2008-08-07T01:00:51 *** phase_bb has joined #openils-evergreen 2008-08-07T01:38:33 *** scottmcd has joined #openils-evergreen 2008-08-07T01:39:23 *** Mark__T has joined #openils-evergreen 2008-08-07T04:46:49 *** Mark__T has left #openils-evergreen 2008-08-07T04:48:02 *** Mark__T has joined #openils-evergreen 2008-08-07T04:53:27 *** kgs has quit IRC 2008-08-07T04:53:30 *** kgs__ has joined #openils-evergreen 2008-08-07T04:53:33 *** kgs__ is now known as kgs 2008-08-07T05:00:55 *** kgs__ has joined #openils-evergreen 2008-08-07T05:02:34 *** kgs has quit IRC 2008-08-07T05:03:36 *** kgs__ has quit IRC 2008-08-07T05:03:37 *** kgs___ has joined #openils-evergreen 2008-08-07T05:03:41 *** kgs___ is now known as kgs 2008-08-07T05:41:46 *** kgs___ has joined #openils-evergreen 2008-08-07T05:44:19 *** kgs____ has joined #openils-evergreen 2008-08-07T05:58:59 *** kgs has quit IRC 2008-08-07T06:02:16 *** kgs___ has quit IRC 2008-08-07T08:32:00 *** dbs has joined #openils-evergreen 2008-08-07T08:36:30 *** pmurray_away is now known as pmurray 2008-08-07T08:37:46 *** kgs_____ has joined #openils-evergreen 2008-08-07T08:37:50 *** kgs_____ is now known as kgs 2008-08-07T08:40:06 *** Slazer has joined #openils-evergreen 2008-08-07T08:41:36 *** kgs____ has quit IRC 2008-08-07T09:15:59 berick: So I guess the answer to avoiding LD_LIBRARY_PATH would either be to set rpath when linking libopensrf (just like pre-autotools days), or to use full paths in opensrf.xml to specify shared libs that don't start with "lib" (because ld cache doesn't contain those) 2008-08-07T09:16:39 * dbs has been reading dlopen man pages 2008-08-07T09:17:05 dbs: interesting, i had not considered using full paths... 2008-08-07T09:17:54 apparently if dlopen() sees a '/' in the filename, it treats it as a relative or full path. 2008-08-07T09:18:28 this assumes that everyone hates the idea of liboils_cstore.so, of course 2008-08-07T09:19:42 i kind of like the explici-tude of using full paths to the libs. 2008-08-07T09:19:55 well, I don't /hate/ it ... the other option is to have autotools set a constant in libopensrf to specify a search directory, and on load failure with the bare name try prepending that path 2008-08-07T09:19:58 you would, sicko 2008-08-07T09:20:02 haha 2008-08-07T09:20:07 berick: :( ... I don't 2008-08-07T09:20:16 but ... it's not the worst thing ever 2008-08-07T09:20:27 or - for more xml-itude - add /openils/lib to opensrf.xml 2008-08-07T09:20:28 and it does make the installation relocatable 2008-08-07T09:21:30 it being full paths 2008-08-07T09:22:09 rpath is relocatable with chrpath, apparently 2008-08-07T09:22:26 (but god help you on aix!) 2008-08-07T09:23:25 haha 2008-08-07T09:23:28 changing time 2008-08-07T09:24:21 chclothes business_casual dan 2008-08-07T09:51:26 argh. why are these mfhd records showing up in my marc dump? bad unicorn. bad. 2008-08-07T09:51:43 time to add a "skip if no 245 field is found" bit to marc2bre 2008-08-07T09:55:57 I also like brandon's suggestion to add the disable/enable trigger to parallel_pg_loader - any objections? 2008-08-07T09:56:59 manitoba greyhound rider beheaded and canabalized fellow passenger? 2008-08-07T09:57:01 that was unexpected 2008-08-07T09:57:11 (sorry, news playing in background) 2008-08-07T09:57:31 berick: yeah - crazy stuff. 