IRC log for #koha, 2006-04-25

All times shown according to UTC.

Time S Nick Message
19:28 kados hey tumer
19:28 tumer hi
19:28 kados tumer: just sent you an email :-)
19:28 tumer could not sleep
19:28 kados heh
19:29 kados they keep me up :-)
19:29 tumer your mail did not come
19:29 kados hmmm
19:30 kados tgarip at the right email addr?
19:30 tumer thats right
19:30 kados strange
19:30 kados sent about 10 minutes ago
19:30 tumer univ server sometýmes goes bezurk
19:30 kados right
19:31 tumer do you know anythýng about thd's new frameworks
19:32 kados yes ...
19:32 kados want me to send you a copy?
19:32 kados they still need some checking
19:32 tumer I have a copy
19:32 kados ahh
19:33 tumer There is a problem if he decides to use 09o instad of 090 fields
19:33 kados of course
19:33 kados he shouldn't do that
19:33 tumer M:F:X can not parse it
19:33 kados right
19:33 kados (but we can of course modify M::F::X as it should be capable of parsing any tag)
19:34 kados (IMO)
19:34 kados tumer: last week I sent you a mail regarding MARC::Charset ... did you recieve that one?
19:34 tumer also all of biblio and addbiblio has to change
19:34 tumer kados: No!
19:34 kados yikes ...
19:35 kados maybe your email server is blocking mail from me
19:35 kados did you get the one I sent about Index Data Support?
19:35 tumer its blocking email depending how busy it is- students you know
19:35 kados wow ...
19:35 kados do you have a yahoo account or something that I can depend on?
19:36 tumer I am still waiting for the indexdata support mail
19:36 kados yikes ...
19:36 kados so at least three mails I have sent ...
19:36 tumer try
19:36 kados let me check if there were others
19:36 kados ok, I'll bounce the messages there
19:37 tumer I want to sort a few more things about Zebra before I come to Marseilles
19:38 tumer By the way I am now using thd's new frame. But 90o instaed of 09o for storing bibnumbers
19:39 kados cool, is it working?
19:39 tumer Yes
19:39 kados 90o ... is that a bug?
19:39 kados tumer: (messages sent to your hotmail account, let me know if you recieved them)
19:39 tumer the problem is tag<10 finds 09o to be lower than 10 so can not be used
19:40 kados ahh
19:40 kados probably bibid/biblionumber should be in 001 eh?
19:40 kados according to standard MARC anyway
19:41 tumer I received 2 mails
19:41 tumer letsd wait and see
19:42 kados 1. Koha Demo of Zebra Integration; 2. MARC::Charset missing Turkish Chars?; 3. Index Data Support Information
19:43 tumer it can be anywhere. And yes 001 is a good choise. ý am merely testing
19:43 tumer I receied 1&3
19:43 kados hmmm, strange
19:43 kados I'll send 2 again just to be sure
19:47 tumer is the 2nd a large file
19:48 tumer it is still not here
19:48 tumer how urgently you need the Zenra-Koha demo setup?
19:49 kados strange
19:49 kados tumer: I need it before the 2nd of May
19:49 kados tumer: as that is when my demo is :-)
19:49 kados tumer: but ideally before then so I can get comfortable using it
19:51 tumer OK. I have to find a way to send it to you. We have a limit on smpt, no ftp server so we have to see what to do
19:51 kados tumer: here's the second email:
19:51 kados tumer: I have an ftp server
19:52 kados tumer: if you tell me your IP address I can enable it for upload
19:52 tumer I can upload there?
19:52 tumer
19:53 kados tumer:
19:53 kados tumer: anonymous user/pass should work
19:55 tumer kados:I'll work on this within 18hrs
19:55 kados tumer: thanks!
19:56 tumer I tried using Zebra1.4 today. But it find 30% less records than 1.35. very strange thing this Zebra
19:57 kados ??
