2011-02-10T00:35:49 *** tildeequals has joined #evergreen 2011-02-10T00:58:31 *** Dmagick has quit IRC 2011-02-10T00:59:36 *** Dmagick has joined #evergreen 2011-02-10T02:18:59 *** atz has joined #evergreen 2011-02-10T02:27:56 *** Guest75972 is now known as jennam 2011-02-10T02:27:59 *** jennam has joined #evergreen 2011-02-10T02:33:50 *** atz has quit IRC 2011-02-10T03:15:00 *** dave-esi has quit IRC 2011-02-10T07:06:27 *** jtbates has quit IRC 2011-02-10T07:08:32 *** StephenGWills has joined #evergreen 2011-02-10T07:33:23 *** sfortin has joined #evergreen 2011-02-10T07:50:48 *** kmlussier has joined #evergreen 2011-02-10T07:52:01 *** sfortin has quit IRC 2011-02-10T07:56:40 *** sfortin has joined #evergreen 2011-02-10T08:00:41 *** rickd_ has joined #evergreen 2011-02-10T08:04:31 *** granitize has joined #evergreen 2011-02-10T08:30:44 *** StephenGWills has quit IRC 2011-02-10T08:33:00 * tsbere has run a full battery of tests on fallthrough, and it works perfectly in all cases. Unfortunetly, dynamic weighting does not work perfectly in all cases in the same tests, so off to testing that today. 2011-02-10T08:42:30 Huh. Easy fix. :D 2011-02-10T08:55:25 *** tater has joined #evergreen 2011-02-10T08:56:00 *** AaronZ-PLS has joined #evergreen 2011-02-10T08:56:26 *** alxp has joined #evergreen 2011-02-10T09:08:27 *** artunit has joined #evergreen 2011-02-10T09:12:38 *** Dyrcona has joined #evergreen 2011-02-10T09:20:17 *** dave-esi has joined #evergreen 2011-02-10T09:33:09 * tsbere pokes eeevil in regards to bug 716447 2011-02-10T09:33:42 I would prefer that make it in before I cut my fallthrough patch, if only to avoid issues for someone trying to apply said patch where I change one of the functions again. 2011-02-10T09:35:05 *** atz has joined #evergreen 2011-02-10T09:39:41 *** atz has left #evergreen 2011-02-10T09:41:04 @later tell dbs thanks for testing the perlmod build. I'll happily chalk my (easily resolved) issues up as a local oddity. 2011-02-10T09:41:04 berick: The operation succeeded. 2011-02-10T09:43:55 *** dbs has joined #evergreen 2011-02-10T09:43:55 *** dbs has joined #evergreen 2011-02-10T09:44:34 tsbere: looking 2011-02-10T09:48:15 eeevil: as I was going to bed last night, I was thinking that it might make sense to keep the SVF stuff in a branch until it's more fleshed out 2011-02-10T09:49:04 dbs: it's nearly fleshed now, but sure, I can do that 2011-02-10T09:50:14 okay. I was thinking about "crap, I'll have to install postgresql 9 on my laptop and set up new VMs" but maybe that's not a big deal 2011-02-10T09:51:03 dbs: fwiw, debian's sloppy backports version of 9.0 and its pg_upgradecluster worked fine for me yesterday 2011-02-10T09:51:33 eeevil: good to know! but can you type unicode characters into psql? :) 2011-02-10T09:52:08 haven't tried ... ugly 'merican here, sir 2011-02-10T09:52:28 debian can handle 8.4 + 9.0 side by side anyway, right? 2011-02-10T09:52:39 * dbs needs to check into fedora's support for 9.0 2011-02-10T09:52:50 grabbing 0483 for tsbere's patch, btw 2011-02-10T09:54:20 tsbere: that's quite the patch! nice 2011-02-10T09:55:00 Wanted to be lazy and just include it in fallthrough. But laziness in testing made it needed to begin with. Figured there was too much laziness to begin with. <_< 2011-02-10T09:55:51 *** rsoulliere has joined #evergreen 2011-02-10T10:10:38 * csharp seeks how to enable the "bad debt" interface in 1.6.1.6 2011-02-10T10:11:21 csharp: what does that look like in 1.4? I don't remember 2011-02-10T10:11:22 csharp: apache config (1.6.1 example apache should contain the spell) 2011-02-10T10:11:52 phasefx: a simple web interface - not used in prod yet 2011-02-10T10:11:55 eeevil: thanks 2011-02-10T10:12:05 ah 2011-02-10T10:16:52 eeevil: Threw a new patch on 635463 for fallthrough. I will work on re-tooling the docs I had been making so that there may be decent ones for all of this at some point. 