2010-01-19T00:27:53 *** pmplett has quit IRC 2010-01-19T00:29:48 *** mck9 has left #evergreen 2010-01-19T00:32:42 I'll poke at this more tomorrow; will try to reproduce on lenny / pgsql 8.3; if it dies there, too, then I start disabling triggers one by one 2010-01-19T00:32:49 *** dbs has quit IRC 2010-01-19T01:17:21 *** pmplett has joined #evergreen 2010-01-19T02:15:39 *** pmplett has quit IRC 2010-01-19T02:39:07 *** pmplett has joined #evergreen 2010-01-19T02:53:46 *** eguest309 has joined #Evergreen 2010-01-19T03:13:08 *** pmplett has quit IRC 2010-01-19T03:44:33 *** eguest309 has left #Evergreen 2010-01-19T04:03:34 *** rsinger has quit IRC 2010-01-19T04:54:53 *** Dmagick has quit IRC 2010-01-19T05:08:08 *** brendan_ga has quit IRC 2010-01-19T05:45:48 *** eguest309 has joined #Evergreen 2010-01-19T06:41:32 *** arianit has joined #evergreen 2010-01-19T07:09:55 *** Dmagick_ has joined #Evergreen 2010-01-19T07:12:34 *** arianit has quit IRC 2010-01-19T07:14:36 *** Dmagick has joined #Evergreen 2010-01-19T07:42:03 *** alxp has joined #evergreen 2010-01-19T07:44:44 *** mck9 has joined #evergreen 2010-01-19T07:51:59 *** gmcharlt has quit IRC 2010-01-19T07:59:55 *** natschil has joined #evergreen 2010-01-19T08:49:48 *** gmcharlt has joined #evergreen 2010-01-19T08:57:09 *** dbs has joined #evergreen 2010-01-19T09:02:42 *** Meliss has joined #Evergreen 2010-01-19T09:03:16 *** bshum has joined #evergreen 2010-01-19T09:12:17 *** jenny has joined #evergreen 2010-01-19T09:20:56 while looking at adding more data to the search results (URIs and call numbers - doing a scoped "limit 3" kind of thing), I was wondering why we fire off one or more calls per bib rather than passing a list of bibs in one call 2010-01-19T09:21:12 there must be some hard-learned lesson behind that 2010-01-19T09:23:17 dbs++ ("clense" made me cringe too) 2010-01-19T09:24:33 senator: heh, it was distracting me while I was trying to focus on finding existing methods that would do what I want, so it was time for it to go :) I don't mind the alias that miker_ re-added 'cause it's for out of tree stuff that I don't have to see :) 2010-01-19T09:24:59 dbs-- # using up daily smiley quotient in one post 2010-01-19T09:25:24 there's a quota? oh dear 2010-01-19T09:25:28 and if i had to guess about why one or more calls are made per bib, based on response i've received to similar questions, i'd say it might have to do with creating a faster feel in the interface (under the assumption that receiving the first couple of results quickly is better than receiving all results at once, later, to some number of users) 2010-01-19T09:26:00 but i'm not fully informed on the matter; that might not apply in this case. 2010-01-19T09:26:00 I thought of that. but the overall latency adds up 2010-01-19T09:26:36 sounds like it's worth an experiment, anyway 2010-01-19T09:26:43 * dbs stumbles on 2010-01-19T09:27:32 rebuilding the trunk database schema on lenny and hoping (sort of) that there are no backend postgresql crashes there 2010-01-19T09:29:12 dbs: fwiw I did that just yesterday evening with no immediately apparent DB problems 2010-01-19T09:29:39 good! sort of. because regressions in pgsql 8.4 would be bad. 2010-01-19T09:30:56 ohhh 8.4... don't know about that 2010-01-19T09:31:06 but i look forward to hearing how this goes 2010-01-19T09:33:58 8.4 had some weird segfaults early on, but most of them seem to have been fixed by 8.4.2 2010-01-19T09:34:05 *** gdunbar has joined #evergreen 2010-01-19T09:34:31 GREAT. ingest with postgresql 