19:57 kados wow, that is strange
19:57 kados (btw: if you want to clean up the marc21 bib framework and commit it you can put it here: misc/marc_datas/marc21_english/structure_def.sql )
19:58 kados tumer: i must leave for about 30 minutes ...
19:58 kados tumer: I'll be back soon
19:59 tumer kados: have you any zebra running with koha?
20:56 thd kados: I think that tumer misunderstands the intended temporary use of 09o.  Following the example of RLIN, XXo and XXr were meant to hold original copy catalogued record content that would otherwise interfere with Koha local use fields until Koha fields were suitably remapped..
20:59 thd kados: Koha's use of a numeric test of < 10 for control fields should be replaced by a regex testing m/^00/
21:03 thd kados: 090 is the second most common place to find LC call numbers, standard for both RLIN and OCLC and formerly for USMARC.
21:07 thd kados: 001 is a single value control field and cannot be substituted for both biblio.biblionumber now in 090 $c and biblioitems.biblioitemsnumber now in 090 $d.
21:12 kados thd: agreed on all accounts
02:43 ToinS hi all
02:43 pierrick hi ToinS
02:44 ToinS pierrick: hello
02:58 hdl hello all
03:06 pierrick hi hdl
03:07 hdl pierrick: hello how are you ?
03:12 pierrick hdl, I'm fine thank you :-)
03:14 pierrick nice WE in Montpellier?
03:14 hdl busy Koha one ;)
03:14 pierrick productive?
03:28 hdl Not much for business, but for friendship yes :)
03:30 pierrick :-)
03:30 pierrick I'm currently trying to understand why Koha is so slow on my computer :-(
03:31 pierrick I've placed some indicators in the Perl code and Perl code execution is fast (less than 100 ms)
03:32 pierrick but when I launch a page, it needs 2 to 5 secondes to get displayed :-(
03:32 pierrick I've tried a PHP script and it displays very fast
03:33 pierrick There must be something with my Apache and Perl
03:41 hdl pierrick: perlversion and apache one ?
03:42 pierrick perl 5.8.7 and apache 2.0.54
04:00 pierrick I've added a time check before and after output_html_with_http_headers, this step is incredibly slow: more than 500ms on my computer
04:01 pierrick hdl, even when I execute the Perl script in command line, it is very slow :-/
04:02 pierrick (not so slow in fact, but more than one second to display the logon page)
04:13 chris pierrick
04:13 chris it might be worth using the profiler
04:13 chris have you used it before?
04:14 chris[…]rlnut/ch06_05.htm
04:14 chris its very useful for finding bottlenecks in code
04:19 chris for eg
04:19 chris on my computer at home
04:19 pierrick hi chris
04:19 chris perl -I /usr/local/koha/intranet/modules -d:DProf
04:20 pierrick I'm installing Apache 1.3.33 to check the problem doesn't come from Apache
04:20 chris hmm
04:20 chris right
04:20 chris dprofpp tmon.out
04:20 chris Total Elapsed Time = 0.305953 Seconds
04:20 chris  User+System Time = 0.285953 Seconds
04:20 pierrick chris, how do you have cookies running script by command line?
04:22 chris hmm im not sure, i know you can pass in variables
04:24 hdl chris: on my machine : koha-2.2 : 0.5 seconds on
04:25 chris %Time ExclSec CumulS #Calls sec/call Csec/c  Name
04:25 chris 27.9   0.080  0.080      4   0.0200 0.0199  C4::Date::BEGIN
04:25 hdl chris: on my machine : koha-dev : 1.3 seconds on
04:25 hdl dprof time
04:25 chris it seems to me that C4::Date
04:25 chris is what is slowing it down
04:25 chris which makes sense since that uses Date::Manio
04:25 chris Date::Manip even
04:26 hdl C4::Date sure, but ZOOM::Exception also.