2008-08-07T09:58:37 of course, miker_ will object because he uses pg_loader for more than just bibs :) 2008-08-07T10:00:08 i suppose i could just check the class hint 2008-08-07T10:03:33 *** Mark__T has left #openils-evergreen 2008-08-07T10:09:24 seeing no objections... 2008-08-07T10:09:59 ha 2008-08-07T10:10:59 dbs: I'm not opposed to a --required-field option 2008-08-07T10:11:25 miker_: Yeah, i think I'll do that too 2008-08-07T10:12:03 just let it be used multiple times, I suppose? 2008-08-07T10:12:04 for the trigger, I'm just going to prepend the disable call to the blah.mfr.sql file and append the enable call to the same file 2008-08-07T10:12:27 well, you can't :) 2008-08-07T10:12:39 whaddya mean? 2008-08-07T10:12:43 blah.mfr.sql is just a copy file 2008-08-07T10:13:03 COPY ... FROM blah.fmr.sql; 2008-08-07T10:13:21 ergh - so this has to happen in separate transactions 2008-08-07T10:13:25 no 2008-08-07T10:13:41 it just has to happen in the wrapper script 2008-08-07T10:14:17 oh - well that's no big 2008-08-07T10:14:26 dbs-- 2008-08-07T10:14:27 so if mfr is in use (there's a hash of hints actually seen) then spit out the disable a the beginning and the enable in the footer blob 2008-08-07T10:14:31 yep 2008-08-07T10:14:40 just a tweak to the path I was headed down 2008-08-07T10:14:51 indeed 2008-08-07T10:15:00 dbs = dUMbASs 2008-08-07T10:15:21 (that was for kbeswick) 2008-08-07T10:15:59 note to self: that may effect my un-committed "wrap mfr in a view for automatic substring indexing" stuff ... look at that later 2008-08-07T10:17:28 biab 2008-08-07T10:36:19 *** gmcharlt has joined #OpenILS-Evergreen 2008-08-07T10:41:46 *** irk has joined #OpenILS-Evergreen 2008-08-07T10:42:06 rk says hello 2008-08-07T10:42:57 *** dbs has quit IRC 2008-08-07T10:59:51 *** dbs has joined #openils-evergreen 2008-08-07T11:01:43 *** chrissharp123 has joined #OpenILS-Evergreen 2008-08-07T11:05:37 hrm, using Perl CGI, is there no way to access the exact URL (aka what the user typed into the browser), complete with query string, regardless of POST params? $cgi->url(-query=>1) returns the full URL, with POST params, but not the original query string... 2008-08-07T11:06:00 * berick knows it's something obvious 2008-08-07T11:08:24 would it be in the apache environment? 2008-08-07T11:08:43 should be 2008-08-07T11:09:16 actually, that's a good point.. i could ask mod_perl what it is 2008-08-07T11:09:16 miker_: it's not the most elegant solution for @required_fields, adding a loop label, but it works for me 2008-08-07T11:10:58 *** gmcharlt has quit IRC 2008-08-07T11:11:34 phasefx++ 2008-08-07T11:11:45 $apache->unparsed_uri 2008-08-07T11:14:41 *** asmodai_ has joined #OpenILS-Evergreen 2008-08-07T11:15:55 excellent 2008-08-07T11:16:54 * dbs gets the feeling that the SD distributor in Ireland probably didn't want that email to go to the list 2008-08-07T11:22:32 *** asmodai has quit IRC 2008-08-07T11:22:32 *** asmodai_ is now known as asmodai 2008-08-07T11:44:29 *** gmcharlt has joined #OpenILS-Evergreen 2008-08-07T12:15:03 dbs: :) 2008-08-07T12:16:24 "rats from a sinking ship" 2008-08-07T12:18:02 any objections to my adding RPATH back to libopensrf.so? 2008-08-07T12:18:31 you have 15 minutes, while I go and get my food, to object 