IRC log for #koha, 2007-10-21

All times shown according to UTC.

Time S Nick Message
12:12 fbcit ping kados
13:26 ansius ping *
13:27 owen Hi ansius
13:27 ansius hard to get anyone at this time as I live in europe
13:28 owen Yeah, Saturday morning here in the U.S., nighttime in New Zealand
13:29 ansius i have had trouble seeting up koha 2.2.9 as it dosnt like mysql5 there is an update cvs but i have no idea how to get it running
13:30 owen Sorry, that's not something I know how to help with
13:32 ansius dang... someone is social, but not skilled ;) ok, i have been told that there is a 2.2.10 version witch works with mysql5 but there is no pkg made :(
13:33 owen Yeah, work in the 2.2 branch is pretty slow now that 3.0 is the priority
13:34 ansius maby any sugestions, cos i really need get running. i could not manage to set up mysql4.1 on debian either
13:34 owen Have you searched through the mailing list archives, both koha and koha-devel ?
13:35 owen If you come back on Monday, you'll find the French developers around, and then later more Americans
13:40 ansius i have searched, haven't found the errors i got... it is something with worng sintax for install script using mysql (mysql5 is more picky) the problem is client wants to get the server runing on monday
13:42 owen There's always the more expensive option:
13:42 owen There's a couple of folks on that list who are on your continent
13:49 ansius thanks paul is on eto be bugged about nev releases (@kados told), right?
13:51 owen Paul is the 2.x release manager, if that's what you mean
13:55 ansius :] ok
14:08 martinmorris that problem you're having is quite easy to solve ansius
14:09 martinmorris it's because one of the table names is also a keyword in mysql5
14:09 martinmorris you just need to put ` (backticks) around it
14:09 martinmorris i think it's require but i can check if you want
14:10 ansius :) ok I can fix it , but mayby you happen to know line nr/file?
14:10 martinmorris give me a few secs
14:13 martinmorris koha.mysql line 780 on my system
14:13 martinmorris change return char(4) default NULL,
14:13 thd` kados: ping
14:13 martinmorris to `return` char(4) default NULL
14:13 martinmorris in the CREATE TABLE issues block
14:14 ansius it is all ready like that :|
14:15 martinmorris ah!
14:16 martinmorris have you posted the error you're getting?
14:17 ansius -------
14:17 ansius ERROR 1064 (42000) at line 1: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near '-koha.* to 'kohaadmin' IDENTIFIED BY 'syll@bu$'' at line 1
14:17 ansius ERROR 1045 (28000): Access denied for user 'kohaadmin'@'localhost' (using password: YES)
14:17 ansius DBI connect('kvk-koha:localhost','kohaadmin',...) failed: Access denied for user 'kohaadmin'@'localhost' (using password: YES) at /usr/local/koha/intranet/modules/C4/ line 411
14:17 ansius Can't call method "prepare" on an undefined value at scripts/updater/updatedatabase line 1553.
14:17 ansius Problem updating database...
14:17 ansius ----
14:18 martinmorris you're SURE you're using the right password?
14:19 ansius dang shure
14:19 ansius ERROR 1064 (42000) at line 1: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near '-koha.* to 'kohaadmin' IDENTIFIED BY 'syll@bu$� at line 1
14:19 martinmorris and that you've set up mysql to allow logins from kohaadmin@localhost and kohaadmin@<whateverotherhost>
14:19 ansius i'll chack is that ok
14:19 martinmorris ok
14:27 ansius thanks, gootit running, install script didn't creat correctly the user
14:27 martinmorris aaaah :)
14:38 ansius :( too soon to be happy, apache config file renders empty...
15:25 thd kados: ping
15:37 owen Poor kados, can't even have a quiet Saturday morning at home eating sugary cereal and watching cartoons.
16:09 fbcit-1 kados:ping
19:50 paul hello kados (working tonight...)
