Evergreen ILS Website

IRC log for #evergreen, 2014-02-04

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat

All times shown according to the server's local time.

Time Nick Message
00:24 dbs @later tell hbrennan If you can create a new directory / file in your custom template folder, you can override the default template
00:24 pinesol_green dbs: The operation succeeded.
00:37 PrabhjyotSodhi joined #evergreen
01:06 shadowspar joined #evergreen
01:30 BigRig_ joined #evergreen
01:32 gdunbar joined #evergreen
01:33 jeffdavi1 joined #evergreen
05:43 mtj_ joined #evergreen
07:06 b_bonner joined #evergreen
07:07 tfaile joined #evergreen
07:11 csharp @later tell hbrennan you should ask your IT dept for a testing server that you can play with too
07:11 pinesol_green csharp: The operation succeeded.
07:25 timlaptop joined #evergreen
07:40 kbeswick joined #evergreen
08:11 collum joined #evergreen
08:26 bshum joined #evergreen
08:29 Dyrcona joined #evergreen
08:39 Shae joined #evergreen
08:41 rjackson-isl joined #evergreen
08:48 StephenGWills joined #evergreen
08:50 csharp jeff: whenever you're around, I have a library who may be interested in adding a local image.  I remember at the Hack-a-Way that you demoed that and that it looked like it got into 2.5.X... do you have any docs for how to add an image?
08:50 Dyrcona joined #evergreen
08:51 jwoodard joined #evergreen
08:55 StephenGWills Morning Y'all.  I *KNOW* I need to upgrade from 2.3 but..  was it common to have Items not show up in the TPAC that would show up in the JSPAC?  I'm having trouble finding LIT on this problem.  I'm running 2.3.5
08:55 dbs csharp++
08:55 dbs StephenGWills: yep, that sounds familiar
08:55 bshum 901c?
08:55 dbs bshum++
08:56 dbs also the inverse; items would show up in tpac that weren't in jspac, due to incomplete checking of various deleted properties
08:57 _bott_ I have 479 copies with action.transit_copy rows, with null dest_receive_time, and current status of something other than "In Transit"    ...good times
08:58 mmorgan1 joined #evergreen
08:58 mmorgan1 left #evergreen
08:59 _bott_ Seems like staff error, combined with floating='t' provides the ability to never finish a transit, and allow a copy to lead a normal productive life
09:00 mmorgan joined #evergreen
09:03 csharp StephenGWills: you have have to reingest all the bibs - at minimum reingest ones with no 901c
09:03 csharp (then vacuum analyze metabib.* tables)
09:04 ericar joined #evergreen
09:04 StephenGWills sharp  I remembered that there is that reingester but on looking for it the other day didn't find it.  I'm just picking this problem back up.  is reinvest a support_script?
09:07 csharp StephenGWills: one sec... there are reingest improvements in 2.4+, so I don't want to give you the wrong info
09:09 jeff csharp: sure, if you want a custom local image for record 12345:
09:09 csharp StephenGWills: http://git.evergreen-ils.org/?p=Evergreen.git;a=b​lob;f=Open-ILS/src/sql/Pg/version-upgrade/2.2-2.3​.0-upgrade-db.sql;h=a6cfd9a93ba2e32b57805b9d593bf​037d717b8d0;hb=refs/heads/tags/rel_2_3_5#l2238
09:09 rfrasur joined #evergreen
09:09 jeff create three dirs /openils/var/web/opac/extras/ac​/jacket/{small|medium|large}/r
09:09 csharp StephenGWills: the lines beginning with \qecho will let you know what you need
09:10 StephenGWills kk looking
09:10 jeff place an .htaccess file in each dir (you might be able to do this at a common ancestor directory -- i've not tried) containing: "ForceType image/jpeg"
09:11 jeff place a jpeg file named "12345" in the large/r/ medium/r/ and small/r/ directories
09:11 jeff if you have multiple app servers without /openils/var/web/opac/extras/ac/jacket being shared storage between them all, you'll need to copy the images to each app server.
09:11 jeff then, test.
09:12 csharp so the ../r directory contains any custom jpegs that are named by the bib id number, correct?
09:12 jeff right
09:12 StephenGWills the bib in question does not have a 901$c, I'm feeling hopeful :)  thanks Chris
09:12 csharp StephenGWills: happy to help!
