Evergreen ILS Website

IRC log for #evergreen, 2018-10-16

| 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
00:48 beanjammin joined #evergreen
07:09 agoben joined #evergreen
07:15 rjackson_isl joined #evergreen
07:50 bdljohn joined #evergreen
08:38 mmorgan joined #evergreen
08:38 bos20k joined #evergreen
08:44 Dyrcona joined #evergreen
09:13 JBoyer If anyone with some eg-grid insight sees this today, I'm getting stuck working with a grid to display overdue items.  The gist is that when the interface first loads several columns are blank, and when the grid is refreshed tjey
09:13 JBoyer they're filled in.
09:14 JBoyer I'd like to have all visible columns displayed on first load and am having a hard time tracking down the issue. (assuming it's a race condition of some kind, which is awesome to track down.)
09:18 JBoyer Looks like that's definitely the case because anything that's hidden in the template isn't fetched, while adding the required or removing the hidden attribute makes a column appear. :/
09:22 yboston joined #evergreen
09:23 Dyrcona JBoyer: I'm not an eg-grid expert but I seem to recall some bugs similar to what you're talking about having been fixed in the past few months. Something to do with adding columns and the values not being filled until a reload was done. If you can find those and look at the code changes, that might help.
09:24 JBoyer Dyrcona++
09:24 JBoyer I'll starting arguing with LP search.
09:24 Dyrcona Good luck.
09:29 rhamby JBoyer: I feel like you should receive a blessing from an old holy man for that quest
09:34 mmorgan JBoyer: possibly helpful? lp 1760997
09:34 pinesol Launchpad bug 1760997 in Evergreen 3.0 "Web Client Holds Pull List - Pickup/Request Library Fields Sometimes Blank" [Undecided,Fix released] https://launchpad.net/bugs/1760997
09:37 JBoyer rhamby, I would appreciate an old man in a cave telling me it's dangerous to go alone and offering up a wooden sword.
09:38 rhamby JBoyer: would MDF be close enough to wood?
09:38 JBoyer mmorgan, that's definitely the same issue, but the chosen way to address it was to sprinkle 'required' all over the place. :/ That works but it's only hiding the issue.
09:38 bdljohn joined #evergreen
09:38 JBoyer rhamby, possibly better for smacking things.
09:38 sandbergja joined #evergreen
09:47 mmorgan JBoyer: Ok, so it's using MDF where wood would be better ;-).
09:47 JBoyer mmorgan++ # :D
10:07 Dyrcona @who had the brilliant idea for release notes in point releases.
10:07 pinesol mnsri had the brilliant idea for release notes in point releases.
10:08 Dyrcona @blame Dyrcona
10:08 pinesol Dyrcona: Dyrcona broke Evergreen.
10:08 Dyrcona :D
10:10 Christineb joined #evergreen
10:16 sandbergja Dyrcona++
10:20 berick JBoyer: if a grid column is hidden its data won't be fetched unless it's required.  for auto/fielder grids, anyway.  otherwise the amount of data per grid could be massive.
10:21 berick if the column is visible, of course it should fetch/render
10:24 JBoyer berick, I understand that, I was just hoping the determination of "is this visible or not" could be made after looking up the user's saved column settings. Otherwise you get a customized grid with a bunch of holes in it.
10:25 berick JBoyer: oh, yeah, that should be the case.  that's the intention, anyway.
10:27 JBoyer I don't know if it's not happening for me on 3.2 because I'm missing something (I basically just cloned and /transit/overdue/g 'd the transit list to start) or if it's not quite ready everywhere.
10:28 * berick agrees forcing fields to be required is not the ideal solution
10:31 beanjammin joined #evergreen
10:32 jeff speaking of required fields...
10:33 berick JBoyer: sorry, just catching up, which UI?
10:33 jeff Does anyone else have interest in a level of "requiredness" just below "required" for patron fields?
10:34 jeff Idea being, warn on a blank field but don't force the user to enter data in it. There might be a common method of doing that now, but I don't think there is short of "recommended", which isn't quite the right fit.
10:34 JBoyer A new one, I copied the transit list UI (/circ/transits/list) and am customizing it to show overdue circs instead.
10:35 JBoyer I gave it a unique persist key, saving columns works, everything works fine after reload, just not custom cols on first load.
10:36 Dyrcona jeff: You mean something like "recommended" or "suggested?" It would be a good idea to fill this field in, but not required to do so?