19:51 paul if I don't mind, you should have a fix from hdl for item editing. are you waiting for chris to push it or have I missed something ?
19:52 paul ok, just seing it now !
19:52 kados hi paul
19:53 kados paul: I just pushed it up
19:53 kados but with a comment
19:53 kados because additem is still badly broken
19:53 paul yep, i've seen the comment.
19:53 kados ok
19:53 paul checking it immediatly to see if I have the same problems
19:53 kados cool
19:57 paul kados : works fine for me. And:
19:57 paul home/jmf/koha/production/in​tranet/modules/C4/ line 432., referer:
19:57 paul is a comment on my code
19:58 owen Hi guys
19:58 paul what does this line contain ?
19:58 paul hi owen
19:58 kados paul:     &logaction(C4::Context->userenv->{​'number'},"CATALOGUING","MODIFY",$​itemnumber,$record->as_formatted)
19:58 paul owen : i've ported menus to ISBD /
19:59 owen paul: have you committed yet? I've been doing the same thing today
19:59 owen Frederic sent me some corrected javascript
19:59 paul nope, just doing it 10mn ago.
19:59 kados paul: I wonder why your bibio is dufferent than mine
19:59 paul no prob, feel free to commit it, i'll take your proposal.
19:59 kados hiya gmcharlt
19:59 gmcharlt hi kados
19:59 owen Yeah, I'll do it Paul. I've done a ton of toolbar work today and yesterday
19:59 paul kados : I think it's CataloguinLog that is ON for yo uand OFF for me.
20:00 kados paul: you have rebased recently?
20:00 kados ahh, could be
20:00 kados well, still, should be identical
20:00 paul kados : yep, 3mn ago ;-)
20:00 paul nope, i've some patches waiting...
20:00 kados ahh
20:00 kados paul: so you think that additem is fine, and that it's the log that is failing?
20:00 paul yep
20:01 kados ok, I'll try turning it off
20:01 owen kados: Are you ready to hear I've changed my mind again? I settled on the YUI toolbar after all.
20:01 kados hehe
20:01 paul owen : I prefer not to have too many js tools.
20:01 paul so this decision ++ for me
20:02 kados paul: ok, you may be right about the log ...
20:02 kados paul: but I still have a serious error:
20:02 paul yep, I have the error when Log=ON
20:02 kados two errors: one is foreign key constraing
20:02 kados the other is ERROR in _koha_modify_item
20:02 kados ahh, same error really
20:03 paul it's probably : C4::Context->userenv->{'number'}
20:04 paul that's it.
20:04 kados mf_koha/items CONSTRAINT items_ibfk_3  FOREIGN KEY ( holdingbranch REFERENCES branches branchcode
20:04 kados ON UPDATE CASCADE) at /home/jmf/koha/production/in​tranet/modules/C4/ line 3760
20:04 paul kados : you've just discovered a big security issue :-D
20:04 kados wonderful :-)
20:04 paul the Get_template_and_user is set AFTER the action.
20:04 kados yea, I knew about that one :-)
20:04 paul so the uservenv is NOT set.
20:04 paul thus the bug.
20:04 kados yep
20:06 paul ok, that works.
20:07 kados yep, for me too
20:07 paul kados : could you explain a little bit more the :
20:07 paul mf_koha/items CONSTRAINT items_ibfk_3  FOREIGN KEY ( holdingbranch REFERENCES branches branchcode
20:07 paul ?
20:07 paul (you commit one or I do ?)
20:08 kados you can
20:08 paul OK.
20:08 kados the constraints error is no longer there when get_template is moved up
20:11 kados paul: I'm still seeing some data loss with repetitive saves
20:13 kados it's to do with subfield 0
20:15 kados paul: any clue why the 'hidden' stuff isn't working for additem?
