes.
14:31 kados I'll bbiab
14:46 chris: you around yet?
14:49 hey russ
14:49 russ hi
14:49 dewey niihau, russ
14:49 kados T-MINUS 20 MINUTES TO MEETING. AGENDA: http://wiki.koha.org/doku.php?[…]ndandnotes06may29
14:49 please add any items you'd like to discuss to the agenda
14:55 rebooting
14:59 bout 1 minute remaining
14:59 russ: anything you'd like to discuss at the mtg?
15:00 hdl: ?
15:00 hdl yes
15:00 russ no the status of the versions is of interest
15:00 hi hdl
15:00 kados hdl: can you add it to the agenda?
15:00 russ: can you go:
15:00 hdl hi
15:00 dewey hi, hdl
15:00 kados /mode +o kados
15:01 give me back my oper status ;-)
15:01 russ oh right
15:01 kados your autoconnect is faster cause you're in NZ :-)
15:02 I'm afraid we don't have enough of a quorum for the Koha Versioning issue
15:02 i think we should have paul, tumer and pierrick in on that one
15:03 esp since paul's RM for rel_2_2
15:03 russ agreed
15:03 hdl yes
15:03 kados so we'll bump that up to next week
15:03 does anyone have anything to say about it for the record?
15:03 (asside from the fact that we need to do something about it? :-))
15:04 who's here?
15:04 russ me
15:04 kados chris up yet?
15:04 russ haven't seen him yet
15:04 hdl I have to say that I am working on authorities. on a special branch.
15:05 Nick me, but I don't matter much.
15:05 kados hdl: how's it coming?
15:05 hdl But it concerns UNIMARC specific authorities management.
15:05 kados right
15:05 hdl A new summary
15:05 amanda morning all
15:05 kados hi amanda
15:05 hdl A new facility to show hierarchies
15:05 amanda hi kados
15:06 kados hdl++
15:07 hdl but still under development and wil have to get adapted to any authority MARC flavor.
15:07 kados ok ...
15:07 I haven't taken a look yet
15:07 but I would be glad to adapt it to MARC21 authorities
15:07 though i suspect tumer will beat me to it :-)
15:08 anything else in "News and Questions"?
15:08 russ: where are you guys at with serials stuff?
15:08 russ i have done a little more work on the wiki
15:08 kados wiki++
15:08 hdl I would be very glad to have tumer's work since he adapted biblio fwk for authorities.
15:08 kados hdl: it's in dev_week
15:08 russ but have got to a stage where i need to review some very old content
15:08 kados hdl: he committed it yesterday
15:09 hdl kados: I shall update and play with it.
15:09 russ i might start firing emails to the list with links to determine if some of the content is still relevant
15:09 kados that'd be great russ
15:09 there's a ton of old old old content there :-)
15:10 russ serials is in final testing with the client who requested the feature
15:10 kados excellent
15:10 russ amanda has written documentation for it, both online help and a manual
15:10 hdl russ: do you itemize serials ?
15:10 kados fantastic
15:10 amanda++
15:10 russ so once we get the all clear on that we will work on committing it to HEAD
15:10 kados great
15:10 russ hdl: itemize?
15:11 hdl barcode issues when they arrive.
15:11 russ ah yes
15:11 amanda yep
15:11 kados russ: how close is rel_2_2 code and your serials stuff?
15:11 russ: is it mostly interface changes?
15:11 russ mostly
15:11 kados cool
15:12 russ our plan is to deploy it not as a replacement but as an alternative
15:12 kados really?
15:12 russ to the current serials module
15:12 kados are you worried that it works differently than the current one?
15:13 russ yep - i dont want to mess around paul and hdl's libraries
15:13 but i guess that is a decision for you :-)
15:13 kados it all depends on how differently they work
15:13 hdl I think we will have to test it and see how it would be accepted here in France.
15:14 kados if yours is based on the current one, it would make more sense to me to implement any changes as system preferences
15:14 russ sure
15:14 hdl And on the differences in database.
15:14 kados it could be potentially confusing to have multiple serials modules that work 99% the same way
15:15 well, we'll wait till we see the commits i suppose
15:15 anything else in News and Questions?
15:15 russ that is all i have
15:16 kados k ... we're gonna skip Koha Versioning
15:16 chris about yet?
15:16 otherwise, we'll have to skip Koha API as well
15:16 chris im here
15:16 kados sweet
15:17 http://wiki.koha.org/doku.php?[…]ndandnotes06may29
15:17 good morning chris :-)
15:17 hdl kados : Do you know about BiblioML ?
15:17 kados hdl: no, what is it?
15:18 russ see you later
15:18 kados bye russ
15:18 hdl http://www.biblioml.org/fr/dtds.html
15:18 kados hdl: it's a set of dtds for dealing with UNIMARC records?
15:18 hdl these are dtds for biblio and authorities management
15:19 kados sweet
15:19 I'll have to take a look at that
15:19 might find something useful for the new Record.pm
15:19 so ... Koha API?
15:19 chris: got a few minutes to discuss the new API?
15:20 chris yep
15:20 what about it ? :-)
15:20 hdl kados: pls detail your points.
15:20 kados well ...
15:21 1. as currently implemented, neither MARC libraries nor non-MARC libraries are fully happy with Koha
15:22 because we're mapping an extensible framework (MARC) to a non-extensible one (old Koha tables)
15:22 so the first issue to discuss is how to resolve this problem so that both types of libraries are happy
15:23 chris and I discussed using an XSLT stylesheet to explain the Koha tables
15:23 (I started one: http://liblime.com/public/koha-tables.dtd)
15:23 sorry ... dtd
15:24 chris and i discussed using a DTD to explain the koha tables
15:24 and I was thinking that a DTD isn't all that different from the MARC frameworks
15:25 if we could implement all the frameworks as DTDs, it would really simplify the process of supporting future record formats
15:25 (I think)
15:26 if we stored all the Koha table data in an XML file
15:26 something like:
15:26 <biblio>
15:26 <biblionumber>1</biblionumber>
15:26 <biblioitem>
15:27 <item>
15:27 etc ...
15:27 hdl what about visibility and values to be linked to authorised_values ?
15:27 kados then we could use zebra to search, etc.
15:28 right
15:28 well, I'm not sure yet, this is just an idea
15:28 I'm no DTD expert either
15:29 in fact, until this weekend, I'd never looked at one ;-)
15:29 but I'm thinking we have:
15:29 scratch that ...
15:29 hdl To my mind, dtds are not that flexible as framework.
15:29 kados maybe not
15:29 chris: what do you think
15:30 chris basically what we want is a way to allow koha to function in a way taht allows people to work with the 3 tier biblio, group, items idea
15:30 kados yep
15:30 chris or deal directly with marc
15:31 so all we are seeking to achieve is a useful abstraction over the marc
15:31 it would be best if the data was only stored once
15:32 kados I disagree that the goal is to abstract the MARC
15:32 that's what we do now
15:32 chris no we dont
15:32 kados no?
15:32 chris we populate a whole other bunch of tables
15:32 kados then I dont' understand what that means :-)
15:33 chris and return data from them
15:33 kados what would be involved in abstraction over the marc?
15:34 chris everything in programming is an abstraction over something, you dont present the data as its stored
15:34 kados ok ...
15:34 chris otherwise ud make everyone work in binary, or flick switches for voltages
15:34 kados :-)
15:35 hdl why not using an existing format then ?