2008-08-07T12:20:22 ha 2008-08-07T12:22:55 dbs: I like rpath better than full paths, so you'll get no objection from me 2008-08-07T12:55:05 don 2008-08-07T12:55:14 done! 2008-08-07T12:55:49 (full paths will still work even with rpath, by my reading of dlopen(), but rpath just lets us be lazy) 2008-08-07T12:56:56 and we've already got the huberis part down 2008-08-07T13:13:23 hubris 2008-08-07T13:13:28 (thanks S) 2008-08-07T13:25:20 *** phase_bb has quit IRC 2008-08-07T13:56:29 *** gmcharlt_ has joined #OpenILS-Evergreen 2008-08-07T13:57:19 *** gmcharlt has quit IRC 2008-08-07T14:19:34 *** gmcharlt has joined #OpenILS-Evergreen 2008-08-07T14:21:17 *** gmcharlt_ has quit IRC 2008-08-07T14:25:34 *** sarabee_ has joined #openils-evergreen 2008-08-07T14:38:43 *** sarabee has quit IRC 2008-08-07T14:38:55 *** miker_ has quit IRC 2008-08-07T14:55:16 dbs: i'm gonna give all the patrons in dev DB your email address so I can test overdue notices. OK? great... 2008-08-07T14:55:49 ha! 2008-08-07T14:55:53 berick++ 2008-08-07T15:00:22 *** kgs has quit IRC 2008-08-07T15:00:25 *** chrissharp123 has quit IRC 2008-08-07T15:03:32 *** kgs has joined #openils-evergreen 2008-08-07T15:10:57 berick: great 2008-08-07T15:11:38 *** gmcharlt has quit IRC 2008-08-07T15:12:54 *** Slazer has quit IRC 2008-08-07T15:44:20 dbs: so ... sboyette has volunteered to start poking at Module::Install-ification ... starting with the opensrf modules, so as to have an end-to-end "standard" install environment (instead of cp -R) for opensrf 1.0 ... objections to a branch in the repo for that? 2008-08-07T15:45:43 none 2008-08-07T15:46:43 * eeevil adds svn://svn.open-ils.org/OpenSRF/branches/sboyette 2008-08-07T15:55:53 *** gmcharlt has joined #OpenILS-Evergreen 2008-08-07T16:32:30 *** pmurray is now known as pmurray_away 2008-08-07T16:48:36 *** gmcharlt has quit IRC 2008-08-07T16:50:04 *** gmcharlt has joined #OpenILS-Evergreen 2008-08-07T16:51:34 @seen gfawcett 2008-08-07T16:51:34 dbs: I have not seen gfawcett. 2008-08-07T16:51:39 @seen grahamf 2008-08-07T16:51:39 dbs: grahamf was last seen in #openils-evergreen 11 weeks, 1 day, 4 hours, 54 minutes, and 11 seconds ago: denials: cool, I'll remember. I'm always grahamf. :-) 2008-08-07T16:51:53 see, _I_ should have remembered that 2008-08-07T16:53:55 *** dbs has quit IRC 2008-08-07T17:00:06 *** irk has quit IRC 2008-08-07T17:07:26 *** gmcharlt has quit IRC 2008-08-07T18:05:49 *** kgs has quit IRC 2008-08-07T18:38:50 *** _dkyle_ has left #openils-evergreen 2008-08-07T18:50:13 *** phase_bb has joined #openils-evergreen 2008-08-07T19:09:53 *** djfiander has joined #OpenILS-Evergreen 2008-08-07T19:32:30 ok, I'm here. now what? 2008-08-07T20:03:56 burritos for everyone! 2008-08-07T20:07:38 *** miker_ has joined #openils-evergreen 2008-08-07T20:59:08 * djfiander reads MIT's receiving manuals 2008-08-07T21:01:11 nice. MIT libraries get a 10% discount on MIT press titles. 2008-08-07T21:01:15 cheap bastard publishers. 