Time  Nick     Message
20:18 hdl      atz : thanks.
20:08 atz      hdl: no
20:07 hdl      atz : do you know who subscribed kohalists on nabble.com ?
20:01 hdl      kados : /me asking savannah guys.
19:38 atz      "The control names/values are listed in the order they appear in the document."
19:37 atz      reference: http://www.w3.org/TR/html401/interact/forms#h-17.13.3.3
19:18 kados    not_having_to_rewrite_koha's_searching++
19:08 atz      kados: you are right
18:36 kados    it's possible, we are swamped with US work at the moment
18:35 indradg  kados, did some people from an organisation called Akshara, from India try to get in touch with you?
18:35 indradg  kados, thats for sure!
18:34 kados    indradg: it's been a while!
18:34 kados    hey indradg !
18:34 kados    atz: is what we use to fetch the list
18:34 kados    atz: http://search.cpan.org/dist/CGI.pm/CGI.pm#FETCHING_THE_PARAMETER_LIST_AS_A_HASH:
18:33 indradg  kados, around?
18:31 atz      i'll check
18:31 kados    specifically if they are named the same, they come back as an ordered array
18:31 kados    yea, and it does actually
18:30 atz      again, my point is that the spec doesn't guarantee the order of params from froms
18:30 kados    ie, you've re-ordered the idx array, which woudln't ever happen
18:30 kados    there you're asking for an author,word list like 'love'
18:29 atz      http://opac.liblime.com/cgi-bin/koha/opac-search.pl?idx=kw&idx=au%2Cwrdl&q=cooper&idx=kw&sort_by=relevance&do=OK&q=love
18:28 kados    identical for me
18:28 atz      interesting... not the same results though
18:28 kados    atz: is what I think you were hoping would trip it up ;-)
18:28 kados    atz: http://opac.liblime.com/cgi-bin/koha/opac-search.pl?idx=kw&idx=au%2Cwrdl&idx=kw&q=love&q=cooper&sort_by=relevance&do=OK
18:24 kados    go for it
18:24 atz      now, in a min. I'll produce a VALID HTTP request based on the same form that will break
18:24 kados    yep, and that will work just fine
18:23 atz      http://opac.liblime.com/cgi-bin/koha/opac-search.pl?idx=kw&q=love&idx=au%2Cwrdl&q=cooper&idx=kw&sort_by=relevance&do=OK
18:23 atz      FF 2 does a GET for this URL, based on adv. search form:
18:22 kados    atz: give me an example of a query CGI that doesn't work
18:22 atz      all that the server gets is the browser request
18:21 kados    regardless of the order
18:21 kados    it works because it keeps track of each array and matches up each query phrase
18:21 kados    :)
18:20 kados    nope
18:20 atz      it works b/c of non-spec browser-dependent behavior
18:20 kados    it works
18:20 kados    go check the code
18:20 kados    I don't know how to explain it in English
18:20 atz      the order of elements in that array is determined by the browser's implmentation, not Koha
18:20 atz      no... they are params w/ the same name, so Koha gets them as an array
18:19 kados    obviously the operators are offset by one
18:19 kados    make sense?
18:19 kados    doesn't matter what order you send them in
18:18 kados    first idx is matched to the first q, etc.
18:18 kados    koha matches up the queries
18:18 kados    yea, that doesn't matter
18:18 atz      ^^ legitimate browser formulation of above query
18:18 kados    ahh, I see
18:18 atz      opac-search.pl?idx=kw&idx=kw&idx=kw&q=murder&q=love&sort_by=relevance&do=OK
18:16 atz      (though most are somewhat reasonable)
18:16 atz      the browser does not!
18:16 kados    ahh, when i researched it, I think CGI does enforce the parameter order
18:15 atz      it relies on parameter order, which is not guaranteed by spec.
