IRC log for #koha, 2004-12-07

All times shown according to UTC.

Time S Nick Message
18:54 genji hey all.
18:57 rach hi
18:58 genji ah, not the person i wanted to talk to, but happy that someone is talking none the less. A good day, rach?
18:58 rach um a bit mental today to be honest
18:59 genji Mental?
18:59 rach lots of stuff happening
18:59 rach our slightly mad neighbour was over earlier chewing my ear about their shed
19:00 rach which isn't really my optimim way to start the day
19:01 genji oh, EVERYONE works from home?
19:01 rach yep
19:01 rach me included
19:01 genji Cool.
19:01 rach
19:01 rach and my home is being renovated at the same time
19:01 rach which has elements of being a bit trying
19:03 genji Impressive, you are, yes.
19:04 rach hence the "mental" today :-)
19:06 rosa How did you withstand the gale, rach?
19:06 rach was ok actually rosa
19:06 rach the board didn't blow in which was good
19:06 rach wasn't too bad here
19:06 rach was worse driving down the road
19:06 rosa that was a relief
19:07 rach yeah driving down was really bad, so I was a bit worried
19:16 rosa Koha will be 5 years old in January
19:20 genji yeah.. congrats.
04:04 genji hey rach, how do i get a job like Katipo's work at home programming jobs?
04:06 rach umm you send me a cv is a good start :-)
04:07 rach and tell me why you'd be a great asset to our team
04:07 rach and what you want to do, and why working for us would be good for you, and good for us
04:08 rach and then I see if we have enough work for another person
04:08 rach basically
04:08 rach you've actually done, what you can do, what you want to do
04:09 rach so if you're self tought that might be fine
04:09 rach um I should add, that living in Wellington Region tends to be a requirement
04:10 rach as we do have actual in person meetings as well as online ones
04:10 rach and you'd miss out if you couldnt' make them
04:10 genji So, living in Palmerston north is not that good?
04:10 rach it's not great - but better than Iowa :-)
04:10 genji Why would anyone live in Iowa..
04:11 rach or maybe Idaho
04:11 rach palmerston north might be ok
04:11 rach depends on what you want to do
04:12 rach and how good at it your are :-)
04:13 rach cool - send me something in the new year
04:14 rach so unless you're handy with a ham there is no hurry :-)
04:14 genji Will do. Think ill pick up a Web design course at open polytechnic.
04:14 rach cool
04:15 rach they are a bit fishy after eating sushi
04:15 genji its the only thing that comes close to web work, and programming in perl. seems the standard language is Visual basic or .NET.
04:52 genji hiya paul!
04:52 genji you active? got a puzzle for ya.
07:00 paul genji, if you are not sleeping, i'm here
07:05 genji im here.
07:05 genji you still here?
07:05 paul yep
07:07 genji okay. deleted EVERYTHING of koha.  mysql data and perl stuff. reinstalled.  installed the provided marc structure. etc. imported my database, and still subjects won't search in search.marc. Only in
07:07 paul maybe the provided MARC21 structure is wrong
07:07 paul may i have test your installation ?
07:08 genji nope. checked the links and bibliosubject.subject is mapped to 500, General Topical description.. sure.
07:08 genji .. default username and password.
07:09 genji thats
07:10 genji logged in?
07:12 paul sorry, was on phone
07:12 paul i'm back
07:13 paul msg genji ??? what are username & password ?
07:13 paul oups...
07:44 paul don't you think ?
07:45 genji yup. okay, this is for the log.
07:45 paul could you summarize here ?
07:45 genji oh, you've got the marc code entering biblioitems in newbiblioitem right?
07:45 paul (i'll copy sub newbiblioitem after)
07:45 paul yep
07:46 genji hey, i remember submitting this bug to koha-devel.
07:46 paul do you remember when ?
07:47 genji 3/12/2004
07:47 genji probably didn't explain myself properly.
