IRC log for #koha, 2005-04-27

All times shown according to UTC.

Time S Nick Message
12:03 kados shaun: so you installed a fresh version of Koha?
12:03 kados redid the DB and everything?
12:03 kados I don't have time to troubleshoot that specific problem
12:04 kados but I can help you uninstall your current Koha and start over
12:04 kados you can probably do that in under 10 mins
12:04 kados drop the koha database
12:04 kados delete /etc/koha.conf
12:04 kados and /etc/koha-httpd.conf
12:04 kados delete from db where user like 'kohaadmin'
12:04 kados delete from mysql where user like 'kohaadmin'
12:04 kados run
12:05 kados (actually, grab 2.2.2b while you're at it
12:06 kados sorry ... that second query is wrong ... should be
12:06 kados delete from user where user like 'kohaadmin'
12:06 kados both of those are after doing 'use mysql' in your mysql client
12:28 shaun sorry, i was away, back now
12:28 shaun has 2.2.2b been rolled yet?
12:30 shaun it's all new - the database is a new one which I created a couple of days ago, each on a completely fresh install of mysql 4.1.11 on a completely fresh install of fedora core 3
12:38 owen 2.2.2b:[…]release_id=318874
12:40 shaun ah, great - what changes are there, without me downloading it?
12:49 shaun does it include my fix to the returns page? :p
12:51 owen I wonder if paul created the 2.2.2b package without updating CVS?  Is that possible?
12:54 shaun i have to ask about the koha release cycle: when preparing for major and minor releases, do we get the cvs head, roll it, and then fix all of the known bugs on that so that release candidates are issued, or is a separate branch maintained for each release, with the new features being developed in head, and the fixes being ported to the release branch, or what?
13:00 owen new features are developed in head, and bugfixes are in the rel_ branch.  So current bugfixes go into rel_2_2.
13:01 shaun so, paul rolled this release from the rel_2_2 branch, then? is this the same with new major releases (e.g. for 2.4, will there be a rel_2_4 branch? | when will it split?)
13:03 owen Yes, this release should have been from the rel_2_2 branch.  When 2.4 is released, the current stable branch will be rel_2_4 and HEAD will be for 2.6 (I guess)
13:05 shaun (so our templates will make it into 2.4, if i commit them to head... nice)
13:05 owen That's right.
13:19 shaun I have a worry: if our templates are not the default in 2.2, and will not be in Koha at all until 2.4, yet we are basing our templates on the 2.2.2 release, we will have to backport all of the changes in the default templates in head between the 2.2.2 release and the time that 2.4 is being alpha'd - got any suggestions for getting around this?
13:20 owen Keep up with updates.
13:21 owen Every time a change is made to a default template in CVS, make a similar change to yours.
13:21 shaun with the cvs mailing list, i suppose - what if the changes are incompatible with 2.2...?
13:22 owen I'm not sure what you mean...If you want your templates to work in 2.2 and 2.4, you'll have to maintain two different versions.
13:25 shaun not so much 2.2 and 2.4 - more 2.2 and head, as our templates came from 2.2, but updates to the non-template stuff (the perl backend) in head could be incompatible with the (stable and released) version of koha we are working on.
13:26 owen Still... you're talking about maintaining two different versions.
13:26 shaun is there much of a change in database structure between 2.2 and the current head?
13:27 owen paul would know if you manage to catch him around here.
13:28 kados shaun: do a dif on updatedatabase between 2.2 and 2.4 to find out
13:29 shaun well - in the same way that there were major changes between 1.x and 2.0, would it be probable that there are any changes between 2.2 and 2.4? (i presumed you would know, as you are working at a library, and therefore have to manage the upgrades to some extent)
13:29 shaun who is updatedatabase? :)
13:30 shaun is it a perl script?
13:30 owen /updater/updatedatabase
13:30 owen If there weren't any changes between 2.2 and 2.4, there wouldn't be a 2.4.  We'd just keep calling it 2.2! :)
13:31 shaun i meant changes to the database specifically
13:35 shaun ah, there are a couple of changes, but nothing that really affects me...
13:37 shaun so, if i work purely on head, how can i ensure that my installed copy is the latest, without rolling, untarring, installing and installing the templates and database again?
13:39 owen Get a fresh copy from CVS?
13:39 kados shaun have you read the documentation on
13:40 kados
13:40 kados specifically that one?
13:42 shaun yes *promptly writes bash script which achieves effect of said document*
13:51 shaun kados, do you have any idea when 2.4 will be out? i really need to demonstrate the system to the library/librarians which will be using it, but if 2.4 is a long way away, i will consider making a custom release of 2.2 with the new templates backported... unless our templates go straight into cvs, which, frankly, they are not ready for yet...
13:52 shaun stability is important, obviously...