20:15 kados ie, hidden fields still show up
20:16 paul (seems that a not hidden subfield sometimes "disappear")
20:16 paul dunno why yet
20:19 paul seems hidden fields are NOT active on the 1st additem call. then, they work
20:20 paul (NOT : they are not hidden, but displayed)
20:20 kados er?
20:21 paul I have a field noted "hidden". when I reach, the 1st time it's not hidden, but shown. then, it's hidden.
20:21 kados owen++
20:25 ansius what is wrong if apache offers to download the *.pl files instead of exucating them, but i have perl.load file in modules-enabled?
20:26 paul ansius: look in your apache log. The exact error is hidden here.
20:26 paul for security reason the error is not displayed on the browser.
20:34 kados paul: is there a functional purpose to having repeatable subfields in additem?
20:34 kados paul: I'm very concerned about those, esp if they are linked to the koha items table
20:34 paul mmm... not sure. can't see one
20:34 kados IMO items.* should be authoritative
20:35 kados owen: you still around?
20:36 kados owen: I'm just wondering if you have any bright ideas about how to improve additem.tmpl when there are a lot of item fields
20:36 kados I'm wondering if this isn't a case where we want the label to appear above the input box
20:38 kados paul: I guess you didn't mean to attach the patch in your last email about additem, but you've sent it to, right?
20:39 paul not yet, but i'll send it before going to bed, for sure
20:39 kados ok :-)
20:39 ansius nothing in the log files... on rebload it says 'Apache/2.2.3 (Debian) mod_python/3.2.10 Python/2.4.4 PHP/5.2.0-8+etch7 mod_perl/2.0.2 Perl/v5.8.8 configured --$'
20:40 kados hiya fbcit-1
20:40 kados fbcit-1: you were after me earlier?
20:40 owen kados: we're just starting to shut things down here at The Plains, so we should talk about it another time
20:40 fbcit-1 kados: yep.
20:41 fbcit-1 kados: I have some questions for my conversion.
20:41 fbcit-1 hold on...
20:41 kados owen: sounds good, have a good weekend
20:46 fbcit-1 kados: pg requires that fk constraints reference values that are unique. So converting some of the fk references will require giving accountlines a pk....
20:47 fbcit-1 there are fk's that reference accountlines.borrowernumber
20:48 fbcit-1 is there any reason that borrownumber would contain duplicates?
20:48 kados paul is the one who did all of the fk stuff
20:48 kados no, borrowernumber should be unique, it isn't?
20:48 paul yes it is
20:48 paul south africa won !
20:48 kados paul: just fyi, fbcit-1 is working on a postgres port
20:49 fbcit-1  `borrowernumber` int(11) NOT NULL default 0,
20:49 kados yikes
20:51 paul there is a unique index on cardnumber, but no unique index on borrowernumber.
20:51 paul we should have a PK on it, undoubtfully !
20:51 paul (an that should be easy, as we have an auto_increment)
20:52 kados paul: can you add that to kohastructure.sql please?
20:52 paul not immediatly : time to go to bed for me.
20:52 paul fill a bug, i'll do it on monday
20:52 paul (i'll bump to I think)
20:52 fbcit-1 So it will not break if I make borrowernumber a pk in my pg port?
20:53 paul not at all fbcit-1
20:53 kados paul: ok
20:53 fbcit-1 paul: tnx!
20:53 paul borrowernumber SHOULD be a PK. it's a design error in mySQL it seems
20:53 ansius ok i looked alll ower the log files, and could not find any error releated that mod_perl vould have nay problems, it just dosen't exucate *.pl files, i think it is something with apache config, tough I don't get where
20:53 ansius sorry for bad english
20:54 kados paul:[…]w_bug.cgi?id=1505
20:54 kados paul: have a good night
20:54 paul thx
20:54 kados ansius: you might have more luck asking folks in #apache or #perl ... I dunno
20:54 kados ansius: it sounds like an apache conf prob
20:55 fbcit-1 kados: well, my conversion generates the table structures, indices, etc. I think only the fk's are left to do... at least in kohastructure.sql
20:55 kados sweet
20:55 kados fbcit-1++
20:55 kados fbcit-1: I hate to say this, but there's a lot more ahead :-)
20:55 fbcit-1 :-P
20:56 kados fbcit-1: so ... having done this in the past ... what do you recommend we do as a project to better support multi-DBMSes
20:56 fbcit-1 this should let me install the app and begin work on the code itself
20:56 kados should we have one kohastructure.sql that works for all DBMSes?