15:35 chris yep we could
15:35 hdl Is there a format you would prefer ?
15:35 chris what i was thinking about (its an early morning idea so might be crap)
15:35 hdl Dublin-Core ?
15:35 chris is that we store the bibliographical data in one place
15:35 and then, we have meta data about that data
15:36 yes maybe in dublin core
15:36 kados hmmm
15:36 chris im trying to avoid the situation we have now, of when we change the title say, we have to change biblio.title, marc_subfield, marc_word etc
15:37 kados righ
15:37 t
15:37 well ... maybe another way to do it
15:37 hdl the problem for us is to keep kind of backward compatibility...
15:37 kados would be to store the koha tables data in XML
15:37 what I'm thinking about
15:38 is the fact that we've had so many people ask if Koha could store non-MARC data
15:38 chris right
15:38 kados if we had a way to define databases and types
15:38 chris yeah
15:38 kados as in types of databases
15:38 this one is 'marc'
15:38 this one is 'dc'
15:38 this one is 'koha'
15:39 chris actually yeah that might work
15:40 for the opac tho
15:40 kados it'd be ideal
15:40 because eventually we're going to want to search multiple databases
15:40 chris you'll still want to have some abstraction over marc
15:40 kados in the opac
15:40 chris or over dc
15:40 etc
15:40 kados hmmm
15:40 yea, true
15:41 I guess ideally something like ISBD if properly implemented
15:41 might be able to serve that purpose
15:41 chris true
15:41 kados or MODS even
15:41 there are XSLT stylesheets for MODS that map between DC and MARC and MODS already
15:41 chris ideally, allowing the library to choose what displays on each page of the search in the opac would be the way to do it
15:41 kados yea, definitely
15:42 hmmm
15:42 chris if the code say
15:42 sorry
15:42 if they could say
15:42 dc.creator = author display that here
15:43 i dunno
15:43 kados yea, it's a tough one
15:43 everyone seen this?:
15:43 http://liblime.com/public/koha-tables.dtd
15:43 chris yep
15:43 hdl I overlooked
15:44 kados woops, posted in the wrong link :-)
15:44 hdl ... I've read DTDs were abandonned in favor of schemas
15:44 kados http://www.loc.gov/standards/mods/
15:44 hdl: MODS Schema, you may be right :-)
15:45 chris that might be what we are after
15:45 kados i implemented a basic marc2mods sub for Record.pm this morning
15:45 Nick The DTDs are getting overcome by schemas for the more elaborate things....
15:45 kados using the xslt on that page
15:46 I'm actually planning to create a new perl module called MODS::Record
15:46 and MODS::File::XML
15:46 to deal with mods records
15:46 but I could use some help knowing how best to implement parts of that
15:46 I've not done much OO programming
15:47 and not sure how best to open files, etc. in a module like that
15:47 chris i dont like OO much at all
15:47 kados no?
15:47 chris nope
15:47 kados it seems to be the way that MARC::* and DublinCore::* are done
15:47 chris yep
15:47 kados as well as Net::Z3950::*
15:48 chris i dont hate it
15:48 kados :-)
15:48 chris i just prefer old school functions and procedures
15:48 kados yea, that's how Koha is written I suppose
15:49 hdl PERL6 is getting more and more OO too.
15:49 chris im yet to be convinced OO does anything other than add confusion
15:49 kados i was gonna ask whether we planned to go OO or not
15:49 chris i dont really mind either way as long as we document the hell out of the methods
15:49 kados yea
15:49 check out Record.pm :-)
15:49 I documented the hell out of it :-)
15:49 chris maybe i dont like OO because usually its really hard to figure out whats going on
15:50 kados yea, I have had that prob
15:50 chris so its less the styles fault, more the people writing in that style :)
15:50 kados just figured it was because I wasn't a programmer
15:50 chris naw
15:50 its just another layer of abstraction
15:50 kados there are two programming questions I've got
15:50 chris it really is all functions in the background
15:50 you are just pretending its an object :)
15:51 kados 1. how do I implement a perl API for a C utility?
15:52 2. how do I open files properly in a module like MODS::Record
15:53 (ie, I've got some XSLT files I need to open, and not where to put them, etc, so they're not hardcoded to my specific directory structure
15:53 chris 1. not sure off the top of my head, i suspect in lots of different ways
15:53 2. yeah thats a tricky one
15:53 hdl 1 is dealt in Porgramming Perl (O'Reilly book)
15:54 s//programing/
15:54 kados hdl: ok, I'll check it out
15:54 chris when you create you mods::record object you could create it with a path
15:54 that you pass to it
15:54 and the files could live under that path
15:54 Nick Chris - if it helps there is another o'reilly book that might help (with the OO thing)
15:55 "Learning Perl Objects, References and Modules" - Schwartz.
15:55 chris the same way you do for html::template
15:55 ahh cool nick
15:55 i like randal schwartz
15:55 he bought me a beer once :-)
15:55 kados heh
15:55 Nick hehe.
15:55 hdl btw I owe you one :)
15:55 chris next time you're in nz hdl :-)
15:57 kados: does that html::template bit help?
15:57 kados hmmm
15:57 kind of
15:57 hdl what about rebuild_zebra_idx...
15:57 kados hdl: what about it?
15:58 I found this great set of record utilities:
15:58 http://www.scripps.edu/~cdputn[…]oftware/bibutils/
15:58 hdl If you could get some parameters and not only biblios directory..
15:58 kados but they're implemented in C
15:59 hence the first question
15:59 anything else to discuss at our mtg?
15:59 chris my $mods = MODS::Record->(path => '/path/to/files');
15:59 kados ahh
16:00 well ... thing is
16:00 I'd like to avoid making people download the files separately
16:00 I'd rather distribute them with the modul
16:00 e
16:00 I tried to figure out how MARC::Charset does it
16:00 but got a bit lost :-)
16:00 chris it'll put them in /usr/local/share or something
16:01 if you can control where the files get put, then you can have a hardcoded ref to them
16:01 kados hmmm
16:01 is hardcoded reference to libraries good programming practice? :-)
16:02 chris the thing with modules like MARC::Charset
16:02 hdl but Linux, windows, MAC seems quite nonesense to have one thing hardcoded.
16:02 chris is that the get installed with a Makefile.PL
16:02 and make and make install
16:03 that can do things like put files somewhere, put the paht of where they are going in the module etc
16:03 kados I see :
16:03 my $path = $INC{'MARC/Charset/Table.pm'};
16:03 yea, I plan to use makefile.PL for MODS::Record
16:04 chris h2xs
16:04 is your friend
16:04 h2xs -XAn MODS::Record
16:04 try that in /tmp
16:04 kados k
16:07 k, I've got some reading / programming to do
16:07 thanks for the help guys
16:07 chris no problem
16:09 ohh nick is nick ... from the olden days
16:09 how the heck are you?
16:11 hdl see you.
16:11 kados bye hdl
16:11 chris night hdl
16:11 kados I've got to get a snack
16:11 bbiab
16:14 Nick Pretty good.
16:14 Looking at getting back to a more normal life.
16:14 chris :-)
16:14 Nick And apologies for my excessive silence!
16:14 Or something like that.
16:14 chris no problem, good to have you back around
16:19 Nick Good to have the chance...
16:24 catch y'all later.
