Evergreen ILS Website

IRC log for #evergreen, 2017-12-04

| 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
00:29 dbwells_ joined #evergreen
06:32 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>
06:57 JBoyer joined #evergreen
07:18 rjackson_isl joined #evergreen
07:30 agoben joined #evergreen
07:49 rlefaive joined #evergreen
08:16 collum joined #evergreen
08:38 tlittle joined #evergreen
08:45 mmorgan joined #evergreen
08:56 bos20k joined #evergreen
09:09 jvwoolf joined #evergreen
09:10 kmlussier joined #evergreen
09:14 jvwoolf1 joined #evergreen
09:25 mmorgan Has anyone successfully created a user activity type for OverDrive?
09:26 yboston joined #evergreen
09:28 mmorgan Documentation suggests adding a row to config.usr_activity_type with correct values for event caller, event type and event mechanism should do it, but I haven't had success.
09:37 mmorgan left #evergreen
09:39 Stompro joined #evergreen
09:42 jeff @later tell mmorgan it depends on how you are handling authentication between OverDrive and the library. What method are you using? EZproxy, HTTP call, SIP2, or other?
09:42 pinesol_green jeff: The operation succeeded.
09:44 mmorgan joined #evergreen
09:44 rlefaive joined #evergreen
09:47 mmorgan jeff: we're using SIP2
09:48 JBoyer joined #evergreen
09:48 jeff okay. i believe there's a variable you need to set on the sip user or institution in order to pass the activity type.
09:49 jeff activity_who attribute on the SIP user in the SIPServer config file.
09:52 mmorgan Ok, will check the SIPServer config file. Is just the ewho enough?
09:53 jeff one sec.
09:58 jeff activity_who should match the ewho value of an entry in config.usr_activity_type whose ewhat is verify and whose ehow is sip2.
10:00 * jeff looks to see how that differs if you're validating passwords via SIP2
10:00 jeff are you requiring users to share their library password with OverDrive?
10:00 mmorgan Ok, does anything need to be restarted?
10:01 mmorgan No, users don't need a password
10:01 jeff ah, that simplifies things.
10:02 jeff you'll need SIPServer to notice the config change. In most cases, that means a restart of SIPServer on the host that OverDrive communicates with.
10:02 mmorgan ok, thanks, will give it a try. running to a mtg right now
10:02 mmorgan jeff++
10:03 mmorgan1 joined #evergreen
10:24 jeff bug 1736173
10:24 pinesol_green Launchpad bug 1736173 in Evergreen "Document activity_who and related options for SIP user activity tracking" [Low,New] https://launchpad.net/bugs/1736173 - Assigned to Jeff Godin (jgodin)
10:42 jeff oh weird. that's not the summary as i had edited it.
10:42 Christineb joined #evergreen
10:42 jeff i'm guessing i edited the search field instead. ah well.
11:12 mmorgan joined #evergreen
11:48 khuckins joined #evergreen
11:51 _adb joined #evergreen
12:47 csharp aebf4a5
12:47 pinesol_green csharp: [evergreen|Bill Erickson] User Activity : SIP activity tracking - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=aebf4a5>
12:48 jihpringle joined #evergreen
12:58 khuckins_ joined #evergreen
13:04 rlefaive joined #evergreen
13:16 rlefaive joined #evergreen
13:32 sandbergja joined #evergreen
13:34 khuckins__ joined #evergreen
13:35 rlefaive joined #evergreen
13:38 sandbergja Hi all!  Does Evergreen only accept EDI INVOICs that have SANs in the format NAD+BY+[SAN]::91?  Our vendor sends them with ::92 at the end, and I want to make sure my understanding is correct before reporting a bug / asking the vendor to use 91 instead.
13:39 sandbergja (I was looking at line 25 of this, btw: https://github.com/evergreen-library-system​/Evergreen/blob/tags/rel_2_12_2/Open-ILS/sr​c/perlmods/lib/OpenILS/Utils/EDIReader.pm
13:45 berick sandbergja: i'm not aware of any.  what vendor, if you don't mind me asking.
13:45 berick for reference, 91 = Assigned by supplier or supplier’s agent, 92 = Assigned by buyer
13:45 sandbergja berick: It's Proquest Coutts
13:46 berick sandbergja: ok.  a vendor I have not encountered.  they all do /something/ different :)
13:46 sandbergja hahaha of course! :-)
13:47 berick an LP bug seems like a good idea.