2011-02-10T10:19:53 csharp: I'm not sure I backported the syntax error fix for the BadDebt.pm module 2011-02-10T10:20:01 csharp: lemme look 2011-02-10T10:20:24 eeevil: I am expecting another round of you saying you want someone else to look at my widget changes, though. 2011-02-10T10:21:29 csharp: perl -c Open-ILS/src/perlmods/OpenILS/WWW/BadDebt.pm says "no, dbs hasn't backported that fix" 2011-02-10T10:21:36 * dbs will do so 2011-02-10T10:21:41 dbs: thanks 2011-02-10T10:24:26 csharp: for a local fix, see http://svn.open-ils.org/trac/ILS/changeset/19343/trunk/Open-ILS/src/perlmods/lib/OpenILS/WWW (adjust to remove /lib/ from the path) 2011-02-10T10:27:47 dbs: much appreciated 2011-02-10T10:29:38 csharp: ahh, now that I've made it back to rel_1_6_0, I see where the syntax error crept in 2011-02-10T10:30:33 rel_1_6_0 had that line as a ternary operator; rel_1_6_1 converted it into an if / elsif / else statement because it added one more condition, but failed to change the : to a semicolon 2011-02-10T10:30:49 csharp: so if you're running rel_1_6_0 or before, you're probably fine 2011-02-10T10:33:36 anyone happen to know what the "right" server address is for AuthorizeNet developer test accounts? 2011-02-10T10:33:51 i'm getting back "desc":"The credit card processor has declined the transaction." 2011-02-10T10:34:02 * 29384: open_tcp_connection: destination host not found: `https://test.authorize.net/gateway/transact.dll' (port 443) (Bad message)\n) (HTTPS headers: ) (Raw HTTPS content: )", 2011-02-10T10:35:36 dbs: good - thanks for that - we'll be in 1.6.1 for production soon (currently 1.4.0.7) 2011-02-10T10:36:17 csharp++ 2011-02-10T10:39:12 also, I completely agree that man pages would be a nice format to offer docs in. writing or extracting (for pydoc, POD & doxygen stuff) is probably step 1 though, followed by getting the content in a wranglable format that can generate HTML & man & PDF & whatever 2011-02-10T10:39:13 anyone anyone? be my hero before this training :) 2011-02-10T10:39:54 * mrpeters-isl realizes he should have asked sooner 2011-02-10T10:39:55 mrpeters-isl: I suspect authorize.net might know the right url ... ;) 2011-02-10T10:40:02 eeevil: this is the one they gave me 2011-02-10T10:40:07 doh 2011-02-10T10:40:21 atheos: any special insider knowledge? 2011-02-10T10:42:42 mrpeters-isl: https://test.authorize.net/gateway/transact.dll is test, https://secure.authorize.net/gateway/transact.dll is production. 2011-02-10T10:42:50 jeff: right 2011-02-10T10:42:54 mrpeters-isl: you can also put your "real" account into "test mode" in the web manage interface. 2011-02-10T10:42:55 so i have test mode set to true 2011-02-10T10:43:10 and im using my login/password frmo developer.authorize.net 2011-02-10T10:43:45 mrpeters-isl: do you have a "real" login/key, or a "developer" login/key? 2011-02-10T10:43:52 developer 2011-02-10T10:44:19 signed up at https://developer.authorize.net 2011-02-10T10:44:31 was hoping this would at least lte me test paying some bills in a training 2011-02-10T10:45:50 tried with secure. as well 2011-02-10T10:46:07 staff client gives - http://paste.lisp.org/display/119548 2011-02-10T10:46:53 mrpeters-isl: Can the server dig/ping secure.authorize.net? 2011-02-10T10:48:04 developer credentials will never work against the live endpoint (secure.authorize...) 2011-02-10T10:48:22 okay - corrected the BadDebt.pm typo, now receiving in apache log: Use of uninitialized value in string eq at /openils/lib/perl5/OpenILS/WWW/BadDebt.pm line 53 2011-02-10T10:48:48 * dbs will looksee 2011-02-10T10:48:51 line 53: my $mark_bad = $cgi->param('action') eq 'unmark' ? 