Time  Nick        Message
12:07 cnighs      g'morning #koha
12:10 paul        kados around ?
12:10 paul        hi cnighs
12:10 paul        does any liblimer have a an idea of the size of a Koha setup with, say, 1 000 000 biblios, and 3 000 000 issues a year ?
12:10 paul        (the disk size, mySQL & zebra)
12:49 gmcharlt    good morning #koha
12:49 paul        hi gmcharlt
12:49 gmcharlt    paul: don't have the numbers offhand - atz, ryan, or cfouts might have them
12:49 paul        cfouts ?
12:52 gmcharlt    cfouts = Clay Fouts, a new sysadmin for LL - not sure if he goes on #koha that often
12:56 slef        paul: I'd expect disk type (RAID is almost essential, but speed and layout matter too) and CPU speed/number to also come into play at that size, based on past experience with webservers.
13:19 kados       paul: with that many transactions you definitely want zebra and mysql on separate disks, probably separate servers
13:19 kados       paul: otherwise zebra's indexing will compete with mysql for disk IO
13:35 paul        kados : thanks for the idea.
13:35 paul        can you tell me for the size ?
15:00 kados       owen: you there?
15:00 kados       owen: 'patrons have been successfuly moved to trash'
15:01 owen        :)
15:01 kados       owen: that just doesn't quite capture what we're doing and if we're going to change that, I think we should change it 'right' :-)
15:01 owen        Doesn't sound very nice for the patrons
15:01 kados       hehe
15:01 kados       how about patron accounts have been deleted
15:01 kados       or anonomyzed even
15:01 kados       however you spell that
15:02 owen        Is this under clean borrowers?
15:04 owen        I see both "deleted" and "moved to trash" in the template, based on <!-- TMPL_IF NAME="trash" -->
15:04 owen        What does that mean?
15:05 kados       paul: ?
15:05 paul        mmm... no idea either. It's something written by toins
15:05 paul        maybe it's :
15:05 paul        1st step (confirm what we will clean) and 2nd step (really delete)
15:05 paul         ?
15:06 owen        What does it mean when patrons are "moved to the trash?"
15:06 kados       deletedborrowers table I bet
15:06 owen        "What do you want to do for deleted patrons?  Permanently delete these patrons   Move these patrons to the trash"
15:06 eric_       anybody knows where i can find a list of what type of marc is used in a country?
15:07 paul        owen: no idea, sorry
15:08 eric_       owen: Could it mean that the patrons is "deactivated", but all its information is still in the database?
15:12 owen        CVS history says cleanborrowers.pl was added by toins
15:13 kados       owen: I bet it moves them to the deletedborroweres table
15:13 owen        Shouldn't that just be the default action?
15:14 kados       if there was one
15:14 owen        I hate to reference "the trash" when there's no mechanism from the interface for retrieving records from "the trash"
15:15 su-erin     any tips for trouble shooting this upgrade from .064 to .094 on ubuntu 7.10:  Can't open blib/ZEBRA_CONF_DIR/etc/passwd for read at rewrite-config.PL line 168.
15:27 eric        su-erin, the first thing to do is to check if the file exists (I guess you check that already, right).
15:27 eric         It should be in the directory where you untar koha
15:45 atz         then check if you can read it  :)
15:47 su-erin     rewrite-config.PL is there and I have opermission to read it; check that yesterday. I think it has something to do with the that ZEBRA_CONF_DIR. I just dont know what directory that should be
15:51 atz         usually in your koha (kohaclone, whatever) directory as   ./etc
15:55 paul        kados ? about your patch UTF-8 to ASCII MAPPINGS. I don't understand why you need equivalent AND map.
