2010-03-09T00:32:18 *** mck9 has left #evergreen 2010-03-09T01:09:50 *** pmp_afk has quit IRC 2010-03-09T02:22:10 *** brendan_bywater has quit IRC 2010-03-09T02:39:17 *** jamesrf has quit IRC 2010-03-09T07:29:53 *** alxp has joined #evergreen 2010-03-09T07:36:26 *** sfortin has joined #evergreen 2010-03-09T07:39:29 *** alxp has quit IRC 2010-03-09T07:46:07 *** rsinger has quit IRC 2010-03-09T07:54:36 *** mck9 has joined #evergreen 2010-03-09T09:01:07 *** sfortin has quit IRC 2010-03-09T09:01:40 *** rsinger has joined #evergreen 2010-03-09T09:02:50 *** Meliss has joined #evergreen 2010-03-09T09:04:06 *** atz has joined #evergreen 2010-03-09T09:06:37 *** gdunbar has joined #evergreen 2010-03-09T09:08:30 *** sfortin has joined #evergreen 2010-03-09T09:12:27 *** bshum has joined #evergreen 2010-03-09T09:13:11 *** jenny has joined #evergreen 2010-03-09T09:33:28 *** grantjohnson has joined #evergreen 2010-03-09T09:33:37 *** atz has quit IRC 2010-03-09T09:37:26 *** dbs has joined #evergreen 2010-03-09T09:37:43 *** Dyrcona has joined #evergreen 2010-03-09T09:39:08 *** sfortin has quit IRC 2010-03-09T09:40:00 Question about marc2are.pl and authority.record_entry: Does ARE expect authority records or full MARC records? 2010-03-09T09:40:23 I am assuming authority records, but I thought I'd ask. 2010-03-09T09:40:26 authority records 2010-03-09T09:40:31 tyvm 2010-03-09T09:45:44 *sigh* continuing reports of double-billing for overdues 2010-03-09T09:45:52 * dbs tries to dig up some examples 2010-03-09T09:45:55 *** ryan__ has joined #evergreen 2010-03-09T09:52:35 *** atz has joined #evergreen 2010-03-09T10:06:17 *** atz has quit IRC 2010-03-09T10:21:12 *** alxp has joined #evergreen 2010-03-09T10:46:55 *** ryan__ has left #evergreen 2010-03-09T10:51:18 *** rjackson-isl has joined #evergreen 2010-03-09T10:58:54 *** BobW has joined #evergreen 2010-03-09T10:59:30 It smells like a 1.6.0.4 release is inching closer... 2010-03-09T11:00:30 *** atz has joined #evergreen 2010-03-09T11:01:32 dbs: re 710/711 ... search cache? did you add -asdlkjffdkjlfkdjls to the end? :) 2010-03-09T11:01:39 * miker_ guesses yes 2010-03-09T11:02:10 -asdlkjffdkjlfkdjls - so *that's* the magic spell ;) 2010-03-09T11:02:38 * phasefx mutters -xyzzy 2010-03-09T11:02:55 gmcharlt: if you do that backwards, you get 30 lives 2010-03-09T11:03:16 and the spraying fireball gun 2010-03-09T11:03:23 but only if you complete it while the nightrider bar is scrolling 2010-03-09T11:03:38 contra++ 2010-03-09T11:03:40 miker_: yes, -fubar and several variants 2010-03-09T11:04:51 miker_++ # back-ports 2010-03-09T11:06:28 dbs-- # assuming that the bib in question was in my search scope. KILL ME NOW, I deserve it 2010-03-09T11:06:36 miker_++ 2010-03-09T11:06:56 it's been a backport frenzy in 1.6.0 land over the past few days 2010-03-09T11:07:00 what, for mis-diagnosing? man, I'd be a kamillionaire! 2010-03-09T11:07:06 oh, for backporting 2010-03-09T11:10:46 *** sfortin has joined #evergreen 2010-03-09T11:11:49 Err, no, different bib. Curses. dbs-- # premature exuberance 2010-03-09T11:20:15 <_bott_> miker_: is that a syntax error on line 496 of 1.6.0.3-1.6.0.4-upgrade-db.sql, missing paren 2010-03-09T11:29:53 *** kbeswick has quit IRC 2010-03-09T11:30:54 _bott_: yes it is - my syntax error. 2010-03-09T11:32:53 and a few other places 2010-03-09T11:43:46 *** kbeswick has joined #evergreen 2010-03-09T11:47:58 *** jamesrf has joined #evergreen 2010-03-09T11:48:45 does anyone know if the in-db circ method supports blocking renews for items on hold? I can't find anything to indicate that it does, but I hope I am wrong. Thanks. 2010-03-09T11:53:51 *** jenny has quit IRC 2010-03-09T12:06:25 I found this for the circ script method: http://list.georgialibraries.org/pipermail/open-ils-general/2009-December/002349.html 2010-03-09T12:09:17 The code behind 'fetchBestHold()' seems a bit involved to be reimplemented directly in the DB. I think this test probably needs to be (optionally) run directly in Circulate.pm. 2010-03-09T12:09:52 Any other opinions on this? 2010-03-09T12:10:57 there are the various hold_stats bits... and the total_copy_ratio vs available_copy_ratio... 