Time  Nick     Message
12:34 owen     Hello #koha
12:46 paul_    hi owen
12:48 owen     paul: maybe you know the answer to this question... Shouldn't the web installer prompt the user for a LibraryName?
12:48 paul     good suggestion.
12:49 paul     should maybe not. but would be better...
12:49 paul     as well as asking for a branchcode & branchname
12:49 paul     maybe a 4th step ;-)
12:58 kados    yea
12:58 kados    that'd be good owen
12:58 kados    morning guys
12:58 paul     morning kados
12:58 owen     kados: would it be helpful for me to file a bug on it?
12:58 kados    well, last night I started working on a new plan for the installer
12:59 kados    and I did add that to my list, so probably you can leave it for now
12:59 kados    I really like this idea of applying specific stylesheets to 'types' of templates, based on paradigms
12:59 kados    login.css, wizard.css, etc.
13:00 kados    those are two easy examples
13:00 owen     I was thinking about stylesheets and javascript yesterday
13:00 kados    but I'm getting hung up a bit in the session management stuff
13:00 owen     I wish we had a better way to manage includes for specific pages/modules
13:00 paul     kados : do you know when you'll release a 3.0beta version ?
13:00 kados    chris re-wrote part of it for the installer to be more secure, but I can't seem to figure out how to add a new sesion param
13:00 kados    owen: yea, I was thinking the same thing
13:01 kados    paul: there are three things holding me back:
13:01 kados    1. installer
13:01 kados    2. re-write members management
13:01 kados    3. a bit more testing
13:01 paul     1. = you mean the tech part of the install, right ?
13:01 kados    paul: yep
13:02 kados    MJ made a start of it, but it's not finished yet
13:03 paul     do you have read my 5hours ago mail about git questions ?
13:03 kados    paul: yep
13:03 kados    cool
13:03 kados    yea, git++
13:03 paul     and enjoy it more and more...
13:04 paul     bye then
13:31 jaron    is there currently a web browseable version of current development online? or is there a public way to check out the latest?
14:42 kados    jaron: didn't I send you the development package at one point?
14:44 jaron    kados: yeah, I just was wondering about the latest and greatest in the repo.
14:44 jaron    no immediate need--just wondering
14:44 kados    http://wiki.koha.org/doku.php?id=en:development:git_usage
14:44 jaron    with the change to git and all
14:44 kados    git.koha.org is what you want
14:44 jaron    ah, ok thanks
14:44 kados    and the wiki page explains how to use it
14:44 jaron    cool thanks
14:44 kados    we haven't done an announcement yet
14:44 kados    because the core of us are still experienceing a few growing pains with git
14:44 kados    but it's getting easier all the time :-)
14:44 jaron    so I've seen ;)
14:45 kados    motto “Every time a patch falls on the floor, a kitten dies�
14:45 kados    :-)
14:47 kados    owen: you about?
14:47 kados    owen: I'm done with my phone call
14:47 kados    :-)
14:48 owen     Okay, first on my list: I still don't have access to the error log on arwen
14:48 owen     I'm trying to add biblios and getting an error from additem.pl.
14:48 kados    ok, give me a sec
15:06 kados    done
15:06 kados    owen: neeext? :-)
15:07 owen     Should Zebra be automatically indexing for me?
15:07 kados    nope that's what zebraquueue.sh is for
15:07 kados    in your home dir
15:08 kados    make sure zebra's running first
15:08 owen     Oh yeah, forgot about that
15:09 owen     How long before records appear in my search?
15:09 kados    should be instant
15:10 kados    ie:
15:10 kados    catalog something
15:10 kados    run zebraqueue
15:10 kados    find it in the opac
15:10 kados    if it's not working, make sure that zebra's running
15:12 jaron    does zebraqueue do the commits from the shadow registry?
