Evergreen ILS Website

IRC log for #evergreen, 2021-12-08

| 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
04:15 alynn26_away joined #evergreen
04:15 alynn26_away joined #evergreen
06:01 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
07:23 rjackson_isl_hom joined #evergreen
07:26 rjackson_isl_hom joined #evergreen
08:38 mmorgan joined #evergreen
08:40 rfrasur joined #evergreen
08:44 mantis joined #evergreen
08:54 Dyrcona joined #evergreen
08:55 jeff jumped from 3.1 to 3.7 last night.
08:59 jeff other than the issues we talked about yesterday (hatch extension needs to be disabled if you don't have the hatch native messaging host installed, sometimes the page doesn't fully load because it doesn't get a response from hatch), it seems we have some workstations where we've needed to remind them of their printers, and patron opt-in doesn't seem to be restricting search results in the way we'd expect.
09:18 collum joined #evergreen
09:18 mmorgan jeff++
09:18 mmorgan Big jump!
09:35 Dyrcona So, it looks like the symspell updates fixed marc_stream_importer.pl on our training server.
09:36 Dyrcona jeff++ We often skip ahead several releases, but not sure I've done that many at once. Must have been a big db upgrade script!
09:38 alynn26 Congratulations jeff
09:44 mantis jeff++
09:46 mantis Had a question about syntax.  Writing in this in a tt2 file for the TPAC.  I have a paragraph inside of a table row that has a link.  I'm trying to wrap them with the translation tag, but when I try to do it in the way I think it should be, it crashes the page.
09:47 mantis ex: [% l('Sentence one') %] <a href=...>[% l('link text') %] </a> [% l('sentence two') %]
09:47 mantis Wanted to see if that syntax was correct
09:58 Dyrcona mantis: That one line looks OK, but the correctness depends on the whole file.
09:59 Dyrcona You might have a missing %] somewhere before that or after.
09:59 Dyrcona END also causes me problems when lots of them are nested.
10:01 csharp_ jeff++
10:05 * Dyrcona forgot just how good Iron Maiden are/were, especially the first Bruce Dickinson period. I have their first 5 albums on vinyl and just repurchased MP3s of the 2015 remastered recordings. (The CD releases suffered from over compression and the "volume wars" of the '90s.)
10:09 mantis Dyrcona: Ok I had an idea that was the case.  This is inside of a table row that is above a map image.
10:09 mantis Looks like there is an END tag here
10:12 jeff You are in a template forest, surrounded by helper functions. There is an END tag here.
10:12 jeff >
10:14 Dyrcona l
10:14 Dyrcona >
10:14 Dyrcona jeff++
10:15 csharp_ I could never get into Iron Maiden - I had friends who were super-fans
10:15 Dyrcona mantis: It might be faster if you could paste the whole file somewhere.
10:17 Dyrcona csharp_: Understandable. It's not everyone's taste. Seems you either LOVE Iron Maiden or you don't like it at all. :)
10:18 Dyrcona jeff: The proper next step is "You are eaten by a Grue." :)
10:20 Dyrcona OK, where was I? ... Right!
10:28 Dyrcona mantis: You also sometimes need to add the END for an included template if it uses WRAPPER(?). I know there are a couple of places in the OPAC where this happens.
10:30 mantis https://pastebin.com/YV2U5yWM
10:33 Dyrcona mantis: Do you localize the catalog or do you plan to do so? If you don't plan to add translations, there isn't much point in using l(). Just sayin'.
10:35 berick mantis: library's
10:35 berick you need to escape the "'"
10:35 Dyrcona mantis: The problem is the apostrophe in "library's catalog" on line 15. Use double quotes for those sentences.
10:35 Dyrcona berick++ Only 'cause I went back to get the line number. ;)
10:36 berick or yeah double quotes are better
10:36 csharp_ syntax color coding helps debugging TT2
10:36 Dyrcona Or, you  could just not translate it.
