2008-07-25T00:01:21 @later dbs ha! I fixed the Vidoe typo with some other stuff, but you beat me to the commit! 2008-07-25T00:01:21 miker_: Error: The "Later" plugin is loaded, but there is no command named "dbs" in it. Try "list Later" to see the commands in the "Later" plugin. 2008-07-25T00:03:31 @later tell dbs ha! I fixed the Vidoe typo with some other stuff, but you beat me to the commit! 2008-07-25T00:03:31 miker_: The operation succeeded. 2008-07-25T01:21:24 *** enculturation has joined #openils-evergreen 2008-07-25T01:21:31 FYI, manually executing 300.schema.search.sql against my database seemed to resolve my problem. I don't know why this didn't work from the build-db.sh script, as I copied and pasted (changing arguments as appropriate) from that script. 2008-07-25T01:23:37 *** enculturation has quit IRC 2008-07-25T01:29:54 *** Mark__T has joined #openils-evergreen 2008-07-25T08:10:27 *** Slazer has joined #openils-evergreen 2008-07-25T08:11:06 *** kbeswick has joined #openils-evergreen 2008-07-25T08:16:52 *** kgs_away has joined #openils-evergreen 2008-07-25T08:24:19 *** kgs_away is now known as kgs 2008-07-25T08:54:31 *** Mark__T has quit IRC 2008-07-25T08:56:46 *** _dkyle1 has joined #openils-evergreen 2008-07-25T09:05:46 kbeswick: i'm gonna send a patch to ils-dev for removing legacy json support, but i'm only editing the makefiles, etc. as much as i have to. i'd like to leave the rest of the cleanup up to you if possible 2008-07-25T09:06:28 for opensrf, that is 2008-07-25T09:08:02 berick: np 2008-07-25T09:09:24 thanks 2008-07-25T09:14:12 *** dbs has joined #openils-evergreen 2008-07-25T09:29:23 kbeswick: one other thing i've noticed with the autotools stuff, is the packages/scripts veriables in src/python/setup.py are getting doubled up. e.g. 2008-07-25T09:29:27 packages=['/home/opensrf/OpenSRF/trunk/src/python//home/opensrf/OpenSRF/trunk/src/python/osrf'], 2008-07-25T09:31:21 berick: fixed already... i should make a patch and send it.. ill do it now 2008-07-25T09:31:31 kbeswick++ 2008-07-25T09:31:42 right after i peel this orange. 2008-07-25T09:31:42 kbeswick: excellent! fly, little birdy, fly out of the nest on your patched wings! 2008-07-25T10:19:38 Cool article via George: http://radar.oreilly.com/2008/07/oscon-day-1-an-open-source-pro.html 2008-07-25T10:19:57 * dbs fears being a blocker 2008-07-25T10:22:06 chromatic is the most curmudgeonly hacker i know 2008-07-25T10:24:14 (respect) 2008-07-25T10:36:11 *** agJohn has joined #openils-evergreen 2008-07-25T10:43:03 dan blocker scott 2008-07-25T10:45:33 kbeswick++ 2008-07-25T10:46:56 arg, big fm_IDL.xml collision in acq merge 2008-07-25T10:51:21 ruh-roh. 2008-07-25T10:58:20 berick: are you cool with moving C source files to c.in for autoconf processing, or would you prefer the SYSCONFDIR thing get defined in a header (.h.in) instead? 2008-07-25T10:59:59 dbs: i wonder if we can just get rid of any instances where sysconfdir is embedded in the code. 2008-07-25T11:00:16 the gateway, for example, doesn't need the default path 2008-07-25T11:00:34 i mean, it'll use it if it has to.. 2008-07-25T11:01:00 i'd prefer not to move the C files to c.in 2008-07-25T11:01:22 Yeah, I suspected as much 2008-07-25T11:01:32 (somewhere, kbeswick is weeping bitterly) 2008-07-25T11:02:18 or sharpening his knives 2008-07-25T11:03:55 *** pmurray_away is now known as pmurray 2008-07-25T11:05:06 IUC, the gateway is the only C files that has embedded SYSCONFDIR 2008-07-25T11:05:11 s/files/file/ 2008-07-25T11:06:56 indeed 2008-07-25T11:08:27 kbeswick: what is the ${prefix}|@prefix@| changing in the gateway? 