IRC log for #koha, 2006-04-15

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

All times shown according to UTC.

Time	Nick	Message
12:51	thd	kados: I just sent you the bzipped file.  Sending had failed to work this morning because the gzipped file had exceeded my quota for attachment size.
12:51	kados	thd: thanks
12:51		thd: I'll take a look asap
12:53	thd	kados: the useful file content is not very large but the HTML log files contained even unused records :)
12:55	kados	thd: very useful explaination
12:55		thd: in the email
12:55		thd: thanks
12:56	thd	kados: I try to avoid being incomplete in explanation :)
14:27	kados	owen: have you worked at all on the barcodes stuff?
14:28		owen: if so, could you commit what you've got?
14:28		owen: I was gonna work on it a bit but wanted to avoid merge conflicts
14:29	owen	I /did/ work on it.  Then I overwrote it with a CVS update. :(
14:30	kados	hehe
14:30	owen	It was just cleanup, though.  Easy to do again. *sigh*
14:30		Say, while you're here...
14:31		Is there any reason why the AmazonContent system pref shouldn't apply to OPAC and Intranet both?
14:31	kados	owen: no, just haven't merged that stuff again
14:31		owen: I'll hack on the barcodes stuff
14:31		owen: get it working first
14:32		owen: then you can beautify it :-)
14:32	owen	Right now it's inside "unless ($in->{'type'} eq "intranet") {"
14:32		Can I just move the 'AmazonContent => C4::Context->preference("AmazonContent"),' outside of that unless?
14:34		Hmmm... looks like I can.
14:54		kados: why doesn't labels-home.pl belong in /barcodes ?
15:06	kados	owen: all the -home stuff should go in the root dir
15:06		owen: ok, cvs is updated as far as placement and links go
15:07		owen: it's all yours
15:07		owen: I've send a mail to mason with a list of remaining tasks
15:07		owen: having to do with db updating, installer, etc.
15:08		owen: re: amacon change, sure
15:08		owen: not sure why it was wrapped in that to begin with
16:04	osmoze	hello
16:39	slef	hello osmoze
16:40		Are the .nz working today or is it a holiday there now?
16:57	kados	i think it's holiday until tuesday
16:57		monday for us
17:20	chris	yep its a holiday, im around for the next 20 mins or so tho if you need something slef?
17:39	slef	chris: mostly curious. Nothing in partic unless you can make si /mode #koha -t ;-)
17:39		kados: what? The US doesn't celebrate Easter Monday?
17:40	kados	slef: no idea, I don't do holidays :-)
17:41	chris	ahh, ill ask next time i see him slef
17:41		easter and christmas nz pretty much shuts down ... you have to get permission to trade
17:42	kados	heh
17:42	chris	and you have to pay your staff double time
17:42		i think its double time .. might be time and a half
17:42		so only gas stations, movie theatres and restaurants generally
17:42		oh and there is a rogue garden store chain
17:42		that opens and gets fined every year
17:43	kados	hehe
17:44	chris	then we have anzac day next tuesday .. lots of public holidays in april
17:44		(the 25th)
17:44	slef	chris: it's like Eng 20 years ago...
17:45	chris	:-)
17:45		20 years ago .. shops didnt open on sundays
17:45	slef	although in the village I grew up, it closed pretty much every weekend
17:45	chris	and there was no advertising on tv on sundays
17:45		in nz
17:46		i forget when 6 oclock closing ended
17:47		1967
17:47		it got changed to 10 oclock closing
17:47		i dont think there is a set time anymore, there are bars that seem to never be closed :)
17:48	slef	we've only just got deregulated opening times for bars
17:48	chris	does mail come on sunday in the US joshua?
17:48	slef	shops have been 24 hours for a while
17:48	chris	ahh yes i remember seeing that on the news slef
17:48	slef	s/deregulated/derestricted/
17:48	chris	right
17:48	slef	still the police object to pretty much all applications for longer hours
17:49		as far as I can tell
17:49		given most the bars in this town, that's probably correct, though :-/
17:49	chris	:)
17:49		whats the drinking age 20?
17:49	slef	England has binge drinking trouble as we adapt to the idea that alcohol is always available
17:49		18
17:50	chris	ahh same as here then
17:50		yeah, binge drinking is still a problem in nz
17:51		mainly young kids
17:52		so they are proposing to raise the drinking age again .. which i suspect wont do a damn thing except hide the problem again
17:52	slef	and you can legally drink younger than that in some situations which I forget
17:52		personally, I think if you're old enough to get it, you're old enough to drink it... sooner you start, sooner you learn when to stop
17:52		as ever, some people are slower learners than others ;-)
17:52		I just hope the English figure out drinking before we drag the rest of Europe down with us
17:52		I think only the Nordics have a worse reputation... and they had even tighter alcohol restrictions!
17:52	chris	IMO there isnt actually anymore drinking going on by 18,19 year olds, its just they can do it in the open now
17:52	slef	for sure
17:53	chris	i forget where in the UK are u slef?
17:53		i have friends who I went to uni with, who are pharmacists in bristol
18:05		hi kyle
18:05		nice work on the template, very snazzy
18:05	kados	hey kyle
18:09	slef	Someone buy .nz a new link to the world, please.
18:10		at least to .uk
18:10		<slef> http://mjr.towers.org.uk/blog/2006/cyclynn has some pictures of just
18:10		outside town
18:10		<slef> yes, it is really that flat
18:10		<slef> no, it's not all like that here (despite what some morons tell you)
18:10		<slef> http://mjr.towers.org.uk/photos/lynn/lynnwide is the town, as seen from
18:10		across the river
18:10		<chris> i have friends who I went to uni with, who are pharmacists in bristol
18:10		*** The server says: ERROR :Closing Link: 50-30-55-213.adsl.realpoptel.net
18:10	chris	ahh cool
18:11	slef	what did I miss?
18:11	chris	nothing really
18:11		i just congratulated kyle on the template he committed
18:11		and kados said hi, thats about it
18:12	slef	Bristol is about 4hrs train or full-speed driving away
18:12		but I'll probably move there soon
18:12		well, near
18:13	chris	ahh ok
18:14	slef	btw, if you mouseover bits of the image, the titles tell you what's what
18:14		on lynnwide that is
18:14	chris	oohh tricky
18:15		thats cool
18:15	slef	I can't identify everything
18:15		there are buildings you see really clearly from Ferry Square that I don't usually notice
18:15	chris	http://photos.bigballofwax.co.[…]llington.jpg.html
18:16		how many people live there?
18:16	slef	10k in the main town, 40k inside the bypass
18:17	chris	right, so quite small in the scheme of things
18:17	slef	(there are four or so villages inside the bypass, slowing running into the town)
18:17	chris	ahh right
18:17	slef	yes and no
18:18		I think it's a very small town, but it's the biggest place for 35 miles by land
18:18	chris	right, its reasonably sized compared to a lot of places in nz too
18:19	slef	most of the stuff until those places to the south and west looks like the pictures from the bridge
18:19		so it has more shops and stuff than a town of 10k usually would
18:19	Genji	hello all.