2008-08-07T21:34:31 *** djfiander has quit IRC 2008-08-07T21:40:12 *** dbs has joined #openils-evergreen 2008-08-07T21:48:36 it looks like etch-backports has pulled postgresql-8.2 2008-08-07T21:49:56 pg8.2.. the pariah 2008-08-07T22:08:19 wtf??? 2008-08-07T22:08:40 8.3 hasn't been out but, like, 6 months 2008-08-07T22:11:27 yeah: i guess it's go hard or go home 2008-08-07T22:11:42 that's so very un-debian-like 2008-08-07T22:11:45 I also had the revelation tonight that I've been thinking about bib import completely ass-backwards 2008-08-07T22:13:40 I've been worrying heavily about keeping all bib ids unique, ensuring that mcmaster's bibs don't conflict with lu's bibs, and thinking about adding a "this bib ID was already in use, so here's what it got remapped to" output hash for subsequent holdings import 2008-08-07T22:14:39 but - blinding revelation - I /want/ the bib ID to match an existing one in the system, if it's a good match (like oclc number, or isbn, or issn) - because I don't want dupe bib records - 2008-08-07T22:14:49 and the holdings will naturally get attached later 2008-08-07T22:14:51 duh? 2008-08-07T22:14:53 heh 2008-08-07T22:17:05 dbs: well ... "naturally" ... 2008-08-07T22:17:19 ESI sells deduplication services ;) 2008-08-07T22:17:20 as long as people don't get to picky / territorial about their bib records :) 2008-08-07T22:17:27 miker_: hah! 2008-08-07T22:19:14 miker_: hell, man, I was using accession number for LU and Mac's bibs because it's the only unique identifier, but of course there's significant overlap 2008-08-07T22:19:57 so I was thinking "crap, I have to alter marc2bre.pl to keep track of the original idfield value and spit out the mapping so that I can get the holdings right" 2008-08-07T22:20:45 but yeah - my brain was obviously soiled by my immersion in LU's "we have 12 partners and there shall be no shared bibs" policy 2008-08-07T22:21:06 miker_: how much do those deduping services cost, anyway? 2008-08-07T22:21:36 dbs: depends on the quality of the incoming data ;) 2008-08-07T22:22:06 they're all unique, time to go home 2008-08-07T22:22:33 heh 2008-08-07T22:22:59 "match on catkey + library policy name" ;) 2008-08-07T22:23:29 dhttp://lists.backports.org/lurker-bpo/message/20080731.141614.a7614dbb.en.html 2008-08-07T22:27:02 geez 2008-08-07T22:27:22 "just use this pgdump wrapper" wtf? 2008-08-07T22:27:24 *** kgs has joined #openils-evergreen 2008-08-07T22:27:25 man ... 2008-08-07T22:27:54 mr fuchs must not have read the release notes for 8.3... 2008-08-07T22:31:48 yeah, that was surprising 2008-08-07T22:44:20 sboyette++ # tests! 2008-08-07T22:44:53 they are a few of my favorite things 2008-08-07T22:48:20 I can't tell you how much I love you. 2008-08-07T22:50:20 you can love me after it's folded back into trunk. although my goal is to get the perl modules installing seamlessly, not to build a test suite, so there will be VERY few tests at that time, most likely 2008-08-07T22:50:34 * phasefx regresses 2008-08-07T22:50:37 but the harness comes for free, and once it's in place, tests can be added one at a time 2008-08-07T22:50:42 sboyette: exactly 2008-08-07T22:51:49 phasefx: exactly :) 2008-08-07T22:52:00 phasefx: soon you're going to be mean 2008-08-07T22:56:55 (then, one day, Devel::Cover) 2008-08-07T22:57:54 * kgs does not believe phasefx will ever be mean 2008-08-07T22:58:33 *** kgs has quit IRC 2008-08-07T23:00:58 nature vs demeanor 2008-08-07T23:09:11 *** kgs has joined #openils-evergreen 2008-08-07T23:35:17 kgs: send me the sample please :) 2008-08-07T23:35:36 dbs: roger roger :) 2008-08-07T23:44:24 *** kgs has quit IRC