2010-08-20T00:14:22 *** dbs has quit IRC 2010-08-20T01:16:53 *** mck9 has left #evergreen 2010-08-20T01:52:15 *** artunit_ has joined #evergreen 2010-08-20T01:53:46 *** artunit has quit IRC 2010-08-20T01:53:53 *** artunit_ is now known as artunit 2010-08-20T02:14:45 *** rickd has quit IRC 2010-08-20T02:15:19 *** rickd has joined #evergreen 2010-08-20T03:01:55 *** magnusenger has joined #evergreen 2010-08-20T04:04:59 *** branflakes has quit IRC 2010-08-20T04:31:13 *** natschil has joined #evergreen 2010-08-20T04:32:15 *** natschil has quit IRC 2010-08-20T04:39:28 *** atz_ has joined #evergreen 2010-08-20T04:42:38 *** atz has quit IRC 2010-08-20T06:33:23 *** Dyrcona has joined #evergreen 2010-08-20T07:21:12 *** collum has joined #evergreen 2010-08-20T07:45:41 *** sfortin has joined #evergreen 2010-08-20T08:02:07 *** r123 has joined #evergreen 2010-08-20T08:05:41 *** mck9 has joined #evergreen 2010-08-20T08:29:46 *** kmlussier has joined #evergreen 2010-08-20T08:36:34 *** Dyrcona has quit IRC 2010-08-20T08:41:13 *** Dyrcona has joined #evergreen 2010-08-20T08:43:47 *** shopkins has joined #evergreen 2010-08-20T08:52:07 *** sfortin has quit IRC 2010-08-20T08:59:25 *** dbs has joined #evergreen 2010-08-20T09:00:36 miker_: ignore my @later, I'm in the process of adding open-ils.cat.authority.records.merge via in_db_auth_merge; never never would have figured out how to json_query that without the bre example 2010-08-20T09:00:47 cargo_cult++ 2010-08-20T09:05:13 *** Meliss has joined #evergreen 2010-08-20T09:08:46 *** dbs has quit IRC 2010-08-20T09:08:52 *** bshum has joined #evergreen 2010-08-20T09:18:35 *** dbs has joined #evergreen 2010-08-20T09:20:05 dbs: ignoring. I was going to mention the bib example 2010-08-20T09:20:45 heh. amazing what 6.5 hours of sleep + forced concentration (on a bus) will do 2010-08-20T09:20:46 grabbing 0376 2010-08-20T09:23:28 and now to hook up the new API to the authority management interface 2010-08-20T09:23:56 *** natschil has joined #evergreen 2010-08-20T09:23:59 *** Dmagick_ has quit IRC 2010-08-20T09:24:55 *** natschil has quit IRC 2010-08-20T09:25:00 *** natschil has joined #evergreen 2010-08-20T09:26:02 *** natschil has joined #evergreen 2010-08-20T09:30:28 *** magnusenger has quit IRC 2010-08-20T09:40:04 *** brendan_bywater has joined #evergreen 2010-08-20T09:40:26 *** yboston has joined #evergreen 2010-08-20T09:45:25 *** Dmagick_ has joined #evergreen 2010-08-20T09:46:42 *** dbs has quit IRC 2010-08-20T09:55:24 *** grantjohnson has joined #evergreen 2010-08-20T09:59:18 *** afterl has joined #evergreen 2010-08-20T10:01:33 *** dbs has joined #evergreen 2010-08-20T10:02:36 we can haz authority record merges via a reasonable UI. yay. 2010-08-20T10:03:54 *** bshum has quit IRC 2010-08-20T10:04:11 w00t 2010-08-20T10:04:19 * berick hands dbs a cheezzzburger 2010-08-20T10:06:18 *** bshum has joined #evergreen 2010-08-20T10:06:57 I was thinking that the authority record merge function should perhaps add one or more "See from" reference into the master record, based on the main entry and any "See froms" in the source record 2010-08-20T10:07:10 Fancy munging. 2010-08-20T10:07:27 hmmm... is this valid javascript: 2010-08-20T10:07:48 2010-08-20T10:07:53 blank is valid 2010-08-20T10:08:29 var [hint, attr] = this._selector(id).getValue().split(":"); 2010-08-20T10:08:55 chrome says no 2010-08-20T10:09:23 whee! in 2.0 we're going to be able to get rid of connect_by() 2010-08-20T10:09:26 specifically, unexpected token '[' 2010-08-20T10:09:53 atz_: yeah, you can't vivicate a list of vars like that 2010-08-20T10:10:03 too perlish for JS's simple mind 2010-08-20T10:10:04 that's what i though 2010-08-20T10:10:05 ;) 2010-08-20T10:10:07 *thought 2010-08-20T10:10:53 * dbs starts chanting "TWO POINT OH! TWO POINT OH!" 2010-08-20T10:14:27 :) 2010-08-20T10:16:50 you can do that in FF/XULRunner 2010-08-20T10:17:02 it wouldn't be good code for the opac but it would be fine for SC interfaces 2010-08-20T10:17:04 js> var [a, b] = [1, 2]; 2010-08-20T10:17:04 js> a 2010-08-20T10:17:04 1 2010-08-20T10:17:08 yeah ;) what senator said 2010-08-20T10:17:47 atz_: ^-- 2010-08-20T10:18:37 can you declare the array var first, then assign, and make it happy everywhere? 