13:52 shaun (brb)
13:52 kados not sure
13:52 owen I think there hasn't really been much done in HEAD since 2.2 came out, so 2.4 isn't looking very close
13:56 owen And dont' forget, shaun, HEAD is for unstable stuff, so you can commit your templates any time, and keep updating as you go.  It's up to you, though.
14:30 shaun *doesn't know what to do...*
14:33 owen Look, if you want your templates to work in 2.2 *and* you want your templates to be part of the 2.4 release, you're going to *have* to maintain two different copies.  Or upgrade your templates all in one go when 2.4 is released.
14:34 owen It's no fun, but no one said the job was easy! :)
14:47 shaun well, if I knew that 2.4 will be released before september, and has the amazing features we were talking about last week (from the argentinians - still don't know the full story...) then I would just commit to head, and make sure that all further template development (right down to the bugfixes) is done on our new ones
14:49 owen We can't predict the future.
14:49 shaun *wonders what else owen might be predicting, aside from the future :D*
14:58 owen You're right, that was so redundant.  I should have just said 'We can't the future.'
15:00 shaun rofl, I can't the future either
15:09 kados well ... unless you're submitting bugfixes you should be committing to HEAD anyway
15:09 kados rel_2_2 is just for bugfixes for 2.2
15:10 owen (and *very* minor additions)
15:14 kados check this catalog out:
15:14 kados
15:14 kados very fancy
15:14 owen Hmmm... I can't seem to connect
15:15 shaun owen: it's a bit slow for me too, it loaded eventually...
15:17 shaun don't see what's fancy about it...
15:18 kados I think the 'suggestions' feature is pretty neat
15:19 kados boy ... it sure is slow!
15:19 shaun yes, but it's at the expense of general usability, and speed...
15:19 kados well ... it wasn't this slow last time I saw it
15:19 shaun i take it they're not running koha on mysql 4 then ;)
15:19 kados hehe
15:19 shaun (burn 'em!)
15:19 kados well actually ...
15:20 kados I've been doing a bit of reading on search methods
15:21 kados I used to wonder what other databases there are besides Relational
15:22 kados I'm starting to understand at least one other: textual
15:22 owen I like how for related terms to 'microtechnology' they offer 'Macbeth--Chronology'
15:22 kados so hehe
15:23 kados so in the case of searching MARC records, rather than using a RDBMS we should be using a textual DBMS
15:23 kados using textual indeces
15:23 kados something like Lucene
15:23 owen Is that more like what Google does?
15:23 kados yep
15:24 kados and every other sane approach to searching texts does something similar
15:24 kados in effect we have two different databases
15:24 kados one is a database 'about' books
15:24 kados the MARC records
15:24 shaun meh... i know a little about implementing lucene in java - what apache forrest has taught me ;) - does anybody else have any idea how to implement it (somehow) in koha?
15:24 owen'd have to program your indexer to give different weight to different areas of the MARC record depending on the search type.
15:24 kados and one is a database 'of' books ... our holdings, status, etc.
15:25 kados I think PLucene is the answer as it's in perl ;-)
15:25 shaun well, yes... i'm not a perl genius (*yet*)
15:26 kados RDBMS are not really designed to handle a textual database and so it's performance and accuracy are actually quite poor
15:26 kados owen: right
15:27 kados you could actually do some really neat stuff with the indexer
15:27 kados
15:28 kados there's that clumsy 'it's' again ;-)
15:28 owen That Lexington catalog may be snazzy, but it's ugly as sin!
15:28 kados ehe
15:28 kados yea
15:30 kados
15:31 shaun are there any commercial/propietary ILS which do this out of the box?
15:31 kados I'm sure most of the big ones do
15:31 kados which would explain why their searches are so damn fast even with HUGE datasets
15:32 kados and even when those datasets include fulltext records
15:33 kados (which is something else we could look ito providing for Koha)
15:33 kados (if we went with PLucene)
15:35 shaun surely that'd be for a 3.0 release?
15:36 owen Right, Wednesday?  Let's get on it, people!
15:36 shaun ??? what happens next wednesday, may i ask?
15:37 kados could be ... redesigning the search may not be very hard though
15:37 owen Koha 3.0!!!  We've got a lot of coding to do!!!11
15:37 kados from the research I've done so far Plucene is quite easy to use
15:37 kados hehe
15:38 kados to do a 3.0 we'd have to do quite a bit of bugfixing too
15:38 shaun what bugfixing is this?
15:39 shaun (i mean, what blockers are there?)
15:39 owen For 3.0?  How about Bug 1235 -- "Plucene functionality still imaginary"?