8.3 / debian lenny worked (phew) 2010-01-19T09:35:42 so I guess I have two options on karmic: build postgresql 8.4.2 from scratch to see if it's a flawed package; if that still fails, build postgresql 8.3.x from scratch to confirm that it's an 8.3 vs. 8.4 thing 2010-01-19T10:00:32 *** eguest309_ has joined #Evergreen 2010-01-19T10:01:03 *** eguest309 has quit IRC 2010-01-19T10:01:04 *** eguest309_ is now known as eguest309 2010-01-19T10:12:33 *** gmcharlt has quit IRC 2010-01-19T10:18:37 *** jenny has quit IRC 2010-01-19T10:29:42 *** moodaepo has joined #evergreen 2010-01-19T10:38:22 *** r123 has joined #evergreen 2010-01-19T10:40:45 *** sfortin has joined #evergreen 2010-01-19T11:31:43 *** gmcharlt has joined #evergreen 2010-01-19T11:41:28 *** chrissharp123 has quit IRC 2010-01-19T11:41:28 *** dbwells has quit IRC 2010-01-19T11:41:28 *** jeff_ has quit IRC 2010-01-19T11:41:38 *** chrissharp123 has joined #evergreen 2010-01-19T11:41:59 *** dbwells has joined #Evergreen 2010-01-19T11:48:19 * dbs is surprised to find out that the buildbot is running on debian testing 2010-01-19T11:49:36 dbs: was the schema in your 8.4 db an upgraded one or fresh, when it was segfaulting? 2010-01-19T11:49:52 sounds like "apt-get install python-lxml" should resolve the problem 2010-01-19T11:49:59 miker_: it was fresh 2010-01-19T11:50:11 the most common cause of segfaults in pg backends (for me) has been mixed plperl/plperlu usage 2010-01-19T11:50:15 hrm... k 2010-01-19T11:50:31 well, everything should be plperlu now, so fresh rules that out 2010-01-19T11:53:26 I also dropped the cluster and recreated it with a C locale to hew as closely as possible to mainline, but still saw the same segfaults 2010-01-19T11:55:34 *** eguest309 has quit IRC 2010-01-19T11:55:52 I could try not adding plperl to see if that makes a difference 2010-01-19T11:56:28 *** eguest309 has joined #Evergreen 2010-01-19T11:57:41 does cstore return the new object on an update? 2010-01-19T11:59:35 *** jeff_ has joined #evergreen 2010-01-19T12:06:59 PG people... I have, say, 10 rows in a table and I want to concatenate the values from a given column for every row in said table into a single value/string 2010-01-19T12:07:12 can that be done with sql or does that require external processing? 2010-01-19T12:08:47 natschil: it does 2010-01-19T12:09:29 miker_: are you sure? it seems to return the id of the new object for me 2010-01-19T12:09:30 berick: SELECT array_to_string(array_accum(column_name),' ') FROM foo; 2010-01-19T12:09:46 natschil: hrm... no, you're probably right ... sorry 2010-01-19T12:10:07 miker_: np 2010-01-19T12:10:10 yes, create returns the new object, update returns the id on success 2010-01-19T12:10:35 thanks, miker_! 2010-01-19T12:10:49 I was thinking that update returned the object ... but, only create as a "quiet" setting (that does anything) 2010-01-19T12:11:04 *** jenny has joined #evergreen 2010-01-19T12:11:25 berick: that's assuming you want ' ' as the joiner, and it doesn't protect against nulls in the column 2010-01-19T12:11:46 to do that, s/column_name/coalesce(column_name,'')/ 2010-01-19T12:12:20 I'm learnding 2010-01-19T12:12:22 thanks again 2010-01-19T12:15:46 berick: alternative to moving the "DROP TRIGGER" stmt outside of the transaction is to use "DROP TRIGGER IF EXISTS..." because I think in-db ingest is targeted at pgsql 8.2 and above 2010-01-19T12:16:05 * dbs learns from miker_ as well 2010-01-19T12:16:35 heh 2010-01-19T12:17:02 I suggested the after-commit to berick, actually, since he was