2008-07-25T11:09:14 dbs: @sysconfdir@ wil expand to ${prefix}/etc 2008-07-25T11:09:28 ahh 2008-07-25T11:09:28 so that just rplaces the ${prefix} to /opensrf/ 2008-07-25T11:09:44 so it should be easy to back out that change, then 2008-07-25T11:10:03 yeah if we don't need to replace anything in the file, its very easy 2008-07-25T11:10:26 all you need to do is take out those sed lines, and remove the line from the .c file 2008-07-25T11:10:56 oh, and rename it back to .c insteado f .c.in 2008-07-25T11:12:03 heh, no renaming (that would keep the SYSCONFDIR beastie in place) 2008-07-25T11:12:47 berick: so removing the #define GATEWAY_DEFAULT_CONFIG directive from the .c file will not cause any problems? 2008-07-25T11:13:26 kbeswick: will need to remove the function that sets the default as well, but it should not cause any problems 2008-07-25T11:13:37 i can try it out real fast to be sure 2008-07-25T11:13:46 berick: k cool 2008-07-25T11:20:39 berick: some warning vaguely pulses in my brain that suggests breaking variable declaration and initialization into two separate statements 2008-07-25T11:21:37 dbs: yeah? 2008-07-25T11:22:29 from a read-ability perspective? 2008-07-25T11:22:30 maybe just reentrant code with xlC on AIX, but I recall (perhaps incorrectly) some problem where the optimizer initialized the variable the first time through for stmts like "char * tmp = NULL;" but on subsequent calls that var would be pointing off to empty space 2008-07-25T11:23:07 well, that's just messed up 2008-07-25T11:23:24 regardless, in this case, the vars were already pointing into empty space 2008-07-25T11:23:29 right :) 2008-07-25T11:23:54 and I don't expect to be compiling this on AIX any day soon. 2008-07-25T11:25:33 kbeswick: oh, removing the GATEWAY_DEFAULT_CONFIG and associated bits works fine 2008-07-25T11:25:44 i'll commit that change 2008-07-25T11:39:32 berick: you'll be happy to know that conify (and by extention the http translator) mostly works in opera 9.51 ... some glitches with dijits, but nothing too bad 2008-07-25T11:40:09 miker_: interesting. are you doing all synchronous requests? 2008-07-25T11:40:29 I'm not certain ... I think there are some streaming requests in there, but maybe not 2008-07-25T11:41:21 that's good news, though, for dojo and the translator 2008-07-25T11:42:09 well... I take that back ... some things are not happy :( 2008-07-25T11:42:20 it renders, but some stuff doesn't end up being displayed 2008-07-25T11:42:22 bah 2008-07-25T11:42:23 those are probably the streaming requests; 2008-07-25T11:44:24 i think i just need to add if(!navigator.userAgent.match(/mozilla/i)) {force non-streaming} 2008-07-25T12:08:01 *** agJohn has quit IRC 2008-07-25T12:25:01 *** agJohn has joined #openils-evergreen 2008-07-25T13:29:10 hmm, didn't think about checking the DateTextBox patch with Opera... one sec 2008-07-25T13:33:19 uh oh... i've worn out the ;/: key on my keyboard 2008-07-25T13:33:26 this does not bode well 2008-07-25T13:33:33 i /need/ that key 2008-07-25T13:33:53 haha .... seriously? 2008-07-25T13:34:38 yes :( i have to hold it in with pressure to get it to work 2008-07-25T13:34:49 oh noes 2008-07-25T13:35:02 darn, i like this keyboard, too. 2008-07-25T13:35:08 * berick digs an old one out of the closet 2008-07-25T13:38:23 too much C 2008-07-25T13:40:09 heh, too much vim, i think 2008-07-25T13:40:09 berick: just type in the ASCII or UNICODE code each time. or remap your PRT-SCREEN key accordingly :) 2008-07-25T13:40:12 * berick :q's 2008-07-25T13:51:32 berick: about the DateTextBox patch - one thing I would obviously do before committing it would be to turn off debug mode :) 2008-07-25T13:52:57 hah! let's leave it! 2008-07-25T13:53:30 "the opac is sassing me!" 2008-07-25T13:53:47 oh, well I guess I shouldn't worry about testing Opera, as the dynamic opac doesn't seem to support it anyway 2008-07-25T13:55:26 oh wait - that appears to be dev.gapines.org is smoked 2008-07-25T13:57:33 hmm, well dev.gapines.org is smoked (FF is barfing) but FF works on demo.gapines.org while Opera complains 2008-07-25T13:58:24 dbs pasted "Opera complaints about OPAC" at http://paste.lisp.org/display/64211 2008-07-25T13:59:58 I'm guessing "parent" is undefined there 2008-07-25T14:00:10 'cause that's the clever kinda guesser I am 2008-07-25T14:07:01 that's odd ... the stanza above should keep the line in question from being evaluated ... or do you mean literally 'undefined' ? 