2008-06-19T01:29:45 *** Mark__T has joined #openils-evergreen 2008-06-19T08:25:06 *** Slazer has joined #openils-evergreen 2008-06-19T08:42:27 *** kbeswick has joined #openils-evergreen 2008-06-19T09:15:34 *** gmcharlt has joined #OpenILS-Evergreen 2008-06-19T09:45:36 kbeswick: okay, a little more feedback on the opensrf autotoolsification 2008-06-19T09:46:19 for options like --with-pid_dir, either use all hyphens (e.g. --with-pid-dir) or just drop the _dir part 2008-06-19T09:46:38 which do you think would be better? 2008-06-19T09:49:16 I'm considering a third option, which would be to do away with the option altogether and just rely on --localstatedir 2008-06-19T09:50:46 that would mean that pid, log, and sock directories would all go in that localstatedir? 2008-06-19T09:50:50 as pid (aka "run" in the FHS) should be in $(localstatedir)/run, sock should be in $(localstatedir)/lock (i think) 2008-06-19T09:51:18 log in $(localstatedir)/log/opensrf perhaps - that one's a bit trickier 2008-06-19T09:51:55 maybe just in $(localstatedir)/log 2008-06-19T09:52:17 FHS = filesystem hierarchy standard, by the way 2008-06-19T09:52:30 yeah i just googled it 2008-06-19T09:52:30 haha 2008-06-19T09:53:33 so take out the options, make those three directories go into $(localstatedir)/directory ? 2008-06-19T09:55:18 that's what makes the most sense to me, but it would be worthwhile 1) testing that it actually works (in case there are hard-coded assumptions in scripts) 2008-06-19T09:55:32 and 2) getting at least a head-nod from berick and/or miker_ 2008-06-19T09:55:59 *** Karen__ has joined #OpenILS-Evergreen 2008-06-19T09:56:33 it will work 2008-06-19T09:56:48 * berick nods head 2008-06-19T09:56:48 well 2008-06-19T09:56:51 wait, what? ;) 2008-06-19T09:56:56 k, ill test it 2008-06-19T09:58:52 kbeswick: okay, here's an example of a hardcoded assumption 2008-06-19T09:59:12 when you install osrf_ctl.sh, the default for pid dir is hardcoded as /tmp 2008-06-19T09:59:18 you need to autotoolize that 2008-06-19T09:59:28 *** Mark__T has left #openils-evergreen 2008-06-19T10:00:44 i could see $(localstatedir)/lock/opensrf and /run/opensrf since there will be several files going into each. 2008-06-19T10:01:12 but not a big deal either way 2008-06-19T10:01:37 it would be easy to do that anyway, so might as well 2008-06-19T10:02:37 likewise, conf file default is currently figuring out the default as $BASEDIR/conf/opensrf_core.xml - that needs to be autotoolized to default to $(sysconfdir)/opensrf_core.xml 2008-06-19T10:02:51 i've noticed when you install apache, for example, if you use the default prefix, it puts all the files just-so (without the extra blah/apache2/...), but when you install it using FHS, it puts the prefixes in automatically. i don't know how big of a pain that would be or how much it matters, honestly 2008-06-19T10:03:12 not sure if that made sense... 2008-06-19T10:06:43 it's been a few years since I last compiled apache (stopped back when php builds got smarter) 2008-06-19T10:07:36 prefixes = application name, like /usr/local/var/log/appname/ ? 2008-06-19T10:08:08 versus /usr/local/var/log/ for default? 2008-06-19T10:09:45 /var/log/apache2 (taking the debian example) is what's used if I set no prefix. if i set a prefix, it uses /log/ 2008-06-19T10:10:35 i think it their case they define a set of defaults that conform to FHS, but if you choose a prefix it plops everything into in the simplest possible way 2008-06-19T10:10:39 s/it/in/ 2008-06-19T10:11:26 not trying to derail, i promise, it just popped into my head 2008-06-19T10:11:33 ah, I get it 2008-06-19T10:13:50 yeah, I dunno if the apache way is recommended or whether it stems from legacy practices 2008-06-19T10:14:07 me either 2008-06-19T10:15:14 :) 2008-06-19T10:15:28 hey, look over there! 