10:36 Dyrcona You have a TT2 syntax highlighter or it's just generic syntax?
10:36 csharp_ in vim you can just do :set syntax=tt2
10:37 Dyrcona I'll have to see if there's a tt2 mode or equivalent for Emacs. :)
10:37 csharp_ yeah, it's very useful
10:37 mantis AWesome the double quotations worked
10:37 mantis Originally we tried / in place for it
10:37 csharp_ but then it's annoying because the HTML is not color coded
10:37 mantis Same result
10:37 csharp_ I don't like those files :-(
10:37 berick csharp_: there's also tt2html
10:38 csharp_ mantis: cleared your cache?
10:38 Dyrcona i should get mixed modes working again. I used to use it with PHP embedded in HTML.
10:38 berick my .vimrc:  au BufNewFile,BufRead *.tt2 setf tt2html
10:38 csharp_ berick: nice
10:38 csharp_ I didn't know about that
10:38 mantis csharp: yeah it's up
10:38 berick don't remember if I had to install it, though..
10:39 mantis Dyrcona++
10:39 mantis csharp++
10:39 csharp_ berick: didn' thave to install it
10:40 csharp_ berick++
10:40 jeff Hrm. So far I'm just fixing them, but I'll try to reproduce later. It sounds like all/most workstations that were set up to print via Hatch have forgotten how to print until we go into Administration -> Workstation -> Printer Settings and pick a printer for each (relevant) context, and apply. Even the default printer that's listed where when we first visit the Printer Settings interface seems to not work
10:40 jeff until we do that.
10:40 csharp_ that's awesome - I've been living with that for years
10:49 jeff Ah. That's... unwanted. Having email hold notices set on a patron makes the patron email address a required field.
10:49 * jeff digs into that change and the intent behind it
10:52 jeff ah, it seems to have snuck in as part of bug 1570072
10:52 pinesol Launchpad bug 1570072 in Evergreen "Hold request update notification preferences on change" [Wishlist,Fix released] https://launchpad.net/bugs/1570072
10:55 Dyrcona jeff: I think we ran into that and patched it, but I forget how.
10:55 jeff I want that entire feature to have a toggle, but it doesn't seem to.
10:55 Dyrcona Or, there was some data trick to fool it.
10:56 Dyrcona Yeah, more switches are good until your settings look like a WW! uboat. :)
10:56 jeff And I'm reading through the (giant squashed commit) to see if there's a reason why it makes the email field required -- if simply changing that will break something else.
11:09 berick jeff: you can just set them to false -- had to do the same here
11:09 berick https://gist.github.com/berick/9​d4155460f370e83a7f6cee46e024636
11:10 jeff ah, but we use them (the booleans)... so what I'll probably end up doing is removing the ng-required bits on default phone and on email.
11:10 berick also commented out the "if (!notify && !(notify === '')) {..."  block
11:11 * berick nods
11:12 jeff when pushing a change, I've been letting ansible do the npm run build-prod and ng build --prod steps followed by a full-on make install. I suspect that there's a more lightweight way to get the webpack bits in place so that clients notice (after a cache expiry/whatnot).
11:14 berick yeah, i have angbuild.yml and angdeploy.yml plays.  very helpful.  build on all => deploy on all
11:14 jeff probably I saw this or something like it recently: https://gist.github.com/berick/0​cada30cad2aa0d7591d2d6017169f0e
11:15 jvwoolf joined #evergreen
11:32 miker Dyrcona: I was looking at 8657435712c9ba870497515af799d8690f280f70 (xml/xpath changes) and I'm not sure I'm following entirely.  do you have a pointer to the pg change (git, changelog, etc) that's requiring all those changes, or what the failure mode was that you were seeing? (the LP number in the commit headline is from another project)
11:43 mmorgan jeff: bug 1901726
11:43 pinesol Launchpad bug 1901726 in Evergreen "Required Fields Based on Hold Notification Preferences Too Strict" [Medium,Confirmed] https://launchpad.net/bugs/1901726
11:43 jeff mmorgan++
11:44 mmorgan Huge disruption at circ desks when that was introduced.
