Evergreen ILS Website

IRC log for #evergreen, 2020-09-11

| 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
00:52 sandbergja_ joined #evergreen
01:17 sandbergja_ joined #evergreen
06:01 pinesol News from qatests: Failed Installing Angular web client <http://testing.evergreen-ils.org/~live//arch​ive/2020-09/2020-09-11_04:00:23/test.29.html>
06:11 kip joined #evergreen
06:26 agoben joined #evergreen
06:37 oleonard joined #evergreen
06:58 JBoyer gmcharlt++ # I thought I had switched everything over to that uri but evidently not.
07:21 rjackson_isl_hom joined #evergreen
07:41 collum joined #evergreen
07:43 JBoyer fyi: I just built a fresh master and didn't see that angular error, so I'm not sure what's up with that, maybe the qa box didn't get all of the Acq commits pushed around 4:30.
08:04 Dyrcona joined #evergreen
08:19 rfrasur joined #evergreen
08:32 mantis1 joined #evergreen
09:01 terranm joined #evergreen
09:07 Dyrcona I'm running something to remove URIs from bib records by vendor and owning library against a Pg 12 database. It appears that updating the URI call number maps is very slow on Pg 12.
09:10 Dyrcona It just took 55 seconds to update 1 record. I used a stopwatch to time it.
09:12 Dyrcona So, 859 hours to finish this batch. I think I'll stop the test.
09:16 terranm Yikes
09:24 Dyrcona I'm going to add code to output the time to update records and run this on our training server which is also Pg 12. Later, I plan to go manually replicate the updates and run them with plprofiler to see where the time goes.
09:25 Dyrcona BTW, running on a Pg 9.6 database on the same server, I just clocked 1 record update at 9.58 seconds, so I definitely think that Pg 12 doesn't like something that we do with bib functions.
09:26 Dyrcona And, FWIW, the Pg 12 database is optimized, the Pg 9.6 database is default configuration.
09:27 Dyrcona However, Pg 12 isn't fully optimized. I left some headroom for the other Pg instances.
09:29 Dyrcona It is still gonna take about 6 days to delete all of the URIs, assuming the time holds up.
09:32 alynn26 joined #evergreen
09:34 Dyrcona The update time on the 9.6 database is down to about 5 seconds, now, but I suspect different records will take different times depending. That's why I am going to add timing code.
09:35 Bmagic JBoyer: the reason the generate_docs.pl used rsync is that I've seen issues when building antora directly into the production directory. Pages get broken during the process. Old pages don't get deleted, etc.
09:37 JBoyer I also wiped out the html_output dir when cleaning tmp_space, so that shouldn't be an issue.
09:37 Bmagic cool
10:07 Dyrcona I am consistently seeing much longer times on the training database than I am on the Pg 9.6 database on lesser hardware with default settings for these updates.
10:07 Dyrcona Like 10x longer on Pg 12.
10:17 Dyrcona Based on what we've seen, I can't recommend Pg 12 in production at this time. :)
10:22 * Dyrcona plans to revert training back to Pg 9.6.
10:49 csharp Dyrcona++ # pioneering
10:56 gmcharlt Dyrcona++
10:58 * gmcharlt claims 1227
10:58 terranm joined #evergreen
10:59 jvwoolf joined #evergreen
11:01 pinesol Showing latest 5 of 8 commits to Evergreen...
11:01 pinesol [evergreen|Bill Erickson] LP1879335 Retrieve linked bibs on demand - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=ec01b3f>
11:01 pinesol [evergreen|Galen Charlton] LP#1879335: (follow-up) tweak sorting of bib list - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=991cbac>
11:01 pinesol [evergreen|Galen Charlton] LP#1879335: (follow-up) improve extraction of thesaurus - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=d6fca5e>
11:01 pinesol [evergreen|Bill Erickson] LP1879335 Tweak bib-list sorting continued - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=acfc0fd>
11:01 pinesol [evergreen|Galen Charlton] LP#1879335: stamp schema update - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=b51b2c3>
11:08 * gmcharlt claims 1228
11:13 pinesol [evergreen|Mike Rylander] LP#1851413: Restore assumed order of full_path - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=cbef8a4>
11:13 pinesol [evergreen|Galen Charlton] LP#1851413: stamp schema update - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=3cd5d35>
11:17 jvwoolf joined #evergreen
12:00 jihpringle joined #evergreen
12:07 akilsdonk joined #evergreen
12:08 nfBurton joined #evergreen
12:23 rjackson_isl_hom joined #evergreen
12:53 sticks joined #evergreen
12:53 sticks Hello! I'm installing through git using the url git.evergreen-ils.org/Evergreen.git
12:53 sticks but it gives me a 404 when I try.