18:15 kados    though I did my best to read the Dragon book and understand recursion and compilers in general :-)
18:15 atz      sorry, I mean the HTTP request form of the query
18:15 kados    a bit beyond my puny skills at the moment
18:15 kados    I didn't have time to write a real query parser
18:14 atz      idx=kw  can't reliably be "paired" to the q terms
18:14 kados    actually, zebra queries are perfect CCL or CQL
18:14 atz      the query structure there doesn't really jive, either
18:13 atz      the results are pretty good from zebra even if not functionally boolean
18:12 kados    I spent a lot of time worrying about it :=)
18:12 kados    really, the user-facing query language needs some peer review
18:12 kados    shoudl work
18:12 kados    op=or
18:12 kados    atz: it's OR or alternatively pipe
18:08 atz      and the op symbol for "OR" is pipe?
18:05 atz      (the slot up top is blank)
18:05 atz      i notice that the 3rd link doesn't reproduce the "display" version of the query
18:04 kados    it's a very reasonable question
18:04 atz      thx
18:02 kados    atz: comparing those results should answer it
18:02 atz      adv. search builds a more complex query, but equiv: opac-search.pl?idx=kw&q=murder&idx=kw&q=love&idx=kw&sort_by=relevance&do=OK
18:02 kados    atz: http://opac.liblime.com/cgi-bin/koha/opac-search.pl?q=money&op=and&q=love
18:02 kados    atz: http://opac.liblime.com/cgi-bin/koha/opac-search.pl?q=money+and+love
18:02 kados    atz: http://opac.liblime.com/cgi-bin/koha/opac-search.pl?q=money%2Blove
18:01 kados    money+love should work too
18:01 kados    adv search should do it
18:01 atz      but there is no OPAC construction for that?  perhaps only adv. search?
18:00 kados    atz: q=money&op=and&q=love
17:56 atz      into a boolean conjunction
17:56 atz      my impression is that it isn't currently structuring, say: money AND love
17:54 atz      how is koha/zebra handling boolean searches?
17:43 kados    I've found #savannah on freenode to be helpful at times too
17:42 kados    hdl: I'm guessing that's another qeustion for savannah-hackers
17:41 slef     also, can you keep the list-id when the lists move, or is mailman still dumb like that?
17:41 slef     savannah-hackers should help I think
17:24 kados    and nabble.com, etc.
17:23 kados    we also need to update koha.org to show the new list info
17:23 kados    maybe the savannah guys will help?
17:23 hdl      we donot have ssh access
17:22 hdl      on savannah ?
17:22 kados    hdl: can't you use the .forward file?
17:20 hdl      It is for redirecting lists/mailman/listinfo into mailman.lists.org
17:16 kados    but maybe savannah guys will help us out?
17:16 kados    maybe we don't have access to mailman.conf?
17:15 kados    hdl: http://www.webmasterworld.com/forum92/3366.htm
17:15 kados    bummer
17:14 hdl      not sure to be able to do that clean.
17:05 kados    cool
17:04 kados    hdl: how can we find out for sure?
16:41 hdl      i think so.
16:21 kados    hdl: so you can also redirect all mail from koha-*@nongnu.org to koha-*@lists.koha.org ?
16:18 kados    hdl++
16:18 kados    hdl: response sent
15:58 hdl      kados: message sent.
15:47 gmcharlt hi fbcit
15:37 fbcit    hi kados, koha
15:37 kados    heh
15:37 fbcit    man, the gang's all here :)
15:36 kados    mornin fbcit
15:36 hdl      OK.
15:36 kados    hdl: (for English readers)
15:36 kados    hdl: if you want me to proofread your announcement let me know
15:36 kados    go for it
15:36 kados    OK
15:35 hdl      But I am much prepared to change savannah's list than french lists who are already on the same server
15:34 hdl      not to me either.
15:34 kados    *shrug* ... it doesn't matter much to me
15:33 hdl      and not koha-fr.
15:33 hdl      first problem was savannah.
15:33 hdl      kados: my understanding was different.
15:32 kados    hdl: does that make sense?
15:32 kados    then we can move koha-*@nongnu.org soon thereafter
15:32 kados    since it's a french koha user list?
15:31 kados    hdl: and perhaps re-name it koha-fr@lists.koha.org as slef recomends?
15:31 kados    hdl: can't we start with koha-infos to test the migration? It should be easiest since you and paul manage that one
15:31 kados    hdl: my understanding was the goal was to move all koha lists to lists.koha.org, right?
15:30 hdl      I thought of all the savannah lists move.
15:29 hdl      infos on koha-fr.org is not a problem for you.
