IRC log for #koha, 2007-12-20

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

All times shown according to UTC.

Time Nick Message
13:59 kados lea: hi there
14:00 morning foxnorth
14:01 foxnorth hey kados
14:01 lea hi kados
14:02 kados i have something to run past you if you have a couple of minutes?
14:03 kados lea: sure
14:07 lea Basically, in the summer my boss got the "online" version of the proprietary system we use for the libraries
14:07 it was an asp.net app and to my disgust we had to setup our very first wndows server
14:08 to cut a long story short, the app doesn't work so he's on the lookout for aomething else
14:08 He's willing to fork out _some_ money to get a new system in
14:08 we have 2 problems as i can see:
14:08 1. Migration
14:08 2. Cost
14:09 Now I'm very comfortable in the Opensource worls and my boss is too. he'd way rather we go for koha than anything else
14:09 kados with you so far
14:10 lea during the summer though, i don't think koha had short loans and holidays which was pretty essential for our 3 libraries
14:10 and from my point of view, i couldn't quite get my head around how to set it all up correctly
14:10 simply because I'm not a librarian
14:11 so given that my boss has earmarked £5K for getting it sorted, what are our options? (If you know of any)
14:12 kados akk, phone call
14:12 just a sec
14:17 lea np :)
14:17 I heard an interesting webcast interview between a liblime guy and talis. Made me want koha even more.
14:32 kados lea: that guy was me! :-)
14:33 lea: did you look at the new 3.0 version yet?
14:52 lea oh thought it might have been you ;)
14:52 I haven't checked out V3 yet no. It's in SVN/CVS?
14:53 right now $5K is about $10K. What would that buy?
14:53 er i mean £5k
14:54 i'd *really* like a koha system here
14:54 kados sure
14:54 well, if you're looking for prices for commercial support, I'd suggest contacting one of the koha companies
14:54 http://koha.org/support/pay.html
14:55 this channel is mainly for discussing development and technical issues
14:55 V3 is in Git: http://git.koha.org
14:55 http://wiki.koha.org/doku.php?[…]lopment:git_usage
14:56 lea ok. Not really looking for commercial support, just trying to work out how to get from A to B
14:56 although I guess there might be some migration work for someone if they want it
14:56 kados well, if you don't need commercial support, you won't need 10K :-)
14:56 koha's completely free
14:56 so you can download and implement it yourself
14:57 lea yeah, but we have semi broken marc21 data
14:57 kados most of the companies won't do data migration without some kind of commercial support agreement in place
14:57 lea right ok
14:57 I'd mended most of it myself, it was just the subject tags
14:57 with regards to implementing it myself, I got pretty close.
14:57 kados but if you're a perl hacker you can use MARC::Record on cpan to clean thigns up
14:58 or MarcEdit, also a free utility (runs on windows)
14:58 lea I used a similar python library
14:58 no, it was that our current lib system doesn't export subject tags (250 i think?)
14:58 kados 650 most likely
14:58 lea ah yeah, 650
14:59 kados if you have ISBNs I can imagine writing a script to pull them off of LOC's Z-server ;-)
14:59 lea it uses "keywords" which is stuck in some custom tag somewhere and is full of, well, crap really
14:59 yeah, I even did that!
14:59 but a lot of the ISBNs didn't match
14:59 if i remember right, a lot were 10 digit instead of 13? something like that
15:00 i experimented with the titles and got some hits with different ISBNs
15:01 kados you could wrap it in a call to OCLC's xISBN service
15:01 lea also, back when i was setting up a koha test server, there was no real guide on how to get it going and i found some things a little complicated. Like how it works :)
15:01 kados to get the 13-digit ones too
15:01 lea I've not heard of OCLC
15:01 kados *cough*
15:01 lea I'm in the uk. Is that a good enough excuse?
15:01 :)
15:02 kados hehe
15:02 oclc.org
15:02 lea yeah, found it
15:03 ok, if this channel is for devel/tech, may i pm you about this, as it's not directly related i guess?
