Time  Nick     Message
11:00 owen     ...something we /didn't/ do with the OPAC and the MARC/ISBD views...
11:00 kados    yea
11:00 kados    that's exactly what I was thinking
11:00 kados    also, regarding the OPAC
11:01 kados    I'm thinking we should just piggy back on a library like YUI at least for layout
11:01 kados    rather than try to roll our own
11:01 kados    maybe for tabs too
11:02 kados    ie, I really like our OPAC design for dev_week, but we could probably make it more browser compliant in 3.0 if we used selected components in YUI
11:02 kados    anyway ... that's my current thinking :-)
11:02 kados    subject to change of course :_)
11:02 owen     Yeah, that sounds sensible if we go with YUI for the intranet.
11:03 kados    I think what we need to do is have a focused meeting to discuss intranet design
11:03 kados    ie, how should this thing actually look :-)
11:05 kados    tina and I were talking yesterday and she had an interesting idea that I was a bit skeptical of but is probably worth mentioning
11:06 kados    rather than have three input boxes on the left-nav, what about a single box that could be used for all three
11:06 kados    and some way (maybe a tab or a drop down) that could switch which function it is performing
11:06 kados    (or an access key)
11:07 owen     I think that's a little too much keyboard interaction for those inputs
11:08 owen     Is she worried about the amount of stuff in that menu?
11:08 kados    yea, a bit
11:08 kados    and in the demos we've done, people tend to be a bit overwhelmed by it
11:10 owen     Yeah, it's overwhelming if you're using an unrestricted user. It's a lot more manageable if you're logged in as a circ-staff user.
11:11 kados    yea, that's true
11:16 owen     Before we get together for a meeting I'd like to have plenty of time to get to know the current status of the 3.0 interface. But yeah, I think getting together to discuss it would be a great idea.
11:19 rch      hi owen
11:19 rch      hi hdl
11:19 owen     hi rch
11:19 kados    hdl: I'm writing an agenda for our 15 minute Installer meeting on the wiki
11:24 kados    hdl: http://wiki.koha.org/doku.php?id=agendandnotes07mar12
11:25 kados    hdl: is now a good time to have a 15 minute meeting to address those issues?
11:26 hdl      yes
11:27 hdl      let's do it
11:27 kados    great
11:27 rch      hdl: do you have userflags sql?
11:27 rch      i didn't see them in fr/
11:28 hdl      Actually, we donot have translated them yet.
11:29 hdl      But we shave to
11:29 hdl      But we have to
11:29 kados    hdl: will that fix the duplication problem I'm seeing?
11:30 rch      kados & i were discussing adding primary key to prevent duplication
11:30 hdl      I donot know. I don't see your problem.
11:30 rch      vs. testing in installer
11:30 rch      -- if you run installer twice :)
11:30 kados    hdl: if I run the installer more than once I get duplicates in user flags and sysprefs
11:31 kados    hdl: it doesn't happen for you?
11:31 hdl      I didnot notice this problem.
11:31 hdl      But since you say there are no primary key, this must happen.
11:32 kados    so is the best solution to add a primary key to both tables?
11:32 kados    if so, who will do it?
11:32 kados    if not, what is another solution?
11:33 hdl      Maybe adding primary keys to Varaible names and userflags flag woud be a solution.
11:33 hdl      You can do it.
11:34 rch      seems easier than testing
11:34 rch      i'll do it
11:34 hdl      Is this the best solution. I donot know.
11:34 hdl      Do we still need bibliothesaurus table ?
11:34 kados    bibliothesaurus table had a fulltext index, is it OK it’s removed?
11:35 hdl      I think no.
11:35 kados    if we dont' need it lets remove it
11:35 kados    who will find out and take care of it?
11:35 hdl      It was to compile subjects.
11:36 hdl      I donot have time
11:36 rch      so it is supposed to be a myisam table
11:37 hdl      I think this table is disposable
11:38 hdl      it was used in Koha1.0 to get subjects.
11:38 rch      ok, so we can drop it from the sql definition
11:39 kados    great
11:39 kados    so next:
11:39 kados    how to signal that the install’s already been done
11:40 rch      the chmod doesn't work for me.
11:40 hdl      Adding a value to koha.conf file would be ok for me.