2008-07-25T14:11:31 huh, opera used to work 2008-07-25T14:14:23 indeed it did 2008-07-25T14:14:37 dbs: ffx is barfing with the dojo stuff? 2008-07-25T14:15:07 miker_: ffx is barfing on dev.gapines.org, the front page 2008-07-25T14:15:18 hrm... so it is 2008-07-25T14:15:22 that's strange 2008-07-25T14:15:27 it is strange :) 2008-07-25T14:15:42 ahh ... autogen 2008-07-25T14:16:16 kinda thought so 2008-07-25T14:16:33 we had some build issues last night :) 2008-07-25T14:16:52 related to the removal of objson and fixin' of cstore 2008-07-25T14:17:14 all better now 2008-07-25T14:17:52 miker_: as for parent, yeah, I meant that Opera seems to be complaining that the preceding line is returning undefined (or null? sigh) and that it's freaking out because the if() clause in 106 is invoking a method on it 2008-07-25T14:18:16 right ... but the lines above check for null and '' 2008-07-25T14:18:31 on thing.parent() before sticking that in var parent 2008-07-25T14:18:51 that's why I was confused ... 2008-07-25T14:19:04 are we in violent agreement? 2008-07-25T14:20:20 * dbs tries out opera's dragonfly (firebug competitor) 2008-07-25T14:20:45 I was not impressed with dragonfly ... 2008-07-25T14:20:54 but, then, I know firebug a bit 2008-07-25T14:44:05 weird. in debug mode in opera, findOrgUnit(x.parent_ou()) returns an object. 2008-07-25T14:44:23 *** kbeswick has quit IRC 2008-07-25T14:45:06 *** Slazer has quit IRC 2008-07-25T14:45:59 *** sydw has joined #OpenILS-Evergreen 2008-07-25T14:46:26 *** phasefx has quit IRC 2008-07-25T14:48:48 *** eeevil has quit IRC 2008-07-25T14:49:44 *** eeevil has joined #openils-evergreen 2008-07-25T14:50:42 *** phasefx has joined #openils-evergreen 2008-07-25T14:51:11 okay, it's barfing on the .children() call 2008-07-25T14:51:30 (yes yes, "duh, dbs") 2008-07-25T15:13:20 sydw pasted "Apache startup Error" at http://paste.lisp.org/display/64215 2008-07-25T15:14:11 does anyone know what to when encountering the above error? OSRFGatewayConfig related 2008-07-25T15:14:51 sydw: is osrf_json_gateway module loaded in your apache config? 2008-07-25T15:15:46 specifically, loaded before eg_vhost.conf is read 2008-07-25T15:15:59 I'll check... 2008-07-25T15:24:19 sydw: it's usually loaded in httpd.conf, fwiw 2008-07-25T15:27:51 thanks all... 2008-07-25T15:28:02 it wasn't loaded :-) 2008-07-25T15:31:36 okay, gave up trying to debug why opera is freaking out about type mismatch and just tried out a simple DateTextBox example; that's fine. 2008-07-25T15:33:18 cool 2008-07-25T15:41:50 you konq, I love opera and all, but at 1.7% market share... ;) 2008-07-25T15:41:53 ha 2008-07-25T15:41:55 konq ... 2008-07-25T15:41:57 konw 2008-07-25T15:56:40 kpwn'ed 2008-07-25T16:00:17 kown ... man ... my keyboard sure is p0wnzd 2008-07-25T16:06:11 aside to berick and his new keyboard: any thoughts about a) DateTextBox patch and b) article would be appreciated 2008-07-25T16:09:13 dbs: have not applied the patch yet. do you notice any difference in page load times? 