15:11 so, I'll try out koha v3. The interface of v2 was mega confusing to me. But then I'm not a librarian i guess
15:16 hdl kados : I sent some bugfix for NZ search on 12.12 But could not find it pushed. did you receive it ?
15:24 fbcit g'morning koha
15:24 atz hello fbcit
15:25 kados hdl: hmmm, not yet
15:25 hdl Did you have problems with it ?
15:25 kados I didn't get it
15:25 maybe chris has it still?
15:57 lea should the README.txt indicate how to install zebra?
16:01 kados hmmm
16:01 lea: what platform are you on?
16:02 lea linux (ubuntu 6.06)
16:02 kados ahh
16:02 lea ?
16:02 kados there was a recent announcement on koha-devel about a new installer for Koha
16:03 that installer should be included in the main git.koha.org repo very soon, next few days most likely
16:03 lea is that different to the one that is in the current repo?
16:04 kados yes
16:04 different in that it actually works :-)
16:05 lea well, i'm never one to complain ;)
16:05 kados hehe
16:05 lea is zebra a perl module? I'd really like to get it installed as "in the next few days" I wont be at work :)
16:06 kados zebra is an application, if you're on ubunto the debian sources for it might work
16:06 indexdata.dk/zebra
16:07 lea ok, is there a "recommended platform". I'm happy to D/L another OS if that'll make it easier
16:07 kados debian etch seems to be the most popular
16:07 and well tested
16:09 lea so, honestly, is it easy to install on etch?
16:09 21 CDs o.O
16:11 kados ahh, use the business card install
16:11 1 CD
16:11 easy, I can't say, easier, yes
16:11 we're working on making it easy :-)
16:14 lea ok got cd1 :)
16:14 I'll attempt to install
16:14 fbcit kados: got a sec?
16:16 gmcharlt: I worked on that script till 2am local...
16:17 gmcharlt fbcit: how'd it go?
16:17 fbcit I cannot get dmake to correctly pass a system call in fix-perl-path.PL to do the mode change
16:18 the call works fine as a one liner via perl -e
16:18 it works fine when you call the script as perl fix-perl-path.PL
16:18 but chokes when called by dmake
16:18 I wonder if it is an escaping issue?
16:19 system qq|attrib -r $pathfile|;
16:20 kados fbcit: do now
16:21 fbcit kados: I ran into a problem with some filenames and subdir names containing white spaces when adapting installer for Win32
16:22 kados really? I wasn't aware of any such files, we shoudl change them if they exist
16:22 fbcit I have changed this so that the filenames/subdirs use a dash instead
16:22 kados yea, good idea
16:22 fbcit I have a patch ready
16:22 kados ok
16:22 fbcit but it might break some current installs
16:22 those that reference these files
16:22 they are gif files in itemtypeimg
16:23 so here comes the patch...
16:24 kados ahh, the images
16:29 gmcharlt fbcit: maybe use Win32::File (and do something to make sure that module is required only on Windows -- perhaps put in a separte script?)
16:59 fbcit: just sent e-mail with example script; it seems Perl functions stat and chmod work OK on my copy of Strawberry running on XP pro
17:02 fbcit gmcharlt: great! maybe my strawberry install is behind...
17:05 gmcharlt fbcit: if I haven't messed anything up in my example, then maybe all fix-perl-path.PL needs to to remove $^O check around the file permissions stuff
17:05 at least until somebody starts work on the VMS port of Koha ;-)
17:06 fbcit gmcharlt: you should have an email w/my current script...
17:07 system call is ~ line 87.
17:12 gmcharlt: I'm stepping out to lunch... brb
17:13 gmcharlt fbcit: ok
17:26 lea bah, by default debian installs postgres. Is that an issue?
