2008-07-16T00:00:17 style question: should the init script be called 'opensrf' or 'evergreen'? All of the daemons self-identify as OpenSRF (router, perl, c), but collectively what it's starting is better described as evergreen, yes? 2008-07-16T00:02:41 asjoyner: well ... the init scripts should really be an opensrf thing, as in, live in the opensrf repository 2008-07-16T00:03:15 evergreen is just a pile of opensrf applications that happen to be started because the opensrf config files says so 2008-07-16T00:03:51 miker_: I'll gladly leave repository location up to someone checking in the change, I was just curious for opinions on what the init script itself should be called (ie. /etc/init.d/{evergreen,opensrf}) 2008-07-16T00:05:16 I could see one for opensrf, and then another for evergreen that check that, say, memcached and postgres are running, and that apache is not, and then starts opensrf ... or .. something 2008-07-16T00:08:15 asjoyner: if you'd like to sketch out a plan on the list and ask for feedback, that might not be a bad idea ... I have less of a dog in this hunt than others, so may not be the best one to answer 2008-07-16T00:08:47 and certainly not the only one that should answer 2008-07-16T00:09:06 sounds like a plan, I'll toss out a suggestion or two and see what others have to say. 2008-07-16T00:09:18 right on ... thanks! 2008-07-16T00:11:25 incidentally, we can tighten up the staff client's application.ini file to restrict itself to certain versions of xulrunner. That could compliment asjoyner's settings tester mod 2008-07-16T00:12:01 phasefx: that would be good ... and might help with semi-automated testing? 2008-07-16T00:12:47 you mean, have a test script try to invoke the staff client and see what happens? 2008-07-16T00:13:44 well, something like that ... but it would certainly help with the issues on hardy clients ... "please install version X" 2008-07-16T00:13:54 yeah 2008-07-16T00:14:04 or invoke version x :) 2008-07-16T00:17:39 if it can do that, sure 2008-07-16T00:36:07 I meant, version X might already be installed, along with version Y. Use version X if you can, end-user 2008-07-16T00:50:17 mildly rambling init script email sent :) I'm headed to bed. G'night, folks. 2008-07-16T01:09:37 * dbs heads to sackville too 2008-07-16T01:09:47 *** dbs has quit IRC 2008-07-16T01:14:57 *** Mark__T has joined #openils-evergreen 2008-07-16T04:12:17 *** cyrus has joined #openils-evergreen 2008-07-16T04:16:25 Hi folks. Any chance of a spot of help with a 1.2.2.1 to 1.2.2.2 upgrade? 2008-07-16T04:18:21 I've just done the bulk of the upgrade, but not the SQL schema upgrade yet. My assumption is that if I did a clean install of 1.2.2.1 initially (not an upgrade), then I won't need to do any other schema upgrades other than 1.2.2.1 - 1.2.2.2 (i.e. nothing to get it up to the level of 1.2.2.1 initially)? 2008-07-16T07:27:38 cyrus: that's correct 2008-07-16T07:35:04 *** atheos_ has joined #OpenILS-Evergreen 2008-07-16T07:35:31 *** atheos has quit IRC 2008-07-16T07:37:40 *** atheos has joined #OpenILS-Evergreen 2008-07-16T07:51:40 *** atheos_ has quit IRC 2008-07-16T08:12:52 *** Slazer has joined #openils-evergreen 2008-07-16T08:20:38 *** kbeswick has joined #openils-evergreen 2008-07-16T08:20:38 *** Slazer has quit IRC 2008-07-16T08:24:41 *** Slazer has joined #openils-evergreen 2008-07-16T08:32:18 *** dbs has joined #openils-evergreen 2008-07-16T08:37:45 *** kgs_ has joined #openils-evergreen 2008-07-16T08:47:36 *** _bott_ has joined #OpenILS-Evergreen 2008-07-16T08:48:31 kgs: the update blog post looks reasonable (and quite familiar - I'm charmed!) to me 2008-07-16T08:50:23 *** kgs has quit IRC 2008-07-16T09:02:13 Iffen it ain't broke, I dudn't fix it ;) 2008-07-16T09:48:52 *** Mark__T has quit IRC 2008-07-16T09:51:42 * dbs goes to move old bed out of bedroom 2008-07-16T10:19:34 *** phase_bb has quit IRC 2008-07-16T10:35:19 dbs 2008-07-16T10:35:28 what up, dude? 2008-07-16T10:36:35 also berick: how big of a problem would it be to rename existing libraries such as osrf_math.so to libosrf_math.so as per the autotools standard 2008-07-16T10:37:17 kbeswick: it woulnd't be a big problem. just have to update the default opensrf.xml.example 2008-07-16T10:37:25 and make that part of the upgrade instructions 2008-07-16T10:37:31 Yeah, I think we had already agreed on that. 2008-07-16T10:37:41 In fact, I might have already done that. 2008-07-16T10:37:57 it hasnt changed in the merge with trunk or anything 2008-07-16T10:38:01 ill do it now 2008-07-16T10:38:10 i think it will fix the problem im currently having 2008-07-16T10:38:12 kbeswick: hold your horses for a second, okay? 2008-07-16T10:38:15 k 2008-07-16T10:39:12 src/c-apps/Makefile.am in the install-exec-hook is the problem 2008-07-16T10:40:38 yep - okay, I had let autotools do its own thing with the core libs like libopensrf (versioning, etc) but didn't do the libosrf_*math.so 2008-07-16T10:40:42 so do it 2008-07-16T10:40:56 k 2008-07-16T10:41:05 i am pretty far along in the make distcheck-age 2008-07-16T10:41:15 should be nearing the end 2008-07-16T10:41:21 excellent 2008-07-16T11:38:22 wait another sec 2008-07-16T11:38:53 those aren't "libraries" (they are collections of fuctions, yes) in the standard sense ... think apache modules 2008-07-16T11:39:02 so ... what does apache do with it's modules? 2008-07-16T11:40:12 of course, a counter example to my yield sign is libdbi-drivers 2008-07-16T11:46:42 apache leaves them as mod_*.so 2008-07-16T11:49:30 apxs has module compilation abilities, though, and that's probably what apache relies on 2008-07-16T11:59:18 i could have apxs compile them instead. i had libtool do it and rename the files before (since libtool requires the prefix 'lib' on files it produces, but renaming made 'make distcheck' break 2008-07-16T12:00:06 s/produces/produces)/ 2008-07-16T12:00:17 well, wouldn't make sense to have apxs compile the opensrf plugins 2008-07-16T12:00:26 but it would for the apache modules (obviously) 2008-07-16T12:01:02 o hyeah, we were talking about two different things... 2008-07-16T12:01:12 *** cyrus has quit IRC 2008-07-16T12:01:15 hmm.. 2008-07-16T12:01:21 kbeswick: that's something else to look out for, actually. we currently get those annoying apache warnings when we run apxs to install the modules.. it'd be cool if it didn't do that anymore.. maybe compiling with apxs is the answer there 2008-07-16T12:02:28 yeah, 2 different topics 2008-07-16T12:17:01 *** kgs_ is now known as kgs 2008-07-16T13:28:37 *** kgs has quit IRC 2008-07-16T14:17:46 *** dbs_ has joined #openils-evergreen 2008-07-16T14:23:00 *** phasefx__ has joined #openils-evergreen 2008-07-16T14:23:55 *** eeevil_ has joined #openils-evergreen 2008-07-16T14:26:42 *** dbs has quit IRC 2008-07-16T14:33:28 *** phasefx has quit IRC 2008-07-16T14:36:02 *** eeevil has quit IRC 2008-07-16T15:07:12 *** dbs_ is now known as dbs 2008-07-16T15:14:44 *** Slazer has quit IRC 2008-07-16T15:41:23 back to the item price discussion, regardless of how you arrive at the price for a given lineitem (vendor order record, vendor site, etc.), there needs to be a well-known location for that number. 