18:19	slef	and more factories... it's just generally a bit wrong
18:20		hello Genji
18:21		how many in Wellington?
18:21	chris	hmm good question
18:21		3 are 4 cities .. wellington itself, porirua, hutt city, and upper hutt ... i think combined its around 600k
18:22		im not sure how many in wellington proper
18:22	slef	look of it reminds me of Toronto somehow
18:22		apart from the wooded hill, which reminds me of Worlebury (small, don't ask)
18:22	chris	heh
18:22	Genji	okay.. im going to be doing chores but can someone please inform me of the current bugs and feature requests on Koha? want to get back into it.
18:24	chris	ah ha
18:24		2001 census 163k in wellington city
18:24		423k in wellington region
18:25	slef	is the region much bigger?
18:25	chris	hmmm, kados might be the best one to answer that genji
18:26		well it includes those other 3 cities .. but from the city of the centre to the outlying cities, furthest one would be 30 mins in a car
18:26		centre of the city even :)
18:27	Genji	when is kados most active?
18:27	chris	http://maps.google.com/maps?f=[…]0.282259,0.839767
18:28		wellington city is at teh southern end of the harbour
18:28		across the harbour to the northeast is huttcity, and upper hutt
18:28	slef	http://tx.mb21.co.uk/gallery/worlebury-hill.asp but it only has a picture of Crook Peak on the other side of the valley (picture with the Hutton label)
18:28	chris	and north is porirua
18:29	slef	heh, maps.google.com is a night view here
18:29		biiig black rectangle
18:29	chris	heh
18:29		thats not that helpful :)
18:29	slef	http://uk2.multimap.com/client[…]000&scale=4000000
18:30	chris	genji: might be a good idea to drop a mail to koha-devel that way lots of eyes willl see it
18:30	slef	next large place is peterborough, 33 miles WSW
18:30	chris	ahh
18:31	slef	next political capital upwards is Norwich, 45 miles E
18:31	chris	right, i have a much better idea now
18:31	slef	forgot I had that map bookmarked
18:31	chris	im not sure why, but i had it my head you were way more to the west
18:32	Genji	we dont have a bugzilla anymore?
18:32	chris	yep we do
18:32		bugs.koha.org
18:32	slef	.uk sites don't like concave coastlines - I search for cinema showings and get results for "Skegness, 22 miles North"
18:32	chris	anything there you are welcome to have a go at
18:32	slef	(~70 miles by road)
18:32	chris	heh
18:33	slef	chris: I'm from near Northampton and spend some time near Bristol.
18:33		went to uni in Norwich and got distracted on the way back
18:34	chris	hehe
18:34	slef	that was, what, 12 years ago now
18:38	Genji	what libs use NPL templates?
18:38	chris	ok im gonna go out for a bit and enjoy some sun
18:39		probably be back later
18:39		NPL do genji .. and i think liblime's clients use variants of it but im not sure
19:04	Genji	how is it that all bugs are assigned to people, even though its status is NEW?
19:14	slef	bugzilla assigns them automatically
19:14		doesn't mean much unless they accept them
19:22	kados	thd: you around?
19:24	thd	kados: yes I am sending you some media type code
19:24	kados	thd: the script's 90% done
19:24		thd: already got the media type codes
19:24		thd: http://www.itsmarc.com/crs/Bib0443.htm
19:24		thd: unless they go above and beyone that list
19:24		thd: the script works like a charm
19:25		thd: just have to add a few more things and I'll be done
19:25		thd: but I can actually insert the files as they are so you can look at them
19:27	thd	kados: I have more media type code than you could determine from that page.
19:27		kados: you may be missing something for books.
19:28		kados: I will send my code in two minutes.
19:28	kados	thd: do they all rely on the leader?
19:29	thd	kados: my code starts with the leader but you have to use both leader positions 06 and 07.
19:30		kados: media type quickly becomes complex after leader but leader code is sufficient for what we need today.
19:30	kados	k
19:31		thd: leader position6 only has two values, right?
19:32	thd	kados: more than two
19:38		kados: you should have the message now
19:41	Genji	kados: media type codes? like my idea i implemented last year?
19:42	thd	Genji: what was your idea?
19:43	Genji	hmm... checking if its still in the cvs...
19:44		the root of koha module has been trimmed extremely....
19:45		okay...
19:46		ah... maybe mediatype is different.. but my implementation, which is in cvs, is mediatype -> itemtype-> itemsubtype
19:47	thd	kados: obviously parts of my code are missing something as I stopped part way through but those issues can be easily fixed.  Some variables certainly need a larger scope or some different treatment.
19:48		Genji: how did you determine media type in your code?
19:50		Genji: kados and I were discussing reading some media type information from the leader, etc. for copy catalogued records.
19:50	Genji	new tables... media type table, itemtype table, itemsubtype table... one linked to the other. for instance... CNFB means Children nonfiction books... ahh.. copying cataloged records.. right. haven't got that far.
19:52	kados	thd:  remember that error you were having when attempting to import record #11?
19:53		thd: these records have some major problems with encoding
19:53		thd: mostlikely because of the way they were saved
19:53		thd: did you write binmode utf8 when you wrote them to file?
19:53		thd: (did php or perl write them to file?)
19:54		thd: (and were they downloaded in raw format?)
19:54	thd	Genji: the media type is a very amorphous term.  MARC generally uses media types contained in standard cataloguing rules but spreads the information all over the record while the leader contains the most important information in positions 06 and 07.
19:54	kados	the script i wrote dies with:
19:54		utf8 "\xEC" does not map to Unicode at /usr/lib/perl/5.8/Encode.pm line 164, <INFILE> line 171.
19:54		on record #11
19:55	thd	kados: they were raw and I mostly had no problems but I did nothing to encode the raw data.  The raw data was the raw data except for display.
19:56		kados: Is record 11 Unicode and not MARC 8?
19:57	kados	no
19:57		NUMBER 10 =>
19:57		LDR 00450cam  2200157   4500
19:57		NUMBER 11 =>
19:57		LDR 01467cam  2200361 i 4500
19:57		at least it claims to be marc-8
19:57	thd	kados: look at 11.html .  Do you see anything wrong?
19:57	kados	however, it's perl that's complaining in this case
19:58		thd: no, not sure what that would tell me anyway
19:58		thd: it's the encoding of a char that's the prob
19:59	thd	kados: why does Perl want to complain about characters that should all be ASCII.
19:59		?
20:00		kados: what is the offending character?
20:00	Genji	offline for chores
20:01	thd	kados: These records should not have encoding issues except for maybe some native Alaskan language characters which I had not noticed in any records.
20:03		kados: does YAZ itself have some encoding bugs?
20:12		kados: PHP wrote to the raw records but did not alter their content except for later in the code where htmlspecialchars() is used for encoding to post the record in a form for manual record saving only after LWP directed automated saving has already happened.  The htmlspecialchars() encoding is removed at the time of manual saving after parsing the post information.