'f' : 't'; 2011-02-10T10:49:05 dbs: thakns 2011-02-10T10:49:07 thanks 2011-02-10T10:49:37 csharp: lemme guess, the URL doesn't have &action=something in it? 2011-02-10T10:50:24 dbs: I set it up as dmcmorris did on our test cluster - as a redirect from /baddebt 2011-02-10T10:51:23 * dbs has reached his limit of knowledge about how BadDebt.pm is supposed to operate, but it's pretty clearly expecting an "action" CGI parameter in that line 2011-02-10T10:51:48 csharp pasted "baddebt redirect in eg_vhost.conf" at http://paste.lisp.org/display/119549 2011-02-10T10:52:28 dbs: thanks for working with me in any case 2011-02-10T10:53:34 it's supposed to be an interface that accepts a file of xact ids to mark their resulting billings as "unrecovered" - meant to meet an auditing requirement for (some of?) our libraries 2011-02-10T10:53:56 * csharp doesn't fully understand its purpose either ;-) 2011-02-10T10:55:35 you could set action=unmark to say "set the unrecovered flag to 'false'", or to anything else to set that flag to "true" 2011-02-10T10:55:47 (just reading the code) 2011-02-10T10:56:31 oh - I see 2011-02-10T10:57:00 aside: the else { die; } in the middle of the cstore transaction is a bit worrisome; not sure that the transaction will automatically get rolled back in that case 2011-02-10T11:06:28 whatever is causing the uninitialized value error prevents the HTML from being created 2011-02-10T11:06:39 * csharp goes back to SVN for clues 2011-02-10T11:07:41 csharp: are you going directly into this via a web browser? 2011-02-10T11:08:02 dbs: yes 2011-02-10T11:08:07 if so, you could just append ?action=markymark to the URL 2011-02-10T11:08:11 there's not a link for it 2011-02-10T11:08:17 dbs: ah - lemme try 2011-02-10T11:11:40 * csharp shrugs and gives up 2011-02-10T11:12:20 csharp: same error? 2011-02-10T11:12:29 can you paste the URL here? 2011-02-10T11:13:45 dbs: it's an internal network 2011-02-10T11:14:05 csharp: yeah, you can obfuscate the hostname / session ID if you want 2011-02-10T11:14:29 when I last looked at this, like a year and a half ago, I was able to just go to /baddebt and it would go straight to the page 2011-02-10T11:15:32 *** shopkins has joined #evergreen 2011-02-10T11:15:56 oh, I just tried https://testbox.pls-hq.org/baddebt?action='mark' but this is where my CGI skills come up short ;-) 2011-02-10T11:26:47 dbs: re else {die}, yes, cstore will roll the transaction back after the timeout (6s by default, IIRC) 2011-02-10T11:28:37 eeevil: Yeah, I knew the timeout would clean things up eventually. And the die ensures it avoids any loop that would tie up lots of backends 2011-02-10T11:28:58 just seems like a rather stark way of handling things 2011-02-10T11:29:03 * dbs heads to french class 2011-02-10T11:32:16 *** tspindler has joined #evergreen 2011-02-10T11:44:05 *** AaronZ-PLS has quit IRC 2011-02-10T11:46:13 *** jenny has joined #evergreen 2011-02-10T11:54:13 *** AaronZ-PLS has joined #evergreen 2011-02-10T11:58:12 csharp: I am looking at the BadDebt.pm file, and it looks like the version in the repo may be incomplete. For instance, the 'show_template' subroutine never gets called anywhere. Do you expect there to be an interface at /baddebt, or are you posting to that URL from some other interface? 2011-02-10T12:02:37 csharp: also, would it be possible to paste the contents of your working BadDebt.pm file from your 1.4 instance? I can make the trunk version 'work', but if your 1.4 version works, it may have other fixes in it as well. 2011-02-10T12:03:43 dbs: there's a mark-bad-debt entry point in the staff client, IIRC, but it looks like line 78 should return show_template() intead of 404 (trunk) 2011-02-10T12:05:47 :q 2011-02-10T12:05:52 d'oh 2011-02-10T12:06:59 csharp pasted "BadDebt.pm" at http://paste.lisp.org/display/119552 2011-02-10T12:07:32 dbwells: yes - I'm expecting an interface - thanks for looking 2011-02-10T12:07:53 eeevil: I don't see the way in via the SC 2011-02-10T12:08:22 maybe a 2.0 addition? 