2008-06-19T10:15:30 * berick shuffles off 2008-06-19T10:15:55 huh where'd he go? 2008-06-19T10:16:06 either way, I think getting rid of the separate --with-pid_dir/--with-sock_dir/--with-log_dir options would be a good simplification 2008-06-19T10:16:13 yeah, already done 2008-06-19T10:16:25 you move fast 2008-06-19T10:16:56 ++ 2008-06-19T10:17:02 its not uploaded yet 2008-06-19T10:17:11 im looking at the hardcoded stuff 2008-06-19T10:18:43 i could have the make file edit osrf_ctl.sh to include the right options set by autoconf? 2008-06-19T10:19:32 doesn't autoconf give you some ability for modifying files when you run ./configure? 2008-06-19T10:20:29 @srcdir@ and crap like that? 2008-06-19T10:20:29 dbs: Error: "srcdir@" is not a valid command. 2008-06-19T10:20:45 err, @srcdir@ and macros like that? 2008-06-19T10:21:55 that may work.. i will look into it 2008-06-19T10:22:53 yeah 2008-06-19T10:23:20 i would make osrf_ctl.sh.in , put the @varnames@ , and it would produce osrf_ctl.sh 2008-06-19T10:23:28 with the correct values 2008-06-19T10:24:36 sounds like a reasonable approach to me. I'm sure there are others - generate an osrf_config that offers the same sort of options as pg_config, and have osrf_ctl.sh read from that 2008-06-19T10:25:03 yeah, that might be a better idea, since i was going to do an osrf_config anyway 2008-06-19T10:32:14 that could help with openils as well - point --with-osrf-config to the osrf_config beast and suck in the appropriate values 2008-06-19T10:38:45 *** rsinger has quit IRC 2008-06-19T10:55:00 *** rsinger has joined #OpenILS-Evergreen 2008-06-19T10:56:58 about to say "hey, how about an osrf_config script that when run polutes the env or explains the compiled layout" ;) 2008-06-19T10:57:20 but then, it's already in the works 2008-06-19T10:57:37 kbeswick++ #you're ok, I don't care what Dan says about you 2008-06-19T11:34:33 Slazer: do you have any more i18n patches queued up? 2008-06-19T11:35:35 dbs: I have a few done.... almost done the current file I'm working on. 2008-06-19T11:35:59 okay, lemme know when patron.patched is updated 2008-06-19T11:36:10 Will do. 2008-06-19T11:53:45 dbs: I don't know what I did wrong but I couldn't push to patron.patched so I pushed to patron.patched.02 2008-06-19T11:53:56 whee :) okay 2008-06-19T12:09:33 arg, lenny doesn't even offer postgres8.2 as an option 2008-06-19T12:17:33 Slazer: line 295-ish of Open-ILS/xul/staff_client/server/patron/holds.js is borked - it's another "append s to the noun if we have plural objects" i18n no-no 2008-06-19T12:17:39 berick: backports? 2008-06-19T12:18:26 oh, lenny doesn't seem to have backports. umm, forwardports? 2008-06-19T12:19:03 crappy 2008-06-19T12:24:34 Slazer: same for line 358-ish 2008-06-19T12:24:49 and 401-ish 2008-06-19T12:25:06 and 443-ish 2008-06-19T12:26:04 (although you did fix up the get_bool(email) bit there - heh) 2008-06-19T12:27:18 and 603-ish and 624-ish 2008-06-19T12:29:38 also, throw a closing parenthesis on the string for staff.patron.info_group.link_patron.error_linking_patron in patron.properties 2008-06-19T12:33:04 @monologue 2008-06-19T12:33:04 dbs: Your current monologue is at least 11 lines long. 2008-06-19T13:04:33 eeevil: dbs: I added a comma to extend-reporter.sql in rel_1_2 to fix a syntax error.. but notice that legacy is mispelled as well. I'll leave fixing table names to you guys :) 2008-06-19T13:05:15 phasefx_: I thought that was fixed ... already 2008-06-19T13:05:31 maybe missed a branch? 2008-06-19T13:14:18 fixed in trunk and 1.2 2008-06-19T13:20:16 *** EGUEST961 has joined #openils-evergreen 2008-06-19T13:22:27 *** EGUEST961 has quit IRC 2008-06-19T13:27:00 dbs (or anyone really) I am confused be something... 2008-06-19T13:27:01 (obj.retrieve_ids.length > 1 ? 