20:13		kados: I specifically avoided passing the raw record over LWP to avoid any possible encoding problems.
20:15		kados: raw is as raw as YAZ provided.  If there are encoding problems those existed in the original record.  Would Columbia University Library really inflict encoding problems on an unsuspecting world? :)
20:27	kados	thd: i found a workaround finally
20:27		thd: 479 records in the file?
20:28	thd	kados: what was the problem?
20:28	kados	thd: bad encoding
20:29	thd	kados: where was the bad encoding?
20:30	kados	thd: in several of the records
20:30		thd: it's impossible to know where
20:30	thd	kados: I think there should be 456 records in the file.
20:30	kados	I have 479 :-)
20:30		importing them into Koha now
20:31		http://library.afognak.org/
20:31	thd	kados: ok more is better, maybe my script miscounted.
20:31	kados	hmmm, some have 0 count
20:31		must be a flaw in my script
20:32		and we've got major encoding probs
20:32		http://library.afognak.org/cgi[…]-detail.pl?bib=12
20:32		for instance
20:33	thd	kados: that is very pretty, are they all like that?
20:34	kados	maybe we only have a few encoding probs
20:34		some of the records didn't conver to to utf-8
20:34		12 EncodingMARC-8
20:34		13 EncodingMARC-8
20:34		128 EncodingMARC-8
20:34		175 EncodingMARC-8
20:34		219 EncodingMARC-8
20:34		299 EncodingMARC-8
20:34		302 EncodingMARC-8
20:34		326 EncodingMARC-8
20:34		330 EncodingMARC-8
20:34		331 EncodingMARC-8
20:34		332 EncodingMARC-8
20:34		333 EncodingMARC-8
20:34		334 EncodingMARC-8
20:34		393 EncodingMARC-8
20:34		407 EncodingMARC-8
20:34		15 to be exact
20:34		I bet everything but those 15 is ok
20:35	thd	kados: what happened to record 11?
20:36		which would not have been 11.html because there were gaps.
20:40		kados: look at the bad record more importantly entirely different record in the MARC view from the detail view for http://library.afognak.org/cgi[…]-detail.pl?bib=12
20:42		kados: the MARC record looks like the bad title only match for "Am salmon"
20:46	kados	thd: explain to me how the records are saved in the file with the tabs
20:46	thd	or rather author title which are also suspect matches
20:47	kados	thd: I think the problem with this whole thing is that you saved binary marc files with all different encodings into a file without specifying or controling the encoding
20:47		thd: so that file has mixed encoding
20:49		nope, I'm wrong about even that
20:49	thd	if (!empty($save_best) && !empty($recQuality) && $bestRec == 1) {
20:49		$fh = fopen($marcCaptureFile, 'a') or die("can't open file");
20:49		fwrite($fh, $rawRec);
20:49		fclose($fh);
20:49		$extraValues[] = "";
20:49		$extraValues = array($progress, $version,
20:49		$recQuality, $recYearMatch, $hostSpec,
20:49		$search_fields, $isbn, $author, $title,
20:49		$publication, $pub_place, $pub_year,
20:49		$l_subject, $l_subject_subdivision,
20:49		$purchased_price, $quantity, $rawRec);
20:49		$extraValuesRow = implode("\t", $extraValues) . "\n";
20:49		$fh = fopen($marcExtraValuesFile, 'a') or die("can't open file");
20:49		fwrite($fh, $extraValuesRow);
20:49		fclose($fh);
20:49		}
20:49	kados	thd: is this perl?
20:50	thd	kados: no Perl uses join for join not implode :)
20:53	kados	thd: so this is php?
20:53	thd	kados: I assume that you would have the same problems if you tried importing the MARC records from the MARC records only file.
20:53	kados	thd: I have even more probs
20:53	thd	yes PHP :(
20:53	kados	hmmm
20:54		I don't know what to say
20:54		hmmm
20:54	thd	kados: what are the more problems?
20:55	kados	how can we salvage this?
20:55	thd	kados: I can send you the scripts and we can rerun everything form your speedy system after we find the problem.
20:55	kados	hmmm
20:56	thd	\kados: what is the problem except for 15 records?
20:56	kados	I don't even know where to begin looking for the problem
20:56	thd	kados: but what problem do you actually see aside from 15 records?
20:57	kados	well, I can't seem to convert the marc-8 to utf-8
20:57		because of the encoding probs
20:57		so if I leave everything as marc-8
20:57		I can import all but 3 or so
20:58		and at least 15 records appear mangled
20:58	thd	kados: I was uncertain about how I added the newline separating the rows but do you see a problem there for . "\n"
20:58		?
20:58	kados	yes that is also a problem
20:58		but I was able to correct that
20:58		in a true marc file
20:59		you wouldn't have \n as the last char
20:59		I was able to chomp() that line to remove it
21:00	thd	kados: yes of course but was there any strange character just before the newline?
21:01	kados	thd: well, there is the end of file that every marc has
21:02	thd	kados: as long as that is a well formed end of file then that is fine.
21:04		kados: what happens if you try to import the records from the MARC only file?
21:04		kados: What workaround did you use for the 11th record?
21:04	kados	thd: it dies on record 171 if I try to import from marc only
21:05	thd	kados: so in Perl you made some character conversion?
21:12	kados	thd: first I tried doing nothing with encoding
21:12		then, I tried converting everything to utf-89
21:12		utf-8 even
21:12		basically the real problem
21:13		is that the records weren't saved correctly
21:13		they must have been re-encoded by php or something
21:14		were they downloaded directly in binary marc or were they scraped off an html page?
21:14	thd	kados: there were direct raw MARC so that we would not have this problem?
21:15	kados	hmmm
21:15		very strange
21:15	thd	s/there/they/
21:15	kados	so how did you acomplish the download?
21:15		sometimes if you don't specify a binary transfer it doesn't do a binary transfer
21:15		is that a possible cause of the problem?
21:18	thd	kados: $rawRec = yaz_record($id[$i],$p,"raw");
21:18		kados: raw is raw
21:18	kados	huh
21:19		then presumably you write that $rawRec to a filehandle?
21:19		huh ... what version of yaz are you running on that box?
21:20	thd	yes, written directly to filehandle
21:24		kados: I have YAZ version 2.1.8-4
21:28		kados: I think I built PHP/YAZ for PHP5 because the Debian package is only for PHP4, and rather old.
21:30		kados: Do you want to try running against the sample 29 records on your system?
21:33		s/running/running the LWP and PHP script/
21:36		kados: are all records affected or only 15?
21:58		kados: the content is ASCII the conversion presumably does nothing but change the indicated encoding in the leader
21:58		s/the/if the/
21:59	kados	ok, so whether or not I convert to utf-8 it crashes on number 171
21:59		I was mistaken that they all imported if they were not re-encoded
21:59	thd	kados: that is the 171st record imported?