17:38 gmcharlt hi lea: it won't hurt; however, pg support is still experimental, so you should install the mysql-server-5.0 package via apt-get or aptitude
17:41 lea done already - thanks :)
17:45 fbcit gmcharlt: sorry, I misunderstood you earlier... you send me an example script... I'll check it out on XP later today/nigh
17:45 s/nigh/night/
17:46 gmcharlt fbcit: ok; if it works OK, then system attrib may not be needed
17:46 fbcit: I took a look at your e-mail
17:46 only reason I can think why the system attrib +/-r isn't working is maybe a different PATH is set in Makefile or used by dmake
17:47 fbcit I did notice that dmake has a per target option to cause dmake to move to a separate memory space prior to executing the target
17:47 gmcharlt and there's some other executable attrib in the revised path
17:47 but if Perl stat/chmod works when you test tonight, hopefully won't need to worry about it
17:47 fbcit right.
17:48 I'm for the easiest out.... after ~6hrs of debug :-(
17:49 gmcharlt understood -- been though that kind of long debug session many times myself
18:43 lea my life is a long debug session
18:43 chris lol
18:53 lea hmm.. I'm following the install guide to the letter and have an issue: the paths in the apache conf are wrong. I've updated document root but I can't seem to find the paths to cgi-bin
18:58 gmcharlt lea: try /usr/lib/cgi-bin/koha/opac for OPAC cgi scripts, /usr/lib/cgi-bin/koha/ for staff interface
18:58 (should be improved when installer changes get merged in)
18:59 lea yeah thanks
18:59 just need to find koha-conf.xml now :)
19:03 gmcharlt /usr/share/koha/etc/koha-conf.xml (yes, things are currently rather scattered :) )
19:06 lea gah, i've lost my template directory now >.<
19:07 yay!
19:07 fbcit kados: try navigating to catalouging and then use the banner search to search for a term you know is not in any title in the db.
19:08 kados k
19:08 o results found
19:08 Biblios in reservoir
19:08 None
19:08 tried a search for 'fizbit'
19:09 fbcit hrmm..
19:09 kados what do you get?
19:09 fbcit Can't use an undefined value as an ARRAY reference at /usr/lib/cgi-bin/kohaclone​1/cataloguing/addbooks.pl line 81.
19:09 kados hrmm..
19:10 lea I'll wait for the installer ;) Thanks all and good evening.
19:10 fbcit cgi-bin/koha/cataloguing/addbooks.pl?q=fizbit
19:10 kados my $total = scalar @$marcresults;
19:10 is line 81
19:11 fbcit not sure what's happening...
19:11 kados could you throw a warn in
19:11 use Data::Dumper;
19:11 my ( $error, $marcresults ) = SimpleSearch($query);
19:11 warn Dumper($marcresults);
19:11 (around line 71)
19:12 gmcharlt fbcit: also warn ($error);
19:13 fbcit: should be undef if you get to line 81, but just in case
19:16 fbcit Dumper says: $VAR1 = undef;
19:20 gmcharlt fbcit: are you running NoZebra?
19:20 fbcit yes... very weird behaviour now...
19:21 NoZebra on this install, yes
19:22 must be something, because once Dumper operates on @$marcresults, the script runs fine and returns a results not found page
19:23 could it be NoZebra that causes @$marcresults to return undef?
19:24 kados that's likely
19:24 paul/hdl or mason should be able to fix tha tup
19:24 fbcit: can you file a bug on that?
19:24 fbcit: and I'd say that's a blocker since it causes a 500 error
19:25 should also be pretty simple to fix
19:25 fbcit right...
19:25 kados :)
19:25 fbcit hehe
19:25 kados in my defense, the searching with zebra is really top notch :-)
19:26 fbcit but somebody has to find the bugs... ;-)
19:26 kados :)
19:27 owen Hi everybody
19:27 kados hey owen
19:27 gmcharlt hi owen
19:27 fbcit hi owen
19:29 kados owen: quick question for ya ... is there any reason to keep all of those images for itemtypes that have spaces in the name?