2008-07-16T15:41:28 sorry, outta nowhere 2008-07-16T15:46:27 *** kbeswick has quit IRC 2008-07-16T15:47:51 and i don't think provider_attr_definition is the right place, since there may not be one (e.g. used the marc from the EG catalog) 2008-07-16T16:04:30 berick: digging through the schema to see what we're looking at 2008-07-16T16:05:02 (aside: "Encumberance" is not a word, "Encumbrance" is, so I'll probably commit a bigass patch to fix that up across the board while it's still nice and early) 2008-07-16T16:05:36 good idea 2008-07-16T16:06:10 dbs: i'm basically referring to acq.lineitem's (bibs/jubs) 2008-07-16T16:06:22 right, that's what i've been looking at 2008-07-16T16:06:35 and there doesn't seem to be any place to hold prices? 2008-07-16T16:07:01 well, there are acq.lineitem_attr's which are created from parsing MARC data 2008-07-16T16:07:34 so, a certain type of those will hold vendor price when you are relying on a vendor record 2008-07-16T16:07:57 in particular, lineitem_provider_attr_definition 2008-07-16T16:07:58 right, from 020c or 022c 2008-07-16T16:08:36 well, there is provider_attr and marc_attr 2008-07-16T16:09:57 but if you're in "grab a bib record from anywhere" mode you can't be guaranteed that you'll get a reasonable value (or possibly anything) 2008-07-16T16:10:02 so, price information can be collected from the marc, but it's not guaranteed to be there 2008-07-16T16:10:05 exactly 2008-07-16T16:10:25 could we check for provider, and fall back to marc (and then fall back to generated, which is meant for users to set directly, not to extract) 2008-07-16T16:10:44 eeevil_: that seems like a good workflow. 2008-07-16T16:10:51 all we need now is the place to store the final value 2008-07-16T16:13:00 so would our current "price" for marc_attr be the default (aka "estimated price"); and we use that for calculations unless another marc_attr like "final_price" is set? 2008-07-16T16:14:12 or a user-entered price exists (which, to me, seems like it should trump any attr prices) 2008-07-16T16:15:05 berick: agreed on trumping 2008-07-16T16:15:11 berick: agreed on trumping, too 2008-07-16T16:15:45 but from a workflow perspective, you may throw the cover price in there if you're just adding an item to a picklist from a catalogue 2008-07-16T16:16:24 later on, when you actually get the results of your order back, the vendor that you purchased from will probably have given you a better price - and that final result has to trump everything 2008-07-16T16:16:33 the "final" price is stored in the fund debit 2008-07-16T16:16:34 (whether the final result is hand-entered or EDIed) 2008-07-16T16:16:55 okay, as long as it's easy to access 2008-07-16T16:17:37 that hangs off of the li_detail in the fund_debit column 2008-07-16T16:17:45 what about the situation where you have user-entered prices, but you are no immediately creating a PO. 2008-07-16T16:17:51 s/no/not/ 2008-07-16T16:18:36 usr_attr_definition=price 2008-07-16T16:18:46 (usr, not generated, from above) 2008-07-16T16:19:17 "usr generated" :) 2008-07-16T16:19:31 lineitem_usr_attr_definition 2008-07-16T16:19:44 * berick forgot that table was even ther 2008-07-16T16:19:50 that'll do it 2008-07-16T16:19:52 :) 2008-07-16T16:20:11 what's the intent of acq.lineitem_generated_attr_definition? 2008-07-16T16:21:05 marc is for "standard", provider is for ... "provider provided", generated is for "other stuff from the marc 2008-07-16T16:21:45 gotcha 2008-07-16T16:22:03 well, this is enough to generate some encumbrances! thanks, guys 2008-07-16T16:22:31 usr (although it does have a non-null xpath field in the definition) is meant to be dirctly settable 2008-07-16T16:22:43 hah! my next question ;) 2008-07-16T16:23:26 so, you can safely supply the empty string there for now, or even default to the empty string but allow xpath if you want to reuse the existing framework 2008-07-16T16:23:51 it's like that because the table inherits from acq.lineitem_attr_definition 2008-07-16T16:24:04 excellent 2008-07-16T16:24:19 and acq.lineitem_attr_definition should not be used directly ... 