Evergreen ILS Website

IRC log for #evergreen, 2016-12-16

| 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
02:58 abowling1 joined #evergreen
05:00 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>
06:28 rlefaive joined #evergreen
07:10 rjackson_isl joined #evergreen
07:23 agoben joined #evergreen
08:01 Dyrcona joined #evergreen
08:31 Dyrcona So, I've got reports of the Home Library select list being empty in edit patron and register patron.
08:31 Dyrcona AFAIK, it is only affecting 3 users: myself and two others at central site.
08:32 Dyrcona If you try to save an updated patron in this state, you get a database error from the home_ou not being set.
08:32 Dyrcona We're using the alt_patron_summary branch.
08:33 Dyrcona I'm about to dive into the code and try retrieving the org tree as my own user with backend calls to see if any errors pop up.
08:33 Dyrcona I thought I'd throw this situation out there in case someone else has encountered it before and has ideas where I should look.
08:36 csharp Dyrcona: is it intermittent for the affected users?
08:36 Dyrcona csharp: Nope. All the time.
08:36 Dyrcona The person who reported it claimed it started at 9:50 am yesterday, but the logs don't show her trying to update a patron until about that time.
08:37 Dyrcona At least, for yesterday.
08:37 Dyrcona I have a hunch that I'm going to investigate before writing any code.
08:37 Dyrcona I'm going to to check in the training database, too.
08:38 Dyrcona Once I download the client.
08:38 csharp autogen.sh is the only thing I can think of, but I've never seen that or had reports of it
08:38 * csharp knocks on wood
08:39 Dyrcona If autogen.sh would fix it, then everyone would have the problem, and the phones would be ringing off the hook. :)
08:39 Dyrcona I think it's permissions/work_ou related.
08:42 StomproJ joined #evergreen
08:43 Dyrcona No training client on updates/manualupdate.html, so I'll just wipe out my workstation and register a new one in production to see if that makes a difference.
08:44 mmorgan joined #evergreen
08:45 csharp oh, yeah - perms would explain that too
08:45 krvmga joined #evergreen
08:45 csharp empty dojo dropdowns usually means a missing perm in my experience
08:47 Dyrcona csharp: Yeah, that's my angle, now. The workstation_ou is probably not the problem.
08:48 Dyrcona I thought maybe it was registered an ou that can't have users or something, but you can't register those, and that's not the case in my account.
08:48 Dyrcona I'll check home_ou on these users, too. Maybe those can't have users.
08:48 Dyrcona To the database!
08:50 Dyrcona And, it's cold in here.
08:50 Dyrcona @weather
08:50 pinesol_green Dyrcona: Methuen, MA :: Clear :: 2F/-17C | Wind Chill: -15F/-26C | Friday: Sunny. High 18F. Winds W at 10 to 15 mph. Friday Night: Partly cloudy skies this evening will give way to cloudy skies with snow developing overnight. Low near 15F. Winds light and variable. Chance of snow 80%. Snow accumulations less than one inch.
08:51 * Dyrcona checks the heat.
08:51 krvmga @weather
08:51 pinesol_green krvmga: Worcester, MA :: Clear :: 1F/-17C | Wind Chill: -17F/-27C | Friday: Sunny. High 17F. Winds W at 10 to 15 mph. Friday Night: Partly cloudy skies this evening will give way to cloudy skies with snow developing overnight. Low 14F. Winds SW at 5 to 10 mph. Chance of snow 90%. Snowfall around one inch.
08:53 Dyrcona "Ahh, but it's cold outside....."
08:53 Dyrcona Anyone else sick of Christmas songs, yet?
08:54 * csharp raises his hand
08:54 * krvmga raises his hand.
08:55 krvmga not that I have anything against Christmas songs but how many times to I have to hear "Santa, baby"?
08:55 * mmorgan was already sick of them in early November ;-)
08:55 Dyrcona Well, krvmga's is working and he's the same profile as two of us for whom it does not work.
08:56 krvmga yes, it is working for me - confirmed
08:56 mmorgan Dyrcona: Just a stab in the dark. Have you tried clearing the client cache?