19:29 owen: can we just use the itemtypecode now for naming them?
19:30 owen We should be able to, although we haven't quite settled on a way to handle item type images, judging from the differences in the templates
19:30 gmcharlt fbcit: I reproduced your bug, btw
19:31 owen It's something that needs an 'audit'
19:31 kados owen: yea, it exists at two levels now, depending on a global syspref
19:31 owen: either at the bib or item level
19:31 yea, it does need an audit
19:48 fbcit hrmm... line 943 of Search.pm says: # IMO this subroutine is pretty messy still --
19:48 owen Yeah, that really should be changed to IMNSHO
19:48 ;)
19:49 fbcit I may have had a better understanding of it @ 2am this morning... :-)
19:51 gmcharlt fbcit: midnight_oil++
19:51 or 2am_oil++
19:52 fbcit actuall after 12am its $time_oil--
20:06 ok
20:06 kados or gmcharlt
20:06 gmcharlt fbict: yep?
20:07 fbcit looks like NZanalyse returns empty results, NZorder operates on empty results and passes all of this back to addbooks.pl which has no provision for handling empty results.
20:07 chris i always read that as New Zealand analyse
20:08 gmcharlt fbcit: looks like it should be return ref to an empty list instead of undef then
20:08 chris yeah
20:08 fbcit k
20:09 gmcharlt also compare with how cataloguing/addbiblio.pl checks results of SimpleSearch
20:14 fbcit gmcharlt: looks to me like addbiblio.pl handles the undef $result by else condition
20:14 shouldn't addbooks.pl take the same approach?
20:14 rather than messing with Search.pm?
20:14 gmcharlt fbcit: that would be safer, I agree
20:15 fbcit where is the code that handles the same case w/Zebra?
20:17 owen Hi tim
20:18 tim Hi owen
20:18 You're probably who can help me.
20:18 Template problem.
20:18 owen Okay
20:18 gmcharlt fbcit: SImpleSearch is (an) entry to both Zebra and NoZebra searches
20:20 tim I'm adding clickable URLs to our OPAC.  Took the code from the CSS template and got it working as good as it does in CSS.
20:20 But it only shows the first URL.
20:21 It's in a TMPL_LOOP, but doesn't seem to loop through the URLs in 856
20:22 Still don't know much about templating to figure what's wrong.
20:23 Looks like CSS is the only template that uses it.
20:23 owen tim, I suspect it's the script's fault, not the template
20:23 Youre talking about opac-detail, right?
20:24 tim Yes.  I guess I should've mentioned that.
20:25 I'm as good with perl as I am with templates, but I'll look into it.
20:25 owen And it's <!-- TMPL_LOOP name="URLS" --> ?
20:25 tim Yup
20:26 fbcit duh...
20:26 s/$results/\@$results/g @ line 206 in Search.pm and things work fine.
20:29 sloppy syntax...
20:35 gmcharlt fbcit: you mean my @$result =  NZorder(NZanalyse($query))->{​'biblioserver'}->{'RECORDS'};
20:35 ?
20:37 fbcit one moment... phone
20:48 gmcharlt: no, I mean return (undef,\@$result);
20:48 $results appears to be an unnamed array.
20:50 I want to pass back a ref to it just as the Zebra version of SimpleSearch does.
20:50 gmcharlt fbcit: NZOrder returns a hashref
20:51 if there are hits, the hashref has 'hits' and a 'biblioserver' key
20:51 'biblioserver' contains hashref that contains 'RECORDS' key, which is an arrayref of biblionumbers
20:52 problem is that if there are no hits, 'hits' key exists but rest doesn't
20:52 so I'm inclined to say that the fix should:
20:52 1. not stack the calls to NZOrder and NZanalyze
20:52 2. store return of NZOrder in a separate var (e.g., $nz_results)
20:52 3. check $nz_results->{'hits'}
20:53 if exists and > 0, set $results = $nz_results->{'biblioserver'}->{'RECORDS'}
20:53 if == 0, set $results = []
20:54 4. ponder that one can occassionally be *too* clever with Perl expressions
20:55 alternatively, change NZOrder so that {'biblioserver'}->{'RECORDS'} always exists in return
20:56 kados someone really needs to clean up the variable names for all the nozebra code
20:57 there's some serious missunderstanding of the terms operator, operand, parameter, index, etc.