2010-03-09T12:11:38 i don't know if that's the only way or not. little-to-no in-db circ experience, just looking at code. 2010-03-09T12:13:42 dbwells: i think you're right - Circulate.pm with YAOUS (yet another OU setting) would make sense 2010-03-09T12:14:17 dbwells: did your search augmentation work out for you? 2010-03-09T12:15:20 dbs: thanks. I'll set about making a patch after lunch sometime and see what the list thinks of it. 2010-03-09T12:15:28 dbwells++ 2010-03-09T12:16:20 dbs: as for search augmentation, I was really just rounding out my understanding after the ISSN spacing issue we discussed yesterday. 2010-03-09T12:17:09 I did read through your serials wiki entry - I wasn't 100% sure how a shelving_unit differs in practice from a binding_unit; also, at a grosser level, it looks like the idea is to completely replace the current fully manual approach? 2010-03-09T12:17:10 hrm ... re renew stopping if a hold could be filled, I think in-db would be better 2010-03-09T12:18:42 action.matrix_test_result I guess? 2010-03-09T12:19:12 err, action.item_user_circ_test? 2010-03-09T12:19:32 config.circ_matrix_matchpoint.stop_renew_on_hold (bool) + ... yeah, test in action.item_user_renew_test 2010-03-09T12:20:12 could be backported to 1.6, but not 1.6.0 2010-03-09T12:22:17 * dbs is no longer sure what distinguishing features are in 1.6 that aren't in 1.6.0: reservations, I guess 2010-03-09T12:22:19 any objection to that? 2010-03-09T12:22:35 dbs: yeah, booking 2010-03-09T12:22:37 new features 2010-03-09T12:22:55 no objection here, we're still script-based circers :) 2010-03-09T12:23:08 for those soon to be on 1.6.0.x and wanting in-db circ but needing stop_renew_on_hold, can it be 'faked' with a script_test in circ_matrix_matchpoint, or is script_test not capable of doing that? 2010-03-09T12:23:16 miker_: the current find_nearest_permitted_hold() function called by fetchBestHold() goes the extra mile of seeing if an item *could* fulfill a hold even if it hasn't been targeted yet. This seems like a lot to re-do in-db, or am I not getting it? (once again) 2010-03-09T12:24:14 dbwells: I want it in-db so it's all configured in one place 2010-03-09T12:24:28 dbs: all i can thing of w/ that acronym is whatsamatta yaous?! 2010-03-09T12:24:34 *think 2010-03-09T12:24:39 atz++ 2010-03-09T12:24:41 one (perhaps heavy-handed approach) would be to see if the copy id is in action.hold_copy_map at all? ;-) 2010-03-09T12:25:11 jeff: that, plus making sure at least one hold is valid, yes 2010-03-09T12:25:30 should be a simple test 2010-03-09T12:25:38 (and that it's not /your/ hold) 2010-03-09T12:25:45 what would be an example of an invalid hold? 2010-03-09T12:25:53 right, not-your-hold important too. 2010-03-09T12:26:01 (though let's not open THAT whole debate again!) 2010-03-09T12:26:03 jeff: invalid is the wrong word .. frozen, etc 2010-03-09T12:26:41 s/valid/active/ ? yay terminology! 2010-03-09T12:26:53 i think i've seen instances of 'stale' data in ahcm, also -- fulfilled/cancelled holds with ahcm entries, i think. 2010-03-09T12:27:17 hold in the database! hold is active! hold is active and targeting! 2010-03-09T12:27:25 2010-03-09T12:27:42 (needs a russian accent to get the full effect) 2010-03-09T12:28:28 It seems unrealistic to me to think that all the circ code should live in the DB. It is certainly possible to have all the configuration in the DB without going to that extra trouble. 2010-03-09T12:30:22 miker_: don't forget the existing hold/renew stuff in trunk 2010-03-09T12:30:31 * atz agrees w/ dbwells 2010-03-09T12:30:40 berick: ? 2010-03-09T12:31:13 miker_: total_copy_hold_ratio and available_copy_hold_ratio 2010-03-09T12:31:31 ahh... yes... right 2010-03-09T12:31:34 hrm. 2010-03-09T12:31:41 for speed, it is a nice goal (making no determination here if it's realistic or not). script_test seems to indicate that you can run any javascript you like to get back some impossible-with-in-db flexibility/functionality... sacrificing speed? 2010-03-09T12:32:00 jeff: that's the purpose, but as yet unimplemented 2010-03-09T12:32:15 Simple: just implement JavaScript in SQL! 2010-03-09T12:32:20 berick: thanks. :) 2010-03-09T12:32:32 I was about to ask if script_test was working somewhere yet :) 2010-03-09T12:33:07 that might be a better option if it has to happen in 1.6, actually 2010-03-09T12:33:17 Maybe I should focus on that instead and not stir up debate. 