08:56 * Dyrcona checks for extra user permissions.
08:57 Dyrcona mmorgan: I've quit the client and even blown out my openils_staff_client XUL profile directory.
08:57 mmorgan That shoud cover it!
08:57 Dyrcona One user wiped out his windows profiles and reinstalled the client.
09:02 Dyrcona No usr_perm_map or usr_grp_map entries for krvmga, and all home_ous are regular branches....
09:02 Dyrcona So, we should have the same permissions....
09:03 mmorgan Dyrcona: When krvmga logs in at the workstation of a user that is seeing the problem, can he see the dropdown?
09:03 Dyrcona mmorgan: I don't think that was done, but I'll ask him.
09:03 krvmga sec, let me check...
09:03 Dyrcona :)
09:06 Dyrcona The other user with the same profile having the problem has a permission that krvmga and I do not have, but having an additional permission would not be it.
09:08 krvmga ok, the user with the problem logged in on a machine of a user who was *not* having the problem; still had problem.
09:08 krvmga seems to be following the login
09:09 Dyrcona So, if the group was granted EVERYTHING at depth 0, why have other permission granted?
09:10 Dyrcona krvmga: Thanks for checking. I'm stumped. :(
09:10 mmorgan Any changes made to perms lately?
09:10 Dyrcona Not by me, but the group that we're has EVERYTHING at depth 0, so it can't be permissions.
09:11 Dyrcona I'm tempted to remove the other permissions on the group and see what happens.
09:13 mmorgan Dyrcona: could lp 1480432 be related?
09:13 pinesol_green Launchpad bug 1480432 in Evergreen "Staff users can have permission at a more restrictive depth than assigned via a permission group" [Undecided,New] https://launchpad.net/bugs/1480432
09:14 Dyrcona But that would affect everyone in the group, no?
09:16 mmorgan Not necessarily. In my experience it's caused problems for individual users that had more than one group assigned.
09:16 Dyrcona And none of the parent groups have any permissions assigned.
09:17 mmorgan If your users don't have secondary groups, then it's probably not your issue.
09:18 Dyrcona But we don't have more than 1 group assigned.
09:18 krvmga i just did the opposite test from the previous one: i logged in as myself on a client where the user wasn't seeing the library dropdown list; saw it as me; didn't see it as the other user.
09:19 Dyrcona krvmga: Thanks, again. ;)
09:19 krvmga same permission group; network administrator
09:19 mmorgan ok. It would be interesting to see what happens if you were to remove the extra permissions from the users seeing the problem.
09:19 Dyrcona I'm going to remove the extra permissions from the group to a separate table.
09:19 Dyrcona :)
09:23 Dyrcona Nope. :(
09:23 Dyrcona Unless memcached has entries that I need to delete.... That's asking too much.
09:25 tsbere Dyrcona: Have you checked to see if there are any horribly broken things in the org tree?
09:25 tsbere Or group tree
09:25 tsbere Or whichever one you are having issues with
09:25 * tsbere hasn't fully read the scrollback
09:25 Dyrcona Well, I don't know what I'm having trouble with, and I would think either of those would cause problems for everybody, not just some people and not others.
09:25 tsbere I once got "things don't load right" in my dev machine with a bad entry or two
09:26 tsbere And they only applied to "Everyone" on the org tree one due to one of the issues being a permission def, so it wasn't loading that entry otherwise
09:26 tsbere or Everything or whatever
09:26 * tsbere is still too cold this morning
09:26 Dyrcona Just to get everyone up to speed: I've got four accounts with the same permission group. It works for two of them and not for the other two.
09:27 Dyrcona That have been definitively reported so far.
09:27 tsbere That...sounds odd. How do their work OU sets vary?
09:27 Dyrcona I've got two other users that say it doesn't work for them, but I don't know their permission group(s).
09:31 Dyrcona tsbere: They vary, but it looks like I can work at every branch, including the one that I signed in at.