10:36 JBoyer Ah, but you can see the same issue in the transits list as-is. Add Author(normalized) and then save the columns. When you first arrive at the transit list the Author column is blank, if you change the dates or otherwise trigger a refresh it will populate.
10:38 jeff Dyrcona: something almost identical to "required", but without the hard stop on creating/saving. A warning dialog allowing the operation to continue or to go back and fill in the required data, as opposed to an error dialog that gives no option to proceed with the operation.
10:39 jeff Dyrcona: some mail clients warn on empty subject or empty body, but you can then say okay and send the message anyway.
10:40 Dyrcona jeff: If it saves without the field filled in, then it sounds like "recommended fields" to me. :)
10:40 berick JBoyer: ok, good, that's helpful.
10:40 jeff recommended saves silently without bringing the missing field to the user's attention.
10:42 * Dyrcona forgot we had that option already.
10:42 jeff so perhaps "warn on missing recommended fields" would suffice, though there are a lot of "recommended" fields that are not that important. i think most of our libraries use it as a means to cut down clutter of "seldom" fields.
10:46 Dyrcona Well, we could add "suggested" or some similar wording for something in between recommended and required.
10:47 Dyrcona That said some people seem to get confused enough with required fields.
10:47 jeff oh, sorry. pretty sure we don't have both recommended and suggested. i was incorrectly mixing both words to describe the same thing.
10:48 jeff a required field where staff does not have the information at hand to complete the field can lead to either 1) not creating/saving the patron, or 2) entering incorrect data -- which is what had me thinking about error vs warn on the subject.
11:17 khuckins joined #evergreen
11:21 jvwoolf joined #evergreen
11:22 mmorgan Recycling a question from Friday: Does anyone have a procedure for globally removing/forgiving fines from patron accounts?
11:26 sandbergja joined #evergreen
11:29 Dyrcona cesardv: You know you have the power to signoff and push the fix for https://bugs.launchpad.net/evergreen/+bug/1796978, right?
11:29 pinesol Launchpad bug 1796978 in Evergreen 3.2 "Web client cannot create two or more copies at the same time" [High,Confirmed]
11:50 plux joined #evergreen
11:51 aabbee joined #evergreen
11:57 JBoyer berick, something I'm noticing is that grid.js calls grid.loadConfig() and then (hah) a .then sets $scope.columns, but that's where that chain ends. So if link() is called before that .then is called, it won't have the right list of columns.
11:58 JBoyer I haven't been able to work my way around to where link() is first called on page load, so if I'm just shouting randomly into the woods don't mind me.
11:59 berick JBoyer: you're def. in the right territory...
11:59 * berick is unable to look at the moment, but keeping an ear open
11:59 JBoyer ++
12:00 plux JBoyer: That rings a bell here. Having an issue with title displaying in staff client where a. is called
12:13 bdljohn joined #evergreen
12:58 nfburton joined #evergreen
13:03 beanjammin joined #evergreen
13:33 yboston joined #evergreen
13:56 jihpringle joined #evergreen
13:58 jvwoolf1 joined #evergreen
14:00 Bmagic It seems that the new search highlighting tt2 code may have ended up removing extended titles in the record summary page. The code prefers to display the varaiable attrs.hl.title over attrs.title_extended
14:12 berick JBoyer: I have a fix for the grid display issue...
14:13 berick as expected, race condition.  it was always there, just more obvious now with the on-server workstation settings
14:16 bdljohn joined #evergreen
14:20 * berick opens bug
14:42 mmorgan berick++
14:53 kmlussier joined #evergreen
14:58 stephengwills joined #evergreen
15:04 mmorgan1 joined #evergreen
15:14 JBoyer berick++
15:14 JBoyer berick++
15:15 JBoyer I've been looking at things on and off; very glad you had time to knock it out.
15:18 JBoyer I'll test that branch out, but something that occured to me when I was looking earlier is that applyControlFunctions will always only look at the default columns since it's called before the columns are loaded.
15:19 JBoyer Not really knowing where those are used I'm not sure how to even verify that at the moment.
15:20 berick JBoyer: that should be OK
15:20 berick that's most setup for future actions
15:20 berick *mostly
15:21 JBoyer Sounds good then. I didn't figure it would affect this, but wondered if it would matter later.
15:30 khuckins joined #evergreen
16:15 JBoyer bug 1798170 is tested, signed, etc. and all around happier with the order of operations.