13:48 sandbergja Okay!  I can go ahead and type that up.  Really, what would need to happen is to change the regex to be a bit more permissive, right?
13:49 sandbergja Actually, now that I look closer, EG is looking for SANs in something that ends in ::31B, not ::91
13:49 sandbergja or am I mis-reading that?
13:50 berick that's right, 31B is hte code for SANs.
13:51 berick so, yeah.. SANs should use 31B.  91/92 are for locally defined values.
13:51 berick like ad-hoc account codes
13:51 sandbergja So, in the meantime, I'll ask the vendor if they can put the SAN in 31B, and file a not-very-urgent LP bug about allowing 92 for account codes.
13:51 sandbergja Thanks so much for your help, berick!
13:52 berick yr welcome
13:55 rlefaive joined #evergreen
14:18 rlefaive joined #evergreen
14:21 rlefaive joined #evergreen
14:40 JBoyer berick, I see all of the existing EDI docs still talk about the webrick install; is there anything about the new A/T EDI setup? (or is it so basic as "enable this A/T, turn off those 2 perl scripts, turn on these two?)
14:47 JBoyer berick, Ah, never mind. Found it in the general 3.0 release notes, I was looking in the more specific directories.
14:48 * berick nods
14:48 kmlussier I'm having trouble building the web client on 2.12. When I run the npm install command, I get the following - https://pastebin.com/n7cBknbg
14:49 berick @band add Weird Error 8
14:49 pinesol_green berick: Band 'Weird Error 8' added to list
14:50 kmlussier :)
14:53 berick JBoyer: more or less what you said.  enable and configure edi-attrs per edi account as desired, cron the new edi_order_pusher.pl script.  if going all-in, disable the legacy edi_pusher script.  no A/T changes are needed, though you could disable the JEDI event-def once fully migrated.
15:00 JBoyer Since we never actually got off the ground in the webrick days I'll just switch off all of that and onward we go.
15:00 JBoyer berick++
15:01 berick nice, that's a good way to start.
15:04 JBoyer kmlussier, that sounds familiar; if you have any node debs installed (i.e. if `dpkg -l | grep nodejs` returns anything) you should remove it and any dependencies and then repeat the prerequisites install to get a more current node installed.
15:26 kmlussier JBoyer: OK, I'll try that. Thanks!
15:26 kmlussier Strange that it hasn't come up before.
15:27 JBoyer I don't recall when it popped up, and if you normally have VMs that don't have nodejs pre-installed you might not normally see it.
15:38 kmlussier JBoyer: Yes, well, that's why it's strange because I am installing it on a newly-built VM.
15:39 JBoyer Ah, very strange then, yes.
15:39 kmlussier node is installed while I run make -f Open-ILS/src/extras/Makefile.install ubuntu-trusty-developer, right? Is it possible we're grabbing an older version of node?
15:42 JBoyer That's when it's installed, yes. It's supposed to install it from source though, so it should always been pretty fresh.
15:42 JBoyer What you're seeing *should* only happen if it's installed via apt-get / aptitude / etc.
15:43 JBoyer But, so long as you're up and going, there's a mystery for another time. ;)
15:47 kmlussier No, not up and going yet. I'm checking master now to see if I come across the problem there too.
15:48 bshum Well, the version of node that gets installed depends on which distro you're on.  Some of them use the packaged apt source, others will download the binary directly
15:48 bshum So I guess my first question is what Linux flavor are you using kmlussier?
15:48 kmlussier ubuntu 14.04
15:49 bshum Okay, 14.04 installs the nodejs from packages
15:51 bshum I did think about suggesting that we use just the nodejs sources instead of whatever distros package with
15:56 bshum Oh nvm
15:56 bshum kmlussier just pointed me out at a patch I did to already make that the approach for 3.0+
16:00 JBoyer Ah, oops. I was looking at my 3.0+ version of the trusty makefile, I forgot about that part.
16:02 bshum It's not hard to port the stuff for using nodejs binary latest LTS instead of the apt packaged legacy stuff.