2010-03-09T12:34:24 not being able to forbid renewal when copy is needed for a hold (using SOME definition of "needed for hold") will prevent us from using in-db circ. 2010-03-09T12:35:00 and anything that can lead to us being able to use in-db circ at least sounds good. :) 2010-03-09T12:37:43 dbs: thanks for the feedback on the serials stuff. I am still aiming to replace as much of the manual process as possible, but that is probably dreaming. As for 'shelving_unit', the idea will be to allow flexibility for managing physical-item-to-item-record as appropriate for a library's needs. 2010-03-09T12:37:50 yeah ... the ratio stuff is a bit too involved to backport ... and it's meant to do what you want 2010-03-09T12:38:38 so, script_test is the best way, IMO 2010-03-09T12:38:46 sorry for the noise/hope ;) 2010-03-09T12:39:10 berick++ # remembering 2010-03-09T12:39:17 dbs: So, a shelving_unit could be a bound_volume, a single issue, or a whole shelf of volumes, whatever. 2010-03-09T12:40:02 dbwells: I think the idea of a binding unit is that it could encompass as few as 1 issues to as many issues as you wanted to collect together in a single beast - the guiding principle being that it maps to a single call number, in any case 2010-03-09T12:40:10 s/call number/barcode 2010-03-09T12:40:51 call number! 2010-03-09T12:40:52 part of the problem is probably that definitions of "do not premit renewal when copy is needed for hold" differ. 2010-03-09T12:41:03 permit, even. 2010-03-09T12:43:19 it doesn't seem that there is any plan for there to be an equivalent to the script-based fetchBestHold() technique 2010-03-09T12:43:43 dbs: Ok, then it is the same thing, I think. :) With binding_unit I naturally thought of bound volumes, and I wanted a table more flexible than that. Gotta run, thanks again. 2010-03-09T12:43:54 Seeya! 2010-03-09T12:44:13 (to touch on what miker_ said about "[ratio stuff is] meant to do what you want" and dbwells' "seems like a lot to re-do in-db") 2010-03-09T12:47:02 *** jenny has joined #evergreen 2010-03-09T12:47:22 oh, but the "check ahcm" approach would probably be "close enough" (or exactly?) for our needs. 2010-03-09T12:47:39 dbs: i'm sure i've asked you this before and forget, but when you guys buy from us vendors, at what point in the process do you translate the price to CAD? 2010-03-09T12:47:58 s/us vendors/US vendors/ ;) 2010-03-09T12:48:03 berick: hahahahah 2010-03-09T12:48:50 berick: sorry. we estimate a CAD price based on the conversion rate (from whatever currency) at the time the order is placed so that we can encumber the fund appropriately 2010-03-09T12:49:36 dbs: ok. so before it's a real order and it's just some lineitems in a list, it retains the pre-conversion price? 2010-03-09T12:50:12 jeff: the ratio stoff allows what is there now (set the total-copys:holds threshold to 0, if it's greater than that then no hold for you!) as well as ratio based tests, like "if there are 2x as many copies as holds, you can place the hold" 2010-03-09T12:50:19 or does it "go into the system" as CAD from step one? 2010-03-09T12:50:48 * berick notes that question may not make much sense with your current work flow 2010-03-09T12:51:27 miker_: oh, right... because the ratio stuff IS the "check ahcm" approach. 2010-03-09T12:52:39 berick: then when we pay the bills after receiving the orders, our finance department sends us the real exchange rate for the applicable bills and we adjust the amounts accordingly 2010-03-09T12:52:39 * berick is trying to decide if we should set the LI price in the local currency as soon as the price is known or at order time 2010-03-09T12:52:40 berick: thinking aqc for a minute, is there anything in the works currently that would retain "cost" of a copy, as opposed to the "price" (which we use as 'list price') attribute on a copy? 2010-03-09T12:53:06 jeff: yes, there will be estimated price and invoiced price (aka cost) 2010-03-09T12:53:37 berick: it's painful 2010-03-09T12:53:52 berick: EDI will want the unit price in known currency... best expressed in the vendor's list currency 2010-03-09T12:54:18 because of course the estimated native currency price might change when you receive the order, too, no? 2010-03-09T12:55:02 dbs: right. it's all just an estimate until the invoice arrives, regardless of currency 2010-03-09T12:55:12 right, and the vendor records you will have loaded will have the same values (probably w/o specifying currency) 2010-03-09T12:55:58 atz: why best expressed in the vendor's list currency? for accuracy? 