Evergreen ILS Website

IRC log for #evergreen, 2019-04-04

| 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
03:17 yar joined #evergreen
05:00 pinesol News from qatests: Failed Running pgTAP tests <http://testing.evergreen-ils.org/~li​ve/test.42.html#2019-04-04T04:59:06,000630798-0400 -0>
07:12 rjackson_isl joined #evergreen
08:05 bdljohn joined #evergreen
08:26 Dyrcona joined #evergreen
08:31 bos20k joined #evergreen
08:43 mmorgan joined #evergreen
09:00 agoben joined #evergreen
09:14 sandbergja joined #evergreen
10:26 pastebot "Dyrcona" at 64.57.241.14 pasted "IDL Parse Error, except xmllint says it's OK" (12 lines) at http://paste.evergreen-ils.org/10836
10:27 Bmagic Dyrcona: stock?
10:27 Dyrcona Bmagic: Not stock, but xmllint doesn't report any issues, and the error suggest a missing </class> which xmllint would catch.
10:28 Bmagic which OS 18.04?
10:30 Dyrcona Nope, 16.04.
10:30 JBoyer joined #evergreen
10:30 Dyrcona I'm reverting the IDL additions to see if that makes a difference.
10:31 Bmagic My thinking is the xmllint version might* matter - or yeah, fiddle with the IDL
10:33 Dyrcona Well, reverting our custom entries seems to have resolved it. Now to figure out what's wrong with our custom entries.
10:34 Dyrcona This didn't happen until I upgraded this VM to 3.2.4 with our customization branches. It was working fine with 3.0 and this patch.
10:34 Bmagic hmmm, that is perplexing
10:35 Dyrcona Though, it is working on the training server....
10:35 Bmagic I had a similiar issue recently but in my case the issue was much more obvious
10:35 JBoyer If it's ok to post the custom IDL in full I'd be curious to see it. At least there are a couple ids in the error message. :)
10:35 * Dyrcona double checks that.
10:37 Dyrcona Nope. Apparently, that commit is missing from our training branch.
10:37 Christineb joined #evergreen
10:40 Dyrcona No conflict markers...
10:48 Dyrcona So, it's looking like something is botching it during installation, actually.
10:50 Dyrcona I'm not sure, though what the \n means in that parser output.
10:54 pastebot "Dyrcona" at 64.57.241.14 pasted "Bmagic: The full IDL" (13098 lines) at http://paste.evergreen-ils.org/10837
10:54 pastebot "JBoyer" at 64.57.241.14 pasted "Dyrcona's message reformatted" (11 lines) at http://paste.evergreen-ils.org/10838
10:55 JBoyer Just to line things up in the error message. You'll probably have to look at the plain text to see what it's doing.
10:56 JBoyer Oh, no, the default view does the right thing re: monospaced fonts.
11:01 Dyrcona cash_report_mimic is the only class that has \n without any ' ' between the attributes. The other classes all have a ' ' or more.
11:02 Dyrcona Sounds like a bug in the XML parser. :)
11:04 yboston joined #evergreen
11:08 Dyrcona Yeahp. Seems to be working, now.
11:09 Dyrcona Dunno if that's a bug in the parser or just one of those things.
11:10 collum joined #evergreen
11:11 Dyrcona And, I need to modify my websocksetd start script to redirect to /dev/null. :)
11:11 Dyrcona The console messages can get in the way.
11:12 sandbergja joined #evergreen
12:08 jihpringle joined #evergreen
12:31 khuckins joined #evergreen
12:33 Dyrcona Anyone else have the user permission editor go unresponsive on 3.2?
12:34 mmorgan Dyrcona: I've seen that on 3.1, it actually behaves better for me on 3.2.
12:36 Dyrcona Well, that interface looks like it lacks intelligence. It seems to update everything listed, so when you have a lot of work_ou_maps, it times out or something.
12:36 mmorgan Actually, let me amend that. It behaves better on 3.1 and 3.2 for me than it did on 3.0.
12:37 Dyrcona Well, I'll blame the slow VM and busy db server for now.