16:02 * jeff laughs at the usage of "nvm" in a conversation about node
16:02 bshum But I guess we didn't backport it cause we didn't expect there to be much change for 2.12
16:02 bshum Heh
16:03 jeff (nvm == node version manager -- not something I think is immediately helpful here)
16:09 Bmagic pg_restore - we get the errors when restoring to a blank DB complaining about functions that do not exist. Should the db have the schema's already created before the restore so that the functions exist ahead of time?
16:10 Bmagic I think the answer is "ignore those function errors and restore to a blank db"
16:10 kmlussier bshum: So we should be backporting bug 1683388 and bug 1718549 to 2.12?
16:10 pinesol_green Launchpad bug 1683388 in Evergreen "Install newer NodeJS binary for Ubuntu Trusty and Debian Wheezy" [High,Fix released] https://launchpad.net/bugs/1683388
16:10 pinesol_green Launchpad bug 1718549 in Evergreen "install NodeJS from source for all supported distributions" [High,Fix released] https://launchpad.net/bugs/1718549 - Assigned to Galen Charlton (gmc)
16:10 rlefaive joined #evergreen
16:11 bshum kmlussier: Possibly.
16:11 bshum Not sure if that's the real problem or not based on your errors you saw
16:11 * kmlussier can try it locally to see if it fixes her issue.
16:13 bshum And I'm not 100% sure what'll happen if you have both nodejs legacy and then install the nodejs binary
16:13 bshum If it'll just figure out to use the newer one, or if it'll be unhappy till you remove nodejs legacy
16:13 bshum I vaguely recall testing that out
16:13 bshum but it's been awhile..
16:14 jeff Bmagic: can you please name the functions in question?
16:14 jeff Bmagic: my first thought is that you may be running into bug 1671150
16:14 pinesol_green Launchpad bug 1671150 in Evergreen "Unqualified references in evergreen.unaccent_and_squash lead to index creation failures with pg_restore" [Medium,Fix committed] https://launchpad.net/bugs/1671150
16:15 Bmagic jeff: I thought of that one too!
16:15 Bmagic that is a problem, but not the only one
16:15 Bmagic ALTER FUNCTION metabib.compile_composite_attr(cattr_def text) OWNER TO evergreen -  ERROR:  function metabib.compile_composite_attr(text) does not exist
16:15 jeff Bmagic: if it's not that, then can you (in addition to the function names or full pg_restore output), i'd be interested to know how you're creating the database.
16:15 Bmagic CREATE FUNCTION compile_composite_attr(cattr_def text) RETURNS evergreen.query_int  ERROR:  type "evergreen.query_int" does not exist
16:16 jeff are you using -e / --exit-on-error with your pg_restore? if not, what is the first error you see in the pg_restore output?
16:16 jeff (also, still interested in how you're creating the database)
16:17 Bmagic not exiting on error... Command was: CREATE FUNCTION compile_composite_attr(cattr_id integer) RETURNS evergreen.query_int is the first error
16:18 jeff okay, metabib.compile_composite_attr() is a plperlu function. does the database you're restoring to have plperlu as a language?
16:19 jeff (\dL from psql)
16:20 miker Bmagic: you need to install the extensions in the empty db
16:20 Bmagic miker: got that
16:20 jeff Bmagic: what method and/or commands are you using to create your empty database?
16:21 Bmagic CREATE DATABASE evergreen TEMPLATE template0 ENCODING 'UNICODE' LC_COLLATE 'C' LC_CTYPE 'C';
16:21 jeff Bmagic: can you share the output of \dL and \dx for the empty database (even if partially-restored-into)?
16:22 pastebot "Bmagic" at 64.57.241.14 pasted "PG creation" (13 lines) at http://paste.evergreen-ils.org/947
16:22 jeff generally you'll want to create an empty database to restore into using proper arguments to eg_db_config.pl or the steps in Open-ILS/src/sql/Pg/create_database_extensions.sql (paying attention to the comments)
16:23 jeff Bmagic: what version of Evergreen for the database dump you're restoring?
16:23 Bmagic CREATE EXTENSION unaccent; is missing from my script.....
16:24 jeff you're missing the unaccent extension, which iirc has been a requirement since 2.11
16:24 jeff yeah.
16:24 jeff also, how are you calling pg_restore?
16:25 Bmagic -C -d
16:25 jeff you don't want -C
16:25 Bmagic alright, I'll try that
16:28 Bmagic still have the issue with missing function calls, how big of an issue is that exactly? In theory I can create the missing function (evergreen.query_int) before restoring?