15:55 atz         paul: i think map is a one-way relationship
15:56 atz         so if I search by e, it will match on acute-e, but not vice versa
15:56 atz         i'm not sure why we would need both though, if there are duplicates
15:57 paul        in french .chr, we use only map, and it work fine
15:57 paul        head/etc/zebradb/lang_defs/fr/sort-string-utf.chr
15:57 atz         su-erin: for example:  ~/kohaclone/etc/zebradb/
15:57 paul        (used also for word-phrase-utf.chr)
16:02 atz         paul: we also ran into differences between wide-bytes and combined characaters
16:02 atz         but maybe we would be fine to just use the mappings you have in the french file
16:02 su-erin     thanks, i'll dig their
16:02 su-erin     *there
16:20 kados       paul: note that only covers sorting, not searching
16:21 kados       paul: do you have a custom word-phrase-utf.chr that you use for French libraries too?
16:21 paul        kados : what cover sorting not searching ?
16:21 kados       sort-string.chr
16:21 kados       it's not used for equivalance in searching, only sorting
16:21 paul        nope, we use the same file content (although we have a word-phrase-utf)
16:21 paul        but it's content is the same as sort-phrase-utf.chr
16:21 kados       then only word-phrase-utf should be used for searching according to default.idx
16:22 paul        right.
16:22 kados       word phrase numeric indeces are separaet from sort indeces
16:22 paul        but the content of both files are the same
16:22 paul        (not sure to understand what you mean...)
16:23 kados       different rules are used for different indexes and therefore, actions
16:23 kados       rules that apply to sorting don't apply necessarily to searching
16:24 paul        yep, I know
16:24 kados       for instance, in English, we want to be able to search for 'The', but we don't want 'The ' to be counted for sorting purposes
16:24 paul        but for us, the same rules are OK. so the files have the same content.
16:24 paul        mmm... the is in empty words, isn't it ?
16:24 paul        so it's discarded ?
16:25 kados       paul: so you use sort-string.chr also for word-phrase-utf.chr?
16:25 kados       because those two files in default Koha are different
16:25 paul        yep
16:25 kados       OK, so that's a local modif you do apart from normal installation
16:25 kados       (just trying to understand clearly)
16:25 atz         kados: so their config references the same file in 2 places
16:25 atz         rather than 2 separate files
16:25 atz         iiuc
16:25 kados       you edit default.idx and point word and phrase indexes to sort-string.chr
16:26 kados       ?
16:26 paul        nope.
16:26 paul        we refer to sort-string.chr & word-phrase.chr
16:26 paul        but those 2 files contains the same thing (map, equivalent...)
16:26 kados       OK, so you copy sort-string.chr to replace word-phrase-utf.chr then?
16:27 kados       paul: the same type of file, yes but in default Koha they have very different rules, esp the French sort-string.chr and default word-phrase-utf8.chr
16:27 paul        I should do an install from scratch to check
16:27 kados       and your sort-string.chr has mappings for French words like Le
16:27 kados       do you want those to be ignored in searching as well as sorting?
16:27 kados       (that would be surprising to me)
16:27 atz         paul: yes, please.  your familiarity w/ this issue will help out considerably
16:28 paul        kados : yes, that's what I want
16:28 paul        (ignore Le when searching & sorting)
16:28 paul        + we always use empty word removal, so "Le" is not sent to zebra anyway for searching
16:28 paul        (so it's usefull only for sorting in fact)
16:28 kados       k
16:29 kados       paul: OK, so I bet that it doesn't work for you because only sort-string.chr is used and only for sorting, not searching
16:29 kados       all searching uses word-phrase-utf.chr
16:30 kados       for english and french it appears
16:30 paul        (I didn't say it don't work for us. I just investigate the consequences of the commit for us ;-) )
16:30 eric        (for the record, when paul talks about <empty words>, he is refering to <stop words>)
16:30 paul        eric : right
16:30 kados       maybe we should move word-phrase-utf.chr to the lang_defs/fr and lang_defs/en
16:30 paul        (empty word is a frenchism..)
16:31 paul        kados : maybe yes.
16:31 eric        Paul, is there a reason why you don't want to sort using the stop words?  I can understand the search, but for the sort, i<m not sure...