21:59	kados	yep
21:59		well, give or take one :-)
21:59		my counting is notoriously off by one :-)
22:00	thd	:)
22:00	kados	thd: I'm going to try eliminating all problematic records from the write
22:01	thd	kados:that sound like an excellent plan
22:03		at least as long as the remaining records are at least one
22:15	kados	I'm going to just have to manually skip the problem records
22:34	thd	kados: I do a character conversion for display from a separate variable.  Maybe a PHP bug creates an upstream problem.  $rec is used for display only.                  $rec = yaz_record($id[$i],$p,"render; charset=marc8,utf-8");
22:34	kados	hmmm
22:34		it might set the charset in yaz-record
22:34		or something
22:36	thd	kados: maybe I should be setting the value of $rawRec first in  $rawRec = yaz_record($id[$i],$p,"raw");
22:36		kados: order that the variables were set had not seemed important
22:37		and it should not be important :)
22:37	kados	agreed
22:37		but it might be
22:40		wow, this is strange
22:40	thd	kados: what is strange?
22:40	kados	in this case, it dies on record 17 even if I delete records 10-25
22:41	thd	kados: delete records 10-25 again
22:42	kados	i did
22:42	thd	and?
22:42	kados	same error
22:42	thd	did it die on 17 again?
22:42	kados	yep
22:42	thd	kados: that is a loop error unless 17 is an unlucky number
22:44		kados: koha has many loop errors in the templates at least.
23:39		kados: after moving the setting of $rawRec for saving before $rec for display, I still have the XML/parser error on the 11th record.
23:50	kados	thd: got 398 records in
23:50		thd: http://library.afognak.org
23:51		thd: a search on 'and' pulls up 298 of them
23:51		thd: seems like quite a few duplicates in there
23:55		and I don't see any with more than one copy
23:58	Genji	kados: you awake?
23:58	kados	Genji: barely :-)
23:59	Genji	kados: good enuf. Im looking at doing some koha devel... have any particular bug / feature enhancement you think I could tackle on my first day back?
23:59	kados	great news!
00:00		know much about encoding?
00:00	Genji	hmm.... not really. what sort of encoding?
00:00	kados	we're currently really hurting in the encoding area
00:00		character encoding
00:00	Genji	utf8?
00:00	kados	iso-8859-1 vs utf-8 vs marc-8, etc.
00:00		we need Koha to be able to handle any encoding we hand it
00:01		well ...
00:01		there's other stuff
00:01		that's just first on my mind
00:01		have you tried out the new zebra plugin?
00:01		there's plenty of work to do on zebra
00:01		in head
00:02		hmmm
00:02		not quite
00:02		we have that already
00:02		in the form of MARC::File::XML
00:03		I've been struggling with a bug
00:03		related to encoding
00:03		in rel_2_2
00:03		if you use a Koha that has iso-8859 encoding in the db
00:03		and you upgrade to rel_2_2
00:04		when you edit existing records
00:04		the special characters get mangled
00:04		I think it's perl's fault
00:04		but I haven't been successful in tracking down exactly where it's happening
00:06	Genji	special characters like accented e's etc?
00:06	kados	yep
00:07	Genji	so... perl is converting the characters before the script has the chance to convert them?
00:08	kados	well, it's complicated
00:08		if you use CVS rel_2_2
00:08		and you updatedatabase
00:08		there's a new syspref
00:08		TemplateEncoding
00:08		set that to the desired encoding
00:08		so if you're running an old Koha
00:08		it should probably be
00:09		iso-8859
00:09		(if it's unimarc that is)
00:09		(in marc21 there are only two valid encodings: marc8 and utf8)
00:09		so ... once you've done that
00:09		look in addbiblio.pl
00:10		ahh ... before that
00:10		you need to upgrade MARC::Record, MARC::File::XML and MARC::Charset
00:10		to the latest sourceforge versions
00:10		the CPAN versions won't cut it
00:10		hehe
00:10		k
00:10		cool
00:13	si	happy easter, joshua
00:13	kados	si: can you oper me?
00:13		si: you too :-)
00:13		go /set +o kados
00:14	si	or indeed
00:14		/mode #koha +o kados
00:14	kados	woot
00:14		thx
00:14	si	no worries
00:15	kados	hehe
00:19	thd	kados: the source file had many duplicates
00:21	kados	thd: right ...
00:21		thd: multiple items are now working
00:21		http://library.afognak.org/cgi[…]-detail.pl?bib=26
00:21	thd	kados: sometimes the same record appears in the source both with and without an ISBN
00:26	si	russ, you might find this useful
00:26		http://video.google.com/videop[…]imoncelli&pl=true
00:27		ack, wrong #
00:27	thd	kados: from that record it seems that you created 690 subfields even when the values for the 690 subfields were empty
00:32	kados	thd: fixing that now
00:33		thd: fixed
00:39	thd	kados: there is a similar issue for cost, purchase price.
00:39	kados	yep, investigating now
00:56	thd	kados: that record is actually a bad match.  OCLC does not have what the record ought to be.  That goes with the unfindable shareholder guides to a native corporation.
01:02		kados: I suppose that any characters in MARC 8 would necessarily exist in Unicode.  I assume MARC::Charset includes every native american language in MARC 8.
01:48	Genji	ok.. back again... on and off doing dishes....
02:44		hiya all
02:47	pierrick	hi Genji
02:53	Genji	pierrick: whats your pet bug/idea for koha?
02:57	pierrick	Genji, sorry, I'm not native englush speaker :-/ what do you mean "pet bug/idea"?
02:57		s/englush/english
03:00	paul	hello all
03:00	pierrick	hello Paul
03:04	ToinS	hello
04:03	thd	pierrick: It seems that Genji missed answering your English question.  Pet something is favourite something, usually a personal favourite.
04:07		pierrick: I think Genji was trying to ask what special idea were you interested in pursuing or implementing in Koha.
04:08		pierrick: a bug in this context is something that keeps motivating you to pursue something.
04:09		pierrick: are you there?
04:14	pierrick	thd, I'm back
04:14		thd, thank you for precision
04:15	thd	pierrick: So what are you specially interested in pursing in or for Koha?
04:15	paul	thd : a quick english question
04:15	thd	yes paul
04:15	pierrick	Genji, I have no "pet idea/bug" for the moment. Maybe I'm interested in tagging biblio from users and presenting something like a tag cloud and related tags
04:15	paul	is "collectivity" a correct word so speak of an professional organisation
04:16		(an institution, a company...)
04:16		because i'm afraid it's a frenchism
04:16		"collectivité"
04:16	thd	paul: Yes that is perfect French:)
04:17		paul: try organisation
04:18		paul: that would be the most interchangeable general term for institution, company, etc.
04:18	paul	pierrick: 1st bug squashing meeting => 18th, april, but which time ?