2008-07-16T16:25:16 *** eeevil_ is now known as eeevil 2008-07-16T16:44:51 *** _bott_ has left #OpenILS-Evergreen 2008-07-16T16:58:20 *** mjg_ has joined #openils-evergreen 2008-07-16T17:16:24 *** kgs has joined #openils-evergreen 2008-07-16T18:15:26 *** kgs has quit IRC 2008-07-16T19:03:50 *** kgs has joined #openils-evergreen 2008-07-16T19:05:51 *** kgs is now known as kgs_errands 2008-07-16T19:06:00 *** kgs_errands is now known as kgs_workout 2008-07-16T19:28:01 *** phasefx__ is now known as phasefx 2008-07-16T19:40:49 *** rsinger has quit IRC 2008-07-16T19:49:58 *** djfiander has joined #OpenILS-Evergreen 2008-07-16T20:22:21 *** mjg_ has quit IRC 2008-07-16T20:26:25 *** kgs_workout is now known as gs 2008-07-16T20:26:29 *** gs is now known as kgs 2008-07-16T20:30:17 * djfiander starts to think that just stopping work until the new Grid is available is probably more efficient than trying to get the current brittle POS working 2008-07-16T20:45:33 berick or miker_ : is there a JUB in the database that has a provider set on it? 2008-07-16T20:48:50 there are 79 of them 2008-07-16T20:48:58 you want an id? 2008-07-16T20:49:13 a picklist id would be great 2008-07-16T20:49:32 picklist 2 2008-07-16T20:50:15 cool 2008-07-16T20:50:25 so, providers are being displayed properly, at least 2008-07-16T21:28:31 god I hate firebug 2008-07-16T21:29:20 djfiander: javascript has turned your soul into a seething pit of negativity 2008-07-16T21:29:29 dbs: you have no idea. 2008-07-16T21:29:40 opening firebug causes firebug to crash 2008-07-16T21:29:48 it's a 'fight club' kinda thing 2008-07-16T21:30:03 ugh 2008-07-16T21:30:39 "Permission denided to call method History.item" appears in the widget bar at the bottom of firefox 2008-07-16T21:30:49 and I can't switch firebug tabs anymore 2008-07-16T21:31:21 infact, cmd-q won't quit firefox. I have to hit the 'close window' decoration on the window before I can get firefox to exit 2008-07-16T21:34:05 Firefox, I don't know how to quit yew 2008-07-16T21:34:11 er, firebug 2008-07-16T21:34:30 I do, but berick's got to implement support for safari in some of the backend stuff. 2008-07-16T21:36:05 with respect to the price discussion: if we have provider_attr's of both "estimated" and "final" (or "list" and "actual", or "estimated" and "actual"), I was assuming we would have usr_attrs to match 2008-07-16T21:36:05 yeah, i can't use firebug at all, it drags things down way too much 2008-07-16T21:36:36 fortunately, berick never makes any mistakes, so he doesn't need a debugger. 2008-07-16T21:36:48 I, on the other hand am barely able to find the keyboard on a good day 2008-07-16T21:36:54 heh, not one that doesn't work ;) 2008-07-16T21:37:04 so a usr_attr of "estimated" would override a provider_attr of "estimated"; but a usr_attr of "estimated" would _not_ override a provider_attr of "actual" 2008-07-16T21:37:33 dbs: that should be simple enough to implement 2008-07-16T21:37:54 yeah 2008-07-16T21:38:06 * djfiander nods like Eddie Izzard 2008-07-16T21:38:09 that sorting makes sense 2008-07-16T21:38:24 *nod*nod*shake*nod* 2008-07-16T21:38:48 I'll take death. 2008-07-16T21:39:00 berick: can you look at my code tree to see if you can figure out why editing providers doesn't work? 2008-07-16T21:39:59 okay. one thing about vmware? 2008-07-16T21:40:04 look at jubgrid.html, Picklist.js 2008-07-16T21:40:18 yessir, jeff? 2008-07-16T21:40:18 djfiander: i'd be glad to. i'm presently staring down the barrel of an upgrade that might take a little time. after that, though 2008-07-16T21:40:27 berick: cool 2008-07-16T21:40:28 upgrading two servers turns into upgrading 29 servers 2008-07-16T21:40:49 I really am stumped 2008-07-16T21:40:54 or at the very least, downtime for 29 servers 2008-07-16T21:41:17 (unless you've got something beefier and more expensive than vmware server) 2008-07-16T21:41:24 jeff: well, depends on what version of vmware you're running; the esx people would say "oh just migrate..." 