2010-03-09T12:56:35 i'm not sure what the MARC21 convention is for USD vs. CAN $ 2010-03-09T12:56:52 if any 2010-03-09T12:56:56 * gmcharlt laughs at atz 2010-03-09T12:57:07 atz: well, when we get MARC from vendor X, we'll know what their currency type is anyway 2010-03-09T12:57:47 right 2010-03-09T13:00:27 We use a bunch of european currencies here, too. whee. 2010-03-09T13:02:20 we'll have to make some decisions for the UI.. if staff enters a price for an item from a foreign vendor, do we assume the price is in the vendor's currency? 2010-03-09T13:02:24 i'm thinking yes 2010-03-09T13:02:49 * dbs thinks so too - matches how our old system worked 2010-03-09T13:03:16 then at po activation time, we do the conversion into local currency 2010-03-09T13:03:42 sounds feasible 2010-03-09T13:04:47 k. some UI clues "Hey, dude, it's in Euros!" will be good 2010-03-09T13:05:00 thanks for the sounding boards 2010-03-09T13:15:48 -= THIS MESSAGE NOT LOGGED =- 2010-03-09T13:18:02 heh 2010-03-09T13:19:50 dbs: this 2-weeks in advance thing for EG10 is gonna kill me too 2010-03-09T13:20:16 half the interfaces I want to show won't even exist that far in advance ;) 2010-03-09T13:20:23 * dbs lolz 2010-03-09T13:21:51 *** jenny has quit IRC 2010-03-09T13:22:42 *** jenny has joined #evergreen 2010-03-09T13:28:23 69.4°F .. so close 2010-03-09T13:28:48 berick: I'm viewing it as motivation :) 2010-03-09T13:28:56 gmcharlt++ 2010-03-09T13:34:41 berick: will the invoiced price be anywhere where we could back-populate it with data for all of the copies we currently have in our system? will it be a copy-level attribute, or stored in an acq lineitem somewhere? 2010-03-09T13:34:49 (sorry, pulled away) 2010-03-09T13:35:06 * bradl double takes, "dbs just 'lolz'"? 2010-03-09T13:35:19 dbs: who are you and what have you done with dbs? 2010-03-09T13:36:06 bradl: welcome to the LOLZONE! 2010-03-09T13:36:16 RETREAT!!! 2010-03-09T13:36:36 jeff: it will be in an invoice table, from which you can find the lineitem_detail (aka copy), from which you can find the copy 2010-03-09T13:36:52 jeff: seems reasonable to populate that data in the real copies at invoice time, though 2010-03-09T13:37:49 i and our auditors would love to see copies have both price (list price, what patrons are charged on lost/etc) and cost (money the library paid to purchase this copy) 2010-03-09T13:38:14 currently we have cost stuffed in copy notes. reporting on it is... subject to limitations. 2010-03-09T13:38:20 (void where prohibited) 2010-03-09T13:38:50 and having the cost populated at invoice time sounds like a good idea overall. 2010-03-09T13:39:25 (of course, then there will come a library that would like to bill patrons cost-instead-of-price when LOST, but that can be just an org unit setting!) 2010-03-09T13:41:26 right 2010-03-09T13:43:36 i think koha models it as cost and replacement cost... where replacement cost can be populated at cataloging time by some rule based cost + X, for example 2010-03-09T13:45:10 yeah, "cost" on the copy in EG is effectively replacement cost, however you want to define that value, be it cataloger or from invoice data 2010-03-09T13:46:24 *** jenny1 has joined #evergreen 2010-03-09T13:48:10 *** BobW has quit IRC 2010-03-09T13:48:57 we could expose price/cost invoice data in the cataloging interface (so catalogers can apply their own rules if necessary), but go ahead and populate the asset.copy.price with the invoiced price or cost, depending on org setting, at invoice time 2010-03-09T13:49:15 *** jenny has quit IRC 2010-03-09T13:58:06 for our auditors, "collection value" == sum(copy_price), and "collection cost" = sum(copy_cost) 2010-03-09T13:58:53 adjust for inflation? :) 2010-03-09T13:59:52 / difficulty of re-acquiring replacement copies based on out-of-print status? need the open-ils.acq.find_ABE_price() method for that 2010-03-09T13:59:56 so if both are copy attributes, and there's the choice of a ou-default "lost == bill {price|cost}", perhaps the "catalogers can apply their own rules if necessary" (if really necessary) is either a third 'replacement' value, or a 'use price/use cost' setting... 