2010-08-20T10:19:52 dbs: it doesn't like declaring the array var first 2010-08-20T10:19:56 by itself 2010-08-20T10:20:23 the errors is "missing = in destructuring declaration" ;) 2010-08-20T10:20:40 somehow "destructuring" is apropos w/ javascript 2010-08-20T10:22:37 so var temparray = this._selector(id).getValue().split(":"); doesn't work, followed by hint and attr being pulled out of it? 2010-08-20T10:23:59 *** grantjohnson has quit IRC 2010-08-20T10:24:10 *** dbs has quit IRC 2010-08-20T10:25:24 or more explicitly... var temparray = new Array(this._selector(id).getValue().split(":")); 2010-08-20T10:26:30 tsbere's way works fine, split() returns an array regardless 2010-08-20T10:26:48 the thing is, is there any expectation that staff client interfaces work in anything other than xulrunner/firefox? 2010-08-20T10:26:56 senator: no 2010-08-20T10:27:20 * senator agrees with berick, as there are plenty of language features (for example https://developer.mozilla.org/en/JavaScript/New_in_JavaScript/1.6 ) that can be taken advantage of 2010-08-20T10:27:22 senator: there is /tons/ of ff/xulrunner-only code in the staff interfaces 2010-08-20T10:27:36 berick: word, i know i'm responsible for some 2010-08-20T10:29:18 however i don't see any reason we should prefer an incompatible syntax. among other things, it will prevent code being shared w/ the opac side as easily. 2010-08-20T10:31:43 atz_: yes, if there's an expectation of sharing. but almost all of the staff (web) UIs interact with xulrunner in a way no regular browser ever could, regardless of JS syntax 2010-08-20T10:31:52 senator, that link talks about js 1.6... not FF-specific JS 2010-08-20T10:33:05 atz_: well, i'm not sure about 1.6, but change the 1.6 to 1.7 and you'll see features including the var [a,b] = [1,2]; syntax. don't know if it's meant to be FF-specific or not, but FF apparently is ok with it where chrome isn't 2010-08-20T10:33:25 copy cost! copy cost! 2010-08-20T10:35:08 jeff: it's a start, anyway 2010-08-20T10:35:39 berick: the cost currently is only recorded in the lineitem, right? 2010-08-20T10:35:55 or is that dealing with the actual acp? 2010-08-20T10:36:45 jeff: lineitem tracks the estimated cost; debits track the encumbrances / spent money. the commit deals with actual newly-created acp's 2010-08-20T10:37:37 *** dbs has joined #evergreen 2010-08-20T10:37:51 and is it storing the cost in asset.copy, and the sql/IDL support for cost just isn't committed yet? 2010-08-20T10:37:54 berick: actually i don't see the same composition. to me "almost all" of the staff client is autogrid table views of something or other. certainly the bulk of day-to-day functionality is in more complex screens, but all the stuff that i view in browser already works fine. 2010-08-20T10:38:27 berick: oh, wait. i think i missed what i was looking for in the IDL. 2010-08-20T10:39:46 berick: yep, nevermind that last question. i see it in the IDL. 2010-08-20T10:42:02 atz_: that's encouraging, but, be that as it may, IMO we shouldn't spend time thinking WWIED (what would IE do *sorry* ;) while wiring up a staff interface. 