09:13 * csharp has been through that battle before ;-)
09:13 csharp jeff: okay - cool - I'll experiment
09:14 jeff csharp: let me know how i goes. in describing it to you just now i think i realized at least one of those things (the ForceType trick) isn't described in documentation yet.
09:14 jeff csharp: if you have a strong need to do both jpg and png and other formats, i'm sure it can be done with mod_rewrite and/or content negotiation. not sure which would be simpler.
09:14 jeff csharp: in our case, we opted for "just jpeg"
09:15 jeff we currently have 22,419 local jacket images.
09:15 bshum jeff and his puppets
09:16 jeff some of them could be deemed superfluous by additional code to search for more URLs in the MARC for jacket images.
09:16 jeff or "use this url with this identifier from the MARC for bibs from this source"
09:16 jeff but we don't have that yet (though there's a wishlist bug for the former)
09:16 jeff (or something like the former)
09:18 csharp wow
09:18 csharp well, we're discussing sustainability at this point - we're pretty sure we can't handle an influx of local image requests at this point
09:18 bshum jeff: I was thinking about custom images the other day as I looked at some random 856 lines.
09:18 csharp but I still want to test it (and possibly script it)
09:19 jeff yeah, we have an interface that staff use to upload jacket images.
09:19 * csharp imagines a staff upload interface that would act like an "incoming" directory
09:19 csharp heh
09:19 csharp well if we have that, I'm game ;-)
09:19 jeff any staff can create a jacket image for any record.
09:19 jeff also, there are leaderboards.
09:20 csharp leaderboards?
09:20 bshum leaderboards++
09:21 * bradl imagines uploading all grumpy cat
09:23 jeff csharp: "all time" and "past 30 days" leaderboards listing staff name ordered by how many jackets they've uploaded.
10:54 serflog joined #evergreen
10:54 Topic for #evergreen is now Welcome to the #evergreen library system channel! | We are publicly logged. | Large pastes at http://paste.evergreen-ils.org
11:04 senator what's a right size for a marc record?
11:05 Dyrcona 0 bytes?
11:05 rfrasur (whatever ebsco wants because they rule the world and we all must bow)
11:05 Dyrcona Seriously, they say that many of our records don't match the size in the LDR, and they sent me some output from their software showing a few records being off by one.
11:06 Dyrcona Of course, they haven't told me which records.
11:06 Dyrcona I didn't think any sane software actually paid attention to the size field.
11:06 senator i had totally forgotten it existed
11:07 Dyrcona They must be using MARC4J.
11:07 Dyrcona It is apparently thousands of our records.
11:07 * berick wonders if it's a unicode bytes vs characters thing
11:08 Dyrcona that's one of my suspicions.
11:09 jbfink joined #evergreen
11:11 jcamins senator: definitely zero bytes.
11:11 jcamins You know, that was so much cleverer when I typed it 7 minutes ago.
11:11 senator heh
11:11 Dyrcona They sent me one example, actually: "ERROR - Record size not valid: Record could not be fixed -- Leader specified length (1981) but rather was (1980)) Errant data (01981cjm  2200433Ia"
11:14 jeff how did you send the records to them?
11:15 jeff in terms of, i'm assuming they're MARC and not MARCXML...
11:15 jeff how were they generated on your end?
11:19 Dyrcona Using my Marque.pm branch and the usual rain dance with MARC::Record.
11:20 Dyrcona Yes, they are MARC, not MARCXML.
11:20 Dyrcona EBSCO prefers MARC, not MARCXML.
11:22 eeevil just a guess, but perhaps m::r is counting the record separator as part of the preceding record, but marc4j is not
11:23 eeevil (also, the marc specs are ambiguous WRT characters vs bytes, and use the terms interchangeably throughout)
11:24 Dyrcona I'm wildly speculating on the marc4j only because I had trouble with it during our migration. It expected the size fields from Horizon to be correct and they were off by miles.
11:25 Dyrcona I don an insert_grouped_field when adding holdings. Should I call the method to recalculate the size after that? I *thought* it was handled automatically.
11:26 Dyrcona s/don/do/
11:28 Dyrcona I'll try that with this Friday's dump. I don't think it will make a difference, but why not?
11:30 Dyrcona Oh. I actually have to know the length to use that method.
11:48 Dyrcona Ah well, another email in the chain with EBSCO....
12:01 akilsdonk joined #evergreen
12:02 afterl joined #evergreen
12:08 jihpringle joined #evergreen
12:22 jboyer-isl jeff, possibly Dyrcona: Some time in the last week or so I was in here asking about an issue that only seemed to effect the conify interfaces and there was some curiosity about how it turned out. What happened was that the wrong hostname was set for OSRFTranslatorCacheServer.