2010-03-09T14:00:05 and estimate auction price and insurance premium if you've got a copy of the Gutenberg Bible 2010-03-09T14:00:12 gonna need a field for appraised value.. heh 2010-03-09T14:00:36 dbs: yeah, check out now-valued $400 art book, declare lost, pay for it == congratulations, you get to sell it! == "bad" 2010-03-09T14:00:37 appraised value and insured value, then 2010-03-09T14:05:08 CREATE TABLE asset.copy_price_data... :P 2010-03-09T14:09:16 createdb copy_price_database 2010-03-09T14:19:28 asset.copy_attr 2010-03-09T14:20:58 now convert all the asset.copy_note entries to the now-more-generic copy_attr table... 2010-03-09T14:21:14 make asset.copy_attr a view on asset.copy_note 2010-03-09T14:21:21 i was just going there :) 2010-03-09T14:21:26 then materialize it 2010-03-09T14:22:07 and bring it full-circle by creating a trigger to update some new "price" and "cost" columns in asset.copy! 2010-03-09T14:22:07 or make title unique on copy_note 2010-03-09T14:22:55 speaking of copy notes, if anyone gets the itch to support notes at batch copy creation, feel free 2010-03-09T14:23:16 or even batch edit 2010-03-09T14:24:07 the problem i have with the data being in copy_note is partially that there's no guarantee about it being "Cost " vs "Cost" vs "Cost:" or even "Cost: $14.95" for the note title 2010-03-09T14:24:48 * phasefx nods, notes are piss-poor for reporting, which is why we pushed stat cats over them. Except stat cats are useless for what you want :-/ 2010-03-09T14:25:00 and partially that copies-without-notes are not included in a report on copies + copy notes 2010-03-09T14:25:40 * phasefx would like to see a free-text toggle on stat cats, so that copy stat cats can get some of the joy and pain that patron stat cats get 2010-03-09T14:26:06 maybe a datatype field for validation would be worthwhile too 2010-03-09T14:26:12 in the case of free text 2010-03-09T14:26:22 i have one audit report that makes use of copy notes and copy stat cats. much pain. 2010-03-09T14:27:01 easier reporting on multiple stat cats at one time would be great, too 2010-03-09T14:27:40 * phasefx likes pet-peeve Tuesdays 2010-03-09T14:28:45 free-text stat cats-- 2010-03-09T14:29:13 user defined typed attributes, then 2010-03-09T14:29:14 hell, free-text-address-types-- 2010-03-09T14:29:39 I spent about half an hour cleaning up variations of "MAILING" "Mailing" "Mailng" "mailing"... 2010-03-09T14:29:43 OCD 2010-03-09T14:29:48 koha has this mechanism for defining authorized values for all sorts of fields 2010-03-09T14:29:49 MailNG 2010-03-09T14:30:06 MailV6 2010-03-09T14:30:11 call steve 2010-03-09T14:43:40 @karma cats 2010-03-09T14:43:40 senator: cats has neutral karma. 2010-03-09T14:46:47 cats++ 2010-03-09T14:53:51 my_cat++ 2010-03-09T14:54:33 i'll increment my cat when she learns to bring me a beer ;-) 2010-03-09T14:57:49 senator++ my_catlikedog++ 2010-03-09T15:05:44 restarting_open-ils.storage_after_adding_a_new_indexing_category++ 2010-03-09T15:05:48 problem solved 2010-03-09T15:14:08 wiki updated 2010-03-09T15:15:40 As some here may have suspected, implementing script_test to handle blocking renews on holds is quite problematic and probably not the right solution. 2010-03-09T15:17:15 left-spectrum approach: as soon as a hold is placed on a given item, set renewals=0 for the current borrower 2010-03-09T15:18:27 the as-yet theoretical script_test only makes sense (to me) as an additional means of testing whether a matchpoint 'matches', not to do additional testing after a matchpoint is chosen. 2010-03-09T15:18:40 two different things 2010-03-09T15:18:58 when the hold is fulfilled, it'd be a pain to go reset the renewals for those folks with the other eligible copies 2010-03-09T15:19:22 s/fulfilled/captured/ 2010-03-09T15:19:49 phasefx: I know, it's grossly simplified, but just trying to come at it from an entirely different angle 2010-03-09T15:20:01 * phasefx nods 2010-03-09T15:20:25 move "renewals remaining" into copy_notes :) 2010-03-09T15:20:38 * phasefx really wants the script_test stuff.. just so we can still have phase of the moon, etc. tests :) 2010-03-09T15:21:02 giving_policy_makers_footguns++ 2010-03-09T15:21:10 otherwise the whole matchpoint mechanism breaks down pretty quickly (that is, if the script_test fails, then what? what if another matchpoint would have matched and passed?) 