16:24 chris cya nick
16:30 kados btw, in case anyone's interested, here's NPL's testing platform for Zebra:
16:30 http://zoomopac.liblime.com/cg[…]catalogue-home.pl
16:30 based on dev-week code, with some local modifs
16:31 chris brb gotta make some phone calls
16:31 kados k, have fun
16:58 rach morning
16:58 chris morning rach
16:59 rosa morning from me too
16:59 rach rosa we had a very exciting morning - ben did his first proper roll from front to back
17:00 rosa oh exciting indeed
17:00 was he surprised?
17:00 rach he was, but quite excited so we did it only a hundred more times :-)
17:01 rosa just to prove it wasn't an accident
17:02 rach yep
17:02 first time the falling over bit was an accident
17:02 but then he very carefully did it some more times
17:03 rosa oh, intelligent, too!
17:03 rach yep - is careful with his head
18:14 kados chris: what's the difference between dynamically loading and statically linking a library?
18:14 chris ahh statically linking a library you include that library in your code
18:15 dynamically linking you link to a version of that library
18:15 for ege
18:15 you can compile apache with mod_perl included
18:15 or you can load it as a module
18:16 kados interesting
18:16 chris /usr/lib/apache2/modules/mod_perl.so
18:16 dynamic library
18:16 kados so i could actually include binutils header files in my BinUtils module
18:16 and it would be statically linked
18:17 ie, the user wouldn't need to install the real binutils
18:17 chris you compile it into your code
18:17 im not sure you can statically link with perl
18:17 kados otherwise, i could dynamically load the header files from the real binutils
18:17 perldoc perlxs
18:17 chris im not sure the concept holds for interpreted languages
18:17 kados perldoc perlxstut
18:18 perlxs seems to indicate either one is possible
18:18 The XS interface is combined with the library to create
18:18       a new library which can then be either dynamically loaded or statically
18:18       linked into perl.
18:18 chris ahh
18:18 kados or is that something different?
18:18 chris It is commonly thought that if a system does not have the capability to dynamically load a library, you cannot
18:18       build XSUBs.  This is incorrect.  You can build them, but you must link the XSUBs subroutines with the rest of
18:18       Perl, creating a new executable.
18:19 if you want a statically linked xsub you need a new perl
18:19 kados ahh
18:19 gotcha, so it's probably better to dynamically load then
18:19 chris yep
18:19 kados k ...
18:19 i also read up on OO
18:20 I think I'm gonna implement MODS::Record as OO just to get some experience working with it
18:20 plus it seems to be the style of choice when dealing with bib records in perl
18:21 this stuff gets complex real fast :-)
18:21 I keep telling myself I need to take some programming classes
18:21 chris :-) what were you seeking to achieve with you linking?
18:21 with your linking even
18:22 kados well ... I've got this tookit called bibutils, written in C
18:22 it does some really killer stuff like:
18:22 #   bib2xml - convert bibtex to XML intermediate
18:22 # copac2xml - convert COPAC format references to XML intermediate
18:22 # end2xml - convert endnote to XML intermediate
18:22 # isi2xml - convert ISI web of science to XML intermediate
18:22 # med2xml - convert Pubmed XML references to XML intermediate
18:22 # modsclean - a MODS to MODS converter for testing puposes mostly
18:22 # ris2xml - convert RIS format to XML intermediate
18:22 # xml2bib - convert XML intermediate into bibtex
18:22 chris ahh cool
18:22 kados # xml2ris - convert XML intermediate into RIS format
18:22 # xml2end - convert XML intermediate into format for EndNote
18:22 etc.
18:22 stuff we'd kill to have in Koha
18:22 chris so you are wanting to build a perl wrapper?
18:22 kados so I want to write a perl API for it
18:22 yea
18:23 chris righto
18:23 kados so then I can just use Wrapper::BibUtils
18:23 in Record.pm
18:23 chris yep
18:23 kados and have access to all that stuff
18:23 chris makes sense
18:23 kados cool
18:24 hopefully by mid-week you'll see MODS::Record and Wrapper::BibUtils as well as some nice modifs to C4::Record
18:24 btw ...
18:24 what's with the Koha C4 directory
18:24 why is it named C4?
18:24 should't we rename it to Koha?
18:25 chris http://search.cpan.org/src/MIR[…]Net/Z3950/ZOOM.pm
18:25 can if you want
18:25 just make sure you replace every instance of C4 with Koha :)
18:25 kados just curious historically why C4 was named C4 :-)
18:25 (yea, of course :-)
18:25 I was thinking it would be cool to have Koha up on CPAN
18:26 chris 2 reasons
18:26 kados rosa's just telling me :-)
18:26 chris cool
18:27 http://opac.bigballofwax.co.nz[…]-detail.pl?bib=12
18:27 if you can find an updated version of this joshua
18:27 kados heh, that's a cool reason for the name C4
18:27 chris its a good theoretical grounding
18:27 kados k, thanks
18:28 chris makes the practical more comprehensible i find
18:28 kados cool
18:28 so was C4 going to be the original name of Koha?
18:28 chris it was just a working name
18:28 kados kinda like C and C++ and C# :-)
18:28 chris until we came up with a real one
18:29 there were a few tossed around before koha was settled on
18:29 i forget the others
18:29 kados what's your thought on CPAN?
18:29 chris i remember sitting in a cafe with rachel, rosa, jo and si .. and probably olwen too talking about names
18:29 i think thats a good idea
18:29 kados think we'll see a day when we can just install Koha from CPAN?
18:30 chris well, most of it
18:30 all the modules anyway
18:30 kados right
18:30 chris its another thing to keep up to date
18:30 BUT
18:30 if we had it in cpan
18:30 then it would be easier to get
18:30 kados yea
18:30 chris lib-koha-perl
18:31 in debian for instance
18:31 kados yea
18:31 chris or
18:31 libkoha-modules-perl
18:32 i think first priority
18:32 fixing the bundle
18:32 kados yea ... i think I actually have write access on PAUSE for that
18:32 i forget now
18:33 why'd you post the zoom link above
18:33 chris oh, cos its the wrapper to the c libs for zoom
18:33 kados as an example of a dynamically linked XS?
18:34 chris yes
18:34 kados yea, I couldn't really figure that out :-)
18:34 chris ok lets find a better one :)
18:35 kados \
18:35 oops
18:35 chris http://search.cpan.org/src/DAN[…]ang-0.07/Slang.pm
18:36 dunno if thats any more comprehensible :)
18:36 kados hmmm ...
18:36 is that all there is to it?
18:36 ie, that's the whole deal?
18:37 chris http://search.cpan.org/src/GIR[…]es-1.13/Curses.pm
18:37 theres a bigger one
18:38 with lots of comments
18:38 kados smallness is good :-)
18:39 chris i *think* thats all there is to it
18:40 you define all the names, which are the same names as in the c
18:40 then use the dynaloader
18:40 and it works
18:40 but maybe thats just wishful thinking :)
18:41 kados there's also the XS file:
18:41 http://search.cpan.org/src/DAN[…]ang-0.07/Slang.xs
18:41 chris right
18:42 id try it with just the version
18:42 or just one function
18:42 kados yea
19:15 hey tumer
19:15 tumer hi kados
19:15 how did the meeting go
19:16 I could not read the logs it seems off at the momemnt
19:16 kados pretty well
19:16 ahh ... too bad
19:16 chris: logbot is off?
19:16 no, seems to be up
19:17 tumer well not the link from wiki
19:17 hi chris
19:17 kados seems to be working:
19:17 http://koha.org/cgi-bin/logs.p[…]saveas=&save=save
19:18 chris it should be working
19:18 tumer kados: regarding my commits
19:18 kados yes?