's ' : ' ') + util.functional.map_list( obj.retrieve_ids, function(o){return o.id;}).join(', ') 2008-06-19T13:27:01 for singular form...is it reduced to: 2008-06-19T13:27:01 util.functional.map_list( obj.retrieve_ids, function(o){return o.id;}).join(', ') or do I need to get rid of anything else? 2008-06-19T13:51:19 Slazer: well, the join(', ') really should be replaced by a getString too 2008-06-19T13:51:46 because some languages don't use commas to concatenate lists of items 2008-06-19T13:54:15 Alright but for the singular form of the whole string... that join should be taken out then, right? And everything else left alone? 2008-06-19T13:57:04 if there's only one object id, then the join won't take place 2008-06-19T13:57:51 Ahh, so it doesn't matter if I leave it there? 2008-06-19T13:58:15 so I would do something like "var id_list = util.functional_map_list(blah);" followed by the if (singular) / else (plural) cases 2008-06-19T13:58:24 passing in id_list as the parameter 2008-06-19T13:59:52 yeah - as I read it, if it's singular then you'll get "id" back, if plural you'll get "id1, id2" 2008-06-19T15:09:34 the example eg.conf for rel_1_2, has optional stuff like CacheRoot, that says it needs Apache 2.2.. I have 2.2.3 and it didn't know what CacheRoot was.. JFYI 2008-06-19T15:11:29 phasefx_: loading mod_cache or whatever? 2008-06-19T15:11:46 phasefx_: you can get rid of the caching 2008-06-19T15:11:53 that's done with memcache now.. 2008-06-19T15:11:56 i'll remove it from the repo 2008-06-19T15:12:12 I got rid of it. Don't know about mod_cache.. was just following the wiki instructions 2008-06-19T15:12:29 and what eeevil said .. 2008-06-19T15:12:45 cache.load is in mods-enabled 2008-06-19T15:14:12 yeah, funny, I had just purged the mod_cache stuff from the wiki install docs a few days back 2008-06-19T15:23:41 dbs: I think I fixed all of the singular/plural problems and pushed it up on the server (still under patron.patched.02) 2008-06-19T15:25:55 Oh nevermind, found one I missed. 2008-06-19T15:26:13 okay. I had just updated, lemme know 2008-06-19T15:28:54 dbs: ok. 2008-06-19T15:28:58 dbs: should be good now. 2008-06-19T15:34:00 still missing parenthesis on staff.patron.info_group.link_patron.error_linking_patron in patron.properties 2008-06-19T15:35:05 and tabs changed to spaces for the refactored hold vs. holds stuff :) 2008-06-19T15:35:47 Use spaces instead of tabs? 2008-06-19T15:36:36 also, you ignored my suggestion to assign the results of util.functional_map_list to the msg variable once, before you get into to the singular vs plural 2008-06-19T15:37:28 that way we only have to see that ugly code once, rather than twice per msg construction 2008-06-19T15:38:04 (everything that you're replacing using tabs for indenting, so ensure that what you replace matches that) 2008-06-19T15:41:42 dbs pasted "alternate construction" at http://paste.lisp.org/display/62505 2008-06-19T15:41:54 okay, bbiab 2008-06-19T15:51:36 hey, that code is bestial.. I mean, beautiful :) 2008-06-19T15:56:16 dbs: changes were made... hopefully I didn't mess anything up. As for tabs and spaces issue. I dunno what's going on. I put "tab" in the editor and 4 spaces get outputted even though I have it set otherwise. =( I'll look at it tomorrow. 2008-06-19T15:56:24 *** Slazer has quit IRC 2008-06-19T15:58:24 *** kbeswick has quit IRC 2008-06-19T16:31:53 *** Karen__ has quit IRC 2008-06-19T16:36:29 Slazer, if you read this later, I cleaned up the tabs and a few other minor things (added some braces for nested if statements for clarity) and committed the patch 2008-06-19T19:05:18 *** djfiander has joined #OpenILS-Evergreen 2008-06-19T19:35:26 ok, dojo is messing with my head. 2008-06-19T19:36:02 when I approve a lineitem, I change the layout of the details grid so that it's read-only. 2008-06-19T19:36:35 this doesn't seem to take, but it really does, because if I switch jubs away and then switch back, the details are, in fact read only. 2008-06-19T19:36:49 I just can't change the layout of the grid that's currently displayed 2008-06-19T19:36:52 which kinda makes sense 2008-06-19T20:52:33 *** djfiander has quit IRC 2008-06-19T23:24:16 *** dbs has quit IRC