12:54 sticks Any ideas?
13:02 sandbergja joined #evergreen
13:03 JBoyer sandbergja, I had a question about your course reserves test server. When I use the upgrade script on an existing server marc edits fail because of the new boolean_facet metabib class. Did you do anything to your server to get around that, or is there drift between the upgrade script and how your server was built?
13:04 berick sticks: according to the site, the URL you want is git://git.evergreen-ils.org/Evergreen.git
13:05 JBoyer sticks, I'm not sure that git.evergreen-ils.org is setup to do http/s pulls with the same urls as git://. Cloning git.evergreen-ils.org/Evergreen.git should work. ALSO, I'd really recommend just using one of the tarball releases on evergreen-ils.org so you can skip all of the optional parts
13:05 Dyrcona berick++ # As I was typing the same thing, more or less.
13:05 JBoyer Eww. I think my client munged that up. Anyway, try the url berick posted, but srsly, grab a tarball instead.
13:06 Dyrcona tarball-- :)
13:07 JBoyer I didn't think you enjoyed npm all that much. :p
13:07 sandbergja JBoyer: oh no!  I built that server using the seed data, rather than upgrading.  I bet there is something wrong with the upgrade script; did you spot anything helpful in the logs?
13:07 Dyrcona I don't but, I have lots of custom modifications.
13:08 sandbergja Dyrcona can get a tarball with just the node_modules folders
13:09 jeff ERROR:  COPY delimiter must be a single one-byte character
13:09 jeff fine. sed it is!
13:10 Dyrcona jeff: sed is Turing complete.
13:10 Dyrcona @praise search Turing
13:10 pinesol Dyrcona: No matching praises were found.
13:10 JBoyer sandbergja, I know the upgrade script has issues, I just pushed a fix to correct some of them. :) but my pg logs do show that it's missing at least metabib.boolean_facet_field_entry though my scratch-built database is also.
13:11 sandbergja JBoyer: oh, I just saw that!  thanks
13:11 JBoyer Though if your db does have that table it would explain why you're able to edit records.
13:12 JBoyer I'm not completely sold on going to all of that trouble to get a yes/no facet either.
13:12 Dyrcona @praise add $who is Turing complete
13:12 pinesol Dyrcona: The operation succeeded.  Praise #18 added.
13:13 sandbergja JBoyer: fair
13:13 sandbergja I'm only like 90% sold on it myself
13:13 sandbergja I could see other boolean facets in the future being helpful (e.g. Published in the last 5 years? y/n)
13:15 JBoyer That one sounds even worse when you consider these things are (normally) only ever updated when the record is. :) We don't need a cron job poking every marc record on a regular basis.
13:15 sandbergja Fair
13:15 sandbergja helpful for patrons
13:15 sandbergja not helpful for us, though
13:15 sandbergja :-)
13:16 sticks berick: your playbook yml files on clone for operation timed out, I've had to comment out the clone and get it manualy.
13:17 JBoyer I suppose the way the information is used is the tricky bit. Facets aren't currently setup in a way that makes that easy, though it is possible to search based on pub year. (I don't recall if it does greater than or only between though.)
13:18 sandbergja JBoyer: I confirmed that my server doesn't have a metabib.boolean_facet_field_entry table
13:18 sandbergja and that I can edit MARC records without issue
13:20 JBoyer is the course_reserves_facet trigger active on your biblio.record_entry?
13:24 sandbergja Yeah, it is
13:24 * JBoyer is confused then.
13:25 sandbergja it's been storing all of the y/n values in metabib.facet_field_entry
13:25 sandbergja ^ metabib.facet_entry
13:27 JBoyer through a one-off special purpose asset.* function, yes. But the rest of the ingest system works from the list of config.metabib_classes and builds the table names automatically.
13:28 JBoyer And one step in that process is DELETE FROM metabib.boolean_facet_field_entry WHERE  ...