19:18 tumer I did not leave behind anything as you thought
19:19 everything about circulation search biblio & authorities management is there
19:20 all other modules I did not commit because of incompatibilities of coding
19:20 kados what other modules?
19:21 :-)
19:21 tumer well serials acquisition etc.
19:21 kados does your serials acquisition differ greatly from what's in rel_2_2?
19:22 is it built on what's there?
19:22 (now we have three versions of serials in Koha!)
19:22 tumer if I commit everything it gets more complicated. Hourly charged reserve section circulation, barcode based reserves ...
19:22 kados (paul/hdl's, katipo's, and tumer's)
19:23 wow ... I didn't realize you'd done that
19:23 chris: hourly charged reserves !
19:23 chris: wonder how similar that would be to what Toledo wants
19:23 chris quite different
19:23 but useful nonetheless
19:23 kados barcode based reserves is what NPL wants
19:24 tumer no not reserves!
19:24 chris oh thats already done
19:24 kados yea in head
19:24 not reserves?
19:24 chris hourly based issues id be interested in
19:24 kados ann ...
19:24 hourly based issues
19:24 chris yeah
19:24 thats something useful for lots of libraries
19:24 tumer hourly charged circulation material which we call reserve section
19:24 kados yea, all that stuff is highly desired
19:25 but I know what you mean about complicating things
19:25 chris maybe we need to merge it in with a systempref or so
19:25 kados we've got to fix this versioning stuff
19:26 tumer its getting more complicated than systemprefs
19:26 we have to merge at one point
19:26 chris yes
19:26 kados my idea would be to have:
19:27 rel_2_2 -> 2.2.6
19:27 dev_week -> 2.4
19:27 head -> 3.0
19:27 chris head is head
19:27 kados yea ... branch when it gets stable
19:27 chris it wont get stable
19:27 :)
19:27 kados hehe
19:27 chris seriously it wont
19:27 we will want to branch when we call feature freeze
19:28 and then make that branch stable
19:28 tumer I want to know whether someone using dev-week can actually setup a system
19:28 it works for me but thats not good enough
19:29 kados NPL's going to be testing it starting at the endo fo this week
19:29 chris yeah if we want 2.4 to come from there we should do a 2.3.0 release (a dev one) from dev_week
19:29 and then a 2.3.1 etc .. until we get it to a point where its good enough for 2.4
19:29 kados huh ... an actual release, eh?
19:30 chris release early release often
19:30 kados tumer: http://zoomopac.liblime.com/cg[…]catalogue-home.pl
19:30 chris we havent done an unstable release in ages
19:30 kados true
19:30 I can't remember one in fact :-)
19:30 chris well leading into 2.2 and 2.0 we did them
19:30 and up to 1.2
19:31 there were a lot of 1.1.x release etc
19:31 you cetainly can spot how far you are away pretty easily
19:32 kados side note:
19:32 tumer: did you see my mail I forwarded to you about attr 3=?
19:32 tumer kados:no
19:33 kados tumer: i don't think you get email from me for some reason
19:33 tumer any mails from you do not come
19:33 kados weird
19:34 I'll have to start using koha-devel to email to you :-)
19:34 tumer very
19:34 well at least use my hotmail and notify me on koha-devel
19:34 kados what's the hotmail again?
19:35 tumer tumergarip@hotmail.com
19:35 kados k
19:36 tumer: forwarded that email from ID
19:36 tumer: on attr 3
19:38 tumer chris:whta time is it there?
19:38 chris 12.38pm on tuesday
19:39 kados tumer: based on Sebs response, I'm wondering if you're thinking of removing the 'starting with' or re-implementing it as he describes?
19:41 tumer: check your spam filter :-)
19:41 tumer kados:since we can not use it I'll remove it
19:42 kados: I'll add the functionality of 'whole of subfield'
19:43 which will mean an exact match say for a title
19:43 kados well ... almost
19:43 it won't catch cross-subfield titles like
19:43 Lord of the Rings: Fellowship of the Ring
19:43 but probably close enough
19:43 tumer just to give opportunity to reduce number of hits when needed
19:43 kados yep
19:44 chris yeah
19:44 tumer its already implemented with authorities -committed
19:44 kados excellen
19:45 t
19:45 I've got to get some dinner
19:45 talk to you guys later
19:45 chris cya later
19:46 tumer and me some sleep at 03:45am
19:46 chris sleep well tumer
19:46 thanks for your hard work
19:46 tumer bye chris
21:12 kados funny ...
21:12 http://dev.perl.org/perl6/architecture.html
21:12 One of the major reasons to revisit Perl was to fix the mess that is XS (the way you extend Perl with C or C++ subroutines).  Perl5 has no API for extension, separate from the functions used to implement Perl, and extending Perl requires hideous amounts of work.  Dan and Larry are aiming to make C extensions as easy as they possibly can be (Brian Ingerson's excellent perl5 Inline modules give some directions for this).  Anyone who has used XS looks forward to its dem
21:13 chris yeah theres lots to look forward to in perl 6
21:14 did you see this
21:14 http://www.perlmonks.org/index.pl?node_id=494927
21:14 (speaking of funny)
21:15 kados hehe
21:25 looks like Inline.pm 's the way to go here
21:25 you can actually run C code without messing with anything
21:26 chris sounds good to me
21:26 kados perl -e 'use Inline C=>
21:26 q{void J(){printf("Just Another Perl Hacker\n");}};J'
21:26 and since conversion is never high-priority for speed
21:26 chris yeah
21:26 kados I figure we can get away with it
21:26 chris does it run compiled c?
21:26 kados haven't gotten that far
21:27 yep
21:27 The most common real world need for Inline is probably using it to access existing compiled C code from Perl. This is easy to do. The secret is to write a wrapper function for each function you want to expose in Perl space. The wrapper calls the real function.
21:27 hehe
21:27 perfect
21:27 chris cool
22:23 kados so, looking at Inline, here's how to access readline() from the GNU ReadLine Library:
22:23 package MyTerm;
22:23    use Inline C => Config =>
22:23               ENABLE => AUTOWRAP =>
22:23               LIBS => "-lreadline -lncurses -lterminfo -ltermcap ";
22:24    use Inline C => q{ char * readline(char *); };
22:24 as described here:
22:24 http://search.cpan.org/~ingy/I[…]Function_Wrappers
22:24    package main;
22:24    my $x = MyTerm::readline("xyz: ");
22:24 I only had one quarter of C way back when, and don't remember much ...
22:24 LIBS => "lreadline, etc'
22:25 would those be the header files? or the .c files?
22:25 (or the bin files?)
22:25 say I've got the function xml2end
22:26 which exists in bin/xml2end.c
22:26 (in the bibutils stuff)
22:27 chris good question
22:27 i havent written any c since 1992
22:27 kados hehe
22:27 it goes on to say:
22:27 We access existing functions by merely showing Inline their declarations, rather than a full definition. Of course the function declared must exist, either in a library already linked to Perl or in a library specified using the LIBS option.
22:27 so it seems I need:
22:28 chris im guessing u are linking to the .so files
22:28 kados 1. the function name and it's 'declarations' (whatever that means)
22:28 and 2. the library where they exist
22:28 .so, eh
22:28 chris thats my guess
22:28 course ld does the linking for ya
22:29 kados http://www.scripps.edu/~cdputn[…]tils_3.24_src.tgz
22:29 is where the source for bibutils lives
22:29 chris it might not be compiled into a dynamically linkable chunk of code
22:29 kados ahh, maybe I need to actually install it first
22:29 chris actually i lie
22:30 its probably the .h
22:30 files
22:30 kados hmmm ...