15:14 kados    it can be configured to use shadow indexes
15:14 kados    well, by default it does actually
15:15 kados    so internally the steps are:
15:15 kados    1. save record to database and update zebraqueue table to indicate an index is required
15:16 kados    2. run zebraqueue script periodically to 'flush' out necessary changes
15:16 kados    that's one way to do it anyway
15:16 kados    or you can opt to have he index update on the fly as you catalog
15:16 kados    (and as you circulate)
15:16 kados    which is fine for most libraries except the largish ones
15:17 jaron    ah, ok.
15:18 kados    establishing a good set of conventions for general use is what's so difficult about doing a project release for a diverse community like ours
15:18 jaron    I was wondering why not do immediate commits but I could see for a large library why it might not work
15:23 paul     jaron: because it seems that, if 2 librarians catalogate simultaneoulsy, zebra may be confused by 2 concurrent "commits", and ... the zebra database is corrupted.
15:23 paul     by corrupted, we mean : impossible to use it anymore for anything, need to rebuild it completly
15:23 paul     happends more than once a week at NPL iirc (kados will confirm)
15:24 jaron    paul: ah, ok
15:24 jaron    good to know
15:24 paul     s/happends/happened/
15:24 kados    well, that was several months ago
15:24 jaron    so is zebraqueue set up as a cron job there and run every night or something?
15:24 kados    the index hasn't crashed at all since the latest release of zebra
15:24 paul     really ?
15:24 kados    jaron: it can run every minute if you want
15:25 paul     even without zebraqueue_start & real time commit ?
15:25 paul     wow...
15:25 kados    paul: NPL still have a nightly update, but Stow and Barberton use zebraque every minute I believe
15:26 paul     IPT has zebraqueue every minute as well, but she finds annoying having to wait up to 60 seconds every time she creates an authority...
15:26 paul     ... to find it in the biblio & create the biblio
15:26 kados    *nod*
15:27 paul     (thus the "auto create autority feature hdl reported last week on koha-devel...)
15:27 kados    I'm pretty sure that zebra's issues with asynchronous updates are fixed
15:28 jaron    so can one insert a record (into the shadow) at the same time the system enters "commit mode" now?
15:28 jaron    because otherwise I could see conflicts happening
15:31 kados    yes
15:54 kados    morning chris
16:48 kados    foxnorth: how's it going?
16:48 foxnorth kados: not bad- i'm testing out that yui rich text editor
16:48 kados    still looking positive?
16:48 foxnorth kados: need to adjust my marceditor stylesheet and see what happens
16:49 foxnorth well, i need to clean up my marceditor html before sticking it in that yui editor, i think
16:49 foxnorth i'll make it match what i tried for that test i did
16:50 foxnorth and use 1 div for a tag, with a span for each subfield
16:50 foxnorth and each subfield delimiter/code
16:50 kados    what about using stock marcxml?
16:50 kados    syntax
16:50 kados    woudl that be too hard?
16:50 kados    not enough block-level elements maybe ...
16:51 kados    I dunno, feel free to ignore me :-)
16:52 foxnorth well, i'll still need to transform that marcxml into divs and spans and what not...
16:52 foxnorth although i could be doing that in javascript as opposed to xslt
16:52 kados    xslt++
16:52 kados    thinking further
16:52 kados    yea, I like div + span
16:53 foxnorth i'm all ears!
16:53 kados    maybe have <div id="leader">
16:53 foxnorth i also give each span a unique id which is reversible back to the marcxml
16:53 kados    <div class="controlfield" class="tag008">
16:53 foxnorth hhm lemme look at a marcxml sample
16:53 kados    div class="datafield" class="tag040"
16:54 kados    http://www.loc.gov/standards/marcxml/Sandburg/sandburg.xml
16:54 kados    might as well use the nomenclature
16:54 foxnorth that's exactly the one i have :)
16:54 foxnorth and yeah, that's what i'm thinking in terms of the divs
16:54 foxnorth and then each subfield is in its own span (so it's inline)
16:54 kados    yea
16:54 kados    sounds like a plan to me
16:55 kados    and you can add new class defs for plugins
16:55 kados    or just walk the DOM (yuk?)