04:26 chris right
04:26 hdl 59.2   0.670  0.670      3   0.2233 0.2233  Smart::Comments::__ANON__
04:26 hdl 7.96   0.090  0.100      4   0.0225 0.0249  C4::Date::BEGIN
04:26 hdl 3.54   0.040  1.141      7   0.0057 0.1630  main::BEGIN
04:26 hdl 2.65   0.030  0.030      9   0.0033 0.0033  base::import
04:26 chris ahh
04:26 hdl 1.77   0.020  1.002      5   0.0040 0.2005  C4::Reserves2::BEGIN
04:26 hdl 1.77   0.020  0.040      8   0.0025 0.0050  Smart::Comments::BEGIN
04:26 hdl 1.77   0.020  0.187      8   0.0025 0.0233  C4::Search::BEGIN
04:26 hdl 1.77   0.020  0.029      4   0.0049 0.0073  C4::Database::BEGIN
04:26 hdl 1.15   0.013  0.686      5   0.0026 0.1373  Filter::Simple::__ANON__
04:26 hdl 0.88   0.010  0.010      2   0.0050 0.0050  AutoLoader::AUTOLOAD
04:26 chris comment out the use smart::comments call
04:26 hdl 0.88   0.010  0.010      1   0.0100 0.0100  ZOOM::Exception::BEGIN
04:26 hdl
04:27 chris it hsould be quite a bit faster with use Smart::Comments; commented out
04:28 hdl Yes. I realized when copying.
04:28 chris :)
04:28 hdl pierrick: Do you have Smart::Comments on your machine ?
04:28 chris its good for testing, not good for production :)
04:29 pierrick hdl, I do
04:29 hdl chris: But ZOOM::Exception coud become another bottleneck.
04:29 chris yep it looks like it could be
04:30 chris maybe we should put profiling code as one of the tasks for the dev week also?
04:30 hdl pierrick: have you seen my dprof and Smart::Comments point ?
04:43 pierrick I've seen, but my problem doesn't come from modules load I think
04:44 pierrick I've tested the total generation time of a page from its beginning to its end and it's good, what's bad is the time to display in my browser
04:44 btoumi hi evrybody
04:44 pierrick hi btoumi
04:44 chris ahh
04:45 chris pierrick, does it make any difference if you change templates?
04:45 pierrick chris, my pages needs several seconds to display (from 2 to 5 at least)
04:45 pierrick working with PROG on HEAD or with default on 2.2 is slow
04:45 pierrick Forbidden
04:45 pierrick You don't have permission to access /cgi-bin/koha/ on this server.
04:45 pierrick Apache/1.3.33 Server at plegall Port 8080
04:46 pierrick :-/ don't understand what I've forgotten in my Apache 1.3.33 configuration
04:47 chris permissions ok?
04:47 pierrick same as with my Apache 2.0.54
04:47 chris pierrick, does it make any difference if you try a different browser?
04:48 pierrick chris, I go back to Apache 2 and test with a text browser
04:48 hdl hi btoumi
04:49 chris oh and hi btoumi :)
04:49 btoumi hi hdl
04:49 btoumi hi chris
04:51 pierrick chris, faster with w3m but still slow
04:52 chris so it looks like its something to do with it running as a CGI
04:52 pierrick chris, this is just a feeling, I have no times displayed
04:52 chris right
04:52 pierrick I would like to test with Apache 1.3.33 before saying it's a Perl problem
04:53 chris right
04:54 chris if its still slow with apache 1.3.33
04:54 chris what i would do is edit
04:54 chris change #!/usr/bin/perl
04:54 chris to #!/usr/bin/perl -d
04:54 chris and add
04:54 chris use Devel::DProf;
04:55 chris then when it runs, it will create a tmon.out file
04:55 chris and you can dprofpp it
04:56 chris and it might shed some light on what is slowing it down
04:58 chris when i do that and run it
04:59 chris %Time ExclSec CumulS #Calls sec/call Csec/c  Name
04:59 chris 41.7   0.020  0.020      5   0.0040 0.0040  CGI::_compile
04:59 chris thats the slowest thing when running it from apache
05:49 pierrick chris, still very slow with Apache 1.3.33 :-/ I'm trying the profiler you're suggesting
05:49 kados hi all
05:49 kados paul around?