2008-07-25T16:09:25 * berick looks at article email 2008-07-25T16:09:50 berick: nothing subjectively, but we all know that's no great test. 2008-07-25T16:10:34 berick: you really don't need to apply the patch - it works, subject to the regression mentioned in the accompanying email (loss of red highlights for invalid field) 2008-07-25T16:12:10 i'd like to apply it just to see how it feels. i'll do it now.. 2008-07-25T16:13:01 arg, using_a_mac_for_work-- 2008-07-25T16:13:03 ah yeah - I should have thrown it onto biblio-dev. argh 2008-07-25T16:13:09 a mac?! 2008-07-25T16:13:19 *sigh* F is borrowing my thinkpad 2008-07-25T16:13:43 and they keyboard is officially useless now 2008-07-25T16:14:54 berick: that sucks 2008-07-25T16:15:30 Windows task manager shows a bump of a few megs in loading the home page for my local system vs. dev.gapines.org 2008-07-25T16:25:02 heh, i love how the browser jitters after you set the date 2008-07-25T16:25:38 berick: that's like tactile feedback 2008-07-25T16:25:46 you're supposed to be using a rumblepad 2008-07-25T16:25:52 hah 2008-07-25T16:26:16 ooh, can use a dancepad to change interfaces in the staff client... 2008-07-25T16:27:03 arg, acq box is being difficult re: placing holds. the interface works fine, though 2008-07-25T16:27:50 berick: oh? I was developing against trunk. hrm. 2008-07-25T16:28:24 dbs: it's the same as far as the opac goes 2008-07-25T16:28:34 thus far, anyway 2008-07-25T16:28:46 so placing holds isn't working for you with this patch? 2008-07-25T16:29:02 no, sorry, placing holds isn't working for other reasons 2008-07-25T16:29:15 i think the patch is fine 2008-07-25T16:29:20 okay, I thought so, but was just being paranoid 2008-07-25T16:32:44 ah, missing circ script path 2008-07-25T16:35:07 *** dbs has quit IRC 2008-07-25T16:48:50 *** sydw has left #OpenILS-Evergreen 2008-07-25T16:51:38 *** _dkyle1 has left #openils-evergreen 2008-07-25T17:06:02 *** pmurray is now known as pmurray_away 2008-07-25T17:51:52 eeevil, I'm curious why asset.stat_cat_entry_copy_map has a stat_cat column, it seems redundant? Was that for performance reasons? 2008-07-25T17:53:30 eeevil, nevermind, I get it now, Only one stat_cat_entry per stat_cat per copy. 2008-07-25T17:59:39 *** dmcmorris_esi has joined #openils-evergreen 2008-07-25T18:00:15 just got the google alert on sydw's paste... 2008-07-25T18:02:23 *** Stompro has quit IRC 2008-07-25T19:42:37 Yo, miker_, you working late, Bro? 2008-07-25T19:43:08 agJohn: yeah, but about to step away. I'll ping back in a few minutes? 2008-07-25T19:43:21 No prob. I'll send an email. 2008-07-25T19:58:57 miker_: I'm assuming you can read an email w/ HTML--otherwise I should maybe try again! 2008-07-25T20:08:09 agJohn: responding now, actually 2008-07-25T20:09:21 Cool beans... I'd like to encourage you to join me in not working as late as we usually do! (My wife spent all day working outside in the yard; gotta take the ol' gal to dinner!) 2008-07-25T20:09:37 haha 2008-07-25T20:10:06 well, 3 migrations ongoing, so not much chance of that. but I appreciate the encouragement! ;) 2008-07-25T20:10:28 3 !!! Do you have all the incompetent help you need? 2008-07-25T20:10:56 heh ... the more the merrier! 2008-07-25T20:11:59 Well, I'll check in with you later. Best of luck with the bits falling where they should. 2008-07-25T20:12:19 thanks :) 2008-07-25T20:22:49 You up for another distraction, miker_? What have I got messed up if I'm getting this on the COPY INTO authority.full_rec table? 2008-07-25T20:22:50 ERROR: could not find tsearch config by locale 2008-07-25T20:23:31 Something subtle in the scripts that add stuff to template1 before creating an Evg DB? 2008-07-25T20:23:51 (Subtle, he says hoping it wasn't dumb!) 2008-07-25T20:24:59 no, that's no biggie 2008-07-25T20:25:09 sec ... searchin' 2008-07-25T20:26:35 I assumed it was related to the trigger on authority.full_rec : authority_full_rec_fti_trigger and the call to the tsearch2 function in it.... 2008-07-25T20:27:44 select set_curcfg('default'); 2008-07-25T20:27:56 it is indeed 2008-07-25T20:29:21 Damn you're good! 2008-07-25T20:29:30 http://www.sai.msu.su/~megera/oddmuse/index.cgi/Tsearch_V2_Notes 2008-07-25T20:29:38 search down for "config by locale" 2008-07-25T20:30:05 if the server's locale is, say en_US.UTF8 2008-07-25T20:30:30 but postgres and tsearch2 were installed with C as the locale 2008-07-25T20:30:55 then you'll get that unless you call set_curcfg (nice and vague, eh?) 2008-07-25T20:30:59 So I wonder where I missed that happening in what I just did. I was running build-db.sh 2008-07-25T20:31:00 Ah so, I make the mistake a long way back.... 2008-07-25T20:31:17 made, I mean (when I installed Postgres) 2008-07-25T20:32:33 no, you're good 2008-07-25T20:32:36 you want C locale 2008-07-25T20:33:25 it bites some distribution/Pg/version combos and not others, and I haven't found a pattern. 