2011-02-10T12:11:54 csharp: there may be a setting to enable it ... recall, this was ~2y ago, and I wrote it quickly to put in place after another effort I wasn't involved with had stalled for quite a while ... I honestly don't remember the details ATM 2011-02-10T12:12:06 csharp: are you sure that is the 1.4 version you pasted? It has the 'reservation' stuff in it, which didn't come until later. 2011-02-10T12:12:46 dbwells: no - I'm working on a test instance of 1.6.1.7 2011-02-10T12:13:12 eeevil: ok - understood - I know it's been awhile ;-) I'll look for the setting 2011-02-10T12:14:25 csharp: I was hoping to see a working version :) In any case, eeevil's fix is also where I would start. On line 78, replace '404' with 'show_template($r)'. Then at least you get the interface, though no promises it will do what it should. 2011-02-10T12:14:52 dbwells: thanks for the guidance 2011-02-10T12:15:10 Hrm. I thought that vandelay could import the JSON bre export -- what it calls "Evergreen Record Entry". My first attempt at importing such a file resulted in -- not success. 2011-02-10T12:20:51 jeff: just export, I think... 2011-02-10T12:22:22 jeff: it autodetects MARC21 vs marcxml ... a patch to detect BRE and handle would be cool, and not too hard, I think 2011-02-10T12:25:52 *** kmlussier has quit IRC 2011-02-10T12:36:00 *** jamesrf has joined #evergreen 2011-02-10T12:44:13 *** rickd_ has quit IRC 2011-02-10T12:50:51 *** tspindler has quit IRC 2011-02-10T12:53:18 *** kmlussier has joined #evergreen 2011-02-10T12:57:44 * tsbere wonders if firefox/xulrunner printing will be fixed anytime soon. Like, before firefox 4/xulrunner 2. 2011-02-10T12:58:20 *** rsoulliere has quit IRC 2011-02-10T13:03:25 maybe if we funnel money to one of their developers 2011-02-10T13:14:16 dbs: re svf topic branch, http://git.esilibrary.com/?p=evergreen-equinox.git;a=summary 2011-02-10T13:14:58 * dbs worked with a number of the Mozilla devs / project team leads in a previous life but that might be a bad thing. heh 2011-02-10T13:15:18 eeevil: cool, danke 2011-02-10T13:18:26 * dbs pulls tt-opac-poc first - wow, looks like berick's been busy 2011-02-10T13:18:40 wow. i should consider upgrading from hardy. python 2.5 doesn't have json in the standard library. 2011-02-10T13:20:28 simplejson to the rescue. 2011-02-10T13:22:32 *** jenny has quit IRC 2011-02-10T13:23:19 dbs checksout svf and diffs against master - looks like eeevil has been hella busy too 2011-02-10T13:23:26 eeevil++ berick++ 2011-02-10T13:23:35 and phasefx++ bshum++ # downloads updating 2011-02-10T13:24:15 dbs: the plan, which I'll explain to the wider community soon, is to quickly build a basic functional t-pac w/ the kcls skin draped over it as a basline. then pause, see what's right, what needs improvement, what we need to do angle it toward a generic skin, etc. 2011-02-10T13:25:55 berick: that sounds good to me; I've been barely functional for the last week (apart from generating overly long emails) so it's not like I would have been able to contribute anything in the short term 2011-02-10T13:28:41 dbs: ah, long emails, la petite mort (but not the good kind). the "then pause" part is where I'm really hoping to elicit broader discussion 2011-02-10T13:30:26 and by "basic functional" i mean the working demo code I shared earlier, skinned 2011-02-10T13:30:31 w/ minor additions 2011-02-10T13:32:09 * berick is very pleased that vandelay accurately reports percent-complete now on uploads 2011-02-10T13:36:16 and then the crazy people can start adding things like microdata and social stuff and fix up the unapi