16:32 kados       eric: 'Le ' is an article and most librarians don't want it to be used for sorting
16:32 kados       s/it/articles/
16:33 paul        eric : "Le Seigneur des anneaux" must appear at "S" for sorting
16:34 kados       paul: maybe we should move word-phrase-utf.chr to the lang_defs/fr and lang_defs/en
16:35 kados       paul: to accomodate your need to not search by articles etc
16:35 paul        kados : didn't you write that 3mn ago already ? and I answered "maybe yes"
16:36 kados       hehe, maybe :-)
16:36 paul        kados : in fact, I don't care having map Le^ ^, for searching, as Le is already removed by Koha before sending the request to zebra
16:36 paul        but I agree it's a good idea.
16:36 paul        (more flexible for the future probably)
16:38 nicomo      yes because you (paul) speculates that the stop-words list will be correctly filled in by the librarian: not a sure bet
16:38 nicomo      so having map Le for seraching might provide a safety net too
16:39 gmcharlt    respecting nonfiling indicators would be better yet, although a 3.2/DOM-mode thing
16:39 nicomo      gmcharlt ++
16:48 eric        anybody knows the usage of 908$a - Put Command Parameter?
17:13 cnighs      hdl about?
17:44 paul        cnighs: hdl was at a client today. should be back tomorrow
17:44 paul        or maybe will jump in for a few minuts in 2 hours or so (although it's already 8PM here)
17:45 cnighs      tnx paul
19:31 atz         acmoore: trying out overdues script for a client...
19:31 atz         all the <<branches.*>> elements seem not to get populated?
19:31 atz         seen that before/
19:31 atz         ?
19:36 acmoore     atz, hmm. No, I haven't. In fact, I have branches.branchphone in the template I've been testing with, and it gets popualted.
19:36 acmoore     I wonder what more I can look at with you.
19:37 atz         let me check their branches table....
19:37 acmoore     atz, do they get removed?
19:37 atz         yeah, they don't appear in the resulting messages, just an empty line
19:38 acmoore     atz, are you supplying a --library flag to the script to limit it to one branch, or running it on all branches?
19:38 atz         running it for all
19:39 acmoore     hmm.. odd. I don't knwo what might be up. I wonder if the same thing happens if you limit it to one branch.
19:39 acmoore     I guess one thing to check is that the field you're looking for is actually populated in the branches table.
19:40 atz         yep, that's why I went to check.  looks normal to me
19:40 atz         mysql> select * from branches;
19:40 atz         +------------+--------------+---------------------------+----------------+----------------+-------------+-----------+----------------+---------+----------+---------------+
19:40 atz         | branchcode | branchname   | branchaddress1            | branchaddress2 | branchaddress3 | branchphone | branchfax | branchemail    | issuing | branchip | branchprinter |
19:40 atz         +------------+--------------+---------------------------+----------------+----------------+-------------+-----------+----------------+---------+----------+---------------+
19:40 atz         | LCC        | Main Library | DeFehr Building 3rd Floor | Kretingos 36   | Klaipeda, LT   |             |           | library@whatever.com |    NULL |          |               |
19:40 atz         | TXT        | Textbooks    | DeFehr Building Basement  | Kretingos 36   | Klaipeda, LT   |             |           | library@whatever.com |    NULL |          |               |
19:40 atz         +------------+--------------+---------------------------+----------------+----------------+-------------+-----------+----------------+---------+----------+---------------+
19:41 atz         (emails changed to protect the innocent)
19:41 acmoore     are you attempting to use branchphone? ;)
19:41 atz         would that prevent branchname from displaying?
19:42 acmoore     ah, no. I just didn't know what field you were trying to use, and branchphone appears to not be populated.
19:42 atz         the odd thing is that *none* of the branches.* fields display
19:44 acmoore     atz, can you run it with the --library=LCC or --library=TXT  set to see if it works? Is this a pain for you to do?