09:34 yboston joined #evergreen
09:42 mmorgan joined #evergreen
09:47 maryj joined #evergreen
09:56 mmorgan joined #evergreen
10:42 mmorgan1 joined #evergreen
10:47 Dyrcona tsbere: Turned out to be an org_unit was changed yesterday for testing. It was assigned an ou_type with the wrong depth when testing was done.
10:47 Dyrcona Why this did not break the list for everybody is beyond me.
10:50 tsbere Dyrcona: I assume it is in part due to work ous. How many of the people it was broken for were logged into or assigned their work OU at/below that org unit versus not at/below it?
10:50 Dyrcona That's a good question. krvmga changed his ous after I last looked. I'm pretty sure that I could work at this ou. I'll check where the others can/can't work.
10:52 bos20k joined #evergreen
10:52 Dyrcona Ah ha! It was the sibling ou.
10:52 Dyrcona We could all work at the ou in question, but only two of us could work at the sibling ou.
10:53 Dyrcona The two ous having different depths and the same parent must have been what broke it.
10:54 Dyrcona Hm... that's not exactly right when I look again.
10:55 Dyrcona krvmga and I could work there, and krvmga added his permission after the trouble started, because id is above X.
10:55 Dyrcona One of the other affected people could work there, but not the first one. Does that even make sense?
10:56 Dyrcona Of the 3 people with the same problem, two of had permission at that ou.
10:56 Dyrcona Well, it's fixed, now, and I know the rough cause.
11:02 Christineb joined #evergreen
11:02 Dyrcona tsbere++ csharp++ mmorgan++
11:03 Dyrcona DBI++
11:03 Dyrcona Data::Dumper++
11:03 Dyrcona diff++
11:03 Dyrcona krvmga++
11:03 Dyrcona perl_hashes_not_being_randomly_ordered++
11:08 berick they are randomly ordered now aren't they?  well, undefined and potentially varried ordering
11:10 Dyrcona berick: They didn't appear to be in my case.
11:10 brahmina joined #evergreen
11:11 Dyrcona I dumped the results of fetchall_hashref on actor.org_unit_descendants(1) from two databases and got practically identical files.
11:11 Dyrcona But, yeah, I think newer Perl does randomize them a bit.
11:12 berick yeah, just sayin i wouldn't count on it
11:12 krvmga Dyrcona++
11:12 Dyrcona Well, then, I'd just have to sort the output by keys before calling Data::Dumper is all. :)
11:13 berick that'll do it :)
11:13 Dyrcona I actually wrote this line yesterday:
11:13 Dyrcona my @results = sort {$b->{score} <=> $a->{score}} sort {$b->{id} <=> $a->{id}} values %merged;
11:15 Dyrcona Perl is so readable.... ;)
11:30 jvwoolf joined #evergreen
11:53 mmorgan joined #evergreen
11:59 mmorgan joined #evergreen
12:03 bmills joined #evergreen
12:05 * mmorgan needs a sanity check on hold matrix matchpoints.
12:05 mmorgan config.hold_matrix_matchpoint.usr_grp seems to apply to action.hold_request.requestor
12:05 mmorgan and
12:05 mmorgan config.hold_matrix_matchpoint.requestor_grp seems to apply to action.hold_request.usr
12:05 mmorgan Am I seeing this right?
12:06 Dyrcona That's not my understanding.
12:08 mmorgan I set up a matchpoint to make items with a certain circ lib NOT holdable for Users.
12:09 mmorgan Then another matchpoint to make a certain permission group able to place holds for items with that circ lib.
12:09 Dyrcona Looking at action.find_hold_matrix_matchpoint the requestor is the requestor and the user is the patron the hold is being made for.
12:10 Dyrcona They can be the same user if the user is placing the hold on their own behalf, like a patron using the OPAC.
12:11 Newziky joined #evergreen
12:11 mmorgan Right. What I'm actually trying to do is fix a ComCat problem.
12:11 Dyrcona What are you trying to do, exactly?
12:11 Newziky left #evergreen
12:12 miker mmorgan: do you have the global flag called circ.holds.usr_not_requestor enabled?