2008-07-16T21:41:29 jinx 2008-07-16T21:41:30 right 2008-07-16T21:42:01 vmotion away, and you're not upgrading a host os as such 2008-07-16T21:42:48 so really, revise my lead-in as "one thing about vmware server" or "one thing about vmware on the cheap" ;) 2008-07-16T21:45:37 I'll agree whole-heartedly with that. 2008-07-16T21:46:00 is vmware server still restricted to just 32-bit vms? 2008-07-16T21:46:17 never was, iirc. 2008-07-16T21:46:33 iirc, it is a 32 bit app running on the host os, though. 2008-07-16T21:47:25 ok, weird. google books was using amicus for "find in library" for a few days. now it's back to worldcat local 2008-07-16T21:47:44 of course, amicus is differently useless from WC local ;-) 2008-07-16T21:48:45 dbs: looks i was wrong. even though vmware server 1 does give you 64 bit options when setting up a guest vm, vmware server 2 beta has this "new" feature: 64-bit guest operating system support:Use 64-bit guest operating systems on 64-bit hardware to enable more scalable and higher performing computing solutions. 2008-07-16T21:50:16 jeff: oh, I know what it was - having to install the 32-bit support libs on gentoo a few years back for vmware server was a PITA 2008-07-16T21:50:47 djfiander: good, because our holdings in amicus are 10 years out of date 2008-07-16T21:50:47 yes 2008-07-16T21:51:11 but i also had to install those to support Dell OMSA on our poweredge servers, so i wasn't as annoyed with vmware :) 2008-07-16T21:51:58 dell-- 2008-07-16T21:51:59 dell-- 2008-07-16T21:52:00 dell-- 2008-07-16T21:53:51 trying to simplify the number of workstations on my desk. four monitors, one windows xp box, one linux box, one vista box, one linux box running vmware server with a few guests, one mac mini. :P 2008-07-16T21:55:02 * jeff wanders way off topic for the channel 2008-07-16T21:55:34 not so far off; I was thinking "hmm, is he testing the staff client on that mac?" 2008-07-16T21:56:05 i should be! 2008-07-16T21:57:01 but currently i'm rdp'd over ssh to the xp box and vnc'd to the mini in order to use ARD to upgrade some staff and public clients :P 2008-07-16T21:57:23 (see previous "simplify" goal) 2008-07-16T22:04:40 *** djfiander has quit IRC 2008-07-16T22:05:38 and the upload of the 1.2.2.3 image begins 2008-07-16T22:06:04 -= THIS MESSAGE NOT LOGGED =- 2008-07-16T22:07:43 dbs: it's in the mail. 2008-07-16T22:19:22 -= THIS MESSAGE NOT LOGGED =- 2008-07-16T22:21:07 Ah well - at least demand for a service shows that the effort is valued by somebody 2008-07-16T22:24:07 demand is the correct term, yes :) 2008-07-16T22:25:11 ;-) 2008-07-16T22:41:47 * jeff installs firefox 3.0.1 2008-07-16T22:44:14 * dbs submits a team registration for Software Freedom Day 2008-07-16T22:44:53 kgs: http://softwarefreedomday.org/ if you haven't seen it yet 2008-07-16T22:46:28 dbs: cool! that sounds wonderful. did you submit a team reg for EG? 2008-07-16T22:46:44 kgs: no, actually, just a team for Sudbury 2008-07-16T22:47:14 ok great! I'll do one for EG 2008-07-16T22:51:28 this is my way of procrastinating 2008-07-16T22:52:06 finding things that are good to do, but not necessarily the right thing to be doing at this moment. sigh. 2008-07-16T22:53:05 * jeff laughs 2008-07-16T22:53:13 isn't that one of the most popular forms? 2008-07-16T22:59:54 jeff: undoubtedly :) 2008-07-16T23:18:50 "Click restart to finish installing the software" 2008-07-16T23:22:16 It's a trick! Don't do it! 2008-07-16T23:58:06 *** dbs has quit IRC