2010-08-20T10:42:17 * dbs is with atz on this one - at least where a reasonably compatible option exists without much more effort. my dumb little authority list works in chrome and firefox 2010-08-20T10:43:07 using dojo as a means of harmonizing things is a good trend I think 2010-08-20T10:43:25 berick: there's GCF for IE to turn it into a real browser, right? not that I've even tried it yet. heh. 2010-08-20T10:43:44 * phasefx has seen an xslt turn xul into xhtml for IE ;) 2010-08-20T10:43:46 ;) 2010-08-20T10:43:49 jeff: *nod* 2010-08-20T10:44:23 phasefx: aye. but then the new problem becomes - some poor sap is going to have to take a stab at updating to dojo 1.5 and fixing all of the brokenness 2010-08-20T10:46:08 dbs: poor blokes. So far all that potential pain seems to be on the dom side. I'm content in xul land where things seem more stable, but not satisfied (grumbles over remote xul being crippled, etc.) 2010-08-20T10:46:29 content, but not satisfied, yeah, I'm a poet 2010-08-20T10:46:46 * phasefx says burn it all and go to ncurses :) 2010-08-20T10:46:48 afaict, fwiw: "Chrome will execute Javascript marked as 'javascript1.7', but it does not support JS1.7 features" 2010-08-20T10:47:16 FF is up to 1.8.5 2010-08-20T10:47:30 FF4 that is 2010-08-20T10:47:45 * dbs laughs at the title of the 5th edition ecmascript spec: "Final final final final draft Standard ECMA-262 5th edition" 2010-08-20T10:48:11 * dbs s at phasefx 2010-08-20T10:48:54 phasefx: have you tackled ncurses-over-ssh printing yet? ;) 2010-08-20T10:49:37 apparently there is some vt100 escape code or something that can print from a terminal 2010-08-20T10:49:39 actually, that's not difficult. 2010-08-20T10:49:40 yes. 2010-08-20T10:49:51 ah, was wondering about that 2010-08-20T10:49:52 "send this next block of text to your printer, please" 2010-08-20T10:49:53 You haven't lived until you'ved merged authorities on a mobile phone 2010-08-20T10:50:06 you then need terminals/emulators that can support it. 2010-08-20T10:50:07 hahaha 2010-08-20T10:50:17 many do, including PuTTY 2010-08-20T10:50:42 and, y'know, DEC vt420s :) 2010-08-20T10:53:30 *** collum has left #evergreen 2010-08-20T10:54:34 atz_: I can imagine someone with a strong enough itch coming up with a cross-browser web interface that could eventually replace the existing client (I think this is what you really want to see), but the main thing against it now is inertia. Evergreen librarians were anti-web-app in the beginning, and even though it's acceptable now, who wants to fund re-inventing it, dealing with new bugs, etc? 2010-08-20T10:56:01 surely the path to that would be made easier if new code was cross-compatible where there was no compelling reason to be xulrunner-only? 2010-08-20T10:56:04 It would make staff client updates a lot easier 2010-08-20T10:56:46 Yep. One by one, replace the XUL interfaces with Dojo interfaces (just like the patron editor), and eventually... voila 2010-08-20T10:57:18 that's how I see it going down 2010-08-20T10:57:51 that makes sense to me 2010-08-20T10:58:05 eventually we'll just need the mozilla bits for weird peripherals and maybe offline mode (if we don't like the offline paradigms other webstacks come up with) 2010-08-20T10:58:20 And then we could work on harmonizing ILS APIs so that one web-based interface could work with every ILS! Easy. 2010-08-20T10:58:45 * dbs can't remember - did dbwells decide to go all XUL for the new serials interfaces? 