07:47 paul yes i just saw it
07:47 paul note that your 1st mail was really too short to be understood (at least by me ;-) )
07:50 genji Okay, for the log. BUG: Koha's erases subjects. Explanation: C4::Biblio::newbiblioitem line 2091 to 2096 uses modsubject to save $biblioitem->{'subjectheadings'}. Of course, this is a null value. due to the nature of modsubject (it deletes all subjects then inserts) it erases subjects. Fustrating after you've spent 10 minutes entering in subjects into only to have it erased. FIX: Delete lines 2091 to 2096.
07:50 genji hows that?
07:51 paul sounds OK for me.
07:51 paul but won't solve the other problem :
07:51 genji other problem. now that i know it, i can do it.
07:51 paul the subjects are not solved properly in the MARC biblio
07:51 paul ok, i'll let you dig.
07:51 paul i'll commit this 1st bugfix soon
07:52 paul sub newbiblioitem {
07:52 paul    my ($biblioitem) = @_;
07:52 paul    my $dbh        = C4::Context->dbh;
07:52 paul    my $bibitemnum = &OLDnewbiblioitem( $dbh, $biblioitem );
07:52 paul    return ($bibitemnum);
07:52 paul ]
07:52 paul (oups...)
07:52 genji wheres the marc code gone?
07:52 paul s/]/}
07:52 paul sub MARCkoha2marc
07:52 paul line 884
07:53 genji eh, no.
07:53 paul ?
07:53 genji my code here has newbiblioitem with marc saving as well.
07:54 genji line 2097 to 2103?
07:54 paul quoted too no ?
07:54 genji you didn't quote.
07:55 paul ???
07:55 genji sub newbiblioitem {
07:55 genji    my ($biblioitem) = @_;
07:55 genji    my $dbh        = C4::Context->dbh;
07:55 genji    my $bibitemnum = &OLDnewbiblioitem( $dbh, $biblioitem );
07:55 genji    my $MARCbiblio =
07:55 genji      MARCkoha2marcBiblio( $dbh, 0, $bibitemnum )
07:55 genji      ; # the 0 means "do NOT retrieve biblio, only biblioitem, in the MARC record
07:55 genji    my $bibid =
07:55 genji      &MARCfind_MARCbibid_from_oldbiblionumber( $dbh,
07:55 genji        $biblioitem->{biblionumber} );
07:55 genji    &MARCaddbiblio( $dbh, $MARCbiblio, $biblioitem->{biblionumber}, '',$bibid );
07:55 genji    return ($bibitemnum);
07:55 genji }
07:56 genji that does it right?
07:56 paul mmm... i thought no.
07:56 paul but after reading, finally, yes
07:56 paul i'm really a good programmer ;-)
07:56 genji you have old
07:56 paul sub MARCaddbiblio won't really create a new biblio when bibid is not null.
07:57 paul kudos to me, i should have do that for such cases :-D
07:57 paul old ? which date ?
07:58 genji you don't have it in yours?
07:59 paul yes i have.
07:59 genji so.. why didn't you quote it? the code works... and taking it out makes it not work.
08:01 paul ???
08:02 genji hmmm.. okay. when you quoted newbiblioitem .. you did not quote the marc adding of biblioitem part.
08:06 genji anyway, new topic. we can put a marc adding of subject at modsubject.. but, how do you delete marc fields? like, i want to erase the marc subjects of a specific biblionumber?
08:06 paul i have to leave for lunch
08:06 genji since thats the behavior of modsubject.. k. seeya soon?
08:06 paul 1H30
08:06 genji ccool. will be here.
08:07 paul but a phone meeting in 2 hours
08:07 paul i let the channel ON anyway
08:07 genji think all i will need from you to finish this bit is 30 minutes.
09:08 paul_lunc genji, i'm back
09:08 genji cool. okay. question. how do you delete a marc field from a biblio?
09:09 paul_lunc in mySQL or in MARC::Record ?
09:09 genji modsubject deletes all subject entries for a given biblio, before inserting new ones.
