Evergreen ILS Website

IRC log for #evergreen, 2021-02-25

| 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
01:08 phasefx joined #evergreen
03:44 khuckins joined #evergreen
04:16 khuckins joined #evergreen
04:56 pinesol News from qatests: Failed Installing OpenSRF pre-requisites <http://testing.evergreen-ils.org/~live//arch​ive/2021-02/2021-02-25_04:00:18/test.7.html>
04:56 book` joined #evergreen
04:57 akilsdonk_ joined #evergreen
06:21 khuckins joined #evergreen
07:26 rjackson_isl_hom joined #evergreen
07:48 collum joined #evergreen
07:49 Dyrcona joined #evergreen
07:54 collum joined #evergreen
08:01 mantis1 joined #evergreen
08:28 mmorgan joined #evergreen
08:52 rfrasur joined #evergreen
10:22 Dyrcona JBoyer++ Looks like our comments crossed in the aether.
10:28 Dyrcona I guess the upgrade script renders my comment about using a Perl script moot. However, what I did was more thorough, since it got all gist indexes, not just the ones on a list of certain tables.
10:28 Dyrcona JBoyer++ I like your solution better than mine overall.
10:28 JBoyer Popular bug today. :) I almost agreed with you initially but then the Internet let slip some secrets.
10:29 JBoyer I initially did all gist also, but I figure if someone has added their own they can choose to upgrade them on their own also. (And while I didn't think to see if GIST / GIN supports it, my script does not support multi-column indexes.)
10:30 Dyrcona I saw the email that you set it to confirmed this morning and thought I'd take a shot at it.
10:30 JBoyer Ah.
10:30 sandbergja joined #evergreen
10:30 sandbergja_ joined #evergreen
10:30 sandbergja__ joined #evergreen
10:31 Dyrcona Based on my "research," your current script is complete for the stock indexes. I haven't added any custom GIST indexes.
10:34 Dyrcona JBoyer: I'm 99% sure that neither supports multicolumn indexes, and GIN only works on columns of type tsvector.
10:34 Dyrcona https://www.postgresql.org/do​cs/10/textsearch-indexes.html
10:35 Dyrcona JBoyer: I'll assign the bug to myself and test your upgrade script on my training database over the weekend.
10:35 JBoyer That's convenient then. If some enterprising committer wants to strip out the last AND clause I wouldn't argue. I also think it's really unlikely that anyone would have their own but you never know.
10:35 JBoyer Dyrcona++
10:41 Dyrcona I agree with you that if someone added their own, then they should take responsibility for updating them.
10:49 collum_ joined #evergreen
11:06 khuckins joined #evergreen
11:17 collum joined #evergreen
11:30 miker Dyrcona: just had a thought and wanted to surface it outside my head, and since you've been testing modern PG versions... one thing to consider, especially WRT query speed differences and particular queries, is the new [NOT] MATERIALIZED clause for WITHs (CTEs) in PG 12+.   We've started using WITH a good bit in hand-tooled queries and the behavioral change to fold CTEs when they appear to be side-effect free and are only referenced once could
11:30 miker potentially hurt us.  so, if you run into queries that are slower /and/ use WITH, I'd personally be very interested in those examples
11:50 jihpringle joined #evergreen
12:03 sandbergja___ joined #evergreen
12:04 Dyrcona miker: OK. I'll keep an eye out for WITH queries. The ones that I've noticed so far have negative logic.
12:05 Dyrcona JBoyer: There's an extra semi-colon in your update script. It's no biggie. I'll fix it. Do you mind if I write a release note?
12:05 gmcharlt joined #evergreen
12:07 Dyrcona miker: I have done almost 0 testing on Pg 11 or Pg 13, fwiw. I've mostly looked at Pg 12. Pg 10 seems to perform about the same as Pg 9.6.
12:41 sandbergja_ joined #evergreen
12:42 JBoyer Dyrcona, +1. I guess I hadn't considered a relnote since no features or anything are actually changing; though it is good to know that things should be faster.
12:42 JBoyer And which semicolon since I'm curious?
12:45 JBoyer Oh, in the FOR select, yeah.
12:45 JBoyer A little copypasta there.
12:52 Dyrcona I think a release note is in order to mention that one might want to audit any custom indexes that they've added.
12:57 mantis1 left #evergreen
12:58 mantis1 joined #evergreen
13:05 JBoyer Dyrcona++
13:19 sandbergja Hi Equinox friends: I'm getting "Service open-ils.courses unavailable" when I try to fiddle around with the course materials module on demo.evergreencatalog.com
13:25 JBoyer sandbergja++ I forgot to update opensrf.xml and opensrf_core.xml for that one, should be good now.
13:26 sandbergja JBoyer++
13:26 sandbergja Thanks so much!
13:44 collum joined #evergreen
13:59 sandbergja_ joined #evergreen
15:06 Christineb joined #evergreen
15:29 mantis1 left #evergreen
15:37 RFrasur_ joined #evergreen
15:37 devted_ joined #evergreen
15:37 BigRig joined #evergreen
15:37 phasefx_ joined #evergreen
15:37 abneiman_ joined #evergreen
15:37 JBoyer_ joined #evergreen
15:37 rjackson_isl joined #evergreen
15:37 dluch_ joined #evergreen
15:37 gmcharlt joined #evergreen
15:37 berick joined #evergreen
15:37 bshum joined #evergreen
15:41 jeffdavis joined #evergreen
15:50 malexander joined #evergreen
15:55 jvwoolf joined #evergreen
16:27 collum joined #evergreen
16:58 sandbergja_ joined #evergreen
17:19 mmorgan left #evergreen
17:57 sandbergja_ joined #evergreen
18:00 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
20:48 sandbergja_ joined #evergreen
21:29 pinesol [evergreen|Bill Erickson] LP1913458 Bucket Add/Delete Item Operations Batched - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=1d0e181>
21:45 sandbergja_ joined #evergreen
22:30 malexander joined #evergreen
23:51 sandbergja_ joined #evergreen

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