13:28 sandbergja oh, this is sounding familiar actually
13:28 sandbergja let me see if I can dust off some memories
13:35 jvwoolf joined #evergreen
13:35 Dyrcona csharp: I know I've asked you before, but what are your plans for Pg 10 in production? Are you planning to upgrade soon? Have you already done so?
13:41 sandbergja JBoyer: On line 1076 of 030.schema.metabib, it should exclude the boolean_facet from the DELETE FROM metabib.[class name]
13:41 sandbergja (which apparently never made it into the upgrade script)
13:41 sandbergja https://git.evergreen-ils.org/?p=working/Evergreen​.git;a=blob;f=Open-ILS/src/sql/Pg/030.schema.metab​ib.sql;h=ac2d7f0eee1a6cc4ba7f0af26ca1917d61343555;​hb=4299088805334a8ceeee402c756a292b9c597779#l1076
13:43 JBoyer That does help explain things. :)
13:43 JBoyer sandbergja++
13:43 sandbergja JBoyer++ # SQL detective
13:44 sandbergja I can throw the updated function into the upgrade script, or I can hold off
13:45 sandbergja If you have a recommendation about a better way to handle those facets, I could do that instead :-)
13:45 JBoyer Well you may as well since the upgrade script as-is leaves systems broken otherwise.
13:46 sandbergja fair enough!
13:46 JBoyer I wish I did have a better way to handle that off-hand, but I don't yet. Knowing how much had to be done to squeeze them in to this one use case has me wanting to find *something* though. ;)
14:00 pinesol [evergreen|Bill Erickson] LP1880726 MARC Batch edit Angular port - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=90c205f>
14:00 sandbergja JBoyer: thanks for working through this.  I just pushed the updated upgrade script to that collab branch
14:00 pinesol [evergreen|Mike Rylander] LP1880726: Support eg2 cookie - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=1e0c4ef>
14:00 JBoyer sandbergja++
14:04 jihpringle joined #evergreen
14:06 sandbergja JBoyer: just so you know, I will probably push one more commit to that branch later today (there are some console errors in the staff catalog when the course module library setting is enabled -- I wanna get rid of those)
14:07 JBoyer Sounds good to me. There are some other branches I can checkout (hey-o) in the mean time.
14:11 berick @rimshot
14:11 pinesol berick: What we have here is a failure to communicate.
14:18 berick @blame [who] for liking the band [band]
14:18 pinesol berick: but I never knew until this day that it was (who [<channel>] <question>) -- Answers <question> with a random nick from <channel>. <channel> is only necessary if the message isn't sent in the channel itself. all along for liking the band Ejabberd Confit
14:18 berick heh
14:20 pinesol [evergreen|Terran McCanna] LP 1884657 Improve Access to Library Info in OPAC - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=60b314e>
15:20 csharp Dyrcona: still on 9.6 - this month our priority is moving from Ubuntu 16.04 to 18.04, then in October or November I was planning to upgrade PG to 10
15:24 Dyrcona csharp: Have you tested the performance of Pg 10 thoroughly? We ran into issues with Pg 12 and performance.
15:30 csharp Dyrcona: nope - not yet - I saw your comments though and noted them with some alarm
15:31 Dyrcona Some things are faster but others seem way too slow.
15:31 Dyrcona I hope to get to do some profiling next week.
15:34 Dyrcona Unfortunately, I spent a lot of this week on https://xkcd.com/1739/
15:38 terranm :)
15:42 mantis1 left #evergreen
15:50 sandbergja JBoyer: pushed my change to the course materials branch
15:50 sandbergja thanks again for reviewing it!
15:50 sandbergja JBoyer++
15:52 * gmcharlt grabs 1229
15:52 JBoyer sandbergja++
15:52 JBoyer I won't have time anymore this week to look into it but happy to help.
15:58 pinesol Showing latest 5 of 10 commits to Evergreen...
