IRC log for #koha, 2007-08-18

All times shown according to UTC.

Time S Nick Message
13:26 Brooke howdy
13:26 dewey what's up, Brooke
13:26 Brooke not much
13:26 Brooke how's it going?
14:00 Brooke howdy
14:03 slef hi all
14:03 slef /names
14:03 [K] *** #koha@FreeNode names: [K] rangi ru55el
14:03 kados hey slef
14:04 kados slef: we didn't announce the git move primarily because we're still leanring how to use it
14:04 kados slef:[…]lopment:git_usage
14:04 slef is CVS still live?
14:04 kados slef:
14:04 kados CVS still exists, but we're not using it
14:04 kados we meaning paul and liblime
14:05 kados looks like sanop have just committed some stuff there
14:05 kados which one of us will have to port to git
14:05 slef you're not using the serene tree which pulls from CVS?
14:05 paul kados : seems that what sanop commited today is just an indenting stuff
14:05 paul (hello kados & slef however)
14:05 slef paul: you mean indent changes only?
14:06 kados paul: ahh, ok, maybe we can just do perltidy?
14:06 paul slef : yes.
14:06 paul kados : maybe (even if I think hard to read perltidyed code)
14:06 paul i'll become more confident in the next weeks I think
14:06 slef (not to koha)
14:06 paul LART ?
14:07 slef Luser Attitude Readjustment Tool
14:07 slef Also known as a 'clue-by-four'
14:07 slef looks surprisingly like a plank
14:07 slef I don't think they'll do it again ;-)
14:09 paul kados : mail sent to your mailbox, about patches management.
14:11 kados paul: checking
14:12 paul hdl coming back from Dakar (senegal) today... (an IFLA conference :
14:12 kados cool
14:13 kados sweet
14:13 slef can anyone tell me whether this URL prompts them what to save as:
14:14 kados slef: I committed one fix to git that was causing that
14:14 kados it was because the content-type was wrong
14:15 kados[…]088b44dac9154fefb
14:16 kados this was the problem line:
14:16 kados +        -type    => 'text/html; charset=utf-8',
14:16 kados -        -type   => 'utf-8',
14:16 kados -        -cookie => $cookie
14:16 kados +        -type    => 'text/html; charset=utf-8',
14:16 kados sorr
14:16 kados sorry
14:16 kados my internet's acting up
14:17 slef owowow Lang just rude into the barrier horizontal
14:19 slef kados: that's a problem, but I don't think it causes this
14:19 slef can anyone tell me whether this URL prompts them what to save as:
14:19 slef[…]=&isbn=0521606322
14:20 kados not in firefox
14:20 slef (by the way, z3950/ seems not to check librarian authentication)
14:20 slef phone
14:20 paul not in ff for me either.
14:20 kados wait
14:20 kados if I click 'search internet' it does
14:24 paul same behaviour on konqueror.
14:24 paul the IES when you click may be an Auth problem...
14:28 slef off-phone
14:28 slef ah, yes, it does for me when I click search
14:32 slef the z39.50 search is erroring and the script is crashing
14:35 slef download is 0-bytes
14:35 slef download seems to be prompted by apache
14:35 slef so everyone's sort-of-right
14:36 kados paul: all your stuff should be pushed to
14:36 kados paul: can you verify?
14:37 paul (but my suggestion about mails is still valid...)
14:37 kados yes I agree
14:37 kados we also need a better test procedure, between chris and I
14:37 slef USA's Leipheimer sets new best time (individual time trial)
14:37 kados because IMO every piece of code should be tested before pushed to
14:37 kados not just eyeballed
14:38 paul that will be a huge huge stuff...
14:38 slef leader Jens Voigt just set a best time at the intermediate check, though...
14:38 slef kados++
14:40 slef hrm, even if I feed my z search garbage, it doesn't break like that
14:40 kados paul: what do you think the role of updatedatabase is?
14:41 paul update the datatabase between 2 versions
14:41 paul minor or major ones.
14:41 paul except "super major" (koha 3.x => koha 4.x)
14:42 kados afaik, the updatedatabase in 3.0 curently is worthless
14:42 paul why ?