20:58 tim Ok.  I could be looking in the wrong direction and sure don't know how to fix it anyway.
20:58 I think I could've found a problem in C4/Search.pm that keeps the URLs from listing.
20:59 At least with our records.
20:59 It looks like it's splitting separate URLs from a string with a pipe '|' between URLs
21:00 Ours each have their own 856u tags
21:00 kados *cough*
21:01 tim: do you have a sub called something like GetMarcUrls ?
21:05 tim Didn't find it.
21:08 kados looks like you've got getMARCnotes and getMARCsubjects
21:08 =head2 my $marcurlsarray = &getMARCurls($dbh,$bibid,$marcflavour);
21:08 :q
21:09 doesn't look like the actual function is anywhere in your installation
21:09 but basically, that's the function you're looking for
21:09 :)
21:10 tim Is it a problem adding it?
21:19 kados shouldn't be
21:19 just need to find it ... should be in the latest rel_2_2 repo
21:19 grep -r getMARCurls *
21:22 fbcit gmcharlt: to further obfuscate... ( $result->{hits} && $result->{hits} > 0 ) ? $result : $result = [];
21:23 fixes the problem
21:23 correctly hopefully
21:24 tim Just get the latest SearchMarc.pm ?
21:24 fbcit btw, I left the stacked calls in place.
21:24 kados tim: I'm not sure if it's a drop in replacement
21:24 tim: but you could try ... just make sure to keep a back-up copy
21:24 tim: of the current one
21:24 gmcharlt fbcit: yep, but perhaps a little too obscure: I suggest at least assign the output of the ? : to a different var, then return it
21:25 i.e., my $search_results = ( foo ? bar : [] );
21:27 fbcit done and done.
21:27 my $search_result = ( $result->{hits} && $result->{hits} > 0 ? $result : [] );
21:27        return (undef,$search_result);
21:27 tim I just tried it on the virtual machine I downloaded.  I think it's 2.2.9 and it hast the getMARCurls and still doesn't work.
21:27 kados hmmm
21:28 tim Just displays the first URL
21:28 kados must not be using getMARCurls then
21:29 fbcit && there goes the patch.
21:29 gmcharlt fbcit: my $search_result = ( $result->{hits} && $result->{hits} > 0 ? $result->{'biblioserver'}->{'RECORD'} : [] ); # perhaps?
21:29 kados tim: sorry I'm not being very helpful :-)
21:30 gmcharlt i.e., does searching on a keyword that *is* present work?
21:30 tim No problem.  I'd like to be half as helpful.
21:31 fbcit gmcharlt: arg!
21:31 gmcharlt fbcit: no, parameter!
21:32 fbcit not quite: Can't coerce array into hash at /usr/share/kohaclone1/C4/Search.pm line 207.
21:34 gmcharlt fbcit: did you change preceding line to be just my $result = NZorder(NZanalyse($query));
21:34 fbcit        my $result = NZorder(NZanalyse($query))->{​'biblioserver'}->{'RECORDS'};
21:34        my $search_result = ( $result->{hits} && $result->{hits} > 0 ? $result : [] );
21:34        return (undef,$search_result);
21:36        my $result = NZorder(NZanalyse($query)); #->{'biblioserver'}->{'RECORDS'};
21:36        my $search_result = ( $result->{hits} && $result->{hits} > 0 ? $result : [] );
21:36        return (undef,$search_result);
21:38 gmcharlt: the search on a valid term breaks badly w/ the latter 3 lines...