complaining about 8.1 ;) 2010-01-19T12:17:55 the IF EXISTS syntax is used in reporter-schema.sql already 2010-01-19T12:18:09 in the definition of reporter.disable_materialized_simple_record_trigger 2010-01-19T12:18:38 *** _dkyle_ has joined #evergreen 2010-01-19T12:18:56 I thought that got changed to use the plpgsql exception mechanism 2010-01-19T12:19:31 that's what I'm seeing in trunk 2010-01-19T12:21:21 * dbs added it in http://svn.open-ils.org/trac/ILS/changeset/14992/trunk/Open-ILS/src/sql/Pg/reporter-schema.sql 2010-01-19T12:22:30 rel_1_6_0 has the exception approach: http://svn.open-ils.org/trac/ILS/changeset/14996/branches/rel_1_6_0/Open-ILS/src/sql/Pg/reporter-schema.sqlhttp://svn.open-ils.org/trac/ILS/changeset/14996/branches/rel_1_6_0/Open-ILS/src/sql/Pg/reporter-schema.sql 2010-01-19T12:22:56 surely we're not trying to maintain 8.1 support in trunk? 2010-01-19T12:24:20 that comment should be s/stragglers using PostgreSQL 8.3 with 1.6/stragglers using PostgreSQL 8.1 with 1.6/ 2010-01-19T12:57:20 *** grantjohnson_ has joined #evergreen 2010-01-19T13:07:33 I don't get it. 2010-01-19T13:13:00 *** natschil has quit IRC 2010-01-19T13:30:09 dbs: agreed, re 1.6/8.1 2010-01-19T13:42:43 phasefx: got a sec for fast-extract? 2010-01-19T13:44:15 *** jenny1 has joined #evergreen 2010-01-19T13:45:11 grantjohnson: what's up? 2010-01-19T13:46:12 :phasefx Could not create file parser context for file "SYSCONFDIR/opensrf_core.xml" 2010-01-19T13:48:20 phasefx: Don't know if this is an environ. variable, php var or I'm missing a config file. 2010-01-19T13:48:33 grantjohnson: probably a literal string in the source code 2010-01-19T13:48:59 looks like a make variable 2010-01-19T13:49:07 a patch to make it smart (and default to /openils/conf/) would be great 2010-01-19T13:49:26 phasefx: doesn't exist here. No such file or directory at /usr/local/share/perl/5.8.8/OpenSRF/Utils/Config.pm line 328 2010-01-19T13:49:54 different perl version? 5.10 maybe? 2010-01-19T13:50:00 checking fast-extract 2010-01-19T13:50:34 you'll need to run it somewhere with OpenSRF, etc. installed. May get faster extract times on your database box, but if it's a dedicated box, might not have OpenSRF on it 2010-01-19T13:51:35 phasefx: it's hard coded in fast-extract. 2010-01-19T13:51:36 my ($config, $delim, $after,$deleted) = ('SYSCONFDIR/opensrf_core.xml', ' | '); 2010-01-19T13:52:08 yeah, change that to /openils/conf/opensrf_core.xml and you should be good 2010-01-19T13:52:30 ya... doh. Thanks 2010-01-19T13:52:46 phasefx: ./configure should fix that 2010-01-19T13:52:52 errr, grantjohnson 2010-01-19T13:52:58 dbs: ah 2010-01-19T13:53:27 if you're copying directly from Open-ILS/examples, then you'll get the unfixed var 2010-01-19T13:54:57 *** gmcharlt has quit IRC 2010-01-19T13:57:30 * dbs will check the autoconf stuff to see if it has fast-extract in its awareness 2010-01-19T13:58:25 yummy 2010-01-19T13:59:20 nope, it doesn't, or else the src file would be fast-extract.in and autoconf would generate fast-extract 2010-01-19T14:00:30 *** phase_bb has joined #evergreen 2010-01-19T14:01:53 *** jenny has quit IRC 2010-01-19T14:01:58 * dbs starts poking at it 2010-01-19T14:06:02 fast-extract seems to be running 2010-01-19T14:17:27 *** rsinger has joined #evergreen 2010-01-19T14:23:48 *** eguest309 has quit IRC 2010-01-19T14:26:19 *** B_Bonner has joined #evergreen 2010-01-19T14:32:24 *** rsinger has quit IRC 2010-01-19T14:35:35 okay, fix committed for fast-extract in trunk 2010-01-19T14:49:06 *** mrpeters-isl has