| Time |
Nick |
Message |
| 06:00 |
pinesol |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
| 07:14 |
|
agoben joined #evergreen |
| 07:17 |
|
rjackson_isl_hom joined #evergreen |
| 08:00 |
|
rfrasur joined #evergreen |
| 08:11 |
|
sandbergja joined #evergreen |
| 08:14 |
pinesol |
[evergreen|Jane Sandberg] Docs: adding information about 3.6 features from the release notes - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=5e85cc2> |
| 08:32 |
|
mmorgan joined #evergreen |
| 08:39 |
|
mantis1 joined #evergreen |
| 09:01 |
|
Dyrcona joined #evergreen |
| 10:59 |
|
nfBurton joined #evergreen |
| 11:18 |
|
jihpringle joined #evergreen |
| 11:33 |
jeff |
hrm. irc.evergreen-ils.org appears to no longer be indexed by Google. |
| 11:44 |
mmorgan |
jeff: http://irc.evergreen-ils.org/evergreen/2020-02-20#i_436933 |
| 11:46 |
jeff |
mmorgan: ah, perhaps for a while now, then. |
| 11:46 |
jeff |
(and yes, the local search had been broken/disabled for Quite Some Time. I don't recall if it was intentional.) |
| 11:47 |
* jeff |
makes note to look into it |
| 11:47 |
mmorgan |
jeff++ |
| 12:05 |
|
sandbergja joined #evergreen |
| 12:12 |
bshum |
Yikes. Well at least mmorgan found the log from that day with those details. Search indexing on the logs hasn't been done since that day though :\ |
| 12:18 |
mmorgan |
bshum: It was easy to find since it was just about the last thing indexed :) |
| 12:18 |
bshum |
Haha mmorgan++ |
| 12:19 |
bshum |
Honestly the biggest problem is that ilbot (the source for serflog) has been dead for a few years now |
| 12:24 |
jeff |
Hrm. My slow patron search issue is back after adding a number of rows in actor.usr_org_unit_opt_in and this time a vacuum analyze doesn't appear to help any/much. |
| 12:24 |
* jeff |
grabs some queries to explain analyze |
| 12:27 |
* Dyrcona |
waits on lunch to arrive. |
| 12:30 |
bshum |
Alrighty then |
| 12:30 |
bshum |
I managed to recreate my notes from Feb on how to reindex things |
| 12:30 |
bshum |
And we're at least up to date again as of Dyrcona's waiting lunch :) |
| 12:31 |
mmorgan |
bshum++ |
| 12:31 |
bshum |
As far as Google excluding irc logs |
| 12:34 |
bshum |
That's a whole other thing. |
| 12:45 |
Dyrcona |
Ha! |
| 12:45 |
* Dyrcona |
is still waiting. |
| 12:45 |
jeff |
JOIN (SELECT id as id FROM actor.usr u WHERE evergreen.unaccent_and_squash(CAST(email AS text)) ~ '^userexamplecom') AS search ON (search.id = users.id) |
| 12:46 |
jeff |
now I'm off to see why we're doing this weird-to-me self-join on actor.usr instead of having an additional WHERE ... AND evergreen.unaccent_and_squash(CAST(users.email AS text)) ~ '^userexamplecom' |
| 12:47 |
jeff |
it may just be coincidence, but the planner does a better job when I move that from a self-join to WHERE, or when I remove the opt-in check. |
| 12:51 |
jeff |
answer: because of how the query is constructed when we're searching addresses |
| 12:52 |
Dyrcona |
I've seen something like that with opt-in checks before. Query is slow with it, much faster without it, but I have no numbers to back that up. |
| 12:52 |
Dyrcona |
Probably need an index somewhere. |
| 13:15 |
csharp |
bshum++ |
| 13:19 |
jeff |
16 seq scans on actor.usr \o/ |
| 13:19 |
jeff |
passing email through unaccent_and_squash each time, i suspect. |
| 13:20 |
jeff |
(and yes, I tried an index on unaccent_and_squash(email) but that does not appear to have helped significantly. looking to see why.) |
| 13:21 |
Dyrcona |
Accents in email? That's unusual. |
| 13:21 |
Dyrcona |
I'm still waiting for lunch. |
| 13:22 |
csharp |
@band add The Dharma Initiative |
| 13:22 |
pinesol |
csharp: Band 'The Dharma Initiative' added to list |
| 13:23 |
Dyrcona |
Why does "The Dharma Initiative" sound like something that I've heard of before? |
| 13:24 |
berick |
@band add The Dharma & Greg Initiative |
| 13:24 |
pinesol |
berick: Band 'The Dharma & Greg Initiative' added to list |
| 13:24 |
Dyrcona |
berick++ |
| 13:25 |
Dyrcona |
Actually, no. I'm thinking of various Dharma Institutes that I have heard of, not the thing from Lost. |
| 13:25 |
Dyrcona |
And, I think lunch is finally here. |
| 13:50 |
|
mantis1 left #evergreen |
| 13:58 |
|
mantis1 joined #evergreen |
| 14:05 |
Dyrcona |
Python feels like being in limbo at the moment. Some packages are still 2.7; others are 3.0+. |
| 14:11 |
rhamby |
If anyone else wishes to join the Release Manager discussion w are in the Hopin event righ tnow. |
| 14:11 |
|
tlittle joined #evergreen |
| 14:46 |
|
jihpringle joined #evergreen |
| 14:52 |
csharp |
jeff: I'm about to test your ancient fix to bug 1013786 - do you still stand by that patch :-) |
| 14:52 |
pinesol |
Launchpad bug 1013786 in Evergreen "tpac: Check for password strength at login" [Medium,Confirmed] https://launchpad.net/bugs/1013786 - Assigned to Chris Sharp (chrissharp123) |
| 14:53 |
* csharp |
missed a question mark in there somewhere |
| 14:58 |
jeff |