15:58 pinesol [evergreen|Kyle Huckins] lp1777677 Security tweaks - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=61e9cf8>
15:58 pinesol [evergreen|Terran McCanna] LP#1777677 Test Notification Method - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=7135654>
15:58 pinesol [evergreen|Jeff Davis] LP#1777677: restrict allowable hooks for test notifications - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=b68eca1>
15:58 pinesol [evergreen|Galen Charlton] LP#1777677: (follow-up) adjust test message buttons in registration form - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=ff1ddd6>
15:58 pinesol [evergreen|Galen Charlton] LP#1777677: stamp schema update - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=ab774e4>
15:59 * Dyrcona *poof*
16:07 Bmagic JBoyer: is there a reason why the public facing documentation is down atm? http://docs.evergreen-ils.org/dev/
16:08 JBoyer No clue; I don't know who runs that server.
16:08 Bmagic Thought I would ask you because I got the sense that you were working on it...
16:08 Bmagic maybe the nightly build broke on master?
16:08 Bmagic since the merge (probably)
16:08 sandbergja Bmagic: I bet it did
16:09 Bmagic yeah, I am just now putting that together
16:09 JBoyer Could be. Looks like it's still hosted at Mohawk college
16:09 sandbergja that's a legacy server we are moving away from
16:09 sandbergja But there are relatively recent versions of the docs here: http://docs-testing.evergreen-ils.org/
16:09 sandbergja (current as of yesterday)
16:10 Bmagic thanks!
16:10 Bmagic any thoughts on how to resolve this? Maybe just change the link to the antora site?
16:13 JBoyer claiming 1230 for analysis
16:14 Bmagic I'm going to try to build a fresh antora set from current master
16:18 Bmagic it worked! - though, the header is busted, not sure where that got lost along the way. I will find it
16:20 jvwoolf joined #evergreen
16:22 pinesol [evergreen|Dan Scott] LP1517298: Basic Matomo Analytics support for tt2 files - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=f62baec>
16:22 pinesol [evergreen|Rogan Hamby] LP1517298: Matomo Support with library org settings - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=3dd8c7d>
16:22 pinesol [evergreen|Jane Sandberg] LP1517298: Simplifying and cleaning up matomo implementation - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=2843ab8>
16:22 pinesol [evergreen|Jason Boyer] LP1517298: Stamp Upgrade Script - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=26bac01>
16:24 JBoyer Bmagic, busted how? I pulled in the partial header.hbs from the ui-lunr stuff, so (as I understand it) that may be overriding the files in eg-antora.
16:24 JBoyer If that file has been significantly changed, that is.
16:24 Bmagic we have some custom stuff up there
16:25 Bmagic dropdown menus and whatnot
16:25 JBoyer What might be best then is to make sure all of the stuff that's needed in ui/ui-lunr is already in eg-antora and tell people they're on their own to make sure that their custom antora-ui repos include the search bits.
16:26 Bmagic JBoyer: this commit tells the most of the story https://git.evergreen-ils.org/?p=eg-antora.git;a=​commit;h=f8dca04b2eccb0dbe4c36f04ebc4d41e0256b1ea
16:28 JBoyer and remove docs/ui from Evergreen proper
16:28 Bmagic I think the Evergreen/docs/ui/ui-lunr/p​artials/header-content.hbs is overriding the one from our eg-antora repo
16:29 JBoyer I know it is. What I'm saying is that anything that we need for lunr should be put directly into eg-antora and the docs/ui directory in Evergreen just thrown away.
16:29 JBoyer So are the other 2-3 files under docs/ui/ui-lunr already in eg-antora with the lunr changes that they need in place?
16:29 Bmagic fair enough - if we delete that file, it breaks the auto-builder?
16:30 JBoyer Not if we remove / comment out the supplemental part and tell people if they want to use a custom antora-ui that they need to include the lunr bits themselves.
16:31 Bmagic I'd like to try that out to make sure - but sure, sounds good to me
16:32 Bmagic the trouble is, since we are building on master, and antora automates the git checkout, any changes I do to the file system is just overwritten by antora during build time. So, the expierement will require updates to master. Or I can make working branch and update site.yml to use that one
16:34 sandbergja JBoyer++ # matomo
16:41 JBoyer Bmagic, it looks like what I was talking about has already been done, there are references to search in eg-antora already, so it may be enough to just take out the supplemental bits and it will work. I'll try it out.
16:41 Bmagic the files are identical?
16:42 Bmagic they're not
16:42 Bmagic also, there is a css file and js files under docs/ui/ui-lunr
16:42 JBoyer Well, it looks like 1 or 2 files have some lunr bits in them, the others are missing.