12:23 jeff Ah. Sorry, should have suggested that first. :-)
12:23 jeff I think the two causes have similar symptoms, but I went with the more subtle possible cause. Sorry. :-)
12:24 jboyer-isl No problem, it's humming along now. (And just in time! upgrade is in 11 days. D: )
12:27 jeff Basic postgres question... I can use the output of a function returning SETOF (in this case, actor.address_alert_matches) using EXISTS (SELECT 1 FROM actor.address_alert_matches([arguments]))... but given a query that returns a number of addresses with an alert, what's a good way to include the actual alert, the data returned by the function?
12:27 jeff Right now I'm discarding it.
12:27 jeff (with SELECT 1 FROM in the EXISTS clause)
12:39 tsbere jeff: Use it as part of a join?
12:41 tsbere SELECT whatever FROM whatever JOIN actor.address_alert_matches(blah) aaam...
12:44 snowball joined #evergreen
12:45 jeff not sure i can do that without an ON
12:45 dbs_ joined #evergreen
12:46 adbowling-isl_ joined #evergreen
12:46 dconnor joined #evergreen
12:46 tsbere jeff: "ON true"?
12:47 jeff heh. it was worth a shot:
12:47 jeff ERROR:  invalid reference to FROM-clause entry for table "au"
12:47 jeff LINE 18: JOIN actor.address_alert_matches(au.home_ou, aaua.street1, a...
12:47 jeff ^
12:47 jeff HINT:  There is an entry for table "au", but it cannot be referenced from this part of the query.
12:48 tsbere jeff: Ahh, you want to use other parts of your query. That complicates matters.
12:48 jeff i'll gist the query as i have it thus far. it might require a completely different approach, or i might just want to punt to creating a new function.
12:48 * tsbere probably needs to know the actual goal at this point
12:49 jeff tsbere: https://gist.github.com/jeff/2e682cecdbf2f7b8af11
12:50 jeff i am reporting on patrons with an address alert
12:50 jeff ideally, i'd like to have the text of the address alert also show.
12:50 jeff (be included in the output of the query)
12:50 jeff i can probably just unroll the actor.address_alert_matches function into my query.
12:51 jeff but the "can i do this with a function returning SETOF in this circumstance" seemed useful to determine an answer to, for future.
12:52 tsbere jeff: Well, I can give you "the first alert in the list" easily. "All alerts in the list" is harder.
12:52 jeff i'll take that for starters.
12:53 * tsbere could be wrong on the "all" front now that he ran a test
12:54 tsbere jeff: Add this to the end of your SELECT DISTINCT line: , (actor.address_alert_matches(au.home_ou, aaua.street1, aaua.street2, aaua.city, aaua.county, aaua.state, aaua.country, aaua.post_code, aaua.mailing, aaua.billing))).alert_message
12:54 jeff ah, see i was trying to think of how to do it with a subselect -- suddenly, easier!
12:56 tsbere I might have an extra ) in there
12:56 tsbere <_<
12:56 jeff you did, i'm pretty sure.
12:56 * tsbere can't actually test this due to MVLC not *having* any such alerts defined right now
12:57 jeff we recently started using them, and i'm trying to make them do double duty for reporting as well as the alerts in the user editor.
12:58 jeff currently takes over a minute to run, but that's reasonable for this report.
13:01 jeffdavis joined #evergreen
13:01 tsbere jeff: Still, that trick is generally useful whenever a "record" is returned from a function, SETOF or no. The extra () around the function (or variable, in some cases) is needed though. <_<
13:08 csharp okay - I'm interested in the practical difference between 'next' and 'return' in perl, in light of commit 174c7db5bb5be07b19d564412fc5651405ba5b1c
13:08 pinesol_green [evergreen|Dan Scott] return, not next, from eval BLOCK - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=174c7db>
13:08 csharp and then commit f040814c7507291c388a35a23c8878293a2524e4
13:08 pinesol_green [evergreen|Liam Whalen] Fixes error with generate_fines for overdues returned after closed date - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=f040814>
13:10 csharp (following up on the fine generator issue we found around our snOMG experience last week)
13:10 Dyrcona csharp: return causes the current code block to stop executing, be that a function or an eval.
13:10 tsbere csharp: eval blocks get processed more as functions, I believe, so return exits that function, compared to next which would be trying to continue the outer loop. I think.