bits 2011-02-10T13:36:41 and really crazy people can try to get rid of the table layout :) 2011-02-10T13:38:12 heh 2011-02-10T13:38:20 berick: fyi the XXX TODO for unapi in record/summary.xml - the reason that's there is because ffx does not handle unapi info in dynamic HTML, so zotero fails if it's not present in the base HTML 2011-02-10T13:38:37 looks like it should be simple to replace with template params though, instead of SSI 2011-02-10T13:38:58 param 1 - date, param 2 - bib record id 2011-02-10T13:39:10 yeah, should be trivial w/ tt 2011-02-10T13:39:28 * dbs sees args.mattype and can't help but think of matt carlson 2011-02-10T13:39:35 heh 2011-02-10T13:39:59 that will be replaced w/ SVF, wee 2011-02-10T13:41:41 * eeevil makes a note to create an in-db unapi branch soon 2011-02-10T14:08:17 In addition to empty tags, lots of "02" tags, "59" tags, etc. I'm suspecting that much of this is people trying to delete a tag and failing, then saving the mangled tag in the record. 2011-02-10T14:17:24 This doesn't seem to be something that MARC::Batch can handle, even with strict mode off. 2011-02-10T14:20:21 *** jenny has joined #evergreen 2011-02-10T14:29:02 wheeee ... perl -c works even where I wouldn't expect it to 2011-02-10T14:34:04 eeevil: yeah 2011-02-10T14:34:33 like against the Publisher sub-modules... 2011-02-10T14:36:09 not all of them 2011-02-10T14:36:21 perl -MFoo may be more useful 2011-02-10T14:36:38 *** jtbates has joined #evergreen 2011-02-10T14:36:47 action.pm for example 2011-02-10T14:37:21 I was only concerned with metabib.pm just now 2011-02-10T14:37:46 *** jtbates has quit IRC 2011-02-10T14:37:55 *** jtbates has joined #evergreen 2011-02-10T14:38:32 metabib.pm doesn't pass for me 2011-02-10T14:39:57 hrm... does here ... oh, I had to PERL5LIB it 2011-02-10T14:40:49 PERL5LIB=~/git/ILS/Open-ILS/src/perlmods/lib/:~/svn/OpenSRF/src/perl/lib/ perl -c 2011-02-10T14:43:04 the ones that fail for me seem to inherit from O:A:Storage:Publisher instead of from O:A:Storage 2011-02-10T14:44:10 whoa ... so, apt-get install git, not happy for me ... crazy 2011-02-10T14:44:18 apt-get install git-core? 2011-02-10T14:45:48 jeff++ # rings a bell 2011-02-10T14:45:59 probly... 2011-02-10T14:46:12 yeah 2011-02-10T14:46:21 silly lenny 2011-02-10T14:47:56 well git in debian predates the existance of git vc, by a long time 2011-02-10T14:48:38 lovely. 67 marcxml records, walk over 'em with MARC::Batch until I get no errors, and i have 49 resulting records. 2011-02-10T14:54:20 jeff: sounds about typical. :) 2011-02-10T14:56:02 *** tildeequals has quit IRC 2011-02-10T15:07:43 *** jamesrf has quit IRC 2011-02-10T15:14:01 *** jamesrf has joined #evergreen 2011-02-10T15:17:03 <_dkyle_> Hey all those using slony: I've got a new editor object with xact => 1, but when I call a pg function that does some updates from that editor via json_query I don't always get the master. 2011-02-10T15:17:40 <_dkyle_> does xact => 1 not give me a transaction? or does that not work with json_query? 2011-02-10T15:20:12 *** tildeequals has joined #evergreen 2011-02-10T15:22:47 _dkyle_: I don't think that works with json_query 2011-02-10T15:23:51 <_dkyle_> dbs: hmm, what method should one use to call a custom function from an editor object? 2011-02-10T15:24:07 <_dkyle_> and have the xact =>1 respected? 2011-02-10T15:30:01 _dkyle_: It's entirely possbile that I'm wrong. My brain is still generating random fuzz 2011-02-10T15:31:44 <_dkyle_> dbs: ++ fuzz generation, thanks, this is a pretty simple little utility thing, I can always go DBI instead of cstore. 