12:12 miker if so, that swaps the meaning of the fields. it's from before we had matrix weights
12:14 Dyrcona Last time I checked there were about 40 org settings that affect holds.
12:15 miker well, yes
12:15 miker but that one in particular is designed to cause exactly what mmorgan is describing
12:15 mmorgan miker: Why yes, that is enabled. Not sure why it is.
12:15 Dyrcona But, yeah, that is a good one to look at. I always forget about it because I've always had it false.
12:15 mmorgan But that certainly explains it.
12:16 Dyrcona miker++
12:16 mmorgan Dyrcona: I was trying to explicitly make ComCat items holdable for only the ComCat permission group as the requestor.
12:16 mmorgan miker++
12:17 Newziky joined #evergreen
12:17 Newziky left #evergreen
12:17 miker mmorgan: if you use a lot of rules that involve groups, untangling that could become a ... job
12:17 Dyrcona mmorgan: Makes sense. I think we might have done that at MVLC. I don't usually deal with the matrix at C/W MARS.
12:17 miker or, rather, lots of matchpoints that have differing usr and requestor values
12:18 Dyrcona Well, even adding something like a Local Use Only group becomes a headache very quickly.
12:19 Dyrcona You quickly end up adding exceptions to other exceptions.....
12:19 Dyrcona Rules aren't rules, apparently. ;)
12:20 Dyrcona But, yeah, changing that setting might lead to some surprises with matrix matchpoints that "work," now.... I think that's where miker is going.
12:20 miker indeed, it is :)
12:20 mmorgan Yeah. sounds like ... a job ;-)
12:21 Dyrcona I have a Perl script that will dump the circ and hold matrices, along with other settings, into an Excel workbook if you think that would be helpful.
12:24 Dyrcona It's on my github: https://github.com/Dyrcona/evergreen_ut​ilities/blob/master/perl/parameters.pl
12:24 mmorgan Dyrcona: Thanks! I'll take a look at it.
12:25 mmorgan Dyrcona++
12:25 mmorgan So it looks like by default that global flag is set to true? http://git.evergreen-ils.org/?p=Ever​green.git;a=blob;f=Open-ILS/src/sql/​Pg/950.data.seed-values.sql#l11536
12:27 kmlussier joined #evergreen
12:27 Dyrcona Indeed it does.
12:28 * mmorgan didn't remember changing it.
12:30 mmorgan So with weighting, sounds like it should be set to false by default.
12:30 Dyrcona I'd agree with that.
12:31 mmorgan I can open a Launchpad bug.
12:31 Dyrcona OK.
12:31 Dyrcona At MVLC, I don't think we used requestor much.
12:33 mmorgan We don't use requestor all that much either. Most of our hold policies are pretty general.
12:34 Dyrcona I do know we turned that flag off.
12:39 Dyrcona If it weren't for "backwards compatibility," I'd vote for just removing that flag.
12:40 Dyrcona Maybe remove it from the base schema, but not remove it via upgrade script?
12:45 mmorgan lp 1650641
12:45 pinesol_green Launchpad bug 1650641 in Evergreen "Global flag circ.holds.usr_not_requestor should not default to TRUE" [Undecided,New] https://launchpad.net/bugs/1650641
12:50 kmlussier mmorgan: I think you missed a 'not' in the description.
12:51 mmorgan kmlussier: Indeed! Thanks for proofreading.
12:51 mmorgan kmlussier++
12:52 kmlussier mmorgan++
12:53 Dyrcona mmorgan++
12:58 miker mmorgan++
12:59 * Dyrcona whips out his trusty git pickaxe to do some digging.
13:02 Newziky joined #evergreen
13:04 Newziky left #evergreen
13:07 JBoyer If memory serves I was desperate for that holds flag, because the NULL / NOT NULL stuff for requestor_grp and usr_grp make no sense and we're never cared who requested anything; only who gets it. (I'm not saying requestor_grp shouldn't be in the holds matrix, I'm just complaining that it's NOT NULL and usr_grp is...)