09:09 genji the marc koha 2 marc onefield only adds fields, doesn't replace.
09:10 genji any ideas?
09:12 paul i'm afraid the only solution is delete > create again
09:13 paul the whole biblio
09:13 genji using modbiblio... you totally sure?
09:13 paul because it's almost impossible to find the right subject in subject entries
09:13 paul probably yes.
09:13 paul I think we should store everything in non-MARC, then call MARC modif (that is a delete / create)
09:14 paul I trieds, 2 years ago, to create an "intelligen" MARCmodbibli
09:14 paul but it's almost impossible to do.
09:14 genji there isn't a way to delete subjects of one biblio entry only?
09:14 paul so many tests that it would cost more CPU to to than a delete / create
09:15 paul iirc no.
09:17 genji wait.. what about MARCdelsubfield?
09:17 genji isn't subjects a subfield of biblio?
09:18 paul unused, should be removed.
09:19 paul & buggy imho, because it doesn't deal with marc_word (& it should)
09:19 genji okay, where is the MARC subjects table stored?
09:20 paul like all other MARC datas :
09:20 paul everything is in marc_subfield_table
09:20 paul & words in marc_word
09:20 genji k. can't MARCdelsubfield be made operational, so we can delete subjects of a specific biblio?
09:22 paul the question being : how to be sure you have found the right MARC data ?
09:23 paul because repeatable fields can be repeated BUT are not "ordered"
09:23 paul so, you could only use the subjectvalue itself.
09:23 paul and that would not be very sure
09:26 genji is there any chance that a client could be searching for your book, at the same time as you are clicking modify, and the record is deleted when they look? or is the whole delete and recreate process fast?
09:27 paul this is possible in theory
09:27 paul BUT
09:27 paul in libraries, biblios are rarely modified.
09:27 paul so the risk is very, very, very, very low
09:27 paul so, it's ignored
09:28 genji right. whats the time like, starting the delete and completing the new biblio? a second?
09:29 paul when I bulkmarcimport lots of notices, I get a 6biblios/seconds.
09:29 paul and that's when disk/CPU is heavily used !!!!
09:29 genji thats deleting and adding?
09:29 paul otherwise, imho, it's 10-25ms
09:29 paul no, just adding.
09:29 paul but as fast as possible.
09:30 genji ah. how long does it take to delete a marc biblio, then?
09:30 paul i just checked that everything is correctly indexed
09:33 genji is there any way that the marc biblio could be different to the old db table, that the marc biblio data is newer? Any way to save the old marc biblio data, and put it in anew, or is every page that modifies a book, modifying the old db at the same time?
09:33 paul every page that modifies a book modify "both" DB
09:34 genji so, why do we have a marc db then? faster searching?
09:34 paul no, more complete storing
09:35 paul MARC has datas that will never be storable in non-MARC DB
09:35 paul even worst : Koha is "multi marc" compliant, so even more impossible
09:35 genji ahh... any datas in marc biblio not storable in old db biblio?
09:35 paul yes, a lot !
09:36 genji so, isn't deleting the marc biblio, and making it anew from old db just to delete subjects and insert new ones, stupid? all that lost data?
09:37 paul no, because if MARC=OFF, all that is in MARC-DB comes from non-MARC DB
09:37 paul it's incomplete MARC, but it's MARC anyway !
09:37 paul changing MARc=OFF => MARC=ON => MARC=OFF is stupid !
09:38 paul because in this case, you will loose datas, you're right
09:38 genji ahh.. but what if MARC is ON and they want to delete subjects?
09:39 paul they edit the MARC biblio,
09:39 paul and when saved, it takes care of all non MARC datas!
09:40 genji okay. but how do they delete the subject subfield of a specific biblio?
09:40 paul yes (unless there's a bug here too ;-) )
09:42 genji ?! i mean how? Youve said that deleting subjects from marc biblio is impossible, yet someone can edit the marc biblio and do so?