04:20		(i should be here, with my tank, 2 fighters, 1 cruiser and at least 5 companies. I'll also get my +5 sword of the paladin and my shield of the deadbug)
04:21	pierrick	paul, I've written a specific mail for BSP
04:22	paul	ok, I missed it
04:22	thd	paul: you need some special potions to keep the bugs from returning from the dead
04:23		pierrick: what is tagging biblios from users?
04:24	paul	tagging biblios is a very interesting idea, if I understand what it is ;-)
04:25	thd	paul: if you understand what is it?
04:25	paul	yes, I think I know what pierrick  is speaking of, and i'm waiting for it's explanations.
04:25		but if i'm right, it's an interesting idea ;-)
04:25	Genji	whats tagging biblios?
04:26	thd	paul: I hope it is even an interesting idea if you are wrong :)
04:26	pierrick	http://lists.nongnu.org/archiv[…]-10/msg00011.html
04:27		I wanted to see if you already had the idea before me, and yes. But it does not seem to have been implemented
04:28	thd	pierrick: How would you describe tagging biblios from users?  I know hat the individual words mean but not the concept that you are attempting to describe.
04:28	pierrick	"Tagging" is a very common features nowadays in applications managing items (any kind of items).
04:29	Genji	hmmm.. users tag.. as in add subjects to their own books?
04:29	thd	pierrick: I have had a few interesting ideas before you were born :)
04:29	pierrick	thd, an example: you're connected to the OPAC, you can add some tag to a biblio
04:30	thd	pierrick: you mean user added notes fields.
04:30	pierrick	thd, yes some kind of
04:31		but maybe having only "tagging" and not "user tagging" would be enough
04:31		The idea is the new navigation way it creates
04:32	thd	pierrick: well the cataloguer can already add notes if that is what you mean by tagging.
04:32	pierrick	thd, why not using it
04:33		what I mean is having this kind of navigation : http://www.modusoptimus.com/bsf/tags.php
04:33	thd	pierrick: what do you mean by why not using it?
04:33	pierrick	I mean I don't really mind were tags come from
04:33		from users, from librarians...
04:34	thd	pierrick: yes records need as many access points and as much content as can be provided.
04:34	pierrick	the origin is a question, but IMO the most important is the navigation
04:35		if users can participate, it can be interesting but not mandatory
04:35	thd	pierrick: such tags seem like user added subject headings.
04:35	pierrick	thd, yes. The admin of the gallery has added the tags manually
04:35		(through metadata in reality)
04:37		having a navigation mode based on chronology would be interesting too
04:38		I think the current OPAC is only a search form while we could provide other navigation modes
04:38	thd	pierrick: I had asked hdl some time ago why that was not already a feature extension to the virtual bookshelves.
04:38	paul	pierrick: the best would be to have a world-wide tag system, based on ISBN
04:38	pierrick	search mode, tag mod, category mode, chronology mode, best rated, most read, etc.
04:39	paul	when a library from Koha network get a tag, it is send to a central server, that can distribute it to all libraries during the night.
04:39	thd	pierrick: that is a very important concept.  Blank search forms are a very limited concept.
04:39	paul	OK guys, I've just commited many many things for borrowers improvements.
04:39	pierrick	paul, that would mean the set of avilable tags is centralized
04:40	paul	it works correctly for me, I think i've commited everything. Could someone check that an update of CVS + updater/updatedatabase make the feature working ?
04:40		you can check what it does at :
04:41		http://i20.bureau.paulpoulain.[…]s/members-home.pl
04:41		(login test/test)
04:41	pierrick	thd, my experience is 5 years of photo gallery development, and IMO managing biblio is not that different from managing photographs
04:41	paul	pierrick: why do you want an available tag list ?
04:41	thd	pierrick: It would not need to be centralised if every library was free to add individual fields and subfields from records at any other library automatically in a distributed network.
04:41	pierrick	thd, in PWG there are several mode of navigation, not only the search one.
04:42	paul	I thought such system where without any available tag list, and the user could enter what he wants.
04:42	thd	pierrick: what is PWG?
04:42	pierrick	thd, PhpWebGallery
04:43		paul, don't you think it would become a real mess if all libraries share their tags?
04:43	paul	in fact, i'm not sure there would be so many ppl entering tags. so having them worldwide could improve a lot their interest
04:43		oups, I must leave NOW
04:44	thd	paul: well one could have both a standard tag thesaurus and free form tagging.  One need not exclude the other.
04:45		pierrick: the user can have filters to protect himself from the mess potential.
04:46	pierrick	thd, of course, it could mandatory that a librarian validate tags before making them public
04:47	thd	pierrick: I favour standard thesauri but I also like the idea of giving users freedom to contribute to the library in any way that is comfortable to the user.
04:48	pierrick	I know an online service blogmarks.net that let each user have public and private tags on their bookmarks
04:49	thd	pierrick: yes there will always be the issue of the librarian needing to protect the library institution form users who may see public tagging as a forum for causing mischief :)
04:49		s./form/from/
04:51		pierrick: I would like to see such a feature implemented in a MARC compliant manner even if it would inevitably lead to breaking the ISO 2709 record size barrier requiring XML.
04:51		pierrick: How is your study of MARC going?
04:52	pierrick	after this discussion, I would answer Genji that I would like to have other navigation mode in the OPAC, not only the search mode
04:53		thd, not very far. I think I understand the tag/subfields structure. I don't bind description to each tag/subfield depending on the MARC flavour
04:55	thd	pierrick: full marks!!  see my brief paragraphs about the alternative to the search paridym at http://www.agogme.com
04:56		s/paradym/paradym/
04:57		pierrick: you should look at a library science textbook to understand the concepts behind what goes into MARC well.
05:10		pierrick: There is a generally well respected book by Chan, although, I unfortunately do not have a copy.
05:12		pierrick: http://www.amazon.com/gp/product/0070105065
05:14		pierrick: I do not know what might be a French equivalent, but would be pleased to know
05:16	pierrick	thanks a lot thd, I don't need a translation, the majority of my technical books are in english
05:24	thd	pierrick: the Chan book is most likely excellent but it may be helpful to consult also a book on French UNIMARC practise.
05:24		pierrick: http://www.amazon.fr/exec/obidos/ASIN/2765405514
05:29	slef	;-)
05:29	thd	pierrick: and volume 2 http://www.amazon.fr/exec/obidos/ASIN/2765408246
05:31		what is this error:  Can't locate object method "as_xml" via package "MARC::Record" ?
05:34	pierrick	slef, what does "puts the boot" means?
05:36	thd	pierrick: Savannah hardware is not fast enough for me to see exactly what slef means yet.
05:37		pierrick: the term may generally mean to impolitely kick another in the trousers for obtaining attention severely.
05:38	slef	thd: it's obtaining possession of the football by kicking the player in the back of the leg
05:39	pierrick	slef, you answered to my mail about forum Vs mailing-list?
05:39	slef	yep
05:40	pierrick	didn't received yet
05:40	thd	slef: I understand what that would mean but that leaves me even more uncertain of your message.