16:43 Bmagic why did the lack of header-content.hbs in docs/ui/ui-lunr/partials break the builder?
16:45 JBoyer I was using the stock antora-ui checkout since that was the only one mentioned in the docs. That file being missing didn't "break" the build, but it didn't have the search stuff either.
16:46 JBoyer So I put the missing file in place, not knowing that would break eg-antora. So the fix is to make sure that eg-antora has everything in it that it needs from the start.
16:59 JBoyer As for testing; using file://path/to/some/repo should be perfectly valid as a git repo path, so nothing needs to change on the git server to test an updated eg-antora repo.
17:07 JBoyer Sigh. Well, one fun thing about my plan is that the lunr folks have apparently never run anything through a linter, so antora's lint step *hates* it.
17:34 gmcharlt I've signed off on the Bootstrap OPAC branch - bug 1778972
17:34 pinesol Launchpad bug 1778972 in Evergreen "Wishlist - OPAC Redesign" [Wishlist,Confirmed] https://launchpad.net/bugs/1778972 - Assigned to Galen Charlton (gmc)
17:34 terranm woo!
17:35 gmcharlt there are undoubtedly going to be a lot of tweaks, but it's installable, now visibale to I18N subsystem, and hasn't broken for me
17:35 gmcharlt now requesting that another committer do a final pass, particuarly on the installation and I18N bits, and merge
17:39 sandbergja I could take a look at it tomorrow
17:39 gmcharlt sandbergja: thanks!
17:39 sandbergja gmcharlt: okay if I assign myself in lp?
17:39 gmcharlt sure
17:39 Bmagic JBoyer: I just had a few minutes and made a branch where I could play. I deleted the header file from the evergreen repo and I can confirm that it solves it for the site at least. https://eg-docs.georgialibraries.org/prod/  - but I know what you are saying. We would prefer to eliminate the override from the Evergreen repo completely, and overlay the files directly on the eg-antora repo. That'll take a little bit of time. doing that now
17:54 sandbergja gmcharlt: I am not sure about the timeline for making the bootstrap OPAC the default in the next major release (nothing against the new OPAC, but it is a lot of work to make sure all local customizations work, etc., especially during a pandemic)
17:54 gmcharlt sandbergja: default != TPAC goes away
17:55 sandbergja I'm wondering if the release notes could be a little more vague?
17:58 gmcharlt tweaking the wording to be explicit that TPAC won't go away in 3.7 is fine by me
17:58 JBoyer Bmagic, I already have it handled, just had to tell eslint to look the other way re: lunr.js.
17:59 JBoyer If I can upload to that repo, that is. Otherwise I can throw you a couple patch files.
17:59 sandbergja gmcharlt: I'll do that, thanks
17:59 gmcharlt but I feel pretty strongly that we need to push for making the new skin default by 3.7 lest we end up in a position where it gets wasted for lack of maintenance
18:00 terranm Unless we find any major issues, we're planning on moving to it in our 3.6 upgrade in January so I'm sure we'll be finding bugs :)
18:00 pinesol News from qatests: Failed Installing Angular web client <http://testing.evergreen-ils.org/~live//arch​ive/2020-09/2020-09-11_16:00:02/test.29.html>
18:01 sandbergja terranm: that makes me feel better!
18:02 gmcharlt sandbergja: (as a bit of historical context, the project did not have a great experience a few years back trying to maintain a parallel OPAC skin for JSPac called "Craftsman")
18:02 * JBoyer shudders at its name
18:02 sandbergja gmcharlt: that is good to know
18:04 sandbergja I still might argue for a longer timeline for getting rid of the current TPAC, but that can wait until a developers meeting
18:07 gmcharlt indeed
18:17 JBoyer Bmagic, apparently I have write access to the eg-antora repo. I've added the necessary bits and committed them with very minor changes. I'll fix up Evergreen to match.
18:21 JBoyer And that's now done. Time to fade away into the weekend.
18:23 pinesol [evergreen|Jason Boyer] Antora build update - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=e1c752d>
18:23 sandbergja JBoyer: enjoy your weekend!
18:55 jvwoolf joined #evergreen
19:04 miker joined #evergreen
19:27 terranm joined #evergreen
22:27 sandbergja joined #evergreen

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