2011-02-10T15:32:33 xact=>1 should work with json_query 2011-02-10T15:33:20 I was just reading through cstoreeditor and coming to that conclusion 2011-02-10T15:33:39 <_dkyle_> berick: ever seen a case where it didn't? 2011-02-10T15:34:06 _dkyle_: no. perhaps you could paste the code? 2011-02-10T15:34:15 based on session() 2011-02-10T15:34:30 * dbs goes home 2011-02-10T15:34:34 *** dbs has quit IRC 2011-02-10T15:34:59 <_dkyle_> berick: will do, thanks for taking a look 2011-02-10T15:38:08 _dkyle_ pasted "xact=>1 not honored?" at http://paste.lisp.org/display/119561 2011-02-10T15:39:22 _dkyle_: use $e->xact_commit instead of $e->commit in this context 2011-02-10T15:39:30 $e->commit is a shortcut for commit + disconnect 2011-02-10T15:40:19 xact_[commit|begin|rollback] give you direct access to db transaction mgmt 2011-02-10T15:41:00 *** alxp has quit IRC 2011-02-10T15:42:13 so, if you want to commit after each json_query, instead of using xact=>1, use $e->xact_begin and $e->xact_commit inside the loop 2011-02-10T15:42:14 <_dkyle_> berick: thanks, but would that explain why I'd wind up with a connection to the non-master db? do I need to call the ->xact_begin even though I passed xact=>1 to new? 2011-02-10T15:42:39 <_dkyle_> berick: nice anticipation! thanks! 2011-02-10T15:42:41 it would explain it on for all loop iteratiosn except the first 2011-02-10T15:42:49 s/on// 2011-02-10T15:43:44 then, after the loop, you'll need to call $e->disconnect to release the cstore backend 2011-02-10T15:43:54 which would have been connected on the first call to xact_begin 2011-02-10T15:45:44 <_dkyle_> berick: got it, mucho gracias 2011-02-10T15:46:02 no prob 2011-02-10T15:47:41 *** shopkins has quit IRC 2011-02-10T15:49:11 oh hey. selecting a bunch of holds and retargeting in the staff client doesn't retarget by request time, so it's likely the most recent hold will end up getting "the item" if there's a single available copy in the system. 2011-02-10T15:49:36 if true, that explains some things. 2011-02-10T15:54:30 it's probably using the list order, however it happens to be sorted 2011-02-10T15:54:36 *nod* 2011-02-10T15:55:06 FIFO users shouldn't notice, at least. But yeah, if you're using this in say View Holds after adding some items to a bib... 2011-02-10T15:55:36 new item added, handed off to circ... doesn't capture because there's no ahcm entries yet for the copy. 2011-02-10T15:55:56 select all holds, retarget... the first hold retargeted gets the new copy set as its current copy. 2011-02-10T15:56:13 even though (in this case at least) the first hold to be retargeted was the newest hold. 2011-02-10T15:56:40 i'd still like to have a "checkin + aggressively attempt to capture" for new items. 2011-02-10T15:57:02 not sure the best approach, but it would get past some annoying workAROUNDflows. 2011-02-10T15:58:18 hrmm, even with a specific current_copy set on a hold, doesn't op-capture still come into play in a more-fair manner? 2011-02-10T15:58:28 but 2011-02-10T15:58:45 no copy map like you said, so falls back to holds with a current_copy 2011-02-10T15:58:48 but 2011-02-10T15:58:54 retargeting builds a copy_map 2011-02-10T15:59:00 jeff: we did that for a while ... it caused retargetting storms that were unmanagable 2011-02-10T15:59:26 that's what i was afraid of. 2011-02-10T15:59:28 retargetting on checkin, I mean 2011-02-10T15:59:50 I think he wants a "retarget on first checkin" 2011-02-10T15:59:57 tsbere: yes 2011-02-10T16:00:06 and retargeting after item creation is risky too with performance? 2011-02-10T16:00:29 retarget if status = In Process 2011-02-10T16:00:32 i was thinking perhaps "select oldest active pending hold on this title, retarget it", with an optional "repeat until end of hold list" if the first hold didn't target the copy for some reason. 