22:30 chris actually i have no idea
22:30 kados heh
22:30 chris i just know that when you are compiling, or making something
22:30 -lsomelib
22:31 means link in that library
22:32 kados ahh
22:33 maybe it's the object files
22:33 well, I'm getting awful tired
22:33 I'll hack on this some more tomorrow
22:33 cheers
03:20 hdl hi
03:20 dewey hi, hdl
03:21 ToinS salut hdl
03:33 qiqo hello everyboday
03:34 anyone home?
03:34 paul hello qiqo.
03:34 qiqo hi paul
03:34 paul poulan?
03:34 paul yep
03:34 qiqo ei how are you
03:34 how was dev week
03:34 paul fine, thanks.
03:35 qiqo any development with 3.0?
03:35 paul devWeek was very very interesting, but also very tiring !
03:35 many. although koha 3.0 is not ready yet
03:35 qiqo how i wished i was there,, but you know, we lack the money
03:35 is zebra doing fine?
03:36 paul tumer (from cyprus), did a wonderful work on this topic.
03:36 qiqo wow
03:36 ei, ive a question, can i remotely access the intranet part of koha even if i am out of that netwoek?
03:37 paul yep, if your apache virtual host permit this !
03:37 qiqo hows that?
03:37 paul it's not a Koha problem, but a TCP-IP / Apache one.
03:37 qiqo ahh..
03:37 so am i to configure apache?
03:38 how will i know if its permitted..
03:38 paul yes, you have to look at apache config file, and maybe your firewall/gateway config.
03:38 qiqo i really have to work outside because i still am at school so i need to access my client's intranet
03:38 ahh alright,, my opac is runing great outside the network
03:41 another question, well as ive told you before, i am planing to have a linux distro with koha integrated with it. How come there are no projects like this before?
03:41 paul I think there are or were some. although not necessary uptodate.
03:42 but none of them is supported officially. This could change if someone want to.
03:42 qiqo actually im done with the System analysis, and right now, i am facing difficulties with packing the perl modules as a debian package, eventually i might surpass these hardships you know
03:43 ohh so to whom shall I ask for its official support?
03:44 you know, an ILS integrated with a linux distro is badly needed in my country since a portion of our archipelago has internet
03:44 and libraries usually are at the remote places..
03:46 geez.. i talk too much, sorry :)
03:46 paul qiqo: Koha just need someone saying "i'll package a distro with Koha included, and maintain it". If, after discussion (on koha-devel probably), it appears the guy saying this is serious and should be able to do this on a long term, then it's donne.
03:47 the guy will be applied "Koha-distro manager" ;-)
03:48 qiqo hmmm.. wow,, i really should have a product here since this is my thesis on my undergrade
03:48 *undergrad..
03:48 hehe
03:48 paul the question is not only "have a product", but also :
03:48 - share the tools used to create it
03:49 - be sure to have a "long-term" update of the product.
03:49 qiqo and i think, i will be obliged to maintain this because i have a line of beneficiaries since this will be implemented to public libraries here in the Phlippines
03:49 :)
03:50 ei is 3.0 available at the CVS?
03:50 paul another point : the tool, to be official, must let you choose between, for example, language and marc flavour. It can't be only 1 language & 1 marc (marc21 or unimarc)
03:50 of course, everything is on CVS.
03:50 qiqo alright
03:50 paul but it don't work at all atm !
03:51 qiqo hmmm.. well linux is very flexible right? so one has the option to choose his/her native language
03:51 and i believe has other language support too, right?
03:52 and i believe koha has other language support too, right?
03:55 well ei gotta go, i will send apport to you a report regarding the development of my project,, thanks paul
03:55 paul your welcome.
03:55 qiqo more power to us
04:21 hdl hi tumer
05:22 paul hello pierrick
05:23 pierrick hello paul
05:23 sorry but I forgot the meeting yesterday :-/ My son had som "difficulties" to sleep
05:24 paul I told everybody we would probably miss the meeting as we were in Lyon. so, they where not surprised !
05:25 chris hi paul and pierrick
05:25 paul hi kiwi.
05:25 chris heh
05:25 paul how are things ?
05:25 ToinS hi all !
05:25 chris hiya toins
05:25 things are good
05:26 busy but good, how are things there?
05:26 paul busy too.
05:26 and a lot of wind today !
05:26 nope
05:26 chris the first time wellington made it to the final
05:27 and most of the game was hidden by fog
05:27 and we lost :(
05:27 pierrick hi chris, hi ToinS
05:28 ToinS hi pierrick
05:31 chris paul: are france touring this year?
05:31 paul touring ?
05:33 chris coming down to the southern hemisphere to play against australia or south africa, or argentina ... i know you arent playing against nz in our winter
05:35 :-)
05:35 nz play brazil in switzerland soon
05:35 thats the only soccer news i know :)
05:38 ok i think i have everything ready for my meeting tomorrow time to go to bed i think
05:38 paul bye chris.
05:38 chris bye
05:38 paul sweet dreams to you & your wife
05:38 (how's she ?)
05:39 chris shes good, shes writing up about the galleries she went to in marseille
05:39 for her work
05:39 paul and baby ?
05:39 dewey well, baby is really fine :-)
07:46 kados morning all
07:47 I guess Context.pm is object-oriented
07:47 or at least it tries to be :-)
07:49 paul hello kados.
07:49 you're right, Context.pm is OO
07:49 + I now have my 1st UN library too : 1 for katipo, 1 for you & 1 for me ;-)
07:50 (a small migration contract)
07:58 kados cool
07:58 which one?
07:58 paul unu-merit
07:58 http://www.merit.unu.edu/
07:59 kados nice
08:06 paul: I'm not familiar with OO programming that much
08:06 tumer has written:
08:06 my $oConnection=C4::Context->Zconn("biblioserver");
08:06 in Search.pm in dev_week
08:06 then, in Context.pm:
08:06 sub Zconn {
08:06    my $self = shift;
08:06    my $server=shift;
08:06 but when I do a warn on the next line:
08:06    warn "SELF:".$self." SERVER".$server;
08:07 SERVER doesn't have 'biblioserver'
08:07 (though $self has C4::Context)
08:07 paul: can you see anything wrong with the syntax above?
08:07 paul what is in self ?
08:07 what says Data::dumper ?
08:08 kados what should I pass to Data::Dumper, @_?
08:09 paul $self and $server 1st I think.
08:09 or @_, it should be useful too
08:13 kados it seems it only has $VAR1 = 'C4::Context';
08:13 so where did "biblioserver" go?
08:16 paul mmm... good question...
08:18 (but suspect there is a problem if you try to connect to 2 different servers : the 1st one will make $context->Zconn exist, and &new_Zconn won't be called.
08:19 so, you can have only 1 connexion here.
08:19 (unless i'm missing something)
08:21 kados the original Context.pm could handle multiple contexts
08:22 at least it claims to in the docs ;-)
08:22 the bottom line question si
08:22 why isn't:
08:22 my $oConnection=C4::Context->Zconn("biblioserver");
08:22 passing "biblioserver" to the Zconn sub as the second argument?
08:23 IIRC, first argument is always class in OO right?
08:33 paul kados : right.