15:29 slef     or koha-...-fr if you want to mimic debianisms
15:28 hdl      start by moving koha-infos ?
15:28 slef     kados: or have koha-fr-... lists
15:28 slef     kados: tell 'em to stop being so silly and use the french one ;-)
15:28 hdl      kados ????
15:27 kados    slef: what do we do when a canadian group wants to create a french list?
15:26 slef     Isn't the spanish already koha-es?
15:26 kados    last time I checked, only latin characters were allowed in email
15:26 slef     ISO-3106 or whatever it is... koha-fr?
15:26 kados    koha-fr-FR ?
15:26 kados    slef: so what, we use RFC4646 codes?
15:25 kados    hehe
15:25 slef     kados: debian sucks at mailing lists, news at 11
15:25 kados    http://www.debian.org/MailingLists/subscribe
15:25 kados    slef: yea, I'm following debian conventions
15:25 slef     kados: French isn't french for french.
15:25 kados    slef: yea, it should be fine
15:24 slef     I'm currently wondering whether rebuild_zebra.pl transfers SERs properly.
15:24 kados    hdl: and perhaps change the name to Koha-French ?
15:24 slef     but zebraqueue_daemon.pl?
15:24 kados    hdl: can you start by moving koha-infos?
15:24 kados    slef: yup
15:24 slef     kados: does opac.ll.c run with record.abs as shipped in 3beta?
15:22 slef     (query was f @attr 1=/record/datafield[@tag="942"]/subfield[@code="c"] SER if anyone wants it)
15:22 slef     so opac.liblime.com is happily indexing SERs
15:22 kados    no, sounds lie your'e doing it
15:22 slef     Number of hits: 100, setno 2
15:22 slef     Search was a success.
15:20 slef     is there more that must be done?  Other bibs show in search results
15:20 slef     because zebradaemon_queue.pl fails noisily
15:20 slef     rebuild_zebra.pl is cronned
15:20 kados    slef: have you indexed the bib records?
15:19 kados    slef: use SRU if you like: opac.liblime.com:9998/biblios
15:19 kados    slef: yup ... running on port 9998/biblios
15:19 hdl      is serialseq stored in marcrecord ?
15:19 slef     kados: this zebra doesn't seem to have knowledge of the bib record for some reason... can I TCP to liblime's demo's zebra?
15:18 kados    just not the planning data
15:18 hdl      Am I misunderstandng ?
15:18 kados    hdl: it has knowledge of the bib record for a serial
15:18 slef     hdl: as we can't simply Add Item them
15:18 slef     hdl: so how are old stocks acquired into the collection?
15:17 hdl      So zebra has no knowledge of it.
15:17 slef     well, searched for 942$c being SER, which seems to be how to spot a serial
15:17 hdl      slef, kados : serials are now stored only in tables, as far as i know.
15:17 kados    s/field/fields/
15:17 kados    what field I mean
15:16 kados    and we need to know aht field specifically
15:16 slef     I'm using whatever's in 3 beta IIRC
15:16 kados    slef: zebra just follows what Koha's record.abs tells it to index
15:16 slef     and they don't show up from yaz-client
15:16 slef     if zebra does... I've a catalogue with serials entered but they don't show up in the catalogue search
15:16 kados    slef: that's a very broad question ... what part of the serial record specifically? to find the answer you'd want to check the record.abs file
15:15 kados    slef: you're asking if Koha indexes serials?
15:15 kados_   hdl: yes, I think we're ready to switch the lists, they seem to be working just fine
14:36 gmcharlt hi slef
14:29 slef     do we have a handy guide to zebra care anywhere?  Like, how to see if the serials are being indexed...
14:27 hdl      hi slef
14:23 owen     Hi slef and everyone
14:21 slef     hi all... how's things?
14:12 gmcharlt hi hdl
14:12 hdl      hi gmcharlt
14:05 hdl      kados_ :
14:04 hdl      Should I send a message on the lists to announce the new lists ?
14:04 hdl      kados :
13:49 nengard  hdl i can't do private messages :( i'm trying to figure that out - i've lost my password - will open skype
13:48 nengard  hdl i am now
13:46 hdl      nengard are you around ?
13:46 hdl      hi america