20:56 kados or one for postgres, one for mysql, .. ?
20:57 fbcit-1 I think a 'one size fits all' approach to the structure will be difficult at least at present.
20:58 fbcit-1 It seems I remember that mysql just went truly relational a short time ago?
20:58 gmcharlt kados, fbcit-1: its always been the oddball -- how long for transactions?
20:58 kados yea, finally got transactions in 5.x
20:59 kados subselects now too
20:59 gmcharlt anyway, gotta go -- see you later
20:59 kados gmcharlt: laters
21:00 fbcit-1 kados: you made a statement the other day which led me to believe that you all use some syntax unique to mysql.
21:00 kados yea, there are definitely some mysql-specific chunks in koha
21:00 kados usually for speed reasons
21:00 fbcit-1 Does mysql not have fulltext search capability?
21:01 kados so for instance, in mysql you can chunk inserts into one statement and it's about 100 times faster
21:01 kados yes, it does have
21:01 fbcit-1 why zebra?
21:01 fbcit-1 is it just faster?
21:01 kados well ...
21:02 kados it's faster and has support for standards-based queries on fielded metadata
21:02 kados the RDBMSes are getting support for fulltext, but it doesn't hold a candle to what's possible with zebra or lucene
21:03 kados plus zebra gives us native support for Z39.50 and SRW/U
21:03 kados which are important standards in library land
21:03 kados that said ... there's still a nozebra option
21:03 kados which you can use if you wan't want to bother with zebra
21:03 fbcit-1 ok. I am not familiar with library standards... :-(
21:04 kados well MARC is the bittie
21:04 kados biggie I mean
21:04 kados
21:04 kados
21:04 kados it's insane, but a prereq for libraries
21:07 fbcit-1 interesting reading
21:09 fbcit-1 back to the db question: is your sql embedded in your code or is it in a seperate module(s)?
21:09 kados it's embedded, though we've talked about pulling it out into a separate module
21:09 kados well ...
21:09 fbcit-1 does mysql support stored procedures?
21:10 kados the perl scripts shoudln't have any sql, but the perl modules in the C4 dir do
21:10 kados yes
21:10 fbcit-1 of course going to sp's would restrict the db selection more I would think.
21:11 kados *nod*
21:11 fbcit-1 sorry, I'm at home and only have the structure script with me so I can't get at the code to look.
21:12 kados no worries
21:13 fbcit-1 I'll probably have more thoughts on this after wading into the thick of the sql in koha.
21:13 fbcit-1 hopefully good thoughts.... ;-)
21:18 fbcit-1 I have another 'unique' field issue....
21:19 fbcit-1 table aqbookfund has a pk composed of (bookfundid,branchcode). neither field is set to UNIQUE.
21:19 fbcit-1 so... is bookfundid a unique value?
21:22 kados fbcit-1: yep, it should be
21:22 kados fbcit-1: if it's not it's likely a bug
21:24 fbcit-1 is it safe to assume that for any given record in aqorderbreakdown, the associated record in aqfund would have the same branchcode and bookfundid?
21:25 fbcit-1 *aqbookfund*
21:25 fbcit-1 sorry
21:26 kados ryan would be the one to ask there
21:33 thd kados: have you seen my patch from last week?
21:36 kados thd: yes, we need to talk abut that
21:36 kados thd: I've finally got the marc21 framework working
21:36 kados you'll want to rebase
21:37 thd ok
21:37 thd kados: does the patch not work now because it is out of synch?