14:42 kados unless you know something I don't
14:42 kados well, #1, we don't have  stable database definition
14:42 kados for 3.0
14:42 kados and #2, it has old database defs from 2.2, etc.
14:43 paul it does 90% of the upgrade from 2.2 to 3.0. The 10% missing being the specific stuff to create marcxml and move some fields.
14:44 kados one library, that uses a small subset of features
14:44 paul Idid EXACTLY what is written on : 22_to_30 wiki page
14:44 kados and only unimarc ;-)
14:44 paul yep, but the unimarc specific stuff is really small.
14:45 kados I bet that:
14:45 kados 1. 2.2 -> updatedatabase -> 3.0
14:45 kados 2. koha < kohastructure.sql (via installer)
14:45 kados 1 != 2
14:47 Brooke grrr
14:47 kados I bet that  2.2 -> updatedatabase -> 3.0 results in a different database structure than kohastructure.sql
14:47 Brooke If I had $ I'd take a French course
14:47 Brooke and a CSI course
14:47 kados heh
14:47 paul (CSI course ???)
14:47 Brooke Informatique
14:48 paul in some indexes & maybe a few foreign keys you're probably right.
14:48 paul but not more than this (or i made something wrong, which is always possible ;-) )
14:49 kados Brooke: would you be willing to test migrating from 2.2. to 3.0?
14:50 kados Brooke: on a non-production system of course
14:50 Brooke what's this non production system you speak of?
14:50 paul I have 3 or 4 migrations from 22 to 30 to come in september...
14:50 kados hehe
14:50 Brooke I don't see why not
14:50 Brooke I can nag me husband to do it
14:51 Brooke are the files out there someplace in the aether?
14:52 kados Brooke: sort of
14:52 kados Brooke:
14:53 kados we don't have a release packaged yet
14:53 kados slef: how's the installer coming?
14:53 slef kados: working on it this afternoon after text/x-perl
14:53 kados slef++
14:53 paul slef +++++
14:53 kados paul: how is re-writing borrower management coming? :-)
14:54 kados paul: and have you seen bug reports from owen?
14:54 paul ??? you're waiting for me to do it ???
14:54 paul yes about the bug reports from owen.
14:54 kados paul: and have you tested the authorities editor, we can't get it working
14:54 kados ahh, I thought you were re-writing borrowers
14:54 paul mmm... works fine on my setup...
14:54 kados I may have missunderstood
14:55 paul http://i15.bureau.paulpoulain.[…]es/
14:55 paul (abel/abel)
14:55 paul no, I thought YOU were rewritting borrowers.
14:55 kados hehe
14:56 kados paul: would you like to see what happens running the installer, importing authorities, and trying
14:56 paul you showed me something few months ago, saying the code size dropped from 1500 to 250 lines iirc
14:56 kados I will give you access to my vhost
14:57 paul hdl should have some time to devote to you next week
14:57 kados 1. remove all features that are broken
14:57 kados 2. delay the release further
14:57 paul 3. release a beta, with the list of what is broken, and delay the stable release a little
14:58 kados I'd consider #3, but not with major features broken
14:58 paul 1. seems impossible to me...
14:59 Brooke I like 4
14:59 Brooke opps
14:59 Brooke 3
14:59 kados if minor parts of a feature are not working, that's fine
14:59 kados but when major features are just throwing 500 errors by default
14:59 kados that's just going to spread FUD
15:00 slef we need a general 500-catcher
15:00 kados slef: people just need to test their code more :-)
15:00 paul wow... 500 errors ? I haven't seen that since months. Maybe it's because i'm working with a 22 => 3.0 database...
15:00 slef kados: what works one place won't work for another - witness this z problem
15:01 kados paul: and that's why we have a problem :-)
15:01 kados paul: that's why I don't trust updatedatabase :-)
15:01 kados I think that kohastructure.sql != updatedatabase
15:01 paul why do you trust more kohastructure.sql than 22+updatedatabase ?