12:46 Dyrcona Group permissions editor seems to just be broken. I tried adding a mapping and I don't even see an attempt to create a grp_perm_map entry in the logs.
12:48 Dyrcona Only thing in the logs matching grp are related to opening conify/global/permission/grp_tree.hml
12:50 mmorgan Dyrcona: I noticed that on a 3.3 system just this morning. Seems to be working ok on our 3.2.4 training system.
12:51 Dyrcona I haven't been told its not working on our 3.2.5 training system. This machine where it's not working is a 3.2.4 test VM. We decided not to go to 3.2.5 after some performance issues with a couple of the fixes.
12:53 Dyrcona So, wonder if it's "one of those things" or something that we backported, but I don't recall backporting any permissions-related things.
13:00 Dyrcona Well, this explains it: POST https://jasontest.cwmars.org/osrf-http-translator 400 (Bad Request)
13:02 Dyrcona I get a number of 404s for dojo.js and openils_dojo.js just opening the page.
13:10 * Dyrcona tries reinstalling OpenSRF rel_3_1, just in case.
13:14 Dyrcona Nope. Doesn't help after rebuilding OpenSRF and Evergreen.
13:19 Dyrcona Well, it's not max_stanza_size, turns out that is 20,000,000 on this machine.
13:28 yboston joined #evergreen
13:29 Bmagic What's the troubleshooting steps for penalties that are not getting applied? It clearly should be getting applied. I've ran open-ils.actor.user.penalties.update on srfsh, it's not deciding that the patron needs the penalty when it clearly does
13:30 Bmagic I'm going for PATRON_EXCEEDS_LOST_COUNT at a threshold of 1. But the patron only has one thing lost, so I thought maybe it needs to have a threshold lower than 1, so I tried .5. No dice
13:31 Bmagic The lost item, patron home library and penalty org unit match
13:32 Dyrcona Try having the patron lose another item.
13:32 Bmagic :)
13:32 bshum Maybe also run the recalculate patron penalty function in the DB
13:33 Bmagic Oh, there is a function in the DB? I couldn't identify one
13:33 bshum Isn't there?  Maybe I'm making it up in my head...
13:33 Dyrcona :)
13:33 Bmagic oh yep actor.calculate_system_penalties
13:33 Dyrcona A lot of the Perl code ends up running db functions.
13:33 Dyrcona Well, a good bit of it, anyway.
13:34 berick Bmagic: any chance the item has been returned and or the xact has been all paid off?
13:34 Dyrcona I'm a bit vague on what the flow is supposed to be with penalties.
13:34 Bmagic bshum: good call - executing that function returns the penalty that I want, however it doesn't insert it into actor.usr_standing_penalties
13:35 bshum Bmagic: Are you sure they don't already have the penalty set then somewhere else?
13:35 Bmagic not according to the database they don't
13:35 Bmagic select * from actor.usr_standing_penalty where usr={myid}
13:36 bshum Okay, just checking cause sometimes I've seen it where we had a lost penalty threshold for higher branch level
13:36 Dyrcona can't argue with the database. :)
13:36 bshum And that resulted in different penalties being assigned
13:37 Bmagic yep, the DB function returns the rows I would expect to see in usr_standing_penalty. So, when do rows get inserted there? Not part of running the DB function?
13:38 Dyrcona probably as part of the back end function.
13:39 Bmagic at least on XUL, I thought that the act of opening the patron would run the calculation
13:40 Dyrcona Bmagic: You don't get a permission error, do you?
13:40 Bmagic not on the UI at least I don't
13:40 Bmagic haven't checked the logs
13:41 Dyrcona When you run it in srfsh, you've logged in and passed in the authtoken?
13:41 Bmagic yep, on srfsh it works without errors
13:41 Bmagic Received Data: "1"
13:41 Dyrcona Did it update the patron in the db?
13:41 Bmagic nope
13:41 Bmagic or at least it didn't apply the penalty that I want it to
13:44 Dyrcona Could be the org_unit that you're using, but the call eventually uses the db function to get the penalties to apply.