21:38 thd kados: I created that patch from a fresh clone
21:38 kados thd: I don't know , it wasn't ever applied?
21:38 kados thd: do:
21:38 kados git fetch
21:38 kados git rebase origin
21:38 kados then remove all patch files and issue:
21:38 kados git format-patch origin
21:39 kados that will tell you if your patch hasn't been applied, because it will show up there
21:39 thd I have not done that today but what was wrong with the one sent?
21:39 kados what did the one you sent do?
21:40 thd it updated the mandatory and optional MARC 21 bibliographic frameworks
21:40 thd in installer
21:42 kados yes, I know that :-)
21:42 kados but what specifically were you updating in the mandatory one?
21:44 thd the missing items column, and the 2006 MARC 21 bibliographic update from LC.  Just what I had already told you
21:45 kados I think that patch was applied
21:45 kados you can check git
21:45 thd kados: the key issue is testing the simple frameworks since I do not have a local build to test for 3.0
21:46 thd kados: I had an email suggesting it was applied but yet it seemed not to have actually been applied when I looked at the file
21:46 kados thd:[…]7f2e17af1;hb=HEAD
21:47 kados big file :-)
21:47 kados thd: we should probably remove all of the commented lines
21:47 kados thd: the ones where you include o9o options, etc.
21:47 thd kados: the comments allow me to understand what I am doing
21:48 kados well, I wonder if you could keep them for your copy, and the project copy wouldn't include them
21:48 kados all the 95k stuff for instance
21:48 thd kados: the comments have been vital for maintaining the integrity of the process
21:48 ansius hi again.. still no luck with apache, i tried to follow manual, the same. changed apache error log to debug, nothing shows up (and sadly I'm only one in #apache ) I guss I doomed  
21:49 kados thd: check the above link, tell me if you see the fields that your patch included
21:49 thd kados: yet if no one ever looks at the comments they are completely harmless
21:51 kados thd: I'm also not clear on the utility of the u* fields
21:52 thd kados: if you read the comments carefully you would know :)
21:53 thd kados: they are from RLIN
21:53 kados thd: can't you keep whatever comments on 95k and etc., in a separate file?
21:53 kados instead of bloating the already HUGE framework file?
21:54 ryan maybe 'framework expansions' ?
21:54 thd kados: yet they are trivial in comparison to the full size of the file
21:54 ryan RLIN being an option?
21:54 kados I don't think RLIN is an option
21:55 thd exactly without such an option there is the potential for some possible user to loose data
21:55 kados thd: you also have some SQL statements in a comment that I don't think are necessary
21:55 kados thd: according to ryan, there's no longer any data loss
21:55 thd those, I just added
21:55 kados thd: for records that are imported
21:55 thd really:
21:55 thd what about in editing?
21:56 thd there never was a data loss problem when using
21:56 kados I suspect including all the RLIN fields in the MARC structure is a workaround
21:56 kados for a bug that should be fixed
21:56 ryan one should be able to get around losing the data now
21:57 ryan but will probably take a fair bit of effoert  on modbiblio.
21:58 kados anyone using RLIN won't be using modbiblio
21:58 thd ryan: what happens when a record is edited and a field is present in the data but missing from the bibliographic framework?
21:58 thd RLIN is becoming OCLC but the transition is not complete
21:59 thd there is also a significant group of libraries which withdrew from RLIN to avoid being absorbed by OCLC
22:00 kados thd: I also was forced to remove some plugins that don't exist in koha yet
22:00 kados thd: because they were causing the editor to fail
22:00 kados thd: I may add them back if I have time to write those plugins
22:01 kados thd: for such libraries, we should move their local use fields to a proper local use field, not the incorrectly used u* fields
22:01 ryan reading modbiblio, it looks like the intent is that you  should NOT  lose data that you haven't explicitly modified or added.