2011-02-10T16:00:44 if you have a high volumn of cataloging, though, you're retargetting the same hold for each of the 50+ copies you add during a day, times (potentially) hundreds of catalogers 2011-02-10T16:01:02 * phasefx nods 2011-02-10T16:01:38 i think with careful thought it could still be done, but i need to give it some more thought and then run it by more eeevil eyes. 2011-02-10T16:01:39 now, one think I could see being possible would be an hourly (or so) A/T that groups by hold target 2011-02-10T16:02:29 *** granitize has quit IRC 2011-02-10T16:02:34 jeff: I agree there are ways to attack it -- at least make it a short delay instead of a full retargetting period 2011-02-10T16:02:41 s/think/thing 2011-02-10T16:06:28 *** kmlussier has quit IRC 2011-02-10T16:07:43 woohoo! I can confirm that the svf branch does not break search .... for basic stuff 2011-02-10T16:09:38 *** jenny has quit IRC 2011-02-10T16:13:04 Is there anywhere I can find documentation or examples of how to form osrfMessages for different tasks such as placing a hold on an item? 2011-02-10T16:13:30 jtbates: have you read dan scott's tutorials on opensrf? 2011-02-10T16:13:57 jtbates: and are you aware of docgen.xsl? 2011-02-10T16:14:24 I had perl module build failure (well, no Build.pl) just now 2011-02-10T16:15:20 *** jenny has joined #evergreen 2011-02-10T16:15:21 This? http://coffeecode.net/archives/180-Fetching-item-availability-from-Evergreen-using-the-OpenSRF-HTTP-gateway.html 2011-02-10T16:15:42 jeff: No, I don't know about docgen.xsl 2011-02-10T16:16:00 jtbates: docgen.xsl will use opensrf introspection to give you the api signature -- at least, for the perl-based services. 2011-02-10T16:16:51 jtbates: two articles here: http://journal.code4lib.org/?s=opensrf 2011-02-10T16:27:44 jeff: Thanks, I will read over those 2011-02-10T16:30:42 *** atheos has quit IRC 2011-02-10T16:30:55 so, MARC::File::XML seems to silently fail to create a record when faced with a record inside of instead of -- my xml namespaces are rusty -- Does this ring a bell, or sound like a "no, that should work" to anyone here? 2011-02-10T16:32:00 i have lots of records starting with jeff: MARC::File::XML itself handles namespaces fine; are you using MARC::Batch to slurp through the file, perchance? 2011-02-10T16:33:05 gmcharlt: i was at first, then also tried MARC::File::XML->in($filename) followed by ->next() on the resulting object 2011-02-10T16:33:08 *** gdunbar has quit IRC 2011-02-10T16:33:20 to the same effect 2011-02-10T16:33:26 hang on a minute 2011-02-10T16:33:38 same issue. the ->next() call returned undef 2011-02-10T16:34:15 replacing marc:record with record in the input file resulted in a record. 2011-02-10T16:34:25 yep 2011-02-10T16:34:31 patch coming in a 'mo 2011-02-10T16:34:34 oh yay. 2011-02-10T16:40:18 Not all records starting with gmcharlt: please let me know if example records would be useful to you. it sounds like you've already identified the issue. 2011-02-10T16:44:22 jeff: I have 2011-02-10T16:44:47 do all of the record elements in your file have the marc: namespace prefix? 2011-02-10T16:46:25 a little over 25,000 records in the database use a marc:record element (out of about 884K records). I first encountered this with records not exporting from OpenILS::WWW::Exporter via the /exporter interface. 2011-02-10T16:47:14 Then, I was trying to iterate over a collection of the marcxml from a single file (mix of record and marc:record in the file, then i was working with single-record-per-file, using the MARC::File::XML->in() method. 2011-02-10T16:48:58 of the of the 247576 bibs i asked to be exported, 67 failed to export. 18 of those use marc:record 2011-02-10T16:49:08 I can't seem to access docgen.xsl. The link I'm trying is: http://dev.gapines.org/opac/extras/docgen.xsl?service=open-ils.cat&all=on&offset=0&limit=25 2011-02-10T16:49:10 and the rest appear to have failed to export due to empty/invalid tags, missing subfields, etc. 2011-02-10T16:50:21 jtbates: do you have a dev/testing system available to you? 2011-02-10T16:50:35 where "system" == "evergreen installation"? 