2010-08-20T10:58:45 yeah, i figure by the time we get down to those last bits, somebody will have a great new solution for us to leverage 2010-08-20T10:59:06 question is, is it worth the effort and reduced functionality to make the staff client a cross-browser app. as a FF-specific app, you could get the best of both worlds 2010-08-20T10:59:08 dbs: he did 2010-08-20T10:59:32 * phasefx is happy enough that Mozilla is cross-platform 2010-08-20T11:00:35 and it's easier to use working conveniences now 2010-08-20T11:00:45 "more cross-browser", yes. "every last browser", probably not. 2010-08-20T11:01:20 and for all that I expect dojo to take over, right now our dojo interfaces feel more brittle to me than the xul ones 2010-08-20T11:01:32 that's true 2010-08-20T11:01:59 That's why dbwells went with XUL, if I recall correctly. To be fair, we're now two major revisions behind Dojo state of the art. 2010-08-20T11:02:09 * phasefx nods 2010-08-20T11:02:20 That said, Dojo's track record has been a bit spotty. heh 2010-08-20T11:02:28 I love these unofficial dev meetings :) 2010-08-20T11:02:33 hence the breakage-gauntlet left to be run 2010-08-20T11:03:18 * phasefx is so happy that xulrunner releases are being put out in lock-step with FF releases now. Don't have to run that particular gauntlet anymore 2010-08-20T11:03:27 *** sfortin has joined #evergreen 2010-08-20T11:03:49 miker_ pasted "how to get rid of connectby()" at http://paste.lisp.org/display/113702 2010-08-20T11:04:10 but thanks to tsbere, we're going to be having the option of self-updating xul apps AND firefox extensions 2010-08-20T11:04:23 paste went wonky some there... odd 2010-08-20T11:05:31 one reason we might care about a Chrome-based staff client interface would be JavaScript performance; I know FF4 is competitive, but imagine if the last mile of the performance piece for KCLS came down to "crap, this patron registration interface doesn't render fast enough / gets mysteriously memory-bloaty after being open for 8 hours" 2010-08-20T11:06:25 Far-fetched maybe. 2010-08-20T11:06:38 that happens in all web apps...browser-based apps suck. 2010-08-20T11:06:49 miker_ annotated #113702 "better paste, I hope" at http://paste.lisp.org/display/113702#1 2010-08-20T11:07:04 What about a browser-based ncurses emulator? eh? 2010-08-20T11:07:36 * phasefx sees green-screens everywhere he goes in the real world, running on windows desktops 2010-08-20T11:07:38 * dbs finishes getting his Debian Squeeze image up to date again - lots has happened in the last 24 hours 2010-08-20T11:07:45 dbs: I've been doing most of my dev in chrome and confirming functionality in FF(3.6) recently, fwiw 2010-08-20T11:07:58 miker_: that's how I've been working otoo 2010-08-20T11:08:09 * Dyrcona prefers gray on black 2010-08-20T11:19:14 *** atz has joined #evergreen 2010-08-20T11:21:41 *** dbs has quit IRC 2010-08-20T11:22:07 *** atz__ has joined #evergreen 2010-08-20T11:22:35 *** atz_ has quit IRC 2010-08-20T11:22:54 sorry... virtualbox install flipping my net cards 2010-08-20T11:23:32 *** jamesrf has joined #evergreen 2010-08-20T11:24:18 *** atz has quit IRC 2010-08-20T11:24:28 *** atz__ is now known as atz 2010-08-20T11:34:34 *** dbs has joined #evergreen 2010-08-20T11:37:26 phasefx: i think i heard you mention it... 2010-08-20T11:49:14 *** dbs has quit IRC 2010-08-20T12:02:54 jfyi, I'm seeing stuff like this with trunk in my apache logs: 2010-08-20T12:02:54 Argument "uris" isn't numeric in numeric gt (>) at /openils/lib/perl5/OpenILS/WWW/SuperCat.pm line 1388. 2010-08-20T12:02:58 [Fri Aug 20 12:02:02 2010] [info] [client 192.168.71.71] OpenSRF JSON Request returned error: osrfMethodException -> *** Call to [open-ils.search.serial.record.bib.retrieve] failed for session [1282320121.878015.128232012118041], thread trace [1]:\nCan't use an undefined value as an ARRAY reference at /openils/lib/perl5/OpenILS/Application/Search/Serial.pm line 176.\n\n, referer: http://192.168.25.141/opac/en-US/skin/defa 2010-08-20T12:03:32 I have no serials data 2010-08-20T12:31:36 how do you purge the AddedContent cache? 