11:41 rch      could update 'koha version'
11:41 hdl      <webinstaller>kohaversion</webinstaller>
11:41 kados    where is koha version stored now?
11:42 hdl      nowhere.
11:42 kados    hmmm, I thought we had a version string somewhere
11:42 kados    it's listed on the about page, right?
11:43 hdl      yes.
11:43 hdl      Was stored in kohaversion.
11:43 kados    can we just use that?
11:44 kados    use a specific string like:
11:44 hdl      But in xml I donot know.
11:44 kados    webinstall-3.0.1 or sth
11:45 rch      i think this sounds ok
11:45 hdl      ok to me
11:45 rch      then user manually removes installer dir
11:48 kados    yep
11:48 kados    will web installer be used to handle upgrades?
11:49 hdl      i think it can.
11:51 rch      hdl: what's the status of the migration_tools/22_to_30/ stuff?
11:52 kados    to use the installer for upgrades we will need some additions
11:53 kados    1. it will need to tell the user that 'koha's installed, there's a newer version, would you like to install it?'
11:55 rch      i'm not sure about 22->3.0 ... seems better to run a series of cmd line scripts.
11:55 kados    yea, I agree
11:55 rch      likely many warnings log stuff
11:55 kados    but for 3.0.0 -> 3.0.1
11:55 rch      right
11:55 kados    we could use the web installer if it was tuned a bit
11:56 kados    so ...
11:56 kados    what's our conclusion on this point?
11:57 kados    I'd like to wrap up this mtg in 3 minutes :-)
11:57 rch      okay, so webinstaller tests version number...
11:58 rch      otherwise, that logic isn't needed until 3.0.1, right?
11:58 hdl      was on phone  with paum.
11:58 hdl      was on phone  with paul.
11:58 hdl      sory.
11:59 hdl      rch : for 22 to 3
11:59 hdl      I agree it is not webinstaller's duty.
11:59 hdl      but console admin tools.
11:59 hdl      (paul greets you all)
12:00 kados    (hi paul)
12:01 hdl      where to put images : in usual images dir
12:02 rch      and how important is platform dependent code?
12:02 kados    it's very important
12:02 kados    otherwise we alienate windows and mac users
12:03 hdl      But I can't give you any good solution.
12:03 rch      but noone's going to run zebra  on those platforms
12:03 kados    tumer does
12:03 rch      really?
12:03 kados    yep, NEU is a windows show
12:04 rch      personally, i don't think we need to support ms.
12:04 rch      tho i agree that the installer's methods could be cleaner
12:04 hdl      If tumer can tell us about good ways to find an discover exe files on a windows bos, ok for me, but I cannot spend to much time on this issue.
12:05 rch      i don't know anything about windows platform.
12:05 hdl      You can make proposition. I would be glad to see something taht you consider clean.
12:05 hdl      I did what I could in the time frame I had.
12:06 hdl      I did what seemed to me the basic assumptions one could do.
12:07 rch      hdl: by 'cleaner', i meant less OS-specific :)
12:07 hdl      Do not presume any software is installed or even installed the simpler way.
12:07 hdl      It works on any Linux *BSD box :D
12:07 rch      yes, to make a really robust installer is a very big job
12:08 hdl      And I donot have any experience in OSX or windows web installation services.
12:09 hdl      Maybe ActiveX for windows could do the job. On OSX, I bet that dev tools can provide ppl with 'find' and 'which' commands.
12:10 rch      right...  well we'll have to as k tumer for windows stuff.
12:11 hdl      kados : is it ok to consider installer can be used for upgrades over 30 jumping to step3 ?
12:11 hdl      I can tweak it to make it work.
12:12 rch      that makes sense to me
12:14 rch      kados & i were discussing defining install files in xml file instead of reading directory contents.
12:14 rch      that might make upgrades easier to manage, too
12:16 kados    yep
12:17 kados    well one problem we have now
12:17 kados    is that the order of the files is not controleld
12:17 kados    it's stored in a hash
12:17 hdl      XML seems to be VERY powerful and hepful. Maybe the way to go.
12:18 kados    so when you install Koha, sometimes it will cause errors because of constraints
12:18 hdl      kados: We could provide numbers as 1st letters of files.