2010-03-09T15:22:17 If (moon == MOON_FULL_PHASE && marc_tag_650.match("werewolf")) { return SILVER_BULLET; } 2010-03-09T15:24:04 if (Math.random() > 0.5 ) { result.loan_duration = 'foo'; } 2010-03-09T15:24:19 dbs-- #for renewals_remain in copy_notes 2010-03-09T15:25:02 jeff: or js2JSON(copy) in copy_notes ;) 2010-03-09T15:25:10 and yeah, what phasefx said regarding post-hold-capture/fulfillment and going back to re-instate renewals :P 2010-03-09T15:25:11 An awesome hack would be to flip renewals_remain to a negative 2010-03-09T15:25:23 _then_ you could restore it, without requiring any other fields 2010-03-09T15:25:35 dbs: but we run into renewals_remain being negative if staff have forced a renewal for whatever reason... 2010-03-09T15:25:42 staff-- 2010-03-09T15:25:45 heh 2010-03-09T15:25:46 still not quite perfect, for the edge case of staff overidden renewals that go negative 2010-03-09T15:25:46 FIRE THOSE PEOPLE 2010-03-09T15:25:46 phasefx: would that mean that only vampires could administer EG? 2010-03-09T15:26:16 gmcharlt: at least it'd be outside of normal business hours 2010-03-09T15:26:25 renewals_remain = renewals_remain - 1 billionty 2010-03-09T15:27:05 I hope I am making sense. To put it another way, many examples I can think of (including my own) needs some kind of script filter at a stage in the process wholly different than the matchpoint selection stage. Probably a more global operation. 2010-03-09T15:27:56 I haven't looked into how in-db circ works yet, alas 2010-03-09T15:27:56 Since that no longer qualifies as an incremental improvement or bug fix, I'm out :( 2010-03-09T15:29:02 dbwells: does this have any bearing? http://svn.open-ils.org/trac/ILS/changeset/15619 2010-03-09T15:29:06 to bridge the gap between rel_1_6_0 and rel_1_6_next, perhaps a utility script to do dbs' hack of tweaking renewal_remain... 2010-03-09T15:31:49 dbwells: the script_test is meant to be a post-matchpoint permit test, just like the circulate test, or the max-per-circ-mod tests 2010-03-09T15:31:57 fwiw 2010-03-09T15:34:45 miker_: thanks. What is the intended behavior if the script_test fails? (I am thinking along the lines of fallback to a less specific matchpoint) 2010-03-09T15:36:06 the idea is that you should be able to drop anything you could put into circ_permit.js into script_test and have it generate events (just like circ_permit.js does) and the surrounding code would then combine those with the result of the in-db call to come to the final event list for the request 2010-03-09T15:36:42 just like the patron parts of JS permit could succeed, but the copy parts could fail (different scripts in the legacy JS world) 2010-03-09T15:37:05 the script doesn't fail, it returns non-success events or it doesn't 2010-03-09T15:37:20 is it just permit stuff, or could you set loan durations, etc. with it? 2010-03-09T15:37:38 phasefx: just permit stuff 2010-03-09T15:37:47 *** alxp has quit IRC 2010-03-09T15:37:47 well, I mean, there's no code yet 2010-03-09T15:38:07 but the design/intention is/was for just permit 2010-03-09T15:38:08 so it could be made to modify the ruleset that the matchpoint chose 2010-03-09T15:38:16 <_dkyle_> would script_test have to be javascript only? 2010-03-09T15:38:24 but I would call that SAAAD 2010-03-09T15:38:39 <_dkyle_> wish I could just do all my circ scripts in perl 2010-03-09T15:39:18 _dkyle_: well, it can load other js (assuming the config file is appropriately configured) and there are JS apis for everything you can do in perl 2010-03-09T15:39:29 we just don't use them in the current circ js 2010-03-09T15:41:32 <_dkyle_> its partly just more comfortable with perl, but I was also thinking of circ script caching, thinking it could be faster 2010-03-09T15:42:04 oh, the compiled js is cached (for a certain number of requests) by the perl wrapper 2010-03-09T15:42:08 but: Open-ILS/src/javascript/backend/libs/jsonopensrfrequest.js 2010-03-09T15:44:13 <_dkyle_> since edits to the circ scripts are seen right away I thought they were not cached, you are saying they are cache then? 2010-03-09T15:50:37 * berick also thinks perl might make more sense, fwiw 2010-03-09T15:51:02 miker_: limiting the script_test to permit tests does help to make sense of it. 2010-03-09T15:52:23 Having lived with the in-db circ setup for a few months now, I do wonder if I should just abandon it. 2010-03-09T15:53:25 tackling it from a different angle, implement the in-db logic in the circ scripts :D 2010-03-09T15:55:03 phasefx: implement the database in Perl. Rosetta! 