Evergreen ILS Website

IRC log for #evergreen, 2019-03-20

| 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
01:21 sandbergja_ joined #evergreen
05:01 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
06:48 agoben joined #evergreen
07:11 rjackson_isl joined #evergreen
07:36 bdljohn joined #evergreen
07:38 bdljohn joined #evergreen
08:21 bos20k joined #evergreen
08:46 mmorgan joined #evergreen
09:29 yboston joined #evergreen
09:38 sandbergja_ joined #evergreen
10:18 Dyrcona joined #evergreen
10:47 csharp things seem a bit happier this morning with round-robin scheduling in place
10:47 csharp we'll see what happens this afternoon
10:48 berick csharp++
10:49 berick csharp: which websocket setup?
11:00 csharp websocketd
11:00 csharp when I confirmed that wasn't the likely cause, I switched back
11:02 berick cool
11:02 * berick leaves his ears on
11:11 pinesol [evergreen|Galen Charlton] LP#1709698: fix error made when stamping DB update - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=f497f41>
11:11 pinesol [evergreen|Galen Charlton] LP#1780639: fix error made when stamping DB update - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=8706a27>
11:27 aabbee this isn't a bug, and i don't know if this sort of comment is considered helpful, but i stumbled across the compare_array function (https://git.evergreen-ils.org/?p=Evergreen.git​;a=blob;f=Open-ILS/web/js/ui/default/staff/cat​/services/holdings.js;h=d78c049bc3fda634839600​fcc802d47d635d3f99;hb=refs/heads/master#l92) and thought the implementation could be cleaner (or at least shorter):
11:27 aabbee function compare_array (x, y, i) {
11:27 aabbee if(i >= x.length || i >= y.length){ return y.length - x.length; }
11:27 aabbee return x[i].localeCompare(y[i]) || compare_array(x,y,i+1);
11:27 aabbee }
11:28 Dyrcona aabbee: That is a bug if you think it would be an improvement.
11:28 csharp gmcharlt++ # eagle eyes
11:29 csharp csharp-- # too hasty sometimes
11:30 aabbee "improvement" is highly subjective. it's not faster or more memory efficient. there's nothing functionally wrong with the existing implementation. i guess that's what i'm asking: does anyone consider this an improvement, or am i just making noise?
11:31 Dyrcona I haven't looked at the current implementation, but I generally think less code is better, unless it is a case of iteration versus recursion and recursion leads to a performance hit.
11:31 aabbee they're both recursive.
11:32 khuckins joined #evergreen
11:33 Dyrcona I think your version is an improvement.
11:34 aabbee hey, thanks! :-) i'll create a wishlist bug for it.
11:36 Christineb joined #evergreen
12:22 jihpringle joined #evergreen
12:43 csharp autogen.sh--
13:13 nfBurton joined #evergreen
13:20 * mmorgan sometimes wishes there was an autogen.sh for humans.
13:21 Dyrcona :)
13:22 JBoyer \me, a curmudgeon: "Hey, weird-o! Reset yourself." The kid on his lawn stands up straight, says "Good day sir!" and vacates the lawn with much haste.
13:22 JBoyer Thanks slashes.
13:50 bdljohn left #evergreen
13:50 bdljohn joined #evergreen
13:53 bdljohn joined #evergreen
14:31 eby joined #evergreen
14:34 sandbergja remingtron++ # closing merged pull requests in Github
14:35 kipd joined #evergreen
14:35 remingtron sandbergja++ #merging pull requests on Github
14:37 sandbergja remingtron++ # teaching me how to merge Github PRs
14:37 sandbergja (I just have to have the last word :-P)
14:39 remingtron sandbergja: I'll never argue with a karma battle
14:44 Dyrcona remingtron++ sandbergja++
14:51 berick finally we get to call the karma police
14:53 khuckins joined #evergreen
14:53 Dyrcona :)
14:53 dbwells ha
15:12 yboston joined #evergreen
15:13 pinesol [evergreen|Bill Erickson] LP1776736 Record merge marc edit repairs - <http://git.evergreen-ils.org/?p=​Evergreen.git;a=commit;h=a2b8f7b>
15:14 csharp things appear much more stable today - we put round-robin into place and upped the max_children for pcrud (reducing the max_children for cstore at the same time)
15:15 Dyrcona csharp: Good to hear. The web staff client uses pcrud a lot more than the old staff client or other services did.
15:24 berick csharp: cool
15:32 dbwells berick: reviewed bug #1776736, but it didn't apply cleanly to 3.1, so a rebase is there if you feel like taking a look.
15:32 pinesol Launchpad bug 1776736 in Evergreen 3.1 "Web client: Edit function in merge screen does not keep edits" [Undecided,Confirmed] https://launchpad.net/bugs/1776736
15:32 berick dbwells++
15:32 berick thanks
15:40 phasefx Stompro: Dyrcona: the call number pull list fix (lp1749502), for me, if you do Print Full List twice in a row, it renders the affix id's and not the labels
15:41 csharp ok, so the next step, I think, is to implement cross-domain registration in a test environment.  Does anyone out there have docs or examples I could use for this?  I have a high-level understanding of what needs to happen, but I'm not sure how to actually do it
15:42 * csharp steps away for a bit but will be back soon
15:49 sandbergja_ joined #evergreen
15:50 berick csharp: take w/ grain of salt cuz I'm not using it, but IIRC you just have to add public and private <router> entries in opensrf_core.xml per each host
15:50 berick that's assuming the bricks can already route to each other and jabber is using non-localhost domains (e.g. private.brick01, public.brick01, etc.)
15:52 Dyrcona phasefx: I see that one some rows but not others. I'm getting -1 on some prefixes.
15:52 Dyrcona Also, doing it a third time, everything looks OK.
15:53 Stompro phasefx, Interesting, I'll test that out.
15:55 phasefx I imagine most folks will only invoke it once per page load, but my chaos powers are compulsive
15:55 Dyrcona Doesn't seem to be much rhyme nor reason when it happens and if it it is prefix, suffix, or both.
15:55 Dyrcona Well, if you do it again and again you get different results each time, at least I seem to.
15:58 Stompro phasefx, I cannot replicate that.  I just loaded the pull list, then printed it (it looked fine), then clicked print again and it looked the same.  Using Chrome.
15:58 phasefx Stompro: with the Print Full List action or a grid print?
15:59 Stompro phasefx, print full list.
15:59 phasefx I'll share a server with you if you want to see it in action
16:01 Dyrcona I'm seeing it on our training server with 3.2 installed.
16:01 Stompro phasefx, now I see it.  If I double click on the print full list... or