13:11 Dyrcona tsbere: We haven't had any complaints of this, have we?
13:11 tsbere I don't recall any. Doesn't mean it hasn't happened.
13:12 csharp in our case, it appears to be going back and charging patrons for library closed dates
13:12 csharp looks like normal hours of operation are okay
13:13 csharp so if we *want* it not to continue evaluating the outer loop, we want return, right?
13:14 csharp next means 'okay next step'?
13:14 jeff one unfortunate aspect of an IDE for reports generation... when your query takes >1m to run, it takes >1m for the IDE to fetch your field names and data types. :-)
13:16 Dyrcona csharp: I looks like the next could be doing what you describe, but I have no empirical evidence to back that up.
13:16 ktomita joined #evergreen
13:16 Dyrcona s/I/It/
13:16 csharp Dyrcona: thanks
13:16 tsbere csharp: I don't think there is a way for the eval block to terminate the outer loop directly, somewhat by design
13:17 csharp our end result is that we don't want fines to be charged for days that have been marked as closed dates, even after the fact
13:18 csharp looks like closed dates are acting more like grace periods with the current behavior
13:19 Dyrcona next breaks from the loop at that point and starts again at the top with the next iteration.
13:21 dbwells csharp: I am pretty sure the code is meant to do what you want it to do.  If I understand your situation correctly, for some reason, actor::org_unit::closed_date->search_where() isn't finding your closed dates.
13:22 csharp dbwells: I can dig for this, but can you see a simple-ish way to test that hypothesis?
13:22 Dyrcona I was about to say something along the lines of what dbwells just said.
13:22 Dyrcona I needed more context. :)
13:23 dbwells csharp: That function basically is a straight DB lookup (I think), so I'd start by looking in the DB and see if everything is in line there.
13:24 Dyrcona csharp: Are you expecting it go by the copy's circ_lib closed day or the circulation's circ_lib closed?
13:24 csharp Dyrcona: circulation circ_lib
13:24 Dyrcona ok. that is what it looks like it should be doing.
13:26 csharp qq
13:28 dbwells csharp: I'd grab one of the bogus bills, figure out the timestamp and circ_lib for that, then do a select from the actor.org_unit_closed table and make certain you are finding what you want.  Who knows, maybe it is something simple like a wrong year, or a closed date not propogated into all the branches.
13:28 csharp dbwells: thanks
13:52 csharp okay - mystery solved - it's not a bug ;-)
13:53 csharp we run our fine generator at 12:30 a.m. and the closing was set later that day
13:53 csharp so fines had already been generated
13:53 * csharp bangs head on desk
13:56 csharp dbwells++ Dyrcona++ tsbere++ # thanks as always for your help
13:59 jeff so we're back to the original "need a new feature for that" conversation of yesterday. :-)
13:59 csharp yeap
13:59 * csharp presses the "forgive all fines for closed day" button and calls it a day
13:59 jeff and i think it could be a useful new feature. i wouldn't want to undertake it until the new-style voids are in, though.
14:00 jeff csharp: wait, you HAVE one of those buttons? :-)
14:00 jeff actor.usr_address.address_type being completely freeform yet having a NOT NULL requirement sure does lead to some interesting things.
14:00 * tsbere wonders if we should all move the fine generator to run *before* midnight so that libraries have the day to add the closed date for snow days and such
14:02 bshum Sigh
14:02 bshum Snow days make me sad.
14:02 csharp tsbere: right - we were just having that conversation here
14:03 tsbere csharp: In your conversation you should note that checkin time fine generation will catch "today" if they are open anyway
14:03 csharp our conversation was very much academic
14:04 csharp I actually think jeff's feature idea is probably better
14:04 csharp snow--
14:04 csharp man I used to love snow days - now they're just a PITA
14:04 tsbere It has been a long time since I loved snow days.
14:05 bshum I think I've been pestering jeff about that "feature" for a few years now.  Every winter, probably.  :)
14:05 tsbere Basically since I was expected to shovel said snow. <_<
14:24 * gmcharlt adds a "snowday" tag to lp1276259
14:24 csharp snow-OMG!2014
14:26 gmcharlt tsbere: the pushme to your pullyou is libraries who want to display accumulated fines on their notices, after the fine generator has run
14:27 gmcharlt though that may be an argument for scheduling it to run at more like 7 a.m., at least for small enough systems where you don't need all of the early morning to process notices
14:31 pinesol_green` joined #evergreen
14:32 Dyrcona Sure, send someone binary MARC records in a docx, why the Hell not!?