05:40	pierrick	but I had hoped you wouldn't answer
05:42	slef	thd: je rigole
05:42	thd	pierrick: FSF is buying new servers for the mail system and improving the routing so the mail queue may not be an endless disc thrashing session in a few months.  They know the delay of the mail queue is a very important issue.
05:44	slef	http://www.fsf.org/blogs/sysadmin/lists
05:48	thd	slef: well yes that was a point of significant discussion and a special presentation at the FSF members meeting.
05:52		slef: at least the FSF mail system is not any worse than I have experienced on Sourceforge.  All Koha devel messages are at least appearing in the log without loss eventually :)
06:34	slef	thd++
07:50	pierrick	kados, are you around?
08:14	kados	pierrick: am now
08:15	pierrick	I'm testing the zebra plugin (still installing it)
08:15	kados	cool
08:15	pierrick	mis/missing090field.pl is not in rel_2_2 but in HEAD
08:15		(and it should not be, I suppose)
08:16	kados	hmmm
08:16		right
08:16		could you commit it to rel_2_2?
08:16	pierrick	I you want
08:16		if you want
08:16	kados	yep
08:18	pierrick	done
08:18	kados	thx
08:25		paul_away: are you here?
08:25	pierrick	pierrickplegall:~/dev/koha/head/misc/zebra$ wc -l  ./unimarc/zebra.cfg ./usmarc/zebra.cfg
08:25		31 ./unimarc/zebra.cfg
08:25		65 ./usmarc/zebra.cfg
08:26		should I suppose unimarc zebra configuration file is not up to date at all?
08:26	kados	pierrick: yep :-)
08:29	pierrick	are there difference between the two files?
08:29	kados	well, yes
08:29	pierrick	(should there be differences?)
08:29	kados	I think so
08:29		lemme look quickly
08:29	pierrick	:-/
08:30	kados	pierrick: no they can be the same
08:30	pierrick	OK
08:30		thx
08:31		kados, you created a "kohaplugin" user on your system?
08:34	thd	kados: what causes this error from bulkmarcimport.pl on any MARC record:   Can't locate object method "as_xml" via package "MARC::Record" ?
08:37		kados: I last updated MARC::XML a few weeks ago but the current CPAN version has a make test error.
08:38		kados: rel_2_2 is now so broken for me that I cannot import even one record :(
08:38	kados	pierrick: no, no kohaplugin user
08:38		thd: you must use the sourceforge version of MARC::File::XML
08:39		thd: the cpan version has a make test error?
08:39	thd	kados: yes I was using the CPAN version
08:40	kados	thd: did you install MARC::Charset?
08:40		thd: check kohadocs.org for instructions on installing the latest sourceforge versions
08:40		thd: my 'installing on debian' document has details
08:40	thd	kados: MARC::Charset and MARC::Record are up to date
08:41	kados	thd: MARC::Record needs to be installed from Sourceforge
08:41		thd: as with MARC::Charset
08:41		thd: the CPAN versions aren't up to date
08:46	thd	kados: I have succeeded slowly in capturing more MARC records by gradually adding targets and searching the correct form of serial titles.  I am a little less than half way through the 615 records.
08:46		s/615/no hits from 615/
08:48	kados	thd: do those MARC records have the same encoding probs as the first batch?
08:48		(out of curiosity)
08:49	thd	kados: I think the only encoding problems were some native Alaskan names.
08:50		kados: I cannot import any record at the moment so I cannot determine encoding problems even for my old ASCII only records :0
08:51		kados: Why is CPAN behind?
08:51	kados	thd: CPAN: it's a long story
08:51		thd: involving some people at follett
08:51	thd	kados: I thought that you were updating CPAN to avoid this problem
08:52	kados	thd: yes, I do have access to MARC::File::XML finally
08:52		thd: but sourceforge version has some untested functions that need to be tested before being put in CPAN
08:52		hey slef
08:53		slef: you've dealt with encoding issues in perl, right?
08:53		slef: utf8 "\xEB" does not map to Unicode at ./afognak2koha.pl line 25, <INFILE> line 1.
08:53		slef: is there any way to get Encode to just warn rather than die on that error?
08:54	thd	kados: what about the possibility that the MARC::Charset mapping or Unicode is incomplete for native Alaskan languages?
08:54	kados	it's possible
08:55		thd: but line 1 is : InDesign CS2 for dummies
08:55	slef	what's afognak2koha.pl line 25 ?
08:55	kados	slef: while (my $line = <INFILE>) {
08:56		where INFILE is:
08:56		open(INFILE, "<:utf8",$infile);
08:56	thd	kados: well if that is the record how could there be any non-ASCII characters?
08:56	kados	thd: I have no idea, there must be something going on somewhere in that yaz or php stuff
08:58	slef	kados: why are you trying to open a non-utf8 $infile with :utf8?
08:58	kados	slef: either way I get the error
08:58		slef: the reason is because there are some wide chars in the file
08:59	slef	what encoding is $infile?
08:59	kados	slef: it's one of the problems with a batch file of marc records - they can lie about their encoding ... claiming to be marc-8 but actually some other encoding
08:59		slef: it seems to be a combo of encodings :-)
08:59		slef: but I'm assuming either mostly 8859 or marc-8
09:00		slef: with a few hundred wide chars thrown in
09:00	slef	8859-which? Can file figure it out?
09:00	kados	hmmm ... not sure how to do that
09:01	slef	file yourbatchfile # on the command line
09:01	kados	k
09:01		file ../alaska_mrc.mrc #
09:01		../alaska_mrc.mrc: data
09:01	slef	you can try <:raw instead, but you may end up outputting gibberish if you can't fix the encoding
09:02	kados	I'll try that
09:02	slef	Have you read man perlunicode?
09:02	kados	still dies on the first record
09:02		yea, but I could probably do a re-read
09:04	slef	hrm, I guess you get to play "guess the encoding" if this is a one-off
09:06	kados	the thing that gets my goat is that it just dies
09:06		I'd be fine with it just warning and mangling a single character in that record
09:07	thd	kados: maybe the issue for that record could be an em dash if "440 0 $a --For dummies", had an em dash.  However it seems to have two hyphens.
09:08		kados: That is the second captured record.  The second captured record had no problem for me.
09:09	slef	kados: well, you're asserting that it's utf8 when it isn't. Could you just open() it and read it in and *then* test it?
09:09	kados	slef: whether I open as utf-8 or not it dies with the same error
09:10		slef: I just tried 'raw' with the same results
09:10	slef	kados: sounds like something is ignoring you. Can you publish script and test data?
09:10	kados	yep
09:11	slef	I'll take a look on the ramfs here
09:12	kados	http://kados.org/afognak2koha.pl
09:12		http://kados.org/alaska_mrc_extra_val.csv
09:13	thd	s/second captured/first captured/
09:13	kados	slef: I typically run it like this:
09:13		./afognak2koha.pl alaska_mrc_extra_val.csv all.mrc alldump.txt
09:13	thd	kados: you did not say last night that you had problems with the first captured record.