19:44 atz         probably this line:   @branches = ('');
19:45 acmoore     I've noticed that some people use a branchcode of '' for a default in their databases, and perhaps leaving it off is suppoosed to be running for that default branch instead of all branches.
19:45 eric        do you know what the issuing column?
19:46 atz         acmoore: that is probably inconsistent... in rules that should be NULL, as in UNMAPPED to branches table
19:46 atz         "" is an illegal value (foreign key violation)
19:48 atz         eric: i imagine that field is meant to designate whether the branch is a circulating one or an internal one (like bindery, cataloguing, etc.)
19:49 atz         i'd bet it isn't used (anymore) though.  much like branchprinter.
19:49 acmoore     atz, oh, I see. well, I wonder how this should work, then. I guess not specifying branchcode to this script should run for all branches, but that appears to be a bug, then?
19:49 atz         acmoore: no luck w/ --library options
19:49 acmoore     atz, when you say no luck, do you mean that it continues to fail to substitute the branches.* fields?
19:49 atz         acmoore: it still runs, and it populates w/ patron and item data.  but the branches.* lines are still blank
19:50 atz         in short, yes.
19:50 atz         :)
19:50 acmoore     odd.
19:51 atz         just rebased this afternoon, btw
19:51 acmoore     the -v flag isn't too great, but if you want me to keep looking at it with you, could you run it with one library and the -v flag and send me the results, please?
19:52 acmoore     I'm just not sure if they'll be useful enough for you to decypher.
19:52 atz         acmoore: it's possible that the *borrowers* are sparsely populated and there isn't a homebranch recorded
19:52 atz         let me check that
19:54 atz         nope... looks like they all belong to the main branch.
19:55 atz         which is odd, b/c I still get results with --library=TXT ... must be because the items that are overdue are TXT's
19:57 atz         well, the error message is on point:
19:57 atz         The following terms were not matched and replaced:
19:57 atz                 branches.branchname
19:57 atz                 branches.branchaddress1
19:57 atz                 branches.branchaddress2
19:57 atz                 branches.branchaddress3
19:57 atz                 branches.branchphone
19:57 atz                 branches.branchemail at /home/liblime/kohaclone/misc/cronjobs/overdue_notices.pl line 383.
20:00 atz         also distressing, since I ran in -n (no effect mode), I see lines like :  debarring 48 Irma BALCIUNAITE
20:01 acmoore     -n is No email.
20:01 acmoore     perhaps that should change, though.
20:06 atz         hmmm.  ok, i have the log from that and can un-debar the 10 or so patrons impacted
20:07 atz         acmoore: so there is no test mode?
20:08 acmoore     i guess not. There's just the mode that sends no emails. It seems to still debar people.
20:09 atz         we definitely need some kind of safe poke-around-but-don't-hurt-nobody mode.
20:09 acmoore     I'd think so.
20:11 acmoore     There was another documentation improvement idea that came up yesterday. I put them together in this bug http://bugs.koha.org/cgi-bin/bugzilla/show_bug.cgi?id=2388
20:11 acmoore     I can open one up to make a test mode of some kind, too.
20:13 acmoore     atz, opened 2389 for that.
20:14 atz         ok, i'll leave it at that for now
20:14 atz         i'll probably just hardcode the branch stuff for the client, since they are the almost the same
20:15 pianohacker (liblime or biblibre people) newbie question: how do you keep clients up to date with the latest git improvements and add their personally requested changes to just their tree?
20:17 atz         pianohacker: the clients have git repos, and you just commit their changes to them
20:17 atz         then when you rebase, git merges everything (or gives errors you have to sort out manually)
20:17 pianohacker OK, thanks
20:30 atz         acmoore: how often should process_message_queue run?
20:34 acmoore     atz, I guess maybe hourly or so? Really, it only needs to be run nightly since the only things that write to it are nightly cronjobs, but perhaps we'll have more things adding to it soon.