16:15 pinesol Launchpad bug 1798170 in Evergreen "Custom grid columns fail to display on initial load" [Undecided,New] https://launchpad.net/bugs/1798170
16:31 * kmlussier shakes her head at JBoyer's Carly Rae Jepsen impersonation on bug 1798170.
16:31 pinesol Launchpad bug 1798170 in Evergreen "Custom grid columns fail to display on initial load" [Undecided,New] https://launchpad.net/bugs/1798170
16:32 berick har
16:37 berick can't help but read it in a Mel Brooks voice
16:38 mmorgan joined #evergreen
16:39 Bmagic what happened to the old copy alert in 3.1. Has anyone migrated the old alerts to the new? or did I miss that step in the postgres upgrade scripts?
16:45 jeffdavis Bmagic: there is a step in the 3.1.0 upgrade script to copy asset.copy.alert_message to the new asset.copy_alert table, 1098
16:53 Bmagic ah!
16:59 jeff yeah, which I recommend XUL-client-users consider deferring. :-)
16:59 berick jeff: do tell...
17:00 jeff going from memory, but I think the new-style copy alerts are a feature that can't straddle XUL / web.
17:00 csharp JBoyer++
17:00 Dyrcona I believe that jeff is correct on that one.
17:03 Bmagic yeah, I am starting to learn that
17:03 Bmagic It seems possible that xul users could make the old style alert though
17:03 Bmagic will the web client respect the old alert?
17:04 jeff and it isn't sufficient to just hold off on using the feature in new ways. the upgrade script migrated legacy copy alerts in a way that made them at least no longer visible / editable by the xul client, even if they continued to fire.
17:05 jeff i do not think we've yet tested if the code supports the legacy copy alert message in the web client, or if we need to revert/change some things there until we're off xul.
17:05 mmorgan left #evergreen
17:05 jeff it was a late discovery in pre-upgrade planning process for us when we jumped from 2.10 to 3.1.
17:06 Dyrcona The test server isn't testing, or is it just not posting to the channel?
17:06 berick jeff: are you just using the browser client for all copy-alert related stuff?
17:07 jeff berick: no, we skipped the migration of copy alerts in the relevant upgrade script, so our legacy alerts remain in-table.
17:07 stephengwills left #evergreen
17:08 jeff i believe that i didn't also need to revert any code, but that's going from memory. i can check notes later this evening.
17:08 berick jeff++
17:08 berick i'm in early stages of plottin a 2.12 -> 3.2 upgrade
17:09 berick and xul will still be in play for some time anway
17:16 * kmlussier confirms that the migration of copy alerts should wait until everyone in the system, particularly catalogers, are using the web client.
17:17 cesardv joined #evergreen
17:18 berick kmlussier: do you recall delaying having negative impacts on the browser client?
17:21 kmlussier berick: Well, neither of the MassLNC consortia are on 3.1 yet, so I don't have production experience. I just know that xul catalogers won't be able to add alerts once they are upgraded. Since cataloging is one area where a lot of people are still using xul, that's a problem.
17:22 berick got it, thanks kmlussier
17:22 jeff I have a local todo to look into it. I have a suspicion that it might be messy. The web client probably sees the alerts at checkin/checkout but probably can't successfully modify them, at least not in a way that doesn't lead to alert message bifurcation. :-)
17:31 cesardv joined #evergreen
17:43 khuckins_ joined #evergreen
17:57 bshum Dyrcona: I'm pretty sure the test server isn't posting to channel because phasefx has to rebuild it out of Wheezy and into Stretch or some other supported distro
17:57 bshum So yeah, no telling how long it's been since the last successful test run :\
17:58 bshum I haven't run one in awhile myself.
18:00 khuckins joined #evergreen
18:36 nfburton joined #evergreen
18:47 Dyrcona bshum: I'm just curious if it is having failures. On rel_3_0, I'm getting live_t/20-hold-targeter.t failing, on master on Ubuntu 18, live_t/09-lp1198465_neg_balances.t fails for me.
18:48 Dyrcona tests+-
19:15 Dyrcona OK. Tests are working for me on master after reloading concerto.
21:51 Bmagic What does it mean when OpenSRF::Transport /usr/local/share/perl/5.22​.1/OpenSRF/Transport.pm:83 Session Error: router@private.localhost/open-ils.serial IS NOT CONNECTED TO THE NETWORK!!! ? Ejabberd is messed up? Which config should I tweak do you think?
22:14 beanjammin joined #evergreen

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