15:01 kados there should be one authoritative place for the sql definition for 3.0
15:02 kados not two
15:02 kados paul: I only trust it more because it's readable
15:02 kados paul: updatedatabase isn't readable without spending 3 hours reading it :-)
15:02 kados I can't just go look up the structure of table X for instnace
15:05 slef you can, but it's in N parts
15:06 kados yea
15:06 kados the point is, we're all working off a different database structure
15:08 paul kados : would it be usefull for you to have ipt sql structure ?
15:08 kados yep
15:08 kados and sanop too
15:08 kados and whatever other koha systems you have installed too
15:11 kados paul: features should work after a stock installation IMO
15:12 paul stock ?
15:12 kados when I run the web installer on a brand new install
15:12 kados all major modules should compile and run the basic functions of the module
15:12 kados that's my definition of a 'beta'
15:12 paul agree.
15:13 Brooke hooray agreement
15:13 Brooke :)
15:13 kados stable is when all major modules compile, and all functions work
15:13 kados when I can run the installer and have all features working, I'll release 3.0 beta
15:13 Brooke sooo emmmm
15:13 Brooke as a whiny user
15:13 Brooke when should we stop getting listened to in favour of the current features getting ironed out before new features are added?
15:14 kados paul: because it was broken
15:14 Brooke I mean, you guys seem too nice to us.
15:14 paul Brooke: that's difficult, because libraries are always OK to fund new features, never to sponsor fixing ones :\
15:14 kados paul: it caused the web installer to crash
15:14 kados paul: you can try it from the CVS version if you like
15:15 Brooke *nod*
15:15 Brooke I realise that you have to eat
15:15 paul how did you build the actual kohastructure.sql ?
15:15 Brooke but should there be a point where folks scream no more new features til we fix the old?
15:16 kados paul: I just reverted and I've been adding minor changes manually since then
15:16 kados IMO, from here on out, the kohastructure.sql should be edited manually
15:17 kados and not the restul of a sql dump
15:17 kados thx
15:18 kados slef: you voiced an opinion on this on the koha-devel list IIRC
15:18 Brooke hmmmm
15:18 Brooke I have a perhaps stupid ide
15:18 Brooke a
15:18 paul i've kdiff3-ed the 2 sql. there are a lot of differences :
15:18 Brooke if Libraries sign on for a new feature
15:18 Brooke can part of the fee automatically go to fixing an old?
15:18 paul - 80% of them are varchar <=> char. not a problem, but strange. Maybe a consequence of a mySQL setup ?
15:19 paul - some are table that can be removes (5-6 on my side, 2-3 on your side)
15:19 paul - some are foreign keys missing on my side
15:19 paul - 3 missing columns on my side (the one recently added by hdl)
15:20 paul only the foreign keys are important I think
15:22 kados paul: how do we know what tables can be removed?
15:22 paul because they are useless :-D
15:22 paul like "websites"
15:22 paul and marc_biblio
15:22 dewey somebody said marc_biblio was where you can find the relationship between marc_subfield_table.bibid and biblioitems.biblionumber in the 2.2 series of Koha
15:22 paul wait, wait...
15:23 paul all my foreign key constraints are at the end of the dump.
15:23 paul I don't know why...
15:23 paul so there may be only a few diffs...
15:23 paul kados : I suggest :
15:23 paul -create database ipt
15:23 paul - mysql ipt <ipt3(2).sql
15:23 paul - mtysqldump ipt >ipt3(3).sql
15:24 paul -mysqldump yourdb >yourdb.sql
15:24 paul - kdiff3 ipt3(3).sql yourdb.sql
15:24 kados *nod*
15:24 paul so you'll have the same dump look...
15:24 paul (you'll see some mysql text in french at the en as well ;-) )
15:26 kados slef: what were the mysqldump options you preferred?
15:26 slef Yes, I think the sql should be hand-maintained along with a changelog written in sql or perl as far as possible
15:26 slef erm
15:28 slef -cdQ at least, maybe --compatible=ansi if you're using mysql-5
15:30 slef does/how does relate to the serene git?
15:30 kados slef: serene git?
15:31 slef git://
15:31 kados slef: I don't think it relates at all
15:31 slef Announced in[…]-06/msg00008.html
15:31 kados slef: you'd ahve to ask chris though
15:31 kados I think we did a cvs convert ourselves
15:32 slef well that makes my life less simple...