joined #evergreen 2010-01-19T14:49:59 all, is there currently any support for holding specific information such as "autographed copy", etc or would that only be available as an item level circulation alert 2010-01-19T14:52:05 depends on what you need that information for. You have copy level notes that don't cause alerts 2010-01-19T14:52:22 and of course, stat cats 2010-01-19T14:52:37 phasefx: i thought so, but somehow, ive never come accross where to access copy level notes 2010-01-19T14:53:34 they're pretty buried (original goal was to de-emphasize notes over stat cats). In the opac, if you drill down to the item level, you'll see notes marked public. In the staff client, you have to go through the item attribute editor 2010-01-19T14:53:54 ok 2010-01-19T14:54:01 and where to copy stat cats appear? in opac as well? 2010-01-19T14:54:14 yes, if public 2010-01-19T14:54:23 ok, so drill down to "copy details" 2010-01-19T14:54:30 and it would appear there, if public 2010-01-19T14:55:01 are those searchable? 2010-01-19T14:55:14 no 2010-01-19T14:55:28 ok, but im sure they could be made searchable for the right price 2010-01-19T14:55:49 I'd imagine so :) 2010-01-19T14:56:11 you could of course embed notes in the MARC, and those could more easily be made searchable 2010-01-19T14:57:02 phasefx: that possible in a consortial setting though, where all libraries may be on the same bib, but only one of the holdings is that special copy with the special note 2010-01-19T14:58:38 yeah, putting stuff in the MARC wouldn't scope it to specific libraries (except in the case of the MFHD stuff) 2010-01-19T15:00:38 or 856's 2010-01-19T15:01:09 Shelving Locations are searchable; you could abuse those (Autograph Shelf) 2010-01-19T15:02:02 mrpeters-isl: if what you're wanting to cater to is vanity searches (donated by X), publishing bookbags with those books could be an option 2010-01-19T15:02:54 *** grantjohnson_ has quit IRC 2010-01-19T15:43:09 *** alxp has quit IRC 2010-01-19T15:48:39 Hey folks, I think I have tracked down a bug in Vandelay and am hoping for some confirmation/help. 2010-01-19T15:49:15 configure.ac now has a template for checking additional default directories for expected files / libdirs (I'm looking at you, Jeroen, and your Fedora) 2010-01-19T15:49:22 In the ingest_bib_items function, we have "FOR item_rule IN SELECT r.* FROM actor.org_unit_ancestors( queue_rec.owner )..." 2010-01-19T15:49:49 dbwells: I know very little about the workings of vandelay, alas - and have to head home. will peek at irc later 2010-01-19T15:50:29 but queue_rec.owner is an actor.usr id, not an actor.org_unit id. This seems to explain why it works for Admin and nobody else. Am I reading it right? 2010-01-19T15:53:04 <_bott_> dbs: I was just looking for something re: your 856 comment. Is that the display without a copy for electronic titles? 2010-01-19T15:53:56 dbwells: sounds like it wants a "select home_ou from actor.usr where id = queue_rec.owner" substituted there 2010-01-19T15:54:06 (off the top of my head) 2010-01-19T15:54:35 _bott_: aye - but it would be a huge stretch to use that for "autographed copies" :) 2010-01-19T15:54:59 dbs: I think so too. I will try to replace the function and see where that gets me. Thanks! 