2010-03-09T15:55:21 simple matter of programming :) 2010-03-09T15:56:23 -= THIS MESSAGE NOT LOGGED =- 2010-03-09T15:56:40 I met the guy at OSCON 2005 at an Apache Derby birds of a feather 2010-03-09T15:57:25 The in-db code makes a few too many assumptions about how things work. Due to various deficiencies we currently have 31 different matchpoints. 2010-03-09T15:58:50 So, at this point, even if the script_test did work, I would need to add it in duplicate to all 31, as there is no longer any sense of global control. 2010-03-09T15:59:59 sorry to complain, but at some point you can't keep it in :D 2010-03-09T16:00:29 letting_it_out++ 2010-03-09T16:00:51 *** sfortin has quit IRC 2010-03-09T16:01:05 it's okay, we all complain about bits and pieces :) take a look at the frankenstein monster that is staff client source code if you want to complain 2010-03-09T16:01:05 I am sure there is some 'grass-is-greener' in my view, as I haven't ever lived with the scripted circ. 2010-03-09T16:01:42 that's a concern of mine as well 2010-03-09T16:02:41 the main points in favor of in-db circ are speed and local control, right? 2010-03-09T16:02:42 as another example, if you need different bahavior for renewals vs. non-renewals, there will be 2 of every rule 2010-03-09T16:03:21 phasefx: basically 2010-03-09T16:03:56 berick: exactly! Maybe I should send you some of my experiences and ideas by email and we can hatch some new ideas going forward? 2010-03-09T16:04:50 if I ever find any real time, that is :) 2010-03-09T16:04:57 *** Meliss has quit IRC 2010-03-09T16:06:03 dbwells: s/me/open-ils-dev/ i'm running on fumes as well :( 2010-03-09T16:06:17 *** frzosima has joined #evergreen 2010-03-09T16:06:22 Perhaps Wednesday at EG2010 should be group therapy day 2010-03-09T16:06:59 * dbs is badly in need of therapy 2010-03-09T16:07:04 <_dkyle_> so sorry to be dense here but was miker telling me that .js circ scripts are in fact cached? 2010-03-09T16:07:41 <_dkyle_> every EG2010 day could be therapy day 2010-03-09T16:07:44 _dkyle_: it's been my experience that they are not cached, but I know there is a caching layer built in 2010-03-09T16:08:12 don't recall off hand when/if scripts are cached, though 2010-03-09T16:08:57 * dbs hasn't had to restart any service after hand-editing JS circ scripts 2010-03-09T16:09:17 <_dkyle_> they don't act like cached scripts, but maybe they get stat'd and automatically reloaded somewhere? 2010-03-09T16:09:26 "Hi everyone, I'm Bill, I'm an EG {hacker|administrator|user}". "Hi, Bill!" 2010-03-09T16:09:31 * berick chain smokes in a circle 2010-03-09T16:09:48 * dbs looking at adding keyword|identifier index for 010a, 020a, 022a and pushing SRU to use keyword|identifier for identifier index... hmm, there must be something wrong with that master plan 2010-03-09T16:09:57 * _dkyle_ pours another 2010-03-09T16:11:53 <_bott_> dbs: HopCat provides great therapy 2010-03-09T16:13:15 -= THIS MESSAGE NOT LOGGED =- 2010-03-09T16:13:49 * dbs will make a suggestion that EG2011 not schedule that at the same time as hackfesting 2010-03-09T16:15:06 right, just come to the hackfest 2010-03-09T16:18:21 <_bott_> Will certainly pass along those sentiments! 2010-03-09T16:21:30 _bott_: it was a follow-on from 2009; not sure where the idea came from, but it's got a bunch of weirdness to it 2010-03-09T16:23:59 Like: can't you talk to a developer at breakfast, or over lunch, or between sessions, or during the various social thingies? 2010-03-09T16:24:01 <_bott_> dbs: I tend to agree. I do believe it was a "this is the way they did it last year" 2010-03-09T16:26:39 hmm, in limited testing keyword|identifier is working quite nicely and should significantly reduce the number of false positives we get between date ranges and ISSNs on a general keyword search 2010-03-09T16:28:18 *** twirlip is now known as twirlip_beverage 2010-03-09T16:30:09 re caching, sorry, you're correct, circ scripts are not cached 2010-03-09T16:30:18 or, they're cached for 1 run ;) 2010-03-09T16:30:26 * jeff grins 2010-03-09T16:30:34 miker_++ # you're human! 