16:28 jeff -C is going to have pg_restore create a new database matching the name of the database in the dump file. that probably isn't the name of the empty database you created to hold your restore.
16:28 jeff i think that's probably solving the wrong problem.
16:28 jeff or attempting to.
16:29 Bmagic it is the same db name
16:29 Bmagic the very tippy top error is complaining about the db already existing, but that is OK because I think I created it just fine
16:30 jeff the behavior in that case is not defined in the manpage. (-C without --clean where the database name in the dump file already exists in the cluster). I haven't tested to see what the actual behavior is. Were you able to share the \dL and \dx output of the database you're attempting to restore into?
16:30 Bmagic that went away when I removed -C
16:30 jeff okay. i still recommend getting out of the habit of using -C, but it sounds like it was not the cause of your issue.
16:31 pastebot "Bmagic" at 64.57.241.14 pasted "\dL \dx" (22 lines) at http://paste.evergreen-ils.org/948
16:33 jeff are you dropping and re-creating the database between pg_restore attempts?
16:36 Bmagic yes
16:36 jeff okay. if you remove -C and add -e to your pg_restore, what does your pg_restore command line look like? feel free to sanitize as long as it's clear what you're sanitizing. :-)
16:37 Bmagic sec
16:38 Bmagic Error from TOC entry 4496; 1255 17114 FUNCTION compile_composite_attr(integer) evergreen
16:38 Bmagic could not execute query: ERROR:  type "evergreen.query_int" does not exist
16:38 Bmagic Command was: CREATE FUNCTION compile_composite_attr(cattr_id integer) RETURNS evergreen.query_int
16:39 Bmagic let me see what it does if that function DID exist
16:39 jeff What version of Evergreen is this database, and what did your pg_restore command line look like?
16:44 jeff the query_int type is part of the intarray extension. i would expect to find both the extension and the type in the public schema. it sounds like your source database had it in the evergreen schema, either due to baseline/upgrade script oddity or something local being different.
16:44 jeff possibly a search_path biting you if you connect as a user called "evergreen", etc.
16:45 Bmagic I thought of search path too
16:46 jeff changing your CREATE EXTENSION for intarray to this might be needed in this case: CREATE EXTENSION intarray WITH SCHEMA evergreen;
16:46 jeff is the source database available for you to query?
16:48 Bmagic weird. I can't find the example of query_int in any of my other test machines
16:48 Bmagic only query_int_wrapper
16:48 jeff it would be a type, not a function.
16:48 jeff \dT
16:48 Bmagic I see
16:48 jeff rather, "\dT query_int" to find it in your search path.
16:48 kmlussier Backporting the commits to install nodejs appears to work. I'll create a branch and post it to LP. bshum++ JBoyer++
16:48 Bmagic yep, \dT shows it
16:49 Bmagic must be a search path thing, let me switch that up
16:49 jeff okay. on the source database does \dT show query_int in the evergreen schema or the public schema?
16:49 Bmagic yes
16:49 jeff that wasn't a yes or no question. :-)
16:50 Bmagic sorry, lol, public
16:51 Bmagic ALTER DATABASE evergreen SET search_path TO evergreen,public;
16:51 Bmagic or public should be first....? I'm not sure where the assumption is made about which schema to look for query_int is made
16:51 jeff i wouldn't go altering your search path without a reason.
16:52 Bmagic if the restore used public.query_int - it would find it, therefore, problem sovled right?
16:53 jeff or if it existed at evergreen.int_array
16:54 jeff does your source database have results for both \dT public.query_int and \dT evergreen.query_int?
16:55 Bmagic here's an idea! let the restore create the db.....
16:56 Bmagic just tried it, so far no errors, lol
16:56 jeff ...which won't work because pg_dump/pg_restore don't do languages or extensions...
16:57 Bmagic missing those would result in errors during restore?
16:57 jeff yup.
16:57 jeff and a failed restore.
16:57 jeff unless your target db already had the correct languages and extensions.
16:58 Bmagic wouldn't it fail pretty quickly?
16:58 jeff probably.
16:58 Bmagic still running......
16:58 mmorgan1 joined #evergreen
17:00 Bmagic I was expecting to see the db during the restore, but it's not available. It must be "in it's head" during the restore process. A giant transaction maybe?