13:44 Bmagic request open-ils.actor open-ils.actor.user.penalties.update "myauth", 571114
13:44 Bmagic I will say that my authtoken is workstationless
13:44 Bmagic at srfsh
13:45 Dyrcona Well, that may be the problem.
13:45 Bmagic can I pass a workstation DB ID using the srfsh# login staff password ?
13:45 Dyrcona Add the workstation and see what happens.
13:46 Dyrcona login <username> <password> [type] [org_unit] [workstation]
13:46 Dyrcona from srfsh help
13:47 Bmagic oh, umm, what do I do for type ? "staff" ?
13:47 Dyrcona you can also use the open-ils.auth.login call.
13:47 Dyrcona yes.
13:49 Bmagic ok, what am I doing wrong
13:49 Bmagic request open-ils.auth open-ils.auth.login {"admin","password", "staff", 285, 1590}
13:50 Dyrcona login admin demo123 staff BR1 BR1-N240WU
13:50 Dyrcona You want the names, quotes not needed.
13:51 Dyrcona srfsh# request open-ils.auth open-ils.auth.login {"username": "admin", "password": "demo123", "type": "staff", "workstation": "BR1-N240WU"}
13:51 Dyrcona Both commands, copy/pasted from my pre-confernce slides. :)
13:51 bshum #shameless advertising
13:51 bshum Dyrcona++
13:53 bshum Just thinking out loud... is this Lost penalty for a single library branch or something?  or is it a higher org tree rule?
13:53 bshum And is the workstation where you're trying to trigger the penalty at the same level?
13:53 bshum *location
13:54 bshum Cause let's say if you had a penalty threshold set for a given library, but you were using the workstation of another library, it wouldn't necessarily trigger the penalty
13:54 bshum Until someone were to retrieve it using the library where the threshold was set
13:55 Bmagic holy cow, that was an ordeal to get logged with the workstation on srfsh. But once I did  and ran the penalties command, it updated the DB
13:55 Bmagic Dyrcona++
13:55 bshum That's one of the reasons why local based blocking rules that apply consortially is not good
13:55 Bmagic bshum++
13:55 Bmagic Conclusion: refreshing the patron on the web client doesn't update the penalties
13:56 bshum Was the penalty set at org unit 1 for the whole consortium or just one place?
13:56 Bmagic at the system level
13:56 bshum Okay, and the workstation used to look up the patron was from that system?
13:56 Bmagic on the web client in my testing, I was logged in at the branch under the system where the penalty is defined
13:56 bshum Hmm
13:57 bshum that sounds suspicious then
13:58 * bshum doesn't know enough about the web client to comment further :D
13:59 Bmagic I didn't try refreshing on XUL - I should have
14:00 * Dyrcona lurks for a bit.
14:00 jeff makes sense. the web client doesn't *have* a refresh option for the patron, right?
14:01 Bmagic cause the browser has one built in
14:02 Bmagic confirmed - XUL applied the penalty
14:02 bshum Running the penalty re-calc every time we retrieve a patron in web client sounds expensive
14:02 bshum Maybe there needs to be a new button for forced recalc
14:03 Bmagic I should have tried that before this whole conversation
14:03 Bmagic The refresh button on XUL is what did it. Opening the patron did not (on XUL)
14:03 bshum Right, again, why recalc every time you open a patron
14:04 bshum Sounds like a fun new feature
14:07 Bmagic The special button (renamed from refresh to Webby friendly terms) might be required to fix this issue. I think* the action trigger MarkItemLost will recalc the penalties at the end so that's something.
14:25 ningalls_ joined #evergreen
15:16 JBoyer joined #evergreen
16:22 Bmagic jeff: RE: auditor table claiming to have status checked out without a matching circulation in action.circulation - Nothing in aged_circulation. also config.copy_alert_type next_status is null 100% of the rows
16:29 jeff I wonder how much work it would take to make the Renew context menu option in Item Status open the Renew interface with the barcodes of the selected items.
16:30 jeff Because then the receipt experience would be a bit better...
16:30 jeff context is for item-present renewals where the patron isn't certain that they want to renew yet, etc.
16:30 khuckins joined #evergreen
16:30 jeff transaction begins with "can you check to see when this is due?"