2010-08-20T12:38:28 * tsbere is looking at his notes for making docs and trying to figure out how to best split it into multiple pages 2010-08-20T12:47:25 * tsbere decided to try something and has found out it doesn't work from the staff_client directory 2010-08-20T12:52:28 phasefx: restart memcache 2010-08-20T12:52:34 re: added content 2010-08-20T12:52:53 berick: can OpenSRF stay running? 2010-08-20T12:53:12 osrf would need restart too 2010-08-20T12:53:18 k, thanks 2010-08-20T12:53:50 * tsbere can't find a reliable way to run the server-xul target without running a full install cycle from the top level makefile outside of manually telling it where webdir is 2010-08-20T12:54:54 *** Dyrcona has quit IRC 2010-08-20T12:56:03 *** kmlussier has quit IRC 2010-08-20T12:58:48 *** jamesrf is now known as jamesrf-afk 2010-08-20T13:12:40 *** dbs_n1 has joined #evergreen 2010-08-20T13:14:30 Dyrcona - the lassos in circ policies idea is noble but capital-H hard. 2010-08-20T13:16:05 *** sfortin has quit IRC 2010-08-20T13:19:29 unless maybe using circ scripts :) 2010-08-20T13:20:45 dbs_n1: Dyrcona left already. 2010-08-20T13:21:01 and already decided it was capital-H hard, I think 2010-08-20T13:31:20 Yep, that's why I was trying to offer consolation 2010-08-20T13:32:05 * dbs_n1 leaves too 2010-08-20T13:32:11 *** dbs_n1 has quit IRC 2010-08-20T13:57:35 ergh, argh - OpenSRF::Utils::JSON->perl2JSON takes a list of integers and stringifies them (e.g. (1, 2, 3) becomes ["1", "2", "3"]) 2010-08-20T13:57:55 ok........ 2010-08-20T13:57:55 not cool when you just want to pass the array to a json_query as an IN clause 2010-08-20T13:58:06 so fix it? 2010-08-20T13:58:35 tsbere: thanks, ever so helpful 2010-08-20T13:59:22 * tsbere has never looked at the perl2JSON routine and thus has no clue what may be wrong 2010-08-20T13:59:22 assuming I fix that, would you like to test all of the uses of perl2JSON to ensure that the fix doesn't break some other functionality? 2010-08-20T13:59:23 denials: but are they really perl ints? if they've been stringified in the past I believe JSON::XS considers them strings ... other thing, does the string version not work in the IN clause? 2010-08-20T14:00:01 they're ints if I take them as $are = int($are) and push them on an array, right? 2010-08-20T14:00:51 push @list, int($are); # if you assign back, perhaps not 2010-08-20T14:00:52 and postgresql doesn't like comparing strings to ints as of 8.3 2010-08-20T14:01:11 it ... should 2010-08-20T14:01:23 ahh - you have a point there, I'll try that 2010-08-20T14:01:31 /that/ cast shouldn't be broken, I mean 2010-08-20T14:02:17 denials: this works for me in 8.4: select id from actor.org_unit where id in ('1','2','3','4'); 2010-08-20T14:03:02 is PG actually complaining about being able to cast from unkown to int? 2010-08-20T14:03:11 or is it avoiding an index, perhaps? 2010-08-20T14:03:43 * denials sheepishly notes the = instead of IN() that he was expecting json_query to generate 2010-08-20T14:04:46 well, even an OR list will surely work... 2010-08-20T14:04:48 * miker_ tests 2010-08-20T14:05:13 works for me: select id from actor.org_unit where id = '1' or id = '2'; 2010-08-20T14:05:20 no, the problem must be in my json_query formulation 2010-08-20T14:05:54 I was going with the JSON tutorial: "where": { 2010-08-20T14:05:55 "parent_ou": [ 3, 5, 7 ] 2010-08-20T14:05:59 ahh... well, if you want an IN list, just say { id => \@list } 2010-08-20T14:06:00 } 2010-08-20T14:06:11 gahhh 2010-08-20T14:06:13 right 2010-08-20T14:06:39 magics everywhere messing with my head 2010-08-20T14:06:41 oh... array ref vs array? 