05:49 pierrick hi kados
05:50 pierrick kados, last week, paul said he would be there the first 3 days of this week
05:50 kados today is monday, the first day of the week, right? :-)
05:50 pierrick right
05:51 ToinS paul will be not around the first 3 day !
05:52 ToinS he 'll come back on thursday
06:05 kados ahh ... I missunderstood pierrick's post :-)
06:05 kados hdl around?
06:06 kados 1. does LDAP support in Koha allow borrowing and reserving or does it only support authentication?
06:06 kados 2. has paul finished the thesaurus support he was working on?
06:55 hdl kados: I'm here now.
06:55 hdl kados: 1) pls detail.
06:56 pierrick kados, sorry I miswrote my post... s/would/wouldn't/ :-)
06:56 kados hdl: I know that LDAP is used for authentication, but I'm not sure if it can be used to check-out, reserve, etc., books to a patron if that patron doesn't exist in the borrowers table
06:58 kados hdl: is that detailed enough?
07:01 hdl kados: you should ask him via email. I know he worked on it, but more on authentication than intranet flags integration.
07:02 kados ok, I will email paul, thanks!
07:02 hdl 2) thesarus support you mean multi level hierarchy ?
07:02 hdl I am working on it.
07:03 hdl (or will be soon)
07:03 kados hdl: (ok, 2, it's enough, thanks)
07:03 kados hdl: mail sent to paul, when you see him, could you ask him to reply asap if possible? :-)
07:05 hdl I shall phone him dureing the week.
07:05 btoumi hi kados i'm bruno from SAN Ouest Provence
07:08 kados hi btoumi
07:08 kados btoumi: good to read you on IRC
07:09 btoumi thanks it's pleasure
07:12 btoumi hdl : paul is here?
07:13 hdl btoumi: No. Not until Thursday.
07:15 btoumi ok i do request to join koha developper group
07:20 kados btoumi: you can ask me :-)
07:20 kados btoumi: do you mean on savannah?
07:20 btoumi yes sorry i believe that u prepare your trip
07:21 kados btoumi: what's your username?
07:21 kados btoumi: on savannah?
07:21 btoumi yes
07:21 kados btoumi: if I have time I"ll add you
07:21 btoumi thanks
07:22 btoumi my username : btoumi
07:22 kados btoumi: ok
07:25 pierrick about my performances problem, using Apache 1.3.33 doesn't improve anything. I've tried using Devel::DProf (can't make it work through Apache) and it revealed Smart::Comments was very slow to load. I commented the "use" statement and now Koha is really faster
07:27 kados hehe
07:27 kados good detective work
07:27 kados Smart::Comments shouldn't be in rel_2_2
07:27 kados if it is, it should be commented out
07:28 kados pierrick: could you remove the 'use Smart::Comments' from CVS?
07:28 pierrick from HEAD?
07:28 pierrick or just in rel_2_2
07:28 kados just in rel_2_2
07:28 kados in HEAD it's ok
07:28 kados since HEAD is not near production-ready
07:29 pierrick I made my test with HEAd, not rel_2_2, I'm going to test the speed of rel_2_2
07:31 pierrick plegall:~/dev/koha/2.2$ find . -name "*.pm" -o -name "*.pl" | xargs grep Smart::Comment
07:31 pierrick revealed nothing :-)
07:32 kados ahh, ok
07:32 pierrick and rel_2_2 is really faster than HEAD with Smart::Comment
07:32 kados wow!
07:32 kados even searching?
07:32 pierrick hum... on rel_2_2, search works
08:59 btoumi msg chris
09:00 pierrick btoumi, I think chris is sleeping now
09:01 btoumi ok it's right i'm forget he leave in nz
09:34 pierrick hdl, are you around?
09:35 hdl yes
09:35 pierrick is it normal reserve can be linked to an item?
09:35 hdl Do they ?