15:32 kados best wait for chris next week
15:32 kados he can tell you how he did it
15:32 kados I wasn't around that day
15:33 slef can you please announce developments to koha-devel?
15:33 slef even if they're not ready/final
15:34 kados slef: I plan to announce git to koha-devel this weekend
15:34 slef else it burns others
15:35 slef but you said we need chris next week to say how it relates to users of the current git?
15:37 kados who uses the current git?
15:37 kados I didn't realize that was a production git
15:37 slef don't know... it's public and I've not been following the logs
15:38 kados paul: importing iptdb now
15:39 kados paul: will mysqldump and diff shortly
15:40 slef hrm, that was probably an error... doesn't look like it's logged
15:44 kados paul: would it be worth having a meeting to go through kohastructure in detail and make choices?
15:45 kados even things like meduimtext vs text should be decided for official 3.0
15:45 kados and what default values should be
15:46 kados and which constraints to use
15:46 kados varchar vs char
15:47 kados there are column differences for some tables too
15:47 kados paul: ?
15:48 Brooke I'm off like a prom dress
15:48 Brooke I've to go to "work"
15:48 kados hehe
15:48 kados Brooke: cya
15:49 paul varchar vs char = I think it's a mysql version choice. iirc, for char <=3, it doen's matter, mysql does always char. otherwise, it's just the way mySQL stores the value internally.
15:49 paul mediumtext vs text = should always be longtext imho
15:49 paul as text is SMALLtext iirc...
15:49 paul I had some problems storing long marcxml (with +100 items iirc)
15:49 kados paul: shouldn't we have a meeting and step through each table?
15:50 kados I see about twice as many constraints in ipt
15:51 kados slef: ?
15:51 paul otherwise, agree about a meeting, and constraint definitions.
15:51 paul twice as many constraints in ipt ???
15:51 kados quite a few more are defined in ipt db
15:56 kados paul:
15:56 kados paul: for instance, look at the constraints on accountlines
15:57 kados paul: compare to
15:57 paul constraints missing should not cause Koha bugs externally. We may have problems with constraints, not without. (but that's fine : we can detect bugs or errors thanks to constraints)
15:57 paul sorry guys, i've been disconnected...
15:57 paul I said :
15:57 paul constraints missing should not cause Koha bugs externally. We may have problems with constraints, not without. (but that's fine : we can detect bugs or errors thanks to constraints)
15:57 paul jmf : your kohastructure is marked "mySQL 4.1.x". Aren't you using mySQL 5 ?
15:58 paul (that may be an explanation too)
15:58 kados the kohastructure is not 'mine'
15:58 paul and havent recieved any answer
15:58 kados I didn't create it
15:58 kados answer?
15:58 dewey answer is usually yes
15:59 paul nothing after my sentence (1mn disconnected, I may have missed something)
15:59 kados can you look at the kohastructure-ipt.sql
15:59 paul where ?
15:59 paul on my desktop ?
15:59 kados maybe irc is acting up :-)
15:59 kados
15:59 kados kohastructure-ipt.sql
15:59 kados oops
15:59 kados
16:00 slef kados?
16:00 dewey kados is probably champion alligator-wrestler in 7 counties
16:00 kados look at the 5 constraints on accountlines for instance
16:00 kados dewey: botsnack
16:00 dewey :)
16:00 kados paul: are all 5 of them meaningful?
16:00 kados it seems to me only 2 are
16:01 kados slef: do you think we should have a meeting to go through the kohastructure?
16:01 slef sounds a good idea, but paul is away 10 days?
16:02 kados yea, maybe when he gets back?
16:02 slef so two weeks today?
16:02 kados that works for me
16:02 kados paul: is that a good date for you too?
16:02 paul kados : right, fk2, fk3, fk4 and fk5 are the same. silly
16:02 paul but probably not causing harm
16:03 paul aug, 31 then ?
16:03 paul OK for me.
16:03 kados cool
16:03 kados I'll announce it on koha-devel
16:04 kados paul: have a good trip
16:04 paul what time ?
16:04 dewey somebody said time was 648
16:04 paul 18GMT ?