17:01 jeff at least with postgresql 9.4, you can't have one extension installed in two schemas in the same database, so i'm puzzled as to why your dump is trying to reference a type of evergreen.query_int while the source db has the intarray extension installed in the public schema.
17:01 mmorgan1 left #evergreen
17:01 jeff unless the evergreen.query_int type was manually created outside of the extension framework, or one or more of our assumptions is wrong.
17:03 * jeff checks some of his own assumptions
17:03 jeff (especially the extensions/languages pg_dump/pg_restore one)
17:10 jeff yup, i'm wrong about the language and extensions. the logic for pg_restore isn't that it skips those on restore when restoring into an existing database, it's that it does a CREATE OR REPLACE on the languages and a CREATE IF NOT EXISTS for the extensions.
17:10 jeff so, does "pg_restore --schema-only dumpfile | grep 'CREATE EXTENSION'" show intarray being created in the public or evergreen schema?
17:11 bshum kmlussier++ # web client building happiness for all
17:11 jeff and what does the output of this look like? pg_restore --schema-only dumpfile | grep "query_int"
17:16 khuckins__ Has anyone here conditionally applied styles to cell content eg-grid-fields? ng-class doesn't seem able to see the item in the current eg-grid-field, whether I make a scope function or do the condition check in the ng-class attribute itself. I can display the data outside the element, so it's a bit odd. Currently working in the bills list
17:20 jeff Bmagic: you've gone silent -- I'm still interested in answers to the various questions above, if only to help update documentation/recommendations and understand what state your database is in and how it may have gotten there. Heading out for a few, back later.
17:20 Bmagic jeff: sorry, back
17:20 jeff heh.
17:21 Bmagic so, just after installing postgres, I did run the typical evergreen install script, then deleted the db and created fresh
17:22 Bmagic I let the eg scripts work it's magic on the fresh postgres install (if that makes a difference)
17:23 jeff i'd be interested in the two pg_restore + grep commands above. it would help understand what the dump file contains with regard to the extension/type in question.
17:23 Bmagic I just ran schema only
17:23 Bmagic no errors
17:24 bos20k joined #evergreen
17:24 Bmagic from what I can tell, it also didn't do anything
17:24 jeff pg_restore --schema-only dumpfile | grep 'CREATE EXTENSION'
17:24 jeff pg_restore --schema-only dumpfile | grep "query_int"
17:24 Bmagic right
17:24 Bmagic maybe I need to throw in the -d evergreen
17:25 jeff i'm interested in seeing the full output of those commands (replacing dumpfile with the name/path of the file you're restoring)
17:26 jeff away for a few, back later this evening!
17:26 jeff Bmagic++
17:26 Bmagic pg_restore  -d evergreen --schema-only dump.dmp | grep "query_int"
17:26 Bmagic jeff++
17:28 pastebot "jeff" at 64.57.241.14 pasted "pg_restore -d evergreen --schema-only dump.dmp | grep "query_int"" (27 lines) at http://paste.evergreen-ils.org/949
17:32 Bmagic sorry khuckins__ - I have not
18:10 jvwoolf joined #evergreen
18:11 JBoyer_alt joined #evergreen
18:30 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>
19:02 beanjammin joined #evergreen
19:04 beanjammin khuckins: You were asking about conditionally adding classes to grid fields?  I just finished doing a little bit of that.
19:06 khuckins beanjammin: Yeah, I've been wracking my brain a bit, specifically trying to make a few glyphicons show up in a field depending on certain values in the item
19:08 beanjammin I banged my head trying to add a class to show overdue items in a Patron's Checked Out items list.  This is what I came up with - http://paste.evergreen-ils.org/950
19:09 rlefaive joined #evergreen
19:15 khuckins Oooh I've done something somewhat similar for a separate part of this - to get row highlighting working
19:16 khuckins Sadly editing the autogrid wouldn't work too well for this particular instance, but thanks :)
19:30 rlefaive joined #evergreen
19:37 Christineb joined #evergreen
22:31 jeff @later tell Bmagic thanks for the paste output attempt, but you seem to have added an argument (-d evergreen). if you still have time and interest, I'd be interested in seeing the output of the command without that added argument.
22:31 pinesol_green jeff: The operation succeeded.
23:15 ohiojoe joined #evergreen

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