09:13	kados	where the 'all.mrc' is the output
09:13		thd: sorry, it's not the first record
09:14	thd	kados: which record is it for you?
09:14	kados	thd: it's number 170 or something
09:14		171 actually
09:14		in this version of the script I've got the whole operation wrapped in 'eval'
09:15		records 170 and 308 throw errors
09:15		in eval
09:15		the rest seem to go in ok
09:15		but then when I try to bulkmarcimport
09:15		rather than 479 records
09:15		I only get 398
09:17	thd	kados: are you certain about the title?
09:17	kados	thd: no, I was incorrect about that record
09:17		thd: sorry about that
09:17	pierrick	kados, zebra plugin is working on my 2.2 :-)
09:17	kados	pierrick: great!
09:17	slef	kados: ok, ready for an annoying thing?
09:17	kados	slef: sure
09:17	slef	kados: the data reading part works OK here.
09:18	kados	slef: meaning you get all the way through the file?
09:18		slef: that's probably because of the eval
09:18		maybe not though
09:18		hmmm
09:18	slef	no, I commented all the MARC stuff as the ramfs box hasn't got MARC::* installed
09:18	kados	ahh
09:18	slef	or at least not MARC/File/XML
09:18	kados	hmm
09:18		I think i was wrong about that reading problem
09:19		it seems to be working now
09:19		grrr
09:20	slef	I think you need to test and encode before
09:20		$record = MARC::File::USMARC->decode($marcrecord); # or warn "not this record\n";
09:20	kados	slef: if you comment out the first eval
09:20	slef	as you say
09:20	kados	record 171 throws:
09:20	slef	MARC::Charset->assume_unicode(1);
09:20	kados	utf8 "\xEC" does not map to Unicode
09:20		yea, but I've tried with and without that lie
09:20		line
09:20		same end result
09:21		slef: so line 171 has a non-mapping character in it
09:21	slef	replace the eval with
09:21	kados	utf8 "\xEC" does not map to Unicode at /usr/local/lib/perl/5.8.4/Encode.pm line 167, <INFILE> line 171.
09:21	slef	if (decode_utf8($marcrecord))
09:22		may need to use Encode 'decode_utf8'; too - I forget.
09:22	kados	k
09:22	slef	if that still works, I'll write something
09:24	kados	utf8 "\xEC" does not map to Unicode at /usr/local/lib/perl/5.8.4/Encode.pm line 167, <INFILE> line 171.
09:24		I did:
09:24		if (Encode::decode_utf8($marcrecord)) {
09:25		at one point I tried
09:25		use Encode qw /WARN_ON_ERR/
09:25		slef: unfortunately, the cpan version of MARC::* don't support unicode
09:25		slef: to work with unicode with MARC::* you need to grab the sourceforge versions
09:26		slef: http://www.kohadocs.org/Instal[…]Debian_sarge.html
09:26		slef: cvs listed there
09:27	slef	this installation sucks atm
09:28	kados	no kidding
09:28	pierrick	kados, could you add the instruction "run # zebrasrv localhost:2100/kohaplugin" in your documentation ?
09:28	kados	pierrick: it's not there?
09:28	slef	uh, isn't there a tarball of marcpm?
09:29	kados	pierrick: I'll tell stephen to add it
09:29		slef: yea, on the sourceforge page you can nab it
09:30	slef	bah, I'm going to have to leave this and get back to work
09:30	kados	k ... thanks :-)
09:32	slef	what's the current state of koha support modules debian packages?
09:33	kados	completely unmaintained to my knowledge
09:34	slef	libyaz-dev seems to be there
09:34	kados	yea, but libyaz-dev what version?
09:34	slef	I'd upload libperl-marc-* if someone has time to build them, or you can wait for me to remember how perl debs work.
09:35		libyaz-dev 2.1.8
09:35		what do we need?
09:35	kados	the very latest :-)
09:35	pierrick	2.1.16
09:35	slef	can you add the versions needed to http://www.kohadocs.org/Instal[…]Debian_sarge.html please?
09:35	kados	yep
09:36		well, it keeps changing
09:36		on like a weekly basis
09:36		esp with all the work being done on MARC::*
09:36	pierrick	kados, the latest version of zebra available on indexdata site is 1.3.34
09:36	slef	can you add them to koha/Makefile.PL in CVS instead then, please?
09:36		it really would help me get the installer working
09:36	kados	I don't have time to track everything down
09:36		pierrick: could you do it?
09:37	pierrick	OK
09:37	kados	thx
09:37		hi paul
09:37		paul: still aiming for the 18th for release date?
09:38	slef	could I just set everything to today's versions?
09:38	paul	kados : no, of course.
09:38	kados	slef: yep
09:38	slef	as in, does it work today? ;-)
09:38	kados	paul: good :-0 I was beginning to worry :-)
09:38		slef: well, not quite
09:38	paul	as there are still some major bugs.
09:38	kados	slef: we still have major encoding troubles
09:38		slef: similar but different to the ones we discussed today
09:38	slef	how about 2.4? what libyaz does that need?
09:39	kados	slef: depends on if you use the zebra plugin or not
09:39	pierrick	standard 2.4 doesn't need zebra
09:39	kados	slef: if you use zebra, it needs the very latest
09:39	slef	standard 2.4
09:39	kados	slef: otherwise, it doesn't need the very latest
09:39		pierrick: but it does need yaz
09:39	pierrick	oups
09:39	kados	slef: but 2.4 does require very latest MARC::*
09:40		slef: and a MARC::File::XML that hasn't been written yet :-)
09:41		what's unstable?
09:41	pierrick	MARC::*
09:41	kados	how so?
09:41	pierrick	using the HEAD CVS is dangerous
09:41		kados, it seems we have no choice if we want zebra working
09:41	kados	right, but MARC::* isn't like Koha, their HEAD almost always works
09:42	pierrick	so my sentence is useless, just to say I think "we are playing with fire"
09:42	kados	yea
09:43	paul	kados: a quick question about openCataloger
09:43	pierrick	so, do I add something to Makefile.PL? (on rel_2_2)
09:44	kados	paul: sure
09:44	paul	ToinS is writing a document to explain what he will work on, and how.
09:44		he seems confident with XUL now !
09:44	kados	great!
09:44	paul	do we create a openCat project on savannah ?
09:45		maybe it could be a good idea to play with subvestion with openCat ?
09:45	kados	paul: sure
09:45	paul	x2 ?
09:45	kados	yep
09:46		I'll warn you that savannah takes forever to register a project
09:46		I tried to register openncip there
09:46		eventually I gave up and went to sourceforge
09:46		a month later they accepted openncip
09:46	paul	ok, then maybe another OSS platform ?
09:46	kados	but requested I change the name to freencip :-)
09:46		the nerv!
09:46		I'm ok with whatever
09:47		you can decide
09:48		paul: did you see my mention of a bug in syncing between koha tables and marc tables?
09:48	pierrick	it took less than a day to register PEM on gna.org
09:48	thd	kados: freedom is better than openness any day
09:48	kados	paul: on koha-devel
09:48	paul	kados: not yet
09:49	kados	paul: it seems that Koha never deletes old holdings entries from marc_word
09:49		paul: I will file a bug report and mark it as blocker
09:50		paul: email was: Apr 12 Joshua Ferraro  (  23) [Koha-devel] Bug in MARC sync in rel_2_2
09:51		bug report created
10:13	pierrick	kados, zebra has worked but I should not have tried to solve the encoding problems... nothing work anymore
10:14	slef	never mind hosting services, you have your own webspace, you have git/cogito, host it yourself
10:38	kados	pierrick: could you explain?
10:41	pierrick	kados, it seem I can't update my zebra database after recreation
10:41	kados	recreation?
10:41	pierrick	drop/create
10:42		I had encoding problem in what was displayed
10:42	kados	drop doesn't work in zebra
10:42		you have to go:
10:42		zebraidx init
10:42		which deletes everything
10:43	pierrick	OK, in fact my problem seems to be my export
10:44	kados	ahh
10:44	pierrick	I've converted my marc_subfield_table to utf8 and now my export is very small
10:45		(it was 6MB before and now it's 0.2MB)
10:45	kados	that can't be good
10:47	pierrick	should I use export.pl in 2.2 or HEAD ?
10:52		wait... I made a mistake in the utf8 conversion procedure
10:54	kados	pierrick: export.pl in 2.2 I think
10:54	pierrick	no, I reversed and my export is still 0.2MB :-/ I really don't understand all this thing :-/
10:55	kados	weird
10:55		pierrick: did you do 'select count(*) from biblio'
10:55		before and after?
10:55	pierrick	what is a normal weight for a isoXXX export ?
10:55		I didn't
10:56		did the export deleted my biblio ????
10:56	kados	no, it shouldn't
10:56		it just exports it :-)
10:56		I mean do the select count(*) before converting to utf8
10:56		but that doesn't make sense
10:56		sorry :-)
10:56	pierrick	no that doesn't
10:57		if it did, we would have a big problem for Koha 3.0 conversion to utf8
10:58		I have 10K biblio
10:59		kados, how do I "select count(*) from zebra" ?
10:59	kados	not sure
10:59		maybe ask koha-zebra?
11:00		I'm interested too :-)
11:00	paul	iirc, I asked indexdata ml, and got an answer.
11:00		it was something strange & long
11:02		jan,9th
11:02		Paul POULAIN wrote:
11:02		> Hi,
11:02		>
11:02		> Is there a tool to have some infos on a zebra DB.
11:02		> something like the number of record in the DB, and other related infos ?
11:02		You can get some info like that out of Zebra by searching the Explain database... Using the YAZ client:
11:02		% yaz-client host:port/IR-Explain-1
11:02		Z> find @attr exp1 1=1 databaseInfo
11:02		Z> form xml
11:02		Z> show 1
11:02		etc.
11:02		The XML representation of Explain records is private to Zebra, but much easier to handle than the more standard alternatives.
11:02		You can retrieve information about a specific database name like this:
11:02		Z> f @and @attr exp1 1=1 databaseInfo @attr exp1 1=3 myDatabaseName
11:02		The contents should be self-explanatory.
11:02		You can also ask for targetInfo, AttributeSetInfo (per database, as before), and possibly other things..
11:02		(my question & sebastian hammer answer)
11:02	kados	cool
11:03	paul	a phpYazAdmin would be something useful ;-)
11:03	kados	I get:
11:03		Sent presentRequest (1+1).
11:03		Records: 1
11:03		[IR-Explain-1]Record type: XML
11:03		<explain><databaseInfo>DatabaseInfo
11:04		<commonInfo><dateAdded>20060401200734</dat​eAdded><dateChanged>20060402182248</dateCh​anged><languageCode>EN</languageCode></com​monInfo><accessInfo><unitSystems><string>ISO</string></unitSystems><attributeSetIds><oi​d>1.2.840.10003.3.5</oid><oid>1.2.840.10003​.3.1</oid></attributeSetIds></accessInfo><n​ame>kohaplugin</name><userFee>0</userFee><a​vailable>1</available><recordCount><recordC​ountActual>148636</recordCountActual></reco​rdCount><zebraInfo><recordBytes>133463477</
11:04		which looks at least in the 'ballpark' :-)
11:04		but it would be very interesting to do
11:04		select count(*) from biblio;
11:04		then install zebra plugin
11:04		then do the above
11:04		and compaare
11:05	pierrick	I'm leaving office now, I'll continue my headache on zebra on tuesday :-)
11:05	kados	pierrick: have a great weekend :-)
11:06		paul: about encoding probs
11:06		paul: (before you leave)
11:06		paul: I'll give you an update
11:06		what I think is happening
11:06		is that Perl is handing the XML parser
11:07		/Perl/ thinks is utf-8.  In other words, perl is mangling the data going into the parser, possibly turning it into valid, correct UTF8, but the parser has been told that this is /not/ in fact UTF8, but ISO
11:07		so I need  need to tell Perl that the data in the $xml variable /is not utf8/
11:07		but I"m not sure how to do this yet
11:09	paul	perldoc Encode don't give you a hint here ?
11:09	kados	I'll check
11:10		but we want to avoid Encode completely I think
11:27	paul	kados: do we call our project OpenCat, OpenCataloger or OpenCataloguer ?
11:28	kados	paul: is OpenCat taken?
11:28	paul	I don't think so
11:28	kados	paul: ok, lets use it then
11:28		paul: wait
11:29		paul: opencat.org is taken
11:29	paul	opencat seems to be something for google
11:29	kados	paul: whereas I own opencataloger.org/com
11:29	paul	ok, so let's start with opencataloger then
11:29	kados	paul: sounds good
11:31	paul	ok, opencataloger registered at gna.org
11:31		waiting for confirmation
11:31	kados	great!
11:34		paul: so re: encoding
11:35		paul: do we agree that Koha will support handling MARC-8, iso-8859, utf-8
11:35		it will only allow storage of utf-8 and iso-8859
11:35		and will not permit mixed encoding
11:35		ie, if 8859-1, only 8859-1
11:36		if utf?
11:36		s /if utfU//
11:37	paul	yes.
11:37		except it's iso8859-15 and not -1
11:37		but the only diff is the €
11:38	kados	ahh, well that would explain something
11:38		maybe it already works perfectly
11:38		I thought it was 8859-1
11:38	paul	8859-1 has been transformed to -15 when EU switches to €
11:38	kados	could you try latest rel_2_2 marc edits of existing data in emn fo r instance?
11:39	paul	but the only diff between them is the € symbol
11:39	kados	after changing to 8859-15 in TemplateEncoding var?
11:39	paul	yes, but not for instance, as I must leave now (7PM soon)
11:43		bye bye world
11:44	kados	mye paul
11:44		bye even :-)

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