2010-01-19T15:55:45 <_bott_> dbs: right, I'm looking for what you demo'd at the conference, but haven't located details 2010-01-19T15:56:59 _bott_: ah - basically, if your 856 has a subfield $9 with the shortname of the org_unit that you want to scope it to, ingest will generate the required asset.uri / asset.uri_call_number_map / asset.call_number entries 2010-01-19T15:57:39 <_bott_> dbs: perfect, exactly what I was looking for! 2010-01-19T15:59:01 *** Meliss has quit IRC 2010-01-19T16:04:13 *** dbs has quit IRC 2010-01-19T16:22:46 *** bshum has quit IRC 2010-01-19T16:32:38 *** jenny has joined #evergreen 2010-01-19T16:41:37 *** atz has quit IRC 2010-01-19T16:49:33 *** jenny1 has quit IRC 2010-01-19T16:59:43 *** gmcharlt has joined #evergreen 2010-01-19T17:02:08 _bott_: only works that way in 1.6, i think. let me know if you find differently. i've got some things where we want to use this soon. :) 2010-01-19T17:06:25 *** moodaepo has quit IRC 2010-01-19T17:07:06 *** moodaepo has joined #evergreen 2010-01-19T17:07:23 *** r123 has quit IRC 2010-01-19T17:16:43 *** destkopPC has joined #evergreen 2010-01-19T17:17:43 hi there guys ! 2010-01-19T17:18:02 i need some support 2010-01-19T17:19:09 destkopPC: hey, what's up? 2010-01-19T17:19:30 i want to setup evergreen 2010-01-19T17:19:34 i want to test it 2010-01-19T17:19:52 im writing project proposal to a local library 2010-01-19T17:21:10 *** B_Bonner has quit IRC 2010-01-19T17:21:21 *** atz has joined #evergreen 2010-01-19T17:21:27 there's a demo server running you could connect to (albeit, one with a very old version of Evergreen). There's also a more recent virtualized image you could download and run inside of either VMWare or VirtualBox. You also have the option of installing the software yourself on a linux server 2010-01-19T17:21:51 i want to have mine 2010-01-19T17:21:58 i have hosting that support perl 2010-01-19T17:22:00 and mysql 2010-01-19T17:22:22 but i never installed perl scripts 2010-01-19T17:23:23 can anyone give my hand on this ? 2010-01-19T17:23:40 Evergreen makes use of postgresql (not mysql), apache, ejabberd, memcached, as well as perl and C. I'm betting a typical hosting company wouldn't be prepared to support such things 2010-01-19T17:25:30 destkopPC: you may want to start by downloading the virtual image and playing with it, just to see what you're in for. Then, if you want to build your own, you could try following the installation instructions on the wiki. Folks here and on the mailing list can help you if you get stuck, but they may not be able to guide you step by step through the entire process (we're all volunteers here, and have other jobs) 2010-01-19T17:27:18 *** B_Bonner has joined #evergreen 2010-01-19T17:28:45 destkopPC, hi 2010-01-19T17:28:58 hi Phurl 2010-01-19T17:29:00 :) 2010-01-19T17:29:12 *** brendan_ga has joined #evergreen 2010-01-19T17:30:01 destkopPC, come to #flossk channel 2010-01-19T17:30:35 *** jenny has left #evergreen 2010-01-19T17:36:39 * phasefx waves, heading out 2010-01-19T17:37:32 i c phasefx 2010-01-19T17:37:45 i have ubuntu server here 2010-01-19T17:38:17 i have a machine, i have installed ubuntu. it support postgresql as well 2010-01-19T17:46:22 destkopPC: You could try this > This Evergreen 1.6.0.0, Ubuntu 9.10 virtual image should run under recent versions of VirtualBox or VMWare. [ http://evergreen-ils.org/~denials/Evergreen-1600-Karmic.zip ] 2010-01-19T17:46:47 or did you want to install Evergreen from SCRATCH : ) 2010-01-19T17:47:13 moodaepo, im very very young tux 2010-01-19T17:47:19 :) 2010-01-19T17:48:23 Well then I would go with the virtual image (which I think phasefx also suggested) 2010-01-19T17:49:26 Download the file I linked and while it's downloading you can install vbox using your package manager 2010-01-19T18:09:43 *** brendan_ga