2008-07-25T20:33:46 the sane thing for /me/ to do would be to call set_curcfg during the schema build 2008-07-25T20:34:02 ... so, I'll fix /that/ now, too :) 2008-07-25T20:34:14 So I missed it in the setup because I was only doing a partial setup (just a new DB)--I like the idea of adding it to the schema build. Cool. 2008-07-25T20:34:21 agJohn: you've found and lead to the fix of 2 bugs tonight! :) 2008-07-25T20:34:55 One of my foster sisters would say, in this situation: "He's NOT just a pretty face!" 2008-07-25T20:35:24 Huhhh. 2008-07-25T20:35:32 :) 2008-07-25T21:19:01 *** dbs has joined #openils-evergreen 2008-07-25T21:25:46 * dbs does a double-take: we supported pgsql 8.0? oh. 2008-07-25T21:25:55 also: I have no life. 2008-07-25T21:27:30 dbs: heh ... not since 1.0 ;) 2008-07-25T21:27:55 but, yeah ... most of EG was developed against 8.0 ... 2008-07-25T21:28:05 well, most of the early stuff 2008-07-25T21:28:50 that makes sense :) 2008-07-25T21:29:36 can't believe how excited I was about 8.0 at the time ... and 8.1? whew! 2008-07-25T21:29:51 LIKE IT WAS FILLED WITH MAGIC 2008-07-25T21:30:22 Now you've got that nice jaded thing going for you - work it, work it 2008-07-25T21:30:47 Have you played with GIN indexes for the FTS much? 2008-07-25T21:31:09 (as I immediately think - or wait, GIST?) 2008-07-25T21:32:08 I have, some. they're faster by a little 2008-07-25T21:32:12 much faster to build 2008-07-25T21:32:16 some 20x 2008-07-25T21:32:35 but you have to be exxxxtra careful querying them 2008-07-25T21:33:12 for indexed tsvecor cols, if you give it an empty tsvector it will throw an error 2008-07-25T21:33:23 er 2008-07-25T21:33:30 empty tsquery, excuse me 2008-07-25T21:34:42 which is annoying 2008-07-25T21:34:57 but avoidable ... just not avoided in every possible way ATM 2008-07-25T21:35:53 hmm. I'll assume that they'll be better in 8.4 then :) 2008-07-25T21:37:08 worth testing ... or protecting against 2008-07-25T21:38:32 tonight's plan: commit Dojo datetextbox stuff. then send article revision to the editor gods. 2008-07-25T21:39:01 kgs: yes, we will have academic reserves in 2.0 2008-07-25T21:40:28 that was ... random 2008-07-25T21:42:09 miker_: knocking off a todo in the fastest possible fashion - albeit potentially ineffective 2008-07-25T21:42:26 haha 2008-07-25T21:52:30 dbs: regarding JVR's question about dojo, I figure we'll just distribute a "matching" version inside the tarball, perhaps inside the repo ... since it's (mostly) BSD ... just need to vet the parts we include in the dist for BSD-ness 2008-07-25T23:45:35 miker_: that's cool. do it as part of the build rather than dumping it in SVN I assume 2008-07-25T23:47:57 well, actually 2008-07-25T23:48:10 I was thinking it would make sense to put what we actually use into svn 2008-07-25T23:48:18 or the whole (bsd) thing 2008-07-25T23:48:27 so that it's stable to a version 2008-07-25T23:49:14 and we don't have to go grab some version X from an archive site to work on 1.4.3 2008-07-25T23:51:54 I could be convinced that that's not good/optimal 2008-07-25T23:53:19 in any case, the end user should not have to go fetch it ... there's basically no chance ever of dojo having a system package ;) 2008-07-25T23:53:37 (and at least there's a chance of that with our other deps)