21:40 gmcharlt fbcit: try this:
21:40        my $result = NZorder(NZanalyse($query))->{'biblioserver'};
21:40        my $search_result = ($result->{hits} && $result->{hits} > 0) ? $result->{'RECORDS'} : [];
21:40        return (undef,$search_result);
21:40 fbcit it returns titles in the biblio reservoir, but not in the catalog
21:40 gmcharlt my bad re strucure of returned hash
21:42 fbcit no good
21:42 now addbooks.pl complains: Not an ARRAY reference at /usr/lib/cgi-bin/kohaclone​1/cataloguing/addbooks.pl line 81
21:43 gmcharlt hmm -- works for me -- is your addbooks.pl back in sync with HEAD?
21:45 fbcit git diff HEAD shows only Search.pm as changed
21:46        my $result = NZorder(NZanalyse($query))->{'biblioserver'};
21:46        my $search_result = ( $result->{hits} && $result->{hits} > 0 ? $result : [] );
21:46        return (undef,$search_result);
21:46 ahh...
21:46 gmcharlt need ->{'RECORDS'} in the ternay
21:46 ternary, rather
21:47 fbcit right, works now.
21:47 both ways
21:49 gmcharlt just checked other intranet catalogue search and OPAC -- don't seem to have broken them with this change
21:50 NZorder is suffering from a bad case of copy-and-pasteitis
21:53 fbcit gmcharlt: sent a corrected patch off, though its mostly your work.
21:53 gmcharlt fbcit: thanks
21:54 fbcit I'm off. I'll try that script on XP tonight.
21:54 gmcharlt ok
21:56 fbcit-away btw kados, I ordered a metrologic scanner today.
21:57 gmcharlt fbcit-away: fyi, kados has done a merge of the installer tree into main
21:58 fbcit-away great
22:31 kados fbcit-away: cool
22:31 fbcit-away: did you get it from posguys? :-)
23:21 fbcit gmcharlt: have you run fix-perl-path.PL from a command line against blib?
23:28 masonj: did you get a chance to look at the position of the barcodes on labes?
23:28 masonj hiya fbcit
23:29 fbcit s/labes/labels/
23:29 masonj no i didnt , things are pretty busy at the moment, with urgent things
23:30 fbcit np
23:30 masonj yep, will do soon tho
23:31 fbcit I'll look later this week too.
23:31 tnx
23:31 masonj np :)
23:45 gmcharlt fbcit: just tried it -- seem to work ok
23:45 test was on Debian, btw
23:49 fbcit maybe it's my machine...
23:49 try it on Win32 if you can
23:50 I have to go, but if you can try it, let me know what happens.
00:09 masonj hdl, u about  on irc?
00:11 must be late in france now...
05:16 [K] *** part FreeNode!#koha: rangi n=chris@203-118-134-114.netspace.net.nz
05:16 *** join #koha@FreeNode: rangi n=chris@203-118-134-114.netspace.net.nz
07:56 hdl masonj: ?
08:12 masonj hi hdl
08:13 hdl hi
08:13 late for you.
08:13 I sent a patch on 12.12 that fixed your problem.
08:13 Indeed -X was a way to weight values.
08:13 masonj ah, i couldnt quite find it
08:14 hdl But when doing an AND it failed.
08:14 masonj could find your patch...
08:14 hdl I think kados hasnot pushed it yet.
08:14 masonj s/could/could not/
08:14 aaah, right
08:15 hdl maybe i can send it to you directly
08:15 masonj cool hdl, thanks for that
08:15 yeah, yes please
08:16 take a look at bug 1677...
08:17 http://bugs.koha.org/cgi-bin/b[…]w_bug.cgi?id=1677
08:18 hdl my patch strips weights value.
08:18 masonj right :),
08:19 so , a change to rebuild_nozebra.pl not needed?
08:19 hdl no.
08:20 masonj ok, ill try that out

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

koha1