09:02 kados so we have a system in Koha, Context.pm, and noone knows how to use it :-)
09:02 great! :-)
09:03 paul lol
09:13 kados hey owen
09:13 how was the weekend?
09:13 owen Hot!
09:14 kados heh
09:15 owen: just noticed a bit of a glitch in the OPAC when displaying long URLs ...
09:15 http://wipoopac.liblime.com/cg[…]etail.pl?bib=4719
09:15 hey tumer
09:15 tumer hi kados
09:15 kados tumer: i discovered the problem with Context.pm
09:15 tumer yes?
09:16 kados tumer: in Search.pm you have:
09:16 my $oConnection=C4::Context->Zconn("biblioserver");
09:16 and in Context.pm:
09:16 tumer hmm
09:16 kados sub Zconn {
09:16    my $self = shift;
09:16    my $server = shift;
09:16 when I add:
09:16 warn "SERVER:".$server;
09:16 I get nothing
09:16 but if I put:
09:17 $server = "biblioserver";
09:17 it works fine
09:17 tumer sure
09:17 kados soe for some reason, it's not passing in the value
09:17 I'm not enough of an OOP geek to understand why
09:17 tumer its my poor perl
09:18 how to call $server and $self ? may be @_ i donno
09:18 kados tumer: I used Data::Dumper
09:18 tumer: and did:
09:19 Dumper(@_);
09:19 well ...
09:19 warn Dumper(@_);
09:19 but all I get is what's in $self
09:19 'biblioserver' doesn't show up in the object
09:19 tumer :-(
09:20 kados it works on your windows box?
09:20 tumer yes it does
09:20 kados really strange
09:20 well, at least we undrstand the prob nw
09:20 now
09:20 I can bug chris about it tonight
09:20 one ommore thing
09:20 more even
09:20 the facility to restart zebra if the server crashes
09:21 has a hard-coded reference to a log file
09:21 (plus I think it won't work on *nix )
09:21 maybe that log file should be stored in the koha.xml?
09:21 tumer oh! that has to be commented aout altogether cause its windwos
09:21 windows
09:21 kados yea, i did comment it out eventually
09:22 finally ...
09:22 the fallback mechanism doesn't work
09:22 if the zebra server goes down (well, the connection at least)
09:22 searching doesn't work
09:23 tumer have to check logs. probably mysql tables not indexed on required fields
09:24 all the old koha tables major fields have to be indexed for it to work
09:24 like title,subtitle author,isbn,subject etc.
09:25 kados hmmm
09:25 tumer I played with record.pl
09:25 kados Record.pm you mean?
09:26 how'd it go?
09:26 tumer lots of error whn it comes to comining chars
09:26 kados ok ...
09:26 tumer combining chars i mean
09:26 kados I suspect the reason is because you've got the PurePerl sax parser
09:26 sec ...
09:26 tumer nop. Expat
09:27 kados run this:
09:27 #!/usr/bin/perl
09:27 use XML::SAX::ParserFactory;
09:27 $parser = XML::SAX::ParserFactory->parser();
09:27 print $parser."\n";
09:27 as a script on the command line
09:27 and tell me the output
09:27 Expat has probs too
09:28 what you need, is compatible versions of libxml2 and XML::LibXML
09:28 tumer all code turned into smileys on my screen i cannot read it
09:28 kados with that setup, on linux, I can handle anything
09:28 ahh
09:28 hang on
09:29 tumer you now that i posted a mail regarding expat long ago
09:29 kados http://liblime.com/public/parser.pl
09:29 tumer regarding problems on windows, lots of reading links
09:29 kados try running that script on windows
09:30 yep ... expat definitely has some probs
09:30 it doesn't work for me either
09:30 the only thing that I've been able to get working in all cases, is compatible versions of libxml2 and XML::LibXML
09:30 tumer whats gets on my nerves is this:
09:30 kados the 'compatible versions' thing is important
09:30 because there are quite a few incompatible versions
09:31 tumer on windows we have a small 13mb utility that converts any marc to utf8 iso8859 and back ,dublincore,xml, and many more reversibly
09:32 and 350mb charset.pm is merely converts from marc8 to uthf8 with problems
09:32 kados heh
09:32 are you on the perl4lib mailing list?
09:32 tumer no
09:32 iam not a perl onger
09:33 monger
09:33 kados right
09:33 well ... it's where the MARC::* guys talk about MARC::* stuff
09:33 and recently i've been sending all kinds of messages about MARC::*
09:33 complaining, giving specific examples of problems, etc.
09:33 finally, i did manage to resolve the issues
09:34 tumer: what's the 13Mb utility?
09:34 tumer MarcEdit
09:34 kados ahh
09:34 tumer from oregon univ
09:34 kados I'd consider using a C library for the marc8->utf8 stuff
09:35 we could use Inline.pm to write a perl wrapper
09:35 tumer its aVB machine, callable from perl as well
09:35 kados if we could fine an open source library out there
09:35 what's a aVB machine?
09:35 tumer VisualBasic
09:36 language that is
09:36 kados ahh
09:36 yea, don't think that'd fly with the koha dev group
09:36 for now, I think we're stuck with MARC::Charset
09:36 actually, it's not MARC::* that has probs
09:36 it's the underlying parser
09:36 tumer no thats why i did not mention it. ALL i am saying others did it
09:37 kados tumer: if you try installing libxml2 and XML::LibXML hopefully it won't wreak havoc on your combining characters
09:38 tumer: you can take a look at the archives of perl4lib:
09:38 http://www.nntp.perl.org/group/perl.perl4lib
09:38 tumer do you remember the issue when searching with combininig chars?
09:38 they show funny on the screen?
09:38 kados yes
09:39 tumer is it resolved?
09:39 kados I'll test and see
09:43 tumer: your turkish records use 90o instead of 090?
09:43 tumer yes
09:44 you do not need my records
09:44 kados no?
09:44 tumer try doing a search on your records but replace normal e with say accented e
09:44 kados ahh, I don't know how to create such a letter on my keyboard
09:45 tumer the search works
09:45 but the search term appears funny
09:45 kados that server isn't running libxml2
09:45 I haven't updated it yet
09:46 I'll try to do so today
09:48 tumer hmm accented characters accented e is Alt+0232 on my keyboard
09:49 kados ™
09:49 ls
09:49
09:49 ahhk
09:49 now I can't see anything :-)
09:49 tumer i can see some parts
09:52 kados tumer: can you point out a utf-8 combining chracter somewhere online so I can use it to test?
09:53 maybe here?:
09:53 http://www.unics.uni-hannover.[…]p/combimarks.html
09:54 weird ...
09:54 tumer on line 0326 the last character s use that instaed of a normal s
09:55 kados now I'm getting 'the server is too busy' error when I search
09:56 tumer is the zebra down?