12:18 hdl      That way, they would be ordered.
12:19 rch      that's what we did...
12:19 hdl      pb would be for required and optional.
12:19 rch      but they're still stored in a hash
12:19 hdl      and hash is not ordered ?
12:20 rch      no guarantees
12:20 rch      it of course can be sorted
12:22 hdl      rch : where is the hash ?
12:22 dewey    the hash is not ordered ?
12:22 hdl      dewey : forget heash
12:22 dewey    hdl, I didn't have anything matching heash
12:22 hdl      dewey : forget hash
12:22 dewey    hdl: I forgot hash
12:23 hdl      rch : OK. seen.
12:25 rch      ah, looking at it now, actually it's an array
12:25 rch      can sort it easily on substring.
12:26 hdl      rch : foreach my $file ($query->param('framework')){
12:26 hdl      That line ?
12:27 hdl      We have to order on basename + directory.
12:29 rch      right
12:29 rch      but if it used xml instead, could it more readily specify upgrade paths as well as install order?
12:30 rch      we are okay for now on ordering.
12:35 rch      so i am still interested in how to redirect user to installer.
12:35 rch      so you can't go straight to opac before installation
12:43 hdl      rch : why not having http redirect in C4::Auth if no kohaversion in Config file ?
12:51 kados    that sounds good
12:52 kados    can kohaversion be a syspref rather than stored in the config file?
12:52 kados    would it be faster to read it then?
12:55 rch      i prefer having original redirect to installer, changed after install.
12:56 rch      perhaps install step 1: install symbolic link in apache to koha-httpd.conf
12:57 rch      at end of install, we rewrite koha-httpd.conf, and instruct to restart web server.
13:29 kados    hdl: you there still?
13:29 hdl      yes
13:29 kados    hdl: rch and I are looking at kohastructure.sql and we were wondering if it would be OK if we use mysqldump to generate it
13:29 kados    the only difference is that constraints are stored inside each table def
13:29 hdl      I think so.
13:30 kados    is there any argument for stroring them at the end instead?
13:30 hdl      kados : about kohaversion stored in the sysprefs, I prefer having it in conf file. Thus, it will be read only once with mod_perl.
13:31 kados    good point
13:31 kados    ok, lets store it in conf file then
13:31 kados    but it means we need to write to the conf file
13:32 hdl      Yes. I think I'l have to cope with it.
13:32 kados    I have another question
13:32 kados    why do we use a system call to mysql instead of just using dbi in the installer?
13:33 hdl      But it could be dangerous to let conf file writable for apache.
13:33 kados    true
13:33 kados    what about a simple installer conf file?
13:33 kados    installer/install.conf or sth?
13:33 kados    it coudl also be used to store directory information
13:34 kados    about what files are in sql-datas, etc.
13:34 hdl      kados : sytemcall to mysql instead of DBI because when using DBI, SOURCING File couldnot be achieved. (Was it because apache is chrooted ? was it for another matter, I donot know.)
13:35 hdl      May be becaus file was too big.
13:35 hdl      I tried at first to use dbh->do("SOURCE $myfile")
13:35 hdl      [14:20:01] <hdl> But that failed miserably.
13:36 kados    hmmm
13:52 kados    ok, so that won't be changed then
13:52 kados    there are two tasks I'm not sure who's taking them
13:52 kados    1. preventing opac and intranet from being visible before installer is run
13:52 kados    2. signal to installer that install has already been done
13:53 rch      hdl: was the failure on source for kohastructure.sql or for frameworks?
13:53 rch      or both?
13:53 dewey    both are pretty secure though
13:53 hdl      kohastructure.sql
13:53 kados    dewey: forget both
13:53 dewey    kados: I forgot both
13:53 kados    so who's taking #1?
13:54 hdl      both are linked.
13:54 kados    yep
13:54 kados    so what is the roadmap to completing them?
13:54 kados    and who is taking on responsibility to complete?
13:54 hdl      I thought I would take them. But if you have time to do them.
13:55 hdl      I won't be able to do it until late this week.
13:55 kados    ok, that's fine
13:55 kados    you can take them for later this week
13:56 kados    I'm going to compose a mail to koha-devel
14:19 kados    summary to koha-devel sent
14:44 kados    hdl: could you explain what 'cities' and 'roads' and 'departments' are in 3.0?