2010-03-09T16:31:10 actually ... hrm ... there was, at one time, a stat test 2010-03-09T16:33:21 not there now, though 2010-03-09T16:33:31 was probably too heavy 2010-03-09T16:33:52 now we just use reset_count 2010-03-09T16:35:58 dbs: there's nothing wrong with your kw|identifier master plan ... in trunk I want to make that mapping configurable via ui right from the (nonexistent?) config.metabib_field config interface 2010-03-09T16:37:37 dbs: btw, with 1.6.0.3, you can add index-specific entries to your search dropdowns 2010-03-09T16:37:39 miker_: yeah, I was thinking that making it a default out-of-the-box would greatly help the SRU / Z39.50 server side of things... and a new config interface would be goodness to tweak (undoubtedly some people will want keyword|issn, keyword|isbn, keyword|lccn for more specific matches ... but they're nuts) 2010-03-09T16:38:27 miker_: I added a "Journal title" search to our dropdown via stypes_selector and some other futzing 2010-03-09T16:38:36 ahh good! 2010-03-09T16:39:00 yeah, just adds &bl=s to a plain old title search. fit the bill nicely 2010-03-09T16:40:30 http://svn.open-ils.org/trac/ILS-Contrib/changeset/817 through 819 was the path 2010-03-09T16:41:09 it's not my prettiest code, for sure 2010-03-09T16:41:41 warren has some good stuff going on for nrcan too 2010-03-09T16:41:54 yeah, but is he sharing it? eh? :) 2010-03-09T16:41:59 heh 2010-03-09T16:42:25 *** bshum has quit IRC 2010-03-09T16:43:34 they teach us to share in kindergarden, but then send mixed messages later :) 2010-03-09T16:44:32 *** grantjohnson has quit IRC 2010-03-09T16:50:42 phasefx++ 2010-03-09T17:07:34 *** Dyrcona has quit IRC 2010-03-09T17:23:51 *** twirlip_beverage is now known as twirlip 2010-03-09T17:28:41 *** jenny1 has left #evergreen 2010-03-09T17:35:16 *** dbs has quit IRC 2010-03-09T17:50:29 *** pmplett has joined #evergreen 2010-03-09T18:38:54 Is the list of OSRF methods available over http available anywhere? 2010-03-09T18:46:30 *** pmplett has quit IRC 2010-03-09T18:52:02 moodaepo: i don't know of it (which is by no means definitive) 2010-03-09T18:52:40 atz: thnx. no worries 2010-03-09T18:54:02 my co-worker thinks there might be methods opening up sensitive info to the worlds and I told him I'm pretty sure evergreen doesn't, trying to find some docs 2010-03-09T18:59:14 it is easy to get that impression, because there are two SRF routing bits 2010-03-09T18:59:22 one is public and the other private 2010-03-09T19:00:17 shouldn't be able to get anything really good w/o an auth token 2010-03-09T19:00:46 *** pmplett has joined #evergreen 2010-03-09T19:01:05 but by all means, it's good to reality-check these things every now and then 2010-03-09T19:02:34 i'd just respond to them "exploitz code or stfu plz" 2010-03-09T19:38:07 *** jamesrf has quit IRC 2010-03-09T19:45:13 dbwells: about before, I'd encourage you to bring up any concerns or ideas you have re in-db circ/hold on the dev list 2010-03-09T19:45:26 dbwells: as berick suggested 2010-03-09T19:54:58 *** jamesrf has joined #evergreen 2010-03-09T21:12:28 *** jamesrf has quit IRC 2010-03-09T21:22:10 *** frzosima_ has joined #evergreen 2010-03-09T21:24:25 *** pmplett has quit IRC 2010-03-09T21:24:58 *** frzosima has quit IRC 2010-03-09T21:24:58 *** frzosima_ is now known as frzosima 2010-03-09T21:29:13 *** frzosima_ has joined #evergreen 2010-03-09T21:32:09 *** frzosima has quit IRC 2010-03-09T21:32:09 *** frzosima_ is now known as frzosima 2010-03-09T21:35:36 *** twirlip has left #evergreen 2010-03-09T21:40:14 *** pmplett has joined #evergreen 2010-03-09T21:42:05 *** jamesrf has joined #evergreen 2010-03-09T22:18:53 *** leed- has joined #evergreen 2010-03-09T22:24:59 *** gmcharlt` has joined #evergreen 2010-03-09T22:25:37 *** leed has quit IRC 2010-03-09T22:25:38 *** dmcmorris_esi has quit IRC 2010-03-09T22:25:38 *** gmcharlt has quit IRC 2010-03-09T22:27:57 *** dmcmorris_esi has joined #evergreen 2010-03-09T22:59:36 *** dbs has joined #evergreen 2010-03-09T23:35:46 *** jamesrf has quit IRC 2010-03-09T23:55:02 *** jamesrf has joined #evergreen