has quit IRC 2010-01-19T18:09:43 *** gdunbar has quit IRC 2010-01-19T18:09:43 *** artunit has quit IRC 2010-01-19T18:11:02 *** sfortin has quit IRC 2010-01-19T18:11:22 *** B_Bonner has quit IRC 2010-01-19T18:13:22 *** gdunbar has joined #evergreen 2010-01-19T18:13:24 *** artunit has joined #evergreen 2010-01-19T18:45:50 *** rsinger has joined #evergreen 2010-01-19T19:43:06 *** rsinger has quit IRC 2010-01-19T19:52:32 *** destkopPC has quit IRC 2010-01-19T20:02:50 *** dbs has joined #evergreen 2010-01-19T20:05:13 dbwells: yeah, i realized on my walk home that the vandelay org_unit ids should probably be work_ous - looks like you came to the same conclusion 2010-01-19T20:25:32 *** Meliss has joined #Evergreen 2010-01-19T20:34:04 dbs: i may have asked you before, but location uris 856 $9 stuff requires 1.6, yes? 2010-01-19T20:34:18 yep 2010-01-19T20:36:27 *** Meliss has quit IRC 2010-01-19T20:37:10 thanks. :) 2010-01-19T21:18:54 * jeff digs into backdated checkins and generate_fines 2010-01-19T21:26:21 has anyone toyed with the idea of having a user-level "max outstanding fines" that works as a cap for overdue fines? 2010-01-19T21:28:24 I thought I heard such a desire expressed, but it may have been from you 2010-01-19T21:29:15 we've probably raised the question before 2010-01-19T21:30:02 i'm thinking a org unit setting that generate_fines consults based on the circ user's home ou 2010-01-19T21:30:43 if defined, it consults the current outstanding bills for overdues... that might be the hard part. 2010-01-19T21:31:14 maybe you could do circ bill balance for any circs not marked LOST... 2010-01-19T21:31:33 hard part is probably determining a logical way for it to work. 2010-01-19T21:34:05 maybe skip the whole thing unless there's an additional argument passed to enable it. 2010-01-19T21:34:38 (so as not to look up an org unit setting every circ you generate fines for) 2010-01-19T21:35:15 though, you should be able to cache that within the sub, so then it would just be one lookup for each unique user ou. 2010-01-19T21:35:43 I'm wondering if it would be easier (though kludgier) to have some cron job go around and tweak circs so that they stop generating fines as needed 2010-01-19T21:35:51 (does opensrf cache that kind of thing on a lower level?) 2010-01-19T21:37:22 i'm trying to determine if i want the fines to resume if the user makes a payment but doesn't return/renew/etc the item(s). 2010-01-19T21:37:37 * phasefx doesn't know.. "I'm not even sure if memcached is an OpenSRF thing or an OpenSRF application thing. I suspect the latter, and that OpenSRF doesn't do any caching on its own 2010-01-19T21:37:39 i think i do, though that would be a little silly of the user. 2010-01-19T21:38:45 yeah, a call to get the ou setting could cache it pretty easily, you'd make the opensrf call to the server, but not hit the db. not sure how much you'd save there, if any. 2010-01-19T21:39:17 probably a good idea to reduce calls where you can :) 2010-01-19T21:39:35 yeah, I'd cache in whatever app I was using 2010-01-19T21:41:05 Oh cool! with the XC NCIP toolkit we could use a REST interface to get XML from Evergreen... it's about time! 2010-01-19T21:41:21 dbs: it's almost like Jangle, without the NCIP :) 2010-01-19T21:42:00 Or unAPI, or supercat, or... 2010-01-19T21:42:35 what they're really saying is that they want a standard cross-ILS api that isn't NCIP 2010-01-19T21:45:16 * jeff learns about money 2010-01-19T21:45:16 :P 2010-01-19T21:49:24 jeff: so are you guys using that payment<->billing mapper thing now for linking stuff within a transaction? 