16:55 kados    the trickiest stuff happens around rules for repeatibility, order, and plugins
16:56 kados    I guess first step is get the interface looking swank :-)
16:56 foxnorth np!  i'm eager for fresh ideas :)
16:57 foxnorth yeah, for repeatibility i added in a a suffix to the tag id, so 500-1, 500-2
16:57 foxnorth but we definately need something sane and understandable for future plugins
16:57 kados    makes sense
16:57 foxnorth so i'm happy to think to about it some more
17:58 foxnorth kados: looks like these resizing handles are produced by a _moz_resizing attribute
17:59 foxnorth kados: they're showing up using the yui editor as well
17:59 foxnorth and this attribute is affecting the <table> the editor is in
17:59 foxnorth i spose i could go with straight divs w/o a table
18:04 foxnorth hhmm...this comment http://codingforums.com/archive/index.php?t-97810.html doesn't bode well for getting rid of resizing handles...
18:11 kados    huh
18:12 foxnorth but going tableless would be ok i think
18:12 kados    tableless+=
18:12 kados    ++ I mean
18:12 kados    hmmm
18:13 foxnorth just need to provide columns for tags, 2 indicators and then subfields...
18:13 kados    how about using <dl> ?
18:14 kados    thought not using tables for tabular data is annoying :-)
18:14 foxnorth yeah...
18:14 foxnorth hhhmm
18:14 kados     ... when you insert an image into the editor ...
18:14 kados    any reason we need an image?
18:14 foxnorth well in my case it's a table that has this _moz_resizing attribute
18:15 foxnorth and that's causing these resizing handles
18:15 kados    ahh
18:15 foxnorth i too tried setting that attr to false with no results
18:15 foxnorth so that post there just confirmed my suspicions....
18:15 kados    right
18:16 foxnorth i think we can get it to work w/o the table though
18:16 foxnorth the tables annoying anyway :)
18:16 kados    heh
18:16 kados    you could do the whole thing in divs and spans I think
18:17 foxnorth yeah i think so too
18:17 foxnorth i've got smth working already
18:17 foxnorth lemme post it so you can check it out
18:19 kados    let me know when you've posted it
18:38 owen     kados, what's the difference between git fetch and git pull ?
18:39 kados    git pull does an auto merge
18:39 kados    git pull is really just a wrapper around git fetch and git merge
18:40 owen     In what situations would you use each one?
18:42 kados    I think we're still working out how we want to do it for the koha project
18:42 kados    there's a lot of flexibility in git
18:45 kados    part of the reason I haven't done an anouncement is because of these kinds of issues
18:45 kados    we need to get a bit more experience with managing git repos
19:02 owen     kados, I guess I'm still confused about when I have to do something to re-synch with the main repository
19:03 owen     I make changes, I commit changes, I make more changes... At some point I think, I'd better re-synch with everyone else's changes... before I submit a patch, right?
19:03 kados    yea
19:03 kados    that's exactly right
19:05 owen     ...and that's when I do git pull ?
19:07 kados    well, we've been using git fetch and git rebase origin
19:10 owen     And that process merges the changes found in the main repo with my repo
19:23 kados    yes
19:25 owen     What kind of information does it provide regarding how things were merged?
19:32 kados    rebase and merge are slightly different
19:32 kados    rebase takes the repo back to the origin and applies all the changes since the last rebase, which brings the repo up to date with the main one ... then it applies local commits as patches
19:33 kados    git merge takes what exists now and merges directly against the origin
19:33 kados    so it's two ways to do the same thing really
19:33 kados    and I haven't used it enough to know when to do one and when to do another
02:14 Brooke   Howdy
02:14 dewey    hola, Brooke