13:07 JBoyer And it's silly to just set requestor_grp=(Users) for every single row.
13:08 tsbere JBoyer: I think that flag actually stemmed from when usr wasn't actually checked by the rules, only requestor, so that flipped the two to allow usr to be checked. <_<
13:09 JBoyer tsbere, implying that now it does nothing but cause confusion, or that it doesn't do much?
13:10 tsbere JBoyer: I think it still does the flip, but now that the hold rules can and do check both anyway...
13:10 tsbere Really all it does now is make the hold matrix lie about which is which when set ;)
13:11 mmorgan ... and it does cause confusion.;-)
13:11 tsbere May still be important for those who have rules that depend on it being true, but an upgrade script that swaps the two in the hold matrix when it is turned on could fix that
13:12 miker tsbere: well, it's more complicated than that, unfortunately. it drives the creation of double rules (one for each direction) for some old-time users
13:12 Dyrcona I will comment on the bug, but I think it should be removed from the base schema, and left untouched in existing installations.
13:13 miker so I wouldn't recommend ripping it out entirely, but we could make it an internal flag to hide it from the UI
13:13 miker and, +1 to false-by-default
13:13 JBoyer I'd be all for a big fix that 1. swaps them or not depending on the setting, 2. deletes the setting, 3. makes requestor_grp nullable, and 4. adds a CHECK constraint to make sure that at least one of them is set. (Or, alternately, make usr_grp the NOT NULL field...)
13:14 JBoyer But I'll admit that I barely ever see or think about chmm anymore, so I'm just grousing because it's "ugly." ;)
13:15 * mmorgan agrees that it should be left untouched in existing installations. I wouldn't want an upgrade script to suddenly change the structure of the matrix.
13:19 mmorgan What about a not null constraint on both usr_grp and requestor_grp, but defaulting both to Users?
13:25 JBoyer That ends up working like the check constraint version; at least one is necessary.
14:02 kmlussier joined #evergreen
14:05 krvmga joined #evergreen
14:06 kmlussier joined #evergreen
14:09 Bmagic @weather 65203
14:09 pinesol_green Bmagic: Columbia, MO :: Fog :: 27F/-3C | Wind Chill: 16F/-9C | Friday: Cloudy skies. Patchy freezing drizzle possible. High 36F. Winds SE at 10 to 20 mph. Friday Night: Cloudy skies this evening. A few showers developing late. Low 34F. Winds SE at 10 to 15 mph. Chance of rain 40%. | Updated: 15m ago
14:10 Dyrcona @weather
14:10 pinesol_green Dyrcona: Methuen, MA :: Clear :: 14F/-10C | Wind Chill: 3F/-16C | Friday: Abundant sunshine. High 16F. Winds SW at 5 to 10 mph. Friday Night: Partly cloudy skies early then becoming cloudy with periods of snow late. Low 14F. Winds light and variable. Chance of snow 80%. Snow accumulations less than one inch.
14:10 Dyrcona It warmed up!
14:48 Dyrcona Question about the make_release script: If I do the -b (build only) option, the only other option that I need specify is -j, right?
14:49 Dyrcona Guess I should specify the version (-v), too.
15:06 Stompro Anyone have any suggestions for web meeting software?  I've used join.me before and it seems pretty simple.
15:06 Stompro In a good way...
15:11 bmills Stompro: we do most of our online meetings with GoToMeeting
15:14 Stompro Thanks bmills
15:15 bmills Stompro: the pricing wasn't so bad for the amount of use we get out of it for committee meetings, council meetings, trainings, etc..
16:02 mmorgan joined #evergreen
16:11 StomproJ joined #evergreen
17:01 jvwoolf left #evergreen
17:02 pinesol_green News from qatests: Test Success <http://testing.evergreen-ils.org/~live>
17:07 mmorgan left #evergreen
17:25 jonadab StomproJ: I believe freenode has a web client.
17:26 StomproJ jonadab, is this in reference to my web meeting software question?
17:30 bmills joined #evergreen

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