22:01 kados IMO
22:01 thd kados: I assumed that the editor would only fail if they were specifically invoked during an editing session
22:02 kados thd: no, it invokes them regardless
22:02 thd kados: RLIN used u and some other letters because there were no local use fields left to map to
22:02 kados well anyway, the current framework (default) in git appears to be working
22:03 thd kados: RLIN, OCLC, legacy CANMARC, etc. have mapped almost every numbered local use field
22:03 ryan on mod, only  framework-defined data is stored.
22:04 thd ryan: that was my reason to include anything which might be in use by any major system in the first place
22:05 ryan i would rather fix the bug than have huge frameworks
22:06 kados me too
22:06 chris ditto
22:06 thd have you tested to see whether the simple frameworks import?
22:07 kados thd: import?
22:08 thd simple is most likely to have introduced an SQL error when I updated for the 2006 MARC 21 bibliographic update
22:08 kados thd: they work when you use the installer if that's what you mean
22:08 thd kados: yes I meant that
22:08 kados well, the default framework does
22:08 kados I have only tested one framework
22:08 kados that's the default one that's on git
22:08 thd kados: I presumed that default worked but simple needs testing for SQL errors
22:09 kados and that one now works for the installer and also for adding/editing biblios and items
22:09 kados thd: the default works with patches that I sent to git
22:10 kados yea, that'd be ideal
22:11 thd kados: simple needs checking because I had to update about 9 frameworks for simple which is 9 chances for error
22:11 thd x however many fields were added per framework
22:13 kados thd: I've got to step out for a bit
22:13 thd kados: when will you be back?
22:13 kados thd: lets have a goal this week of finishing all the frameworks
22:14 thd kados: we need to discuss UNIMARC English
22:14 kados thd: I won't be back until tomorrow
22:14 kados thd: yes, we do
22:14 thd that is how it is ever an install option
22:14 kados yep, definitely something to discuss
22:15 thd kados: or you need to decide the design of that :)
22:15 kados thd: if you have some time, check out the default framework on git
22:15 thd have you changed it that is removed missing plugins?
22:15 kados I have removed missing plugins
22:15 kados the one on git is the most current working one
22:15 kados[…]7f2e17af1;hb=HEAD
22:16 kados here is the history of changes:
22:16 kados[…]7f2e17af1;hb=HEAD
22:16 thd kados: did you remove them from simple also?
22:16 kados it looks like your patch was applied 44 hours ago
22:16 kados based on the history
22:16 thd kados: you changed it after committing so I can read the diff in git?
22:17 kados yes, you can even read the diff on gitweb
22:17 kados at this link:
22:17 kados[…]7f2e17af1;hb=HEAD
22:17 kados ok, time for me to go
22:17 kados talk to you soon
22:17 thd kados: did you change simple also?
22:18 thd kados: when will you be back tomorrow?
22:18 kados thd: I didn't change simple
22:18 kados thd: I can't promise a time
22:19 thd which means almost never :)
22:19 kados thd: I fly to Hartford, at 4pm, so before 4pm and after 7pm I hope to be around
23:18 ansius help needed, the log post: sript returns with ERR 500
23:32 ansius who would have thought, world need permisions to read koha.config
23:42 thd ansius: the webserver user needs to be able to read koha.conf, however, it should not be accessible to a web client.
23:43 ansius no web cliet cant get to that dir
23:43 ansius after 3 days off @#$% i have got koha running, huh now starts second part configuring
23:44 thd ansius: having the web user read the config file should be much safer than having root start Koha :)
23:47 ansius i'm not a guru in *nix systems, usually i just set up CMS make a template, this ha been a stretch for me.
09:39 martinmorris morning, is there any update on when 2.2.10 and/or 3 will be out - I've not been following things for a bit
10:09 CGI913 hello all
10:09 CGI913 any one can help me installing koha on my windows xp machine??
10:11 CGI913 hello

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