15:41 owen ;)
15:45 shaun[…]loc_type=allwords
15:45 shaun substr&field0-0-0=noop&type0-0-0=noop&value​0-0-0=&cmdtype=doit&order=%27Importance%27
15:45 shaun long link ;) -- bugzilla needs a little bit of cleaning
15:46 kados owen: hehe
15:47 shaun brb
15:48 kados so for using plucene I think the thing to do is
15:48 kados export the marc records as xml
15:48 kados using a free tool (we'll need to find one)
15:50 kados then extract the meta tag for insertion into the indexer
15:50 kados tags that is
15:51 kados then modify our search code to grab bibids from plucene's indexes before passing it off to find out item-specific info, status, etc.
15:52 kados chris around yet?
15:55 kados maybe it would be better to use MARC::Record to extract data directly and insert it into the index via the indexer...
15:56 shaun "export the marc records as xml" -- when? with an index daemon/cronjob or at the time of searching, directly?
15:56 kados see my revision ^ ;-)
15:56 kados this would happen as records were added/deleted
15:57 kados so ... not very often
16:02 shaun wouldn't this be quite slow (particularly on pre-mysql 4.1 - *brings back performance topic*)?
16:02 kados wouldn't what be slow?
16:03 shaun searching, indexing or adding, particularly following a large marc import
16:03 kados btw: plucene also supports variations and inflections of a word
16:04 kados well ... searching will be really fast since we won't be using mysql
16:04 kados indexing and adding can be backgrounded ... it might take a while
16:04 kados but it doesn't happen nearly as much as searching eh?
16:05 shaun why not mysql? will it be using xml entirely?
16:06 kados no ... disregard my xml comment
16:06 shaun (*very confused*)
16:08 shaun please explain more, as i don't understand how we could not be using mysql, yet still searching the database
16:12 shaun ?
16:21 kados
16:21 kados podcast on LibLime ;-)
19:11 rach russ?
02:57 hdl hi
03:40 Sylvain hi
04:32 Sylvain ah, we may have a candidate in the next 2-3 weeks ...
04:32 Sylvain is seems interesting :)
10:25 owen paul... didn't there used to be a 'delete' button on the addbiblio screen?
10:30 owen hi shaun
10:31 shaun hi
10:32 shaun still haven't got any book data :( (and ben seems to have gone off in a huff...)
11:03 owen Someone should build a Mozilla toolbar for Koha.  You could put the whole navigation menu in and free up space onscreen.
11:05 hdl what a killer feature.
11:05 owen Anyone here know XUL? :)
11:05 hdl No, but having some docs on it.
11:05 owen For that matter we could probably build a whole Koha interface in XUL
11:06 shaun i know a little (jack of all trades ;)) - i'll look into it
11:06 hdl But then Koha wouldn't be that easily installable on multiple machines ;)
11:07 owen True, but it'd still be cool B)
11:07 paul hdl : could be a good idea for intranet opacs
11:07 shaun | do you know a way of getting two installs on one box? the two installs seem to fight over /etc/koha.conf ;)
11:07 paul yep shaun, very easy. I have at least 10 ;-)
11:07 paul just add :
11:08 paul SetEnv KOHA_CONF /etc/ANOTHER_koha.conf
11:08 paul into your virtual host
11:08 paul (in both opac & librarian interface)
11:09 paul note that the installer just check for /etc/koha.conf.
11:09 paul so if you mv /etc/koha.conf /etc/koha2.conf
11:09 paul you can install it again !
11:09 hdl and also SetEnv PERL5LIB /usr/local/koha/modules
11:09 paul yep, hdl is right.
11:09 shaun the opac is what i am thinking - for my implementation (in a school), it would be useful to have a restricted, fullscreened firefox, with limited navigation - i can imagine it being good for self checkout
11:09 shaun | k, tries that, thx
11:10 paul shaun : the self checkout already exists in Nelsonville
11:10 paul should be commited soon (isn't it owen ?)
11:10 owen Yes, self-checkout is in use at one of our branches.  I think Joshua just needs to find time to clean it up and commit it.
11:12 hdl paul : a propos de la visite médicale ?
11:12 shaun i know it exists, what we would like to see in the school library is the restricted terminal (stops users from going to another site etc)
11:12 paul (pas appelé, tant pis pour eux...)
11:12 hdl Je viens ou pas.
11:13 hdl s/./?/
11:13 paul je réfléchis...
11:13 paul tu as bien avancé sur le dictionnaire ?
11:14 hdl Je vois à peu près les traitements pour la recherche. Mais je ne sais pas trop quoi faire des résultats. Une liste de checkboxes pour les intégrer à la recherche ? ou autre...?
11:14 paul ok, alors viens, comme ca on fera ensemble.
11:14 paul le rdv est à 11H
11:15 paul donc tu peux venir dans la matinée
11:15 hdl donc pas forcément aux aurores ;)
11:15 paul pas la peine de te lever à 5H par contre ;-)
11:15 hdl ok.