09:56 kados no
09:56 it has:
09:56 10:56:04-30/05 zebrasrv(2) [session] Starting session 2 from tcp:127.0.0.1 (pid=29910)
09:56 10:56:04-30/05 zebrasrv(2) [log] User perm for perm.anonymous: r
09:56 10:56:04-30/05 zebrasrv(2) [request] Auth none
09:56 10:56:04-30/05 zebrasrv(2) [request] Init OK - ID:81/81 Name:ZOOM-C/YAZ Version:1.70/2.1.18
09:56 10:56:04-30/05 zebrasrv(2) [log] ResultSet '1'
09:56 10:56:04-30/05 zebrasrv(2) [log][app2] zebra_register_open rw = 0 useshadow=0 p=0x80b88e0,n=,rp=(none)
09:56 10:56:04-30/05 zebrasrv(2) [log] profilePath=.:/usr/share/idzebra​/tab/:/koha/zebradb/biblios/tab cwd=/koha/zebradb/biblios
09:56 10:56:04-30/05 zebrasrv(2) [log] user/system: 0/0
09:56 10:56:04-30/05 zebrasrv(2) [request] Search ERROR 125 1 1+0 RPN: @attrset Bib-1 @attr 2=102 @attr 5=1 @attr 6=3 @attr 3=1 @attr 4=1 @attr 1=1016 �
09:56 10:56:04-30/05 zebrasrv(2) [session] Connection closed by client
09:57 10:56:04-30/05 zebrasrv(2) [log][app2] zebra_register_close p=0x80b88e0
09:57 I get a completely different response when I try a real search
09:57 instead of just that accented A
09:57 I'll tro 0326
09:57 try even
09:58 tumer well those characters i see are not defined in our sort-utf.chr
09:58 but s is
09:58 kados ahh, that would expain it then
09:58 I get ? for s
09:59 you too?
09:59 tumer you mean at search? or the result
09:59 kados at result
09:59 when I copy/paste in the Ş (special S with tail)
10:00 but, libxml2 isn't installed, so I'll withold judgement for now :-)
10:01 tumer i tried and i get another character instead of s with cedilla
10:01 not readable character
10:02 did you realise that we had to talk about sort-utf file as well?
10:03 do we have to define all available characters in this table? I had to do it with turkish chars
10:04 and if that is the case whats the use of having Charset?
10:04 kados hmmm
10:04 tumer its as if writing char_decode again under a differnt name
10:04 kados tumer: you're a MARC21 library
10:04 tumer: as such, there are only two valid encodings for your records
10:04 tumer yes but all utf8 now
10:04 kados tumer: MARC8 and UTF-8
10:05 tumer: but if you download records from outside, you still need to get them in MARC-8 format
10:05 tumer we download marc8 but create utf8
10:05 kados tumer: and convert to UTF_8
10:05 tumer yes char_decode does that for us
10:05 kados we will always need a mechanism to convert from marc-8 to utf-8
10:06 tumer the current char_decode in biblio.pm (modified) always done that for us
10:06 kados but can you imagine how large char_decode will become?
10:06 every time a new language is added
10:06 it's better to use the codetables.xml file provided by LOC
10:07 which has a complete mapping already defined from marc-8 to utf-8
10:07 for all cases
10:07 tumer what i am saying is if we have to write it for zebra isnt it the same problem
10:07 kados no, we don't need to search marc-8 in zebra
10:07 just utf-8
10:07 internally, everything's utf-8
10:08 what's sort-utf file anyway, zebra's 'how to sort' chart?
10:08 tumer in zebra you have to define the alphabet you use how they sort and which maps to what base char
10:09 in tabs folder we have a sort-something-utf.chr file
10:09 which i committed with utf8 characters for turkish
10:10 thats why you can search with s instead of s with cedilla to recah the same record. Like mysql does
10:10 s/recah/reach
10:11 kados right
10:11 I guess I need to update my sort-something file
10:11 since i can't search on S with cedilla
10:11 tumer you should already have it
10:12 can you point me tou your server i wanna do asearch with accented chars if possible
10:13 kados http://zoomopac.liblime.com/cg[…]catalogue-home.pl
10:14 tumer any term suggestions?
10:14 english
10:15 kados book
10:16 tumer no it does not search on accented characters.
10:16 unless you have some other sort-blabla its strange
10:20 kados I dont' think I do
10:20 it's got equivilent lines
10:20 with aa<withaccent>
10:20 weird
10:22 tumer and thtas the file used by zebraidx when indexing the records, ie in same folder as our record.abs
10:22 oh another thing
10:22 kados it's in /koha/zebradb/biblios/tab
10:22 tumer default.idx
10:22 kados yea
10:23 it's in there too
10:23 tumer same as committed ie points to sort--
10:23 kados also in /koha/etc/tab
10:24 yep, just copied from cvs
10:24 tumer veryy strange
10:25 in sort file there are lines called equivalent
10:26 with some accented characters equalling to unaccented forms
10:26 kados yep, in mine too
10:26 tumer they all work for me
10:27 i noticed that i dont have accented e in there which i should add and reindex my zebra
10:27 kados i wonder if the accented chars aren't getting passed unharmed to zebra on my system
10:27 ie, maybe they're being mangled along the way
10:27 tumer but for you i am stunned
10:28 kados this worked when we set it up on dev-week
10:28 unless ...
10:28 maybe it has to actually have a database with the chars in it in order for the search to work
10:28 that sounds like something ID would do :-)
10:28 tumer no
10:28 kados and it would take us weeks to figure it out :-)
10:29 tumer well here is another thing about ID
10:29 infact 2
10:29 their new zebra 1.4 does not work. Im tired i did not bug them
10:30 second say you index on additional authors or some other field
10:31 and the records you entered did not have that field up to that day
10:31 but you know you'll have them
10:31 kados sigh
10:31 tumer try and do a sort on taht field
10:31 it actually crashes
10:31 kados tumer: well, I've paid for support, so they have to fix a reproducible bug within 10 days
10:32 paul at leas, we will know if it was worth the price !
10:32 kados paul: right!
10:32 tumer: how can I reproduce it on my system?
10:33 tumer yes put a sort in your record.abs on afield that you know does not contain data yet
10:34 kados while zebra's running?
10:34 tumer no
10:34 do this and reindex zebra
10:34 then using yaz client do a search
10:34 and a sort on that field
10:34 kados ok
10:35 lets go step by step here ... what field shall I try in record.abs?
10:35 how about a field that does not exist at all in the data
10:35 tumer which field you are sure doews not have data
10:35 kados field 0f0
10:35 tumer i donno your records
10:35 kados $a
10:36 tumer: how about this:
10:36 melm 0f0$a      False:s
10:36 tumer just sec finding a bib1-att
10:36 kados tumer: added to record.abs ... do I also need to change another conf file? bib-1 or something?
10:38 tumer use melm 0f0$a Dewey-classification:s
10:38 kados in fact ... it won't even index I think
10:38 ok, I'll try that
10:38 now i re-index the records
10:39 indexing now
10:39 tumer do a search f book and then sort 1=13 i<
10:40 13 means dewey- classification
10:40 kados can you specify the RPN? I'm not up to speed on that yet
10:40 or is it just :
10:40 f book
10:40 sort 1=13
10:40 tumer sort 1=13 i<
10:41 kados Z> sort 1=13 i<
10:41 Received SortResponse: status=failure
10:41 Diagnostic message(s) from database:
10:41    [207] Cannot sort according to sequence -- v2 addinfo ''
10:41 Elapsed: 0.000427
10:41 is that the same as what you get?
10:42 tumer yes
10:42 kados in the zebra log i get:
10:42 11:41:37-30/05 zebrasrv(1) [log] user/system: 0/0
10:42 11:41:37-30/05 zebrasrv(1) [request] Sort  ERROR 207 (1)->1
10:42 tumer but its a bug
10:42 kados why?
10:42 tumer if you have at least one record with taht data it stops
10:42 kados (i don't quite understand why it's a bug since there are no values to sort by)
10:43 it stops?
10:44 ahh, you mean it claims to have sorted correctly
10:44 doesn't throw the error
10:44 hmmm ...