15:04 hdl      cities are for towns
15:04 hdl      roads should be Street, Boulevards, ...
15:05 hdl      departments are Bouche-du-Rhône (around Marseille), Hérault (around Montpellier) They have different zipcodes :
15:06 hdl      13000 for Marseille
15:06 hdl      34XXX for Hérault
15:06 hdl      75XXX for Paris
15:07 hdl      kados : is it clear ?
15:09 tnb      hdl: hey :)
15:09 tnb      If I send you an email, could you give me a few stats on SAN-OP?
15:09 tnb      collection size, # borrowers, etc?
15:10 hdl      tnb: send it to btoumi. He should be more able to provide you with figures.
15:11 tnb      what is his address?
15:11 tnb      never mind, i found it :)
15:11 tnb      thanks!
15:13 hpassini SAN-OP have 150 000 notices or biblios
15:13 hdl      kados : department stands for library department : Youth Albums, Adults, Science
15:13 hpassini 250 000 items
15:14 hpassini and 20 000 borrowers and btoumi is on hollidays
17:38 kados    rach: howdy
17:40 hdl      kados : do you have a sorting pb with results on more than 2 pages ?
17:40 hdl      There is one on rel_3_0.
17:41 hdl      selected order is not kept.
17:41 hdl      Could you verify on a devweek box ?
17:41 chris    sounds like a template problem?
17:41 chris    not passing through the order variable?
17:41 hdl      It is.
17:42 chris    and hi hdl :-)
17:42 hdl      but I think there also can be a mix.
17:42 hdl      hi chris.
17:42 hdl      How are you and kahungari ?
17:42 chris    good thanks, a bit tired today, he woke a couple of times last night
17:42 chris    but mostly good :)
17:43 hdl      mix between order Selection in template and selected in combobox.
17:43 chris    ahh right
17:44 hdl      combo box forgets wht has been selected.
17:46 hdl      + hardcoded labels in displaySortBy.
17:47 kados    we don't have a proper query parser
17:47 kados    if you want you could just use zebra
17:47 kados    instead of the query parser I slapped on top
17:47 hdl      This should not happen.
17:48 kados    hdl: feel free to file some bugs
17:48 kados    hdl: I'd be happy to work on better query parsing
10:01 kados    anyone else having trouble committing to savannah?
10:13 owen     Hi kados
10:13 kados    hey there
10:14 kados    fix the index, break the index, fix the index, break the index :-)
10:14 kados    this time it looks to be a problem with the query that's done in addbiblio
10:14 kados    to find duplicates
10:53 kados    hdl: you around?
10:53 hdl      yes
10:53 hdl      hi
10:53 kados    hdl: could you send a link to the record that you are having trouble with ?
10:53 kados    hdl: otherwise it will be impossible to troubleshoot for the perl4lib guys
10:53 kados    (hi :-))
10:54 hdl      I sent Ed Summers the XML record.
10:54 kados    also, just FYI, I've discovered that sending an attachment of a MARC file often corrupts it
10:54 kados    excellent
10:54 hdl      It is basically failing on what appears as µ
10:55 hdl      But should be ae with ligature.
10:55 kados    could you send it to me too? :-)
10:55 hdl      No. ASCII for mu greek letter.
10:55 kados    ahh, that's likely the prob then
10:55 kados    don't think MARC::Charset can handle anything but UNICODE or MARC-8
10:56 kados    hdl: ok, second question
10:56 kados    hdl: did you see my "questions for french team"?
10:56 hdl      Yes.
10:56 hdl      But cannot answer most of them.
10:56 kados    who can? san-op?
10:57 hdl      Yes and paul.
10:57 kados    ahh, ok
10:57 kados    thanks
10:57 kados    rch and I are continuing our code audit today
10:57 kados    it would be very useful to have those answers asap
10:58 hdl      I find it quite dangerous to rename sysprefs, since you have to make code audit to know where it is used.
10:59 hdl      a grep C4\:\:Context\-\>preference can help. But I don't know if ppl always use this facility to gain access to sysprefs
10:59 kados    well grep -r SysPref can help
10:59 kados    but I agree, it should be done carefully if at all