11:47 Dyrcona miker: Pg 11 and Pg 12 release notes, plus previous IRC discussion. Guess I typoed the commit header. I'll fix it if I get around to it.
11:48 Dyrcona Failure mode is database functions don't work.
11:48 Dyrcona Perl and Pgtap test fail.
11:48 miker heh, well, I figured /that/ much :)
11:49 Dyrcona Well, the commits fix the tests, so there you go.
11:49 Dyrcona I figured this would wait until after next week's dev meeting anyway.
11:49 miker I was looking for an LP bug as I thought that might have error messages, is all. there's a mix of namespace addition and local-name() and I'm trying to understand the details
11:50 miker I mean, sure, it can. I just happened to be in some similar areas
11:51 Dyrcona miker: bug 1937294
11:51 pinesol Launchpad bug 1937294 in Evergreen "Updating Evergreen for Newer PostgreSQL Versions" [Undecided,In progress] https://launchpad.net/bugs/1937294 - Assigned to Jason Stephenson (jstephenson)
11:52 miker thanks
11:53 Dyrcona Also, some of the code was already using local-name(). I went with what seemed simpler/less noisy at the time I did it. If you want it to be all one way, then we can change it.
11:59 jeff on some workstations, I'm finding that we needed to re-enable "use hatch for printing" (extension and native messaging host were all there and good to go). in most cases I think we're also verifying correct printer on each context and hitting apply.
11:59 miker well, local-name() seems simpler for folks not familiar with the details of how xmlns works than extra params for ns mapping, but I don't have a strong /personal/ preference (though, local-name is probably less typing in our use case ;) )
12:00 jeff with other worktations, use hatch for printing is just fine, all set, but we're needing to re-apply printers to print contexts. in many/most cases, the correct printer is associated with the correct context in the UI, but they don't work until we visit that UI (and hit Apply?). Some/all seem to have lost settings like margins.
12:01 jeff I'll try to reproduce later, but it might just be a quirk of the large jump we made.
12:01 rfrasur joined #evergreen
12:01 Dyrcona miker: OK. I'll switch it to local-name(), retest, rebase, and push a new branch later this week. In some places, I thought xmlns was "simpler," but it does look like more code.
12:02 miker Dyrcona: and, I see where ./ -> // comes from, the root-node -> document-node change. that's annoying, as both ./ and // should have worked before, but now only // works ... but it's more correct, so, whatevs. that'll be easy enough to check for both in any code (as you've done ... should others check for more?) and for data in tables
12:03 miker Dyrcona: oh, no, I don't want to put work on your plate right now. just discussing ATM, and I'll pitch in if I end up creating more churn :)
12:03 mmorgan jeff: I think we had reports similar to what you describe after a recent upgrade. Not a big jump like yours, it may have even been a point release.
12:04 miker (fwiw, ./ -> // was the least obviously needed change, to me)
12:05 * mmorgan remembers talking to one of our pc support techs, and hatch printer settings looked fine, but until they hit the Apply, printing failed.
12:05 mmorgan jeff: ^^
12:15 jonadab joined #evergreen
12:18 jeff We also ditched the sms_check constraint on action.request, as we don't need or care about sms_carrier.
12:23 jeff and needed to change some bits on the place hold screen, because sms-related settings that weren't being shown were preventing the form from being submitted. :-)
12:36 Dyrcona miker: I got the obvious things that have tests. More eyes are welcome.