2010-08-20T14:07:25 no, I was using perl2JSON to convert the array ref to [1,2,3] (after assigning int($are) to a new var and pushing that var onto the array) 2010-08-20T14:07:43 oh, no, just pass the array ref 2010-08-20T14:07:44 and stuffing that directly into the where clause 2010-08-20T14:07:48 right, I get that now 2010-08-20T14:07:51 :) 2010-08-20T14:08:44 this is where I agree with myself from about a day ago that the JSON tutorial is great, but translating it into other languages can bring madness :) 2010-08-20T14:08:49 thanks miker_ 2010-08-20T14:10:05 like my mama always used to say, sometimes you're the forehead and sometimes you're the desk 2010-08-20T14:14:06 json_query makes a good effort to DWIM ... if you don't think too hard about it, or something ... 2010-08-20T14:15:48 yeah. my mistake was trying to match the example in the tutorial - heh 2010-08-20T14:15:54 got it working, much thanks for the help 2010-08-20T14:16:42 soon the authority management interface will show a list of the bibs linked to each authority record. oh yes. 2010-08-20T14:38:28 *** shopkins has quit IRC 2010-08-20T14:40:01 *** gdunbar has quit IRC 2010-08-20T14:41:42 btw, my forehead? It's _made_ of desk. 2010-08-20T14:42:47 that's qotd material right there 2010-08-20T14:54:15 denials: fwiw, there's no injection risk with IN via json_query, but you can certainly make the query fail ;) 2010-08-20T15:02:05 *** kmlussier has joined #evergreen 2010-08-20T15:04:25 denials: lamely, the test call worked again... 2010-08-20T15:08:18 *** brendan_bywater has quit IRC 2010-08-20T15:09:23 I think one of either tabs or spaces should never have been invented. Or tab-to-space settings shouldn't be configurable. @curse this flexible world 2010-08-20T15:11:47 I think the mistake was treating them both as whitespace 2010-08-20T15:14:08 ah, i assume that's in reference to my last commit. doh. 2010-08-20T15:15:13 shall I add a vim:noet:ts=4:sw=4: ? 2010-08-20T15:18:00 make a prorject .vimrc file 2010-08-20T15:18:08 *project 2010-08-20T15:19:01 or a couple, if the styles are different in different parts 2010-08-20T15:20:43 atz: would vim find the .vimrc if it was in a parent directory of whatever project file you were editing at the time? 2010-08-20T15:20:59 you can pass it as an option 2010-08-20T15:21:23 gotcha 2010-08-20T15:21:25 or use it to replace/supplement your own 2010-08-20T15:26:34 Some .vimrc additions recommendations may be in order 2010-08-20T15:27:04 like autocmd BufNewFile,BufRead /path/to/openils/src/* set tabstop=4 shiftwidth=4 etc 2010-08-20T15:41:30 *** jamesrf-afk has quit IRC 2010-08-20T15:59:38 *** Meliss has quit IRC 2010-08-20T16:01:12 *** kmlussier has quit IRC 2010-08-20T16:03:13 *** bshum has quit IRC 2010-08-20T16:10:34 *** rjackson-isl has left #evergreen 2010-08-20T16:22:06 *** afterl has quit IRC 2010-08-20T16:30:53 *** r123 has quit IRC 2010-08-20T16:44:09 *** jamesrf has joined #evergreen 2010-08-20T17:17:19 *** pmplett has joined #evergreen 2010-08-20T17:46:41 *** atz has quit IRC 2010-08-20T17:58:26 *** natschil has quit IRC 2010-08-20T17:59:45 *** yboston has quit IRC 2010-08-20T18:05:50 *** atz has joined #evergreen 2010-08-20T19:17:35 *** jamesrf has quit IRC 2010-08-20T20:24:34 *** pmplett has quit IRC 2010-08-20T20:25:17 *** pmplett has joined #evergreen 2010-08-20T20:48:14 *** Callender has quit IRC 2010-08-20T20:49:08 *** phasefx has quit IRC 2010-08-20T20:49:32 *** leed has quit IRC 2010-08-20T20:50:02 *** Dmagick_ has quit IRC 2010-08-20T20:55:02 *** phasefx has joined #evergreen 2010-08-20T20:56:14 *** leed has joined #evergreen 2010-08-20T20:56:25 *** Dmagick has joined #evergreen 2010-08-20T21:03:03 *** Callender has joined #evergreen 2010-08-20T22:01:19 *** pmplett has quit IRC