16:31 jeff Also, there doesn't seem to be a way to print a single item receipt from Item Status.
16:32 jeff Bmagic: interesting! i don't remember all of the other details, but I think so far I'm stumped, barring a direct db update or copy template shenanigans.
16:33 Bmagic right! I'm exporing an idea, be back shortly
16:33 Bmagic exploring
16:37 mmorgan Bmagic: It was possible in the xul client to change an item's status to Checked out in the editor under certain circumstances.
16:37 Bmagic oh yeah?
16:37 Bmagic That almost has to be it
16:38 mmorgan bug 1616980
16:38 pinesol Launchpad bug 1616980 in Evergreen 2.12 "Magical statuses not so magical - possible for staff to edit items into or out of magical statuses" [Medium,Fix released] https://launchpad.net/bugs/1616980
16:38 Bmagic mmorgan++
16:39 Bmagic but it wasn't possible in 3.0+ ?
16:39 Bmagic The date on this is this year and the system was on 3.0.9
16:40 mmorgan That fix was for the web client, it was never fixed in xul.
16:40 Bmagic sweet! Thanks
16:40 Bmagic I was about to start reading the code
16:40 mmorgan bug 1695921
16:40 pinesol Launchpad bug 1695921 in Evergreen "Possible for staff to edit items into or out of magical statuses (XUL)" [Undecided,Won't fix] https://launchpad.net/bugs/1695921
16:41 mmorgan @decide read code or read launchpad
16:41 pinesol mmorgan: go with read launchpad
16:41 mmorgan :)
16:41 jeff was status the only thing that changed, or did the edit timestamp and other bits change as well, as you'd expect from an edit operation in the client?
16:41 Bmagic is it plausable that the staff switched it to checked out manually (even though the item was returned and on the shelf) - then the system marked it lost and billed the patron that last circed it?
16:41 jeff i can't remember if that question was asked/answered in the previous description of the mystery. :-)
16:43 Bmagic edit date changed as well as status changed time
16:43 Bmagic auditor did record who did it
16:48 jeff ah good. also be interesting to example copy templates for that user, as those were another way to force a status change.
16:48 jeff as for your question about the system billing the previous patron who circulated the item... that would be common under a Damaged workflow, but I'm not sure that you'd end up with that for Lost.
16:49 jeff gmcharlt: i just dusted off bug 1437106 to put a comment on it, containing a question for you if and when you have a moment.
16:49 pinesol Launchpad bug 1437106 in Evergreen "keyboard shortcuts in the web staff client" [Undecided,Confirmed] https://launchpad.net/bugs/1437106
16:50 * gmcharlt sneezes
16:50 jeff :-)
16:53 Bmagic Another clue here is: the item was automatically marked Lost but the patron never received the related billing. That's because the circulation wasn't open.
16:53 jeff which process marked the item lost?
16:54 mmorgan Bmagic: could bug 1758975 possibly be relevant?
16:54 pinesol Launchpad bug 1758975 in Evergreen "Item status can be set to Lost and Paid inappropriately when closing a transaction for a Long Overdue or Lost returned item" [Medium,Confirmed] https://launchpad.net/bugs/1758975
16:54 Bmagic I was thinking that the system did the right thing by not billing the patron, because the circulation was closed
16:58 pastebot "gmcharlt" at 64.57.241.14 pasted "for jeff: the Evergreen side of the experiment" (22 lines) at http://paste.evergreen-ils.org/10840
17:00 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
17:03 mmorgan left #evergreen
17:08 jeff gmcharlt++ i'll keep my eyes out for the Mousetrap side, or let me know if there isn't one. Thanks!
17:37 jeffdavis Are we going to have a proper dev meeting this month? I know there was an informal quasi-meeting on Tuesday...
18:11 Dyrcona We usually have one at the conference, at least among the committers.
18:29 jamesrf joined #evergreen
18:56 jeffdavis Yeah, that makes sense.
20:03 sandbergja joined #evergreen
21:44 sandbergja joined #evergreen
22:05 Dyrcona Oh! I'm still signed in?

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