2010-01-19T21:59:40 mmm, ISSNs are getting truncated at their first 4 digits in reporter.materialized_simple_record in trunk 2010-01-19T22:01:45 prolly cause that's how they're getting shoved into metabib.real_full_rec... 2010-01-19T22:01:47 * dbs keeps digging 2010-01-19T22:08:54 hmm, naco_normalize, you not being my friend? 2010-01-19T22:12:21 yeah, removing the \- from the pertinent tr/// line in public.naco_normalize avoids that issue. Not sure what ramifications that would have otherwise, though 2010-01-19T22:13:39 dbs: it deviates from the naco normalization standard :) 2010-01-19T22:14:21 how about a special case, like the controlfield stuff, to skip 022? 2010-01-19T22:14:36 was thinking about that 2010-01-19T22:15:00 my unicode is getting corrupted, too, might need to add Encode + decode_utf8() in there 2010-01-19T22:15:23 e.g. é was _not_ becoming e, but garbled 2010-01-19T22:15:42 in naco_norm? 2010-01-19T22:17:40 yeah 2010-01-19T22:17:56 but nope, that doesn't fix the problem. Weird. 2010-01-19T22:17:58 hrm... pg should be passing the data in as utf8 2010-01-19T22:18:49 oh. hmm. 2010-01-19T22:19:02 phasefx: i don't think we need that to do what i'm thinking with fines right now, but yes... i hope to test that Real Soon Now! 2010-01-19T22:19:35 jeff: yeah, I was just curious 2010-01-19T22:19:40 (hi miker_!) 2010-01-19T22:19:40 biblio.record_entry.marc has "Québec" already, mayhaps naco_normalize is corrupting that? 2010-01-19T22:21:59 jeff: hey :) 2010-01-19T22:22:56 apologies to all in advance, as i will likely be full of fines-related questions over the next week or two. :-) 2010-01-19T22:23:45 dbs: I'm not in a position to debug right now, but if you don't find it then I'll look tomorrow morning 2010-01-19T22:24:06 no problem miker_ 2010-01-19T22:24:14 this is what I get for not testing before :) 2010-01-19T22:27:01 :) 2010-01-19T22:28:58 anyone here have access to syndetics api docs / developer docs? 2010-01-19T23:26:27 hmm. I'm thinking that lc() is behaving mighty poorly, even with utf8 strings, because in the postgresql environment the locale is C 2010-01-19T23:48:11 so far from my reading the syndetics ac handler should be easy to teach to look at UPC 2010-01-19T23:48:35 openlibrary and amazon can do it, just not as easily/blindly 2010-01-19T23:50:02 (1x1 images with 200 OK are evil) 2010-01-19T23:50:13 absolutely agreed 2010-01-19T23:50:45 i'll bet openlibrary would be open to a tweak to give a clearer 404 2010-01-19T23:51:04 i think amazon already does, but i haven't played with their api since they started requiring the hashes 2010-01-19T23:51:37 syndetics appears to just be "give us the isbn and the upc and the oclc and we'll either give you an image or give you a 1x1" 2010-01-19T23:51:58 which is good for "get me a cover by any means from syndetics", but bad for stacking ac handlers 2010-01-19T23:55:37 yeah, I'm sure openlibrary would be willing to do the right thing 2010-01-19T23:56:07 even if it were an additional parameter so as not to break things that want the current behavior 2010-01-19T23:56:16 strict404=true or something 2010-01-19T23:56:56 i think you can still give an image AND a 404 2010-01-19T23:57:05 but that trips onError on an img, iirc. 2010-01-19T23:58:37 anyway, i'm going to try to get better docs, but i can get syndetics to look up with upc, now i just need to tweak the opac to extract and pass more keys. need to see if bibtemplate will be required / easier there.