09:43 paul in MARC editor, Koha completly ignore what the fields "means". Koha just know that the user want to set up a 650$a field with value "reincarnation"
09:43 paul anything can be deleted/modified
09:43 paul but without knowing "what it means"
09:43 paul it's the MARCmarc2koha that "translate" values into non-MARC (=meaningful) values
09:43 genji can't the link be used, backtracked to the subfield?
09:44 paul no,
09:44 paul for instance, when a biblio is modified, it's in fact deleted / created again
09:44 paul because many "meaningful" values are not linked
09:44 paul and can't due to MARC specifications
09:45 genji but.. when a biblio is modified.. doesn't it lose its extra data from marc?
09:45 paul (we could track, but with non standard MARC fields)
09:45 paul ??
09:45 genji okay. MARC=ON
09:45 paul when MARC=ON,
09:45 genji User has lots of extra data in marc biblio
09:45 paul everything is stored in MARC
09:45 genji user modifies marc biblio
09:45 paul all what can be "understand" is stored in non MARC
09:46 paul yes, and ?
09:46 genji marc biblio is deleted, and made new.
09:46 paul right
09:46 genji extra data is deleted?
09:46 paul which extra data ? there is no
09:47 paul a general question : i'll have to leave soon
09:47 paul (phone meeting in 15mnà
09:47 paul i can release a 2.2.0RC4 now
09:47 paul with a lot of bugfixes
09:47 genji MARC has datas that will never be storable in non-MARC DB
09:47 genji correct?
09:48 paul do you think i have to wait for your bugfix for subject linkings ?
09:48 paul (correct)
09:48 paul for example, notes (300 in UNIMARC)
09:48 paul there are >20 differents notes
09:48 paul notes on author,
09:48 paul on translation
09:48 genji think so. its a showstopper for some libraries... not being able to search for subjects.
09:48 paul on aterial
09:48 paul ...
09:49 paul (that won't mean a 2.2.0RC5 won't be made, I could add the "known bug" to release notes)
09:49 genji okay.. so, what happens to this unstorable data when marc biblio is deleted and made anew?
09:49 paul (the question is do you think you could fix the problem in a day or 2 ?)
09:50 paul it's in the MARC editor, so it's stored again in MARC-DB
09:50 genji ahh.. apache has it stored as params?
09:50 paul no, but when MARC=ON, adding a biblio open a "MARC editor", not
09:51 paul and all the MARC biblio is here
09:51 genji hmm..  okay.
09:52 genji hey, i could write in script to carry over bibliosubject to marc, skipping through the database, but without that subject deleting thing, i can't build it into modsubject.
09:53 paul imho, the best solution would be to store everything into non-MARC DB
09:53 paul then call MARCkoha2marc
09:53 paul to build a complete MARC::Record
09:53 paul the
09:53 paul (then)
09:53 paul call MARCmodbiblio($dbh,$record...)
09:53 paul to store it correctly in the MARC-DB
09:54 genji what happens to the unstorable-in-non-marc data thats in the MARC-DB that MARCmodbiblio deletes?
09:55 paul it doesn't exists
09:55 paul as MARC=OFF !!!
09:55 paul should only exist when MARC=ON
09:55 paul but if MARc=ON, we NEVER arrive here !
09:55 genji ah.. right.
09:55 genji modsubject isn't used in non-marc?
09:56 genji err
09:56 genji wrong way.
09:56 genji modsubject isn't used in marc=on?
10:03 genji if modsubject isn't used when marc=on (i can make a check for the parameter just in case) then i can do the fix, completed in 24 hours, plus routine to put in upgradedatabase. okay?
10:03 genji gotta sleep... so back in 8 hours.
10:04 paul genji : ok, send me the fix when available
10:05 GenjiBed how do you check a value of a parameter? i.e check that marc is infact off?
10:06 paul C4::Context->systempreference(MARC)
10:06 paul or something like that
10:06 GenjiBed cool. night.

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