09:35 pierrick in my imagination, it was linked to a biblio, not a specific item
09:36 pierrick in reserves table, you have biblionumber and itemnumber
09:36 hdl In fact, at first sight, it should be linked to a biblio.
09:36 pierrick the two can be set
09:36 hdl BUT.
09:36 hdl when you have a multiple branch library with independatBranches.
09:37 hdl you doNOT want ppl to reserve items on an other branch.
09:38 hdl Thus, reserves mustnot be only on biblio but must be available for an item ?
09:38 hdl .
09:38 hdl Is it ok for you ?
09:38 pierrick not at all, I don't understand what you say
09:38 pierrick on passe en francais si tu veux bien
09:38 hdl ok.
09:39 pierrick chaque branche est liee a une liste d'items
09:39 hdl Une bib a 3 sites.
09:39 hdl 3 sites indépedants.
09:39 pierrick continue...
09:39 hdl Elle ne veut pas que le lecteur du site1 puisse emprunter de livre du site2.
09:40 hdl Or la notice biblio n'est pas attachée à un site.
09:40 pierrick evidemment, ce sont les items qui sont attachés à un site (selon le bon sens)
09:40 hdl Si Site1 et Site2 ont chacune des exemplaires de Harry Potter.
09:41 hdl Il faut que ce soit lexemplaire du site 1 (si le lecteur est du site1) qui soit réservé.
09:41 pierrick je continue à ne pas comprendre la logique
09:41 hdl Explique ce que tu ne comprends pas alors.
09:41 pierrick une bibliothèque a 2 sites
09:42 pierrick le site 1 possède 3 copies du biblio 4321
09:42 pierrick le site 2 possède 4 copies du biblio 4321
09:43 pierrick le borrower 987, attaché au site 2 souhaite réserver le biblio 4321
09:43 pierrick ce serait malheureux qu'il doive attendre un item précis alors que 4 items sont disponibles sur ce site
09:44 pierrick je prends le problème à l'envers, mais je pense que le modèle de données est incohérent
09:45 pierrick je constate cette incohérence en tentant de corriger le bug 668
09:46 pierrick parce que lorsqu'on veut emprunter un item réservé, on supprime la réservation sur l'item alors qu'un réservation est théorique sur un biblio, donc l'annulation de résa ne marche qu'une fois sur 10.
09:49 pierrick a mon avis, avoir l'itemnumber dans la résa a du sens dans certains très très rare, mais si c'est pour résoudre les problèmes de branches, c'est pas la bonne façon de faire à mon avis
09:50 pierrick hdl, je loupe un truc ou ce que je dis a du sens?
09:51 pierrick hi tume
09:51 pierrick s/tume/tumer
09:51 hdl hi tumer.
09:52 tumer hi everybody
09:52 hdl pierrick: Certes, c'est peut-être dommage.
09:52 tumer pardon je ne parle pas français
09:53 hdl Mais mon cas, c'est Harry Potter : 1exemplaire dans le site1, Un autre dans le Site2
09:53 pierrick sorry tumer, we had a conversation specific to reserve
09:53 hdl tumer: sorry we are talking about reserves.
09:53 tumer I thought it was for me sorry
09:54 pierrick hdl, n'avoir qu'un exemplaire par site n'est il pas un raccourci réducteur à la solution des branches?
09:54 tumer Can anyone give me the name of hotel in Marseilles so I can book
09:54 tumer when you are free offcourse
09:54 pierrick tumer,[…]ooking_online.php
09:55 hdl C'est assez fréquent en Bib de recherche.
09:56 hdl En fait, la réservation des exemplaires m'a été demandé au moins trois fois.
09:56 pierrick hdl, je ne vais pas m'acharner, à mon sens c'est une erreur de conception et si ça ne te choque pas, alors je n'en ferai rien :-)
10:19 tumer pierrick: which hotel everyone else is staying
10:24 pierrick tumer, the one I gave you the URL
10:25 tumer pierrick:thanks

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