16:04 paul (6PM GMT)
16:04 paul ok for me, when you want
16:04 kados 18:00 GMT works for me
16:05 kados slef: ?
16:05 kados earlier for slef?
16:05 paul okfor me
16:05 slef before 16GMT if possible, but I'll do 1800 if there's no better time
16:05 paul except 5GMT,because it's dinner time.
16:05 paul 4PM GMT is OK
16:05 kados for me too
16:05 slef paul: you eat dinner before breakfast? ;-)
16:06 paul :-D
16:07 kados slef: I'd like to talk abotu the installer when you have a chance
16:07 kados slef: is today good for that?
16:08 slef yes, as soon as I've finished slapping this disk and then kicked the other storage
16:08 paul bye bye kados, see you in 2 weeks then
16:08 slef bye paul... safe journey
16:36 slef ok, I think that's all working now
16:36 slef I know I've an email about the installer awaiting reply, by the way
16:37 slef but I'll talk anyway
16:37 kados yea, I wasn't sure how to respond to that email actually
16:37 kados lemme pull it up
16:38 kados what's the subject again?
16:38 kados Development Schedule? that one?
16:39 slef searching...
16:42 slef Message-ID: <>
16:43 slef sorry for the delay
16:44 kados my mail client doesn't work wtih message ids
16:44 kados subject?
16:44 slef the one you said
16:44 slef Development Schedule
16:44 kados k
16:44 slef oh, I see another tab
16:45 kados I did respond btw ;-)
16:45 slef seems old
16:45 kados Jul 9th
16:45 slef "more coming soon"?
16:45 kados yea
16:45 kados this is it mate :-)
16:45 slef is this the more?
16:45 slef ok
16:45 slef cool
16:46 kados so what do you think about the comments first off
16:46 slef I've stopped it running the tests and I disagree that it's the installer's job to be the INSTALL instructions
16:46 kados ahh, ok
16:46 kados so lets identify the elements of an install and roles of those elements
16:46 slef but anyway, if we make this CPAN-friendly, the installer should be able to trigger the installation of the prereqs itself
16:46 kados to make sure we have expectations set properly
16:47 kados triggering would be nice
16:47 kados there are some things that can't be installed from CPAN obviously
16:47 kados like zebra
16:48 slef indeed.  should be a 1% job to add those to the packaging
16:48 kados do we test for those things currently?
16:48 slef not directly, but could
16:48 kados there is a list of them in the file
16:48 kados but I'm not sure if hdl's method of testing is portable
16:49 slef what would I look for? A zebra binary?
16:49 kados and whether there is a portable way of doing that anyway ...
16:49 kados that's a good question
16:49 kados I'm not sure how to handle that in a portable way
16:49 slef might be safer to leave it to the packaging system
16:49 slef (s)
16:51 kados slef: ping me when you're back
17:13 thd kados: have we switched to git?
17:13 kados thd: we have
17:14 kados thd: announcement forthcoming
17:14 thd kados: so my CVS commits will have no effect?
17:15 kados well, someone has to manually merge them with git
17:15 kados so far there have only been 5-6 or so
17:17 thd I am about to commit another 6
17:20 kados hmmm
17:20 kados thd: do you have git installed?
17:25 slef phone
17:25 thd not yet
18:33 thd kados: so I have committed to CVS.  I will have to learn git later.  marc21_simple_bib_frameworks.sql now has authority links in CVS.
18:38 kados thanks thd
18:39 kados thd: does it also have the extra column for default value?
18:39 thd kados: the copy in HEAD does
18:42 thd kados: all 10 frameworks files for rel_2_2 and HEAD have been updated
18:43 thd s/10/10 MARC 21/
18:44 thd kados: the only thing missing is last year's MARC 21 bibliographic format changes which were announced a few months ago.
18:53 kados thd: I'll update git
18:53 kados thd: thanks!
20:06 slef hahahaha
20:07 slef ---- Unsatisfied dependencies detected during [K/KO/KOHA/koha-2.9.2007081719.tar.gz] -----
20:07 slef Shall I follow them and prepend them to the queue of modules we are processing right now? [yes]
20:07 slef now that's what I call music!
04:15 [K] *** join #koha@FreeNode: ru55el

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