11:15 hdl les embouteillages finissent à quelle heure ?
11:16 hdl ou bien ils ne finissent pas ???
11:16 paul vers 9H30 tu dois être tranquille
11:16 hdl ok.
11:16 shaun DBD::mysql::st execute failed: Invalid default value for 'aqbudgetid' at scripts/updater/updatedatabase line 1061. DBD::mysql::st execute failed: Invalid default value for 'id' at scripts/updater/updatedatabase line 1061.
11:16 shaun -- in, when installing on mysql 4.1 - anybody else noticed this?
11:17 paul iirc shaun, you should also have a warning just before or after the update, saying "don't worry, it's not really a problem"
11:17 owen Yes, paul, is chris's
11:19 shaun paul: there is no such message...
11:20 paul mmm... anyway, i'm almost sure it's a problem you can ignore.
11:32 Sylvain ohlala, je viens de répondre pour les règles de prêt et entre temps deux réponses, elle va être débordée de réponses la pauvre :)
11:33 Sylvain bon on a tous 3 répondu à peu près pareil, ça va :)
11:33 paul ;-)
11:33 Sylvain mais je pense que le pb qu'elle évoque mériterait réflexion
11:33 Sylvain j'avais aussi pas mal galéré
11:34 paul moi, je ne considère pas ca comme un bug.
11:34 paul (ou alors il y a un truc qui m'échappe)
11:34 paul si on met 0,0 dans une *, ca veut dire "pas de droit"
11:34 Sylvain oui
11:35 Sylvain si on ne met rien ça veut dire on veut pas remplir cette case
11:35 Sylvain mais si c'est dans une colonne/ligne *
11:35 Sylvain ça veut pas pour autant dire 0,0 pour toute la ligne/colonne
11:35 Sylvain (enfin ça fait qques temps que j'avais rencontré le pb je suis plus exactement sur du pb)
11:35 paul donc ton idée serait qu'il faudrait mettre automatiquement le max sur les * s'il n'y a rien ?
11:35 paul (par exemple)
11:36 Sylvain bon, j'ai peur de dire des conneries, c'est vieux tout ça
11:36 Sylvain attends
11:36 paul mauvaise idée à la réflexion : si on calcule automatiquement, on aurait "nb max = plus grande des valeurs. Or pour beaucoup, c'est la somme des valeurs)
11:36 paul 5 livres, 3 CD et 8 n'importe quoi.
11:37 hdl_away en fait, pardon d'intervenir en partance, mais c'est peut-être que tout le monde c'est peut-être la définiton des droit de tous les autres, à l'exclusion de ceux qui sont définis.
11:37 paul par contre, pour les * par type de lecteur, c'est plutôt le plus grand, effectivement.
11:39 hdl désolé, j'ai dit une bêtise.
11:40 Sylvain ah, je viens de refaire mes petits tests
11:40 Sylvain en fait, quand on a rien dans une case, c'est bien ça, il créé un enresgitrement avec le nombre de prêt NULL
11:41 Sylvain alors maintenant, pourquoi est-ce génant déjà ? je me rappelle plus
11:45 Sylvain bon bein je sais plus trop en fait :(
11:45 Sylvain mais bon, il me semble avoir entendu Pascale dire quelquechose a ce sujet, peut-être qu'elle va réagir sur la ML
11:45 Sylvain toujours est-il qu'il y a un truc quand mm :)
11:45 Sylvain sur ce, c'est l'heure de rentrer sous la pluie parisienne ...
11:45 Sylvain a demain
11:45 paul on va attendre et voir si Pascale et Carole réagissent
11:45 paul à demain
11:46 paul (tu bosses sur quoi en ce moment au fait)
11:46 paul (bon, on verra demain)
11:46 paul ;-)
11:46 sylvainOu j'ai 2 secondes ;)
11:46 paul hdl a commencé le "dictionnaire"
11:46 sylvainOu en fait j'étais parti sur dautres choses et j'essaye de finaliser les commandes depuis le réservoir
11:46 paul carrément touchy pour gérer à la fois les bibs avec listes d'autorité et celles sans.
11:47 paul (mais on a trouvé comment faire, ca va être mignon tout plein)
11:47 sylvainOu mais avec ces biblio biblioitems, OLD, new MARC ..
11:47 sylvainOu bon bein cool pour vous
11:47 paul en fait, on ne cherche rien sur biblio/biblioitems...
11:47 sylvainOu ça va être intéressant !
11:47 paul on utilise uniquement les champs marc
11:47 paul avec l'API de, on peut trouver ou chercher les choses.
11:48 paul (restera à valider les performances, mais je suis pas trop inquiet)
11:48 sylvainOu ok
11:48 sylvainOu sur ce, c'est vraiment le départ
11:48 sylvainOu ++

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