10:44 tumer I want to be able to sort on fields which will have data later on even if they have null in then now not a database-error
10:45 kados how about a temporary solution
10:45 which is to create a single record with all values in it
10:46 I see what you mean though
10:46 is this a priority bug?
10:46 tumer kados: its more serious than you think
10:46 kados ahh, yea I can see it being a real problem
10:46 tumer is paul around?
10:46 kados especially for small collections
10:46 paul yep
10:47 tumer did you folow this
10:47 whats your opinion
10:47 kados ok, I'll file a bug report
10:47 tumer: I'll cc your hotmail account
10:47 paul tumer: no, i didn't follow the thread
10:48 tumer the problem is if there is no data in the sort field zebra throws and error and does not even return the unsorted results
10:49 and dont forget we have to pre-define all aour sort fields beforehand
10:50 kados tumer: was this something that was supposed to be fixed in 1.4? or is it a new bug we discovered?
10:50 tumer new thing
10:50 kados k
10:50 filing bug now
10:50 paul could we have a sort order that contains something for sure ?
10:51 tumer say i sort on title and subtitle
10:51 but my records up until then did not have subtitles yet
10:52 searching the records with sor order title,subtitle returns an error
10:52 thats a stupid bug
10:53 when a single record enters with subtitle error gone
10:59 kados:for your bug report=> we are building the sort into our search we are not sorting after we receive the results
11:00 kados right
11:02 tumer do you wnat me to expand on that?
11:03 kados sure
11:03 tumer you searced: f book
11:03 got results
11:04 then did a sort
11:04 got error
11:04 kados right, I get that
11:04 how can we write the whole thing in one go?
11:04 tumer my code does like mysql does
11:04 find book sort something
11:04 kados lets do a specific example for the report
11:05 tumer so the records fetched once and sorted
11:05 kados lets sort by title first and then by 1=13 or whatever
11:05 in yaz-client
11:05 lets sort by title first and then by 1=13 or whatever
11:05 in yaz-client
11:05 tumer just a sec this server does not like irc much
11:06 i donno how to do this with yaz client
11:06 i think it was 7=13
11:07 but my search code has it in zoom
11:07 kados how do we do two sorts?
11:07 tumer so asking for sorted records throws out an error and returns failiure
11:07 kados ie, first sort by title, then by author?
11:08 if possible, we should try to get a one-liner for the bug report
11:08 tumer two sorts 1=4 i< 1=13 i<
11:08 kados yea, that fails too
11:11 tumer: ok sending it off, cc to you
11:11 if you have more to add, cc me
11:11 tumer ok
11:12 kados sent
11:12 so, it should be fixed by June 10 :-)
11:12 amazing how that works :-)
11:14 tumer: any other bugs we can try and reproduce?
11:14 tumer well have used the lates zebra from snapshot or cvs?
11:15 have you used i mean
11:15 kados ahh ... did we ever find the snapshot?
11:15 I'm running the debian package
11:15 lemme check the version
11:15 1.3.34
11:16 is what I'm running
11:16 tumer good stable one
11:16 kados didn't we go looking for the latest snapshot and it didn't exist?
11:16 :-)
11:16 or did you eventually find it?
11:17 tumer the cvs did not exist indexdata.dk/taz has got a link to it now
11:17 kados cool
11:17 tumer not taz /yaz
11:17 kados yep
11:17 tumer well it does not even index now
11:19 kados what doesn't?
11:20 tumer zebraidx stopped working with 1.4 major bug i believe but since its still roduction did not report it
11:22 kados ??
11:22 if you don't report it how will it ever get fixed ? :-)
11:22 well ... nevermind ...
11:22 tumer well tired of ID
11:22 kados what about 1.3.34 ... any bugs there?
11:23 yea, sorry about that
11:23 tumer i have been using 1.4 for the last month
11:23 kados don't know why, but a lot of folks have problems dealing with them
11:23 I haven't personally, but have heard many reports of problems
11:23 i can try installing 1.4 and reproduce a bug
11:24 do i need a new version of yaz?
11:24 tumer version 2.1.19
11:24 same place has it
11:25 kados ok
11:25 hmmm ...
11:25 before I do this
11:25 are there any bugs in 1.3.34?
11:25 we should try to reproduce?
11:25 tumer not that i know of except the one i reported and they produced 1.3.35
11:26 for windows
11:27 you remember the bug we reported at dev-week?
11:27 kados the attribute=3 one?
11:27 (btw: did you recieve the message I forwarded to your hotmail sent by seb?)
11:27 (about that one)
11:27 tumer they corrected it and now i cannot even use 1.4 cause it does not even index
11:28 kados they corrected the attribute=3 one?
11:28 in 1.4?
11:28 tumer no the one about finding less records in 1.4 than in 1.3
11:29 the one adam sent a reply about some rank thing
11:29 kados ahh ... I don't remember that one
11:29 tumer you read it to me
11:29 kados ahh ...
11:29 now I remember
11:29 tumer its on their zebralist
11:29 kados yea
11:30 well ... maybe we should stick with 1.3.34?
11:30 until 1.4 is more stable?
11:30 tumer yep
11:31 1.3.35 for me
11:31 kados right
11:31 paul_away bye bye & see you tomorrow
11:31 kados bye paul_away
11:31 tumer bye paul
11:31 kados so I'm going to go ahead and test edits and adds, etc.
11:31 I've updated cvs on my test box
11:31 to the latest dev_week
11:32 tumer k
11:33 have to go now
11:33 kados heh
11:33 it forwards me to the search screen
11:33 tumer what does?
11:33 kados adding a new record
11:33 no results found
11:33 I assume that means it failed :-)
11:34 tumer let me check dev-week
11:34 kados nothing interesting in the apache log
11:35 doesn't look like zebra saw any activity
11:35 tumer and mysql?
11:35 kados dunno ... nothing in the logs
11:35 tumer by the way zconnauth has a $server in context.pm as well
11:36 kados ahh, maybe that's the prob
11:37 tumer that is the more complex one with user password etc.
11:38 kados hmmm ...
11:39 added $server = "biblioserver" but still nothing
11:39 no errors anywhere
11:39 I'll have to throw some warns in I suppose
11:39 tumer going back to search was strange anyway
11:39 kados so addbiblio.pl right?
11:40 or haven't you modified that at all
11:40 is everything changed in Biblio.pm?
11:40 tumer i did not change addbiblio.pl at all but everything in biblio.pl
11:40 kados ok
11:40 tumer addbiblio.pl is what it was there
11:42 kados could you explain briefly the order of operations?
11:43 is all the zebra stuff in the MARC* routines?
11:44 tumer addbiblio.pl calls for Newaddbiblio which i think calls MARCaddbiblio which adds a marc record to mysql and calls zebraop
11:44 kados hmmm
11:44 i do get this error:
11:44 DBD::mysql::st execute failed: called with 2 bind variables when 0 are needed at /koha/intranet/modules/C4/Search.pm line 687.
11:44 DBD::mysql::st fetchrow failed: fetch() without execute() at /koha/intranet/modules/C4/Search.pm line 735.
11:45 when I add a new item
11:45 new record I mean
11:45 and if I go:
11:45 warn "SERVER:".$server;
11:45 in sub zebraop in Biblio.pm
11:45 it never gets called
11:46 no warn shows up
11:47 tumer i'll have to set up a test version on server and try it i2ll come back
11:47 kados k
11:47 thanks for your work on this tumer

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

koha1