I didn't even think I had code on that bug. Looking! |
| 15:00 |
csharp |
ha! |
| 15:01 |
csharp |
the commit right before yours on that branch references JSPAC removal - such were the times |
| 15:02 |
jeff |
indeed. |
| 15:04 |
jeff |
I think the approach in that branch is still ...mostly sound? |
| 15:05 |
jeff |
I'm trying to think of what's changed since then. I do think making it check an org unit setting (other than the password regex) at this point might be a good idea. |
| 15:07 |
jeff |
I also suspect that the forced redirect has potential to break some (already fragile) things, but I'm not sure there's a great way around that. Maybe just make it a noisy <div>? The code doesn't force a new password... |
| 15:10 |
jeff |
I'd want to see how it works with the bootstrap opac. |
| 15:11 |
jeff |
I'd like to make similar functionality work in the staff client, but that can be a distinct bug. |
| 15:14 |
csharp |
jeff: I rebased your branch and have it up and running on csharp-master.gapines.org (admin/demo123) |
| 15:14 |
csharp |
just in case you're interested |
| 15:15 |
csharp |
there was a conflict in EGCatLoader.pm, but it otherwise picked cleanly |
| 15:53 |
|
sandbergja joined #evergreen |
| 15:59 |
|
jihpringle joined #evergreen |
| 16:01 |
|
mantis1 left #evergreen |
| 16:24 |
bshum |
So admittedly I've been out of the loop for a bit, but I guess we've changed from asciidoc to asciidoctor syntax for some of our docs? Including the README? |
| 16:25 |
bshum |
And that's why I get errors when I try to generate an html using asciidoc against it |
| 16:28 |
jeff |
Antora uses Asciidoctor to parse AsciiDoc, and while they don't come right out and say that there are two different flavors, "Asciidoctor aims to be compliant with the AsciiDoc syntax, but there are some differences to keep in mind." https://asciidoctor.org/docs/asciidoc-asciidoctor-diffs/ |
| 16:28 |
bshum |
Yeah |
| 16:28 |
bshum |
It could be my server |
| 16:29 |
bshum |
But it complained about two lines in README that it didn't understand. And googling led me to asciidoctor |
| 16:29 |
bshum |
So once I installed asciidoctor and ran it against README, it went through cleanly to make a file |
| 16:29 |
|
sandbergja joined #evergreen |
| 16:30 |
bshum |
I'll poke a bit more, just getting the latest and greatest copy of the docs |
| 16:30 |
bshum |
Well just the install readme |
| 16:30 |
bshum |
So rusty :D |
| 16:36 |
pinesol |
[evergreen|Mike Rylander] LP#1879790: Give carousels their own context org - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=ffd0739> |
| 16:55 |
Bmagic |
oh, IRC... right |
| 17:09 |
|
mmorgan left #evergreen |
| 17:15 |
|
dbwells joined #evergreen |
| 17:16 |
|
dbwells_ joined #evergreen |
| 17:44 |
jeffdavis |
seems like the suggested systemd setup for websocketd doesn't work as-is on Ubuntu 18.04 |
| 17:45 |
jeffdavis |
I get this error msg: "The unit files have no installation config (WantedBy, RequiredBy, Also, Alias settings in the [Install] section, and DefaultInstance for template units). This means they are not meant to be enabled using systemctl." |
| 17:45 |
jeffdavis |
using this: https://wiki.evergreen-ils.org/doku.php?id=dev:websockets:gateway:websocketd#optional_systemd_setup |
| 17:48 |
Dyrcona |
Looks like it needs to be updated for more recent systemd versions. That said, I've never used it. I have a shell script to run systemd in the background. |
| 17:53 |
jeffdavis |
Adding an [Install] block to the websocketd-osrf.service file containing "WantedBy=multi-user.target" seems to do the trick, but I'm unsure if that is the correct solution. |
| 17:54 |
Dyrcona |
jeffdavis: That's more or less correct. You could specify other targets, probably make it start before opensrf or apache or whatever. |
| 18:01 |
pinesol |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
| 18:01 |
jeff |
jeffdavis: I think you use org unit opt-in. Have you encountered slow patron searches, especially when searching by email? It may or may not be something that's being caused by our specific ratio of opted-in to not-opted-in users, etc... but thought I'd ask. |
| 18:03 |
jeff |
Where "slow" is > 30 seconds to complete a patron search by email when opt-in is enabled. |
| 18:05 |
jeffdavis |
jeff: we do use opt-in. I don't think patron search is notably slow. I just did a consortium-wide email search and got results in about 4 seconds. |
| 18:06 |
jeffdavis |
About 3% of users have an entry in actor.usr_org_unit_opt_in |
| 18:08 |
jeffdavis |
no custom indexes or anything on that table |
| 18:16 |
jeff |
Thanks for checking! |
| 18:34 |
csharp |
@band add Bladder of Wine |
| 18:34 |
pinesol |
csharp: Band 'Bladder of Wine' added to list |
| 18:34 |
|
dbwells joined #evergreen |
| 18:38 |
|
sandbergja joined #evergreen |
| 22:01 |
|
JBoyer joined #evergreen |
| 23:17 |
|
sandbergja joined #evergreen |
| 23:23 |
|
laurie joined #evergreen |