2011-02-10T16:51:04 jtbates: there are virtual machine images that you can download and run yourself -- having access to the logs can be an incredible help when trying to write a client. 2011-02-10T16:51:29 (it would also allow you to install your own copy of docgen.xsl that would match the version that you're testing against) 2011-02-10T16:51:31 jeff: No, I don't. Is docgen.xsl not publicly accessible? 2011-02-10T16:51:49 jtbates: it's not installed by default, and wasn't installed on the handful of public-facing dev systems i just checked. 2011-02-10T16:53:14 jeff: Ok, I'm downloading a VM image now 2011-02-10T16:53:49 gmcharlt: for what it's worth, the marcxml collection in a single file was created by outputting the bre marc fields in question to a single file and wrapping it in -- that may have been naive of me, with respect to namespaces 2011-02-10T16:54:40 Once I've got the VM running, how do I install docgen.xsl? 2011-02-10T16:57:03 Also, is there a Python implementation that uses OpenSRF-over-HTTP rather than XMPP? 2011-02-10T16:57:10 yes 2011-02-10T16:57:16 (i believe so) 2011-02-10T16:57:32 *** bshum has joined #evergreen 2011-02-10T16:58:05 jeff: Any idea where to look for it? 2011-02-10T16:58:56 jtbates: it should be included in the OpenSRF distribution. I don't know how stable/complete the python client's http transport is, but berick is likely the one to pester. He might also correct my earlier statement if I'm wrong. 2011-02-10T17:01:24 traditional gateway support is pretty well tested and has been used quite a bit 2011-02-10T17:01:36 osrf-http-translator support is more minimal at this point 2011-02-10T17:02:06 You can do quite a bit with the gateway. 2011-02-10T17:02:58 looks like only gateway is currently supported 2011-02-10T17:03:21 i think started stubbing out translator support, but never got around to finishing it 2011-02-10T17:04:35 Yes, I see the http_translator.py now 2011-02-10T17:06:16 that's actually proof-of-concept server-side apache module code 2011-02-10T17:06:22 that was eventually ported to C 2011-02-10T17:06:45 Ah 2011-02-10T17:23:34 *** jtbates has quit IRC 2011-02-10T17:25:00 jeff: have a look at http://marcpm.git.sourceforge.net/git/gitweb.cgi?p=marcpm/marcpm;a=commitdiff;h=d733685ff9b30ee1438e0aa7e3a95b654e83b71c 2011-02-10T17:28:00 *** Dyrcona has quit IRC 2011-02-10T17:38:59 jeff: assuming that tests out OK for you, I'll cut a release of MARC::File::XML 2011-02-10T17:46:24 *** jenny has left #evergreen 2011-02-10T17:51:21 *** dbs has joined #evergreen 2011-02-10T17:51:21 *** dbs has joined #evergreen 2011-02-10T18:00:31 *** sfortin has quit IRC 2011-02-10T18:24:37 ok ... with search and ingest working (nominally) on the svf branch, I'm calling it a night. 2011-02-10T18:59:26 gmcharlt: thanks! willtest and let you know. 2011-02-10T19:14:24 *** dbwells has quit IRC 2011-02-10T19:16:26 *** dbwells has joined #evergreen 2011-02-10T19:17:05 *** jamesrf has quit IRC 2011-02-10T19:26:02 *** rsinger has quit IRC 2011-02-10T20:00:42 *** rsinger has joined #evergreen 2011-02-10T20:01:15 *** dbs has quit IRC 2011-02-10T20:36:15 *** bshum has quit IRC 2011-02-10T22:25:10 *** sylvar has quit IRC 2011-02-10T22:26:14 *** sylvar has joined #evergreen 2011-02-10T22:30:21 *** tildeequals has quit IRC 2011-02-10T22:30:22 *** wjr has quit IRC 2011-02-10T23:01:49 *** tildeequals has joined #evergreen 2011-02-10T23:01:49 *** wjr has joined #evergreen 2011-02-10T23:21:50 *** jennam has quit IRC 2011-02-10T23:27:20 *** jennam has joined #evergreen 2011-02-10T23:27:47 *** jennam is now known as Guest26712 2011-02-10T23:33:21 *** tildeequals has quit IRC