12:37 * miker fires up sed
12:52 mmorgan jeff: bug 1933381
12:52 pinesol Launchpad bug 1933381 in Evergreen 3.6 "Angular: place hold fails with a generic error when Notify by SMS is selected without a carrier" [Low,Fix released] https://launchpad.net/bugs/1933381
12:54 mmorgan Continuing the white screen discussion from yesterday, I have been able to reliably reproduce the white screens. I added the steps to reproduce to the doc I shared yesterday:
12:54 mmorgan https://docs.google.com/document/d/1d5dC1LhgZgPb​XeFDWcmbCjRCotFl3fZSC5-GlJVyBic/edit?usp=sharing
12:54 mmorgan Hoping others can try these steps and see if they can reliably reproduce the white screens.
12:55 Dyrcona Does check in with amnesty mode remove lost bills? I could check the code, but thought someone might just know.
12:56 mmorgan Dyrcona: I believe so, depending on settings.
12:59 Dyrcona mmorgan, thanks! I'll give it a whirl. I'm writing a script to check in and delete long lost copies for one of our members. It should be generic enough to share with the community later.
13:00 * mmorgan checks in a Lost item and confirms that the bill is removed. In our case, the bill is "removed" by an adjustment payment.
13:00 mmorgan Dyrcona++
13:02 jeff I'm fairly certain that today I've seen the "white screen" / "page partially loads, seems not to have received an expected response from hatch" that were were looking at yesterday.
13:04 mmorgan jeff: Can you reliably reproduce it with the steps in the doc I shared a few minutes ago?
13:06 jeff looking!
13:07 jeff I'll give those steps a try. I am fairly certain that in at least one case today I saw the issue on a workstation that *did* have "Use Hatch for printing" enabled.
13:09 mmorgan jeff: I'm pretty sure it does happen with Use Hatch for printing enabled, because the folks we hear from about this are generally on the circ desk.
13:09 Dyrcona Q for the developers in here: When making a cstore request with a flesh, can you specify a filter like when doing a json query join? I want to flesh the circulations on a copy but only the one with an undefined checkin_time.
13:15 Dyrcona To answer my own question, it looks like "No."
13:16 Dyrcona A filter option isn't present in doFieldmapperSearch. Looks like it could be added, but....
13:18 Dyrcona S'allright, I'll just loop over the circulations. There's likely only going to be one anyway since these items were set to lost 7 to 9 years ago.
13:31 Dyrcona Heh. "Amnesty mode" is called "void_overdues" in the back end. :)
13:31 Dyrcona I either never noticed or forgot.
13:39 mmorgan1 joined #evergreen
15:09 Dyrcona Bad arg length for Socket::pack_sockaddr_in, length is 0, should be 4 at /usr/lib/x86_64-linux-gnu/perl/5.26/Socket.pm line 157.
15:10 Dyrcona Got that while running autogen.sh. I don't think that's an Evergreen problem, but I do suspect that my VM may be hosed.
15:10 Dyrcona Maybe my configuration is bad somewhere....
15:11 Dyrcona osrf_control --diagnostic says everything is running.
15:12 Dyrcona Meh. I'll just blow it away and remake it later. I've got others I can use.
15:19 mmorgan left #evergreen
15:39 * Dyrcona wonders if it is wise to use cstore to search for thousands of copies at a time.
15:43 Dyrcona I'm not doing that, but I think my cstore session is timing out before all of work finishes.
15:44 Dyrcona Yeah, definitely. Maybe I should try getting them all in one go. I'm streaming results already.
16:00 Dyrcona Yeah, no. Search for 2,779 barcodes all at once is a bad idea.
16:01 Dyrcona I think a cstore editor is a better idea. It doesn't make a connected session.
16:03 Dyrcona Probably faster to swithc to AppUtils->simplereq.
16:30 jamin joined #evergreen
16:35 jamin Are there any good docs for making proper use of user activity types?  As in creating a new one?  (Which I've done, and have III Patron API working for one org, but... just making up values to stuff in the fields doesn't seem right, and am not surprised to see nothing end up in the user activity table.)
16:36 jvwoolf left #evergreen
18:01 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>

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