14:32 bshum Are we putting that in Marque.pm?  Cause if so, then I think we may have gone too far now.... ;)
14:32 Dyrcona EBSCO sent me this.
14:33 Dyrcona Or, rather sent it to a library director who sent it to someone else who finally sent it to me today.
14:34 bshum Sounds like the usual chain of screaming to me.
14:35 dbs bshum: Marque.pm.docx?
14:36 csharp docquex
14:36 Dyrcona csharp++
14:36 senator \Dyrcona: at least it wasn't printed and faxed somewhere along the way
14:37 Dyrcona I think I've seen that or very close, a printed an faxed screen shot of the MARC editor came my way once just after migration.
14:37 * csharp often gets "screenshots" taken with a smarthphone camera
14:37 csharp smartphone, even
14:37 Dyrcona bshum: screamshots. :)
14:38 Dyrcona Well, I can figure out the record IDs, at least.
14:38 Dyrcona Of course, the records have probably changed in the past month.
14:39 Dyrcona I like this error message though: Record could not be fixed -- Leader specified length (2545) but rather was (2544))
14:39 kbutler joined #evergreen
14:39 Dyrcona So, you figure out what the length *should* be, but you can't fix it?
14:39 dbs Dyrcona++ # screamshots
14:40 csharp @quote add < Dyrcona> Sure, send someone binary MARC records in a docx, why the Hell not!?
14:40 pinesol_green` csharp: The operation succeeded.  Quote #77 added.
14:41 csharp @quote random
14:41 pinesol_green` csharp: Quote #37: "bkuhn: ... my wife is watching the Les Miserables anniversary concert... and it was just at that point where that Jonas brother comes on as Marius and he's really bad. :)" (added by Dyrcona at 10:16 PM, December 01, 2012)
14:41 csharp bkuhn++
14:43 jeff landscape formatted report, approx 32 patrons per page, we have 29 pages worth of patrons having an address matching an address alert.
14:44 jeff berick++ for using regular expressions for address alerts.
14:44 tsbere jeff: With an unknown number of duplicate patrons due to matching multiple alerts?
14:45 jeff tsbere: correct, though i already eliminated a large number of those that were due to an alert overlap. :-)
14:47 jeff i have 16 active address alerts configured for 16 addresses -- not doing anything clever like trying to match multiple addresses with one alert.
14:48 jeff so, using regex only to make the matching a little more fuzzy-yet-not-too-fuzzy, we have 175 distinct strings matching those 16 alerts. :-)
14:49 Dyrcona berick wins the prize! The problem I am having with EBSCO appears to be multibyte characters. They're counting 1 character, we're counting 2 or more bytes.
14:49 Dyrcona jeff: Now, you've got two problems. :)
14:50 jeff 175 goes down to 126 when case is no longer a consideration.
14:51 berick yay, i guess
14:52 jeff i found these to be most helpful: "^123 +(W.*)? *(Eleventh|11th)" and "^456 +(E.*)? *(Eighth|8th)"
14:53 jeff match 123 W Eleventh St and 123 West Eleventh Street and 123 11th but not 123 W 11th Street, etc.
14:58 Dyrcona I wonder what they'll do if I send them UTF-8, instead of trying to convert to MARC-8.
15:08 ktomita_ joined #evergreen
15:17 AaronZ-PLS Hmmm, apparently my earlier message didnt go through.
15:17 AaronZ-PLS Is there a permission to control if a user (or group of users) can perform pre-cataloged checkouts?
15:18 csharp AaronZ-PLS: I think it's just CREATE_COPY
15:18 * csharp may be wrong
15:22 Dyrcona Cannot decode string with wide characters at /usr/local/lib/perl/5.14.2/Encode.pm line 176.
15:22 Dyrcona Does that mean it is already decoded?
15:23 AaronZ-PLS csharp: I dont think that is the case as we have circ users who can create pre-cat copies, but they dont have the CREATE_COPY permission
15:23 Dyrcona AaronZ-PLS: Actually, I was looking at that just the other day, lemme check the code again.
15:23 AaronZ-PLS Dyrcona: Thanks
15:25 bradl_ joined #evergreen
15:27 Dyrcona There are no explicit permission checks that I can find for creating a copy with a precat checkout.
15:29 finnx_ joined #evergreen
15:30 Dyrcona CREATE_NONCAT_TYPE && DELETE_NONCAT_TYPE appear to be the permissions involved.
15:30 Dyrcona When I say I was looking at that, I really mean I was looking at circulation in general.
15:34 AaronZ-PLS Dyrcona: Thanks. I think that those are for non-catalouged circulation types. What I am lookign at is making it so that when they misscan a barcode or type it in wrong and it comes up saying "Mis-scan or non-cataloged item. Checkout as a pre-cataloged item?" they do not have the option to continue
15:35 AaronZ-PLS (ie: the screen seen in page 2 of: http://www.biblio.org/documentation/wp-con​tent/uploads/2011/04/05Circ234CheckOut.pdf )
15:36 Dyrcona Oh, that's a different question from what I was answering. Just remove those permissions and it should fail.
15:38 AaronZ-PLS Dyrcona:The "circ" users do not have the "CREATE_NONCAT_TYPE", "DELETE_NONCAT_TYPE" or "CREATE_COPY" permissions. Is there another permission that I need to remove?
15:39 Dyrcona I didn't see anything other than the permission that allows them to circulate in the first place.
15:39 Dyrcona Are you just trying to make the window go away?
15:41 AaronZ-PLS Various libraries are having problems with staff checking out items as pre-cat items and then they have a pre-cat item with a barcode such as "0000000029849284u892309" sitting on their "bad stat cat report"
15:42 Dyrcona Yes, what do you want to do.
15:42 AaronZ-PLS We would like to make it so that the circ users cant "accidently" create pre-cat items
15:42 csharp AaronZ-PLS: not a whole lot to do about that other than deleting the errant copy
15:43 * tsbere keeps thinking about, but not doing anything about, adding a "pre-cat barcode regex"
15:44 tsbere Theory being "if the barcode doesn't match said regex don't let them make it a pre-cat" to help distinguish "not in the DB" from "horrible mis-scan"
15:44 csharp tsbere: that seems like the only solution software-wise
15:44 tsbere That and maybe a permission for creating the things. >_>
15:44 csharp we treat it as a training issue and/or inevitable annoyance ;-)
15:44 AaronZ-PLS csharp: The problem being that (in the cases with barcodes that have leading zeros) the libraries open up the report in Excel, it deletes the leading zeros and they cant access the item
15:45 Dyrcona The copies are created without a permission check.
15:45 AaronZ-PLS Dyrcona: So it would appear.
15:45 Dyrcona I just assumed that everything in NonCat.pm was involved in circulation, but I guess it isn't.
15:46 Dyrcona Looks like creating noncat types is orthogonal to circulating noncat copies.
15:47 Dyrcona Only thing you can do right now is remove their circ privileges, and I'm sure that is out of the question.
15:47 Dyrcona What you want requires code modifications.
15:47 gmcharlt I've looked at it as well, and I agree with Dyrcona
15:48 AaronZ-PLS Dyrcona: Removing their circ previliges would be slightly counterproductive :D. Should I create a bug report/wishlist for this?
15:48 kbutler joined #evergreen
15:48 gmcharlt yeah, it's a wishlist item
15:48 * Dyrcona agrees with gmcharlt.
15:48 gmcharlt currently, the bias is towards getting the circulation done and the patron out the door
15:49 csharp patron_out_the_door++
15:49 Dyrcona We had a copy with barcode of '3' after migration that changed libraries on a regular basis.
15:50 AaronZ-PLS gmcharlt: I understand, the issue we are running into is misscans of items in the catalog which then cost much more staff time to fix than forcing the checkout staff member to re-scan the item and also mess with circulation stat cats
15:51 gmcharlt AaronZ-PLS: yeah, I'm not disagreeing with you, just expressing the current state of affairs
15:51 AaronZ-PLS Is there a section of the code that I should point to in the wishlist/bug report?
15:52 gmcharlt I suggest keeping it broad -- several different ways to address it have been suggested during this conversation, and the bug report probably shouldn't over-determine it
15:52 Dyrcona Circ/Circulate.pm and Circ/Noncat.pm is where I'd make changes.
15:53 Dyrcona I'd be inclined to add a NONCAT_CIRC.override permission and handle it that way.
15:54 gmcharlt there would also necessarily be some XUL changes
15:54 gmcharlt and ITEM_NOT_CATALOGED would be what one would grep for regarding pre-cats, as opposed to non-cats
15:54 Dyrcona Yes, most likely.
15:55 AaronZ-PLS Dyrcona:I agree. It would be useful to allow certian users to create pre-cat circs but not require that they have the same permissions as a cataloger.
15:55 Dyrcona gmcharlt: I want to change the subject, and since you're responsible for MARC::Charset, you hopefully know the answer.
15:55 * gmcharlt does have a preference, unless a patch is accompanying it right away, that bug reports list problems, maybe suggest solutions, but don't try to overdetermine the implementation of said solutions
15:55 gmcharlt Dyrcona: I'm all ears
15:55 * Dyrcona agrees that bug reports should detail problems and not define solutions.
15:56 Dyrcona gmcharlt: I'm trying to count characters, not bytes in a multi-byte string in Perl.
15:56 Dyrcona I've googled and tried the suggestion of using decode(), but that doesn't work.
15:56 csharp AaronZ-PLS: also, the "tabular output" in the reports display should show leading zeroes, jsyk
15:57 Dyrcona In fact, it spits out what I pasted above about not being able to decode a string with a multi-byte character.
15:57 Dyrcona If I encode, I get the same result as from just using length() alone, so that can't be right.
15:57 AaronZ-PLS csharp: We use a custom reporting interface which saves the data as a .xls file for the staff to use
15:58 gmcharlt Dyrcona: well, decode() or decode_utf8() is in fact what needs to happen to get a UTF8 string, to make CORE::length count characters and not octets
15:58 gmcharlt the question would be what the characer encoding of the octets are at the point where you're doing the decode
15:59 Dyrcona It's a MARC record from our database, they should already be UTF-8.
15:59 gmcharlt AaronZ-PLS: is the reporting interface capable of specifying that all selected fields should be text, not general?
15:59 gmcharlt (when it outputs the spreadsheet)
15:59 Dyrcona And the character in question is a musical flat symbol, which looks good on my UTF-8 terminal.
16:00 gmcharlt Dyrcona: does 'use Encode; $str = decode_utf8($your_octets);' work any better for you?
16:00 hbrennan joined #evergreen
16:02 Dyrcona It does, and I added a print to stderr that shows I'm not using set_leader_lengths correctly.
16:03 Dyrcona Or, does as_usmarc call set_leader_lengths again?
16:03 Dyrcona I'm trying to mangle the size of my MARC record in case you couldn't tell.
16:05 gmcharlt as_usmarc does in fact call st_leader_lengths
16:05 gmcharlt and if it's working correctly, should always be setting the leader and directory fields to be based on octets, not characters
16:06 Dyrcona gmcharlt: It is, but ebsco apparently counts characters, not octets.
16:07 gmcharlt bless their hearts
16:07 Dyrcona gmcharlt: When I get the character count, I get the same size that they tell me the record *should* be.
16:07 gmcharlt would they prefer to accept MARC8 instead?
16:08 Dyrcona gmcharlt: They would, but then they tell me the record size is wrong also, and I have a hunch I might be doing the conversion incorrectly.
16:10 Dyrcona so, UTF8 sets the leader size to 02546, MARC8 to 02545, and length() on both tells me 02544.
16:11 Dyrcona Leader specified length (2545) but rather was (2544)) is what EBSCO had to say about the record that I sent.
16:11 gmcharlt could you past the original MARCXML from your database?
16:13 Dyrcona I can, but I'm seeing consistent results with the other records.
16:14 AaronZ-PLS gmcharlt: I'll have to check if it can be set to text.
16:14 Dyrcona The length() of the record in characters matches what EBSCO says it should be, and the header on the MARC8 record matches what they said we said it was.
16:14 AaronZ-PLS I created the bug report: https://bugs.launchpad.net/evergreen/+bug/1276328
16:14 pinesol_green` Launchpad bug 1276328 in Evergreen "Add a way to block "pre-cat" item creation based on permissions or settings" (affected: 1, heat: 6) [Undecided,New]
16:15 * bshum wonders if it's time to change pinesol_green back to just simple ol' pinesol.
16:15 * Dyrcona would love to email EBSCO and say, you're counting the record length incorrectly, but I doubt that they would listen.
16:20 pastebot "Dyrcona" at 64.57.241.14 pasted "MARCXML for EBSCO" (199 lines) at http://paste.evergreen-ils.org/11
16:21 Dyrcona gmcharlt: That is not straight out of our database because I am also exporting item info in 852 tags.
16:29 dbwells Dyrcona: probably just a past problem, but your example record has an extra </datafield> after the 852
16:29 dbwells s/past/paste/
16:29 * Dyrcona doesn't care enough to bother with that.
16:29 Dyrcona Straight out of the MARC documentation: Computer-generated, five-character number equal to the length of the entire record, including itself and the record terminator. The number is right justified and unused positions contain zeros.
16:30 Dyrcona So, MARC doesn't say it it is octets or characters.
16:30 Dyrcona "length. A measure of the size of a data element, field, or record and is expressed in number of octets."
16:31 Dyrcona But they do, and EBSCO is wrong!
16:32 Dyrcona I am now done with this conversation, and I'm making no changes for EBSCO.
16:32 Dyrcona Thank you all for putting up with me.
16:40 eeevil Dyrcona++
16:41 eeevil Dyrcona: for your next trick, convince B&T a library might want to use all 35 bytes available to name itself in EIDFACT, not just 5. ;)
16:42 eeevil see also: item identifiers
16:45 roses joined #evergreen
16:46 roses Several months ago (almost a year) I was working on setting up a test Evergreen server and someone here told me about where to find some fake/test records to import into my catalog?  Where can I find something like that?
16:50 bshum roses: If you're just setting up some sort of demo system, you can always use the stock concerto dataset that comes with Evergreen 2.4+
16:50 roses bshum:  I installed 2.3.7
16:51 bshum That step gets described a bit in step 9.1 of the 2.5 README
16:51 bshum http://evergreen-ils.org/documentation/ins​tall/README_2_5.html#_loading_sample_data
16:51 bshum roses: Ah, well, 2.3 doens't have the concerto options.
16:51 bshum Also, 2.3 is dead to me.
16:52 roses bshum: crud - yep I will be updating it soon to 2.5 - in fact it might be sooner!
16:52 bshum roses: Yeah I think official support of 2.3 ended a couple months ago.
16:53 roses bshum: thanks - will be updating tomorrow
17:07 stevenyvr joined #evergreen
17:15 mmorgan left #evergreen
17:27 phasefx ooh, Civilization humble bundle
17:46 smyers_ joined #evergreen
18:12 sseng_ joined #evergreen
18:15 hbrennan We're planning an upgrade from 2.3.4 to 2.5... Can someone confirm that we need to first upgrade to 2.4.6 before upgrading to 2.5.2?
18:32 gmcharlt hbrennan: no, it's not necessary to pause
18:32 gmcharlt during the process, the database would in some sense be at 2.4.x, but only for a moment
18:32 hbrennan Oh, okay. I read something about not skipping versions, but it was old stuff
18:32 hbrennan wonderful
18:33 hbrennan so 2.3.4 to 2.5.2 is no problem
18:33 gmcharlt yeah, that was about the DB schema updates that are part of the process, where you do have to make sure that you don't skip intermediate versions (although I'm simplifying a bit)
18:33 gmcharlt the software itself would get straight from 2.3.4 to 2.5.x
18:34 hbrennan http://docs.evergreen-ils.org/2.5/_​upgrading_the_evergreen_server.html
18:34 hbrennan Are these instructions still good then?
18:35 gmcharlt mostly; http://docs.evergreen-ils.org/2.5/_upgr​ade_the_evergreen_database_schema.html would be a bit different to include some bits referenced by the 2.4 upgrade instructions
18:36 hbrennan good to know
18:36 hbrennan We're all nervous about it, but if we wait until we're not nervous... we'll never upgrade
18:36 hbrennan I'm trying to convince everyone to just jump with me :)
18:37 gmcharlt there's no kersplat at the end, just new functionality ;)
18:38 hbrennan And TONS of it
18:38 hbrennan our catalogers will be so happy to have the 2.4 stuff, and I'm excited about some of the bits from 2.5
18:40 dluch joined #evergreen
18:41 hbrennan Someone has to be on board who knows next to nothing about databases, git, etc
18:41 hbrennan That's me
18:42 hbrennan so so so excited about Day 1 of the conference!
18:43 sseng joined #evergreen
18:46 dcook joined #evergreen
19:33 sseng joined #evergreen
19:42 jbfink joined #evergreen
19:50 artunit joined #evergreen
20:08 jbfink joined #evergreen
20:28 jbfink joined #evergreen
20:41 jbfink joined #evergreen
21:00 pmurray_away joined #evergreen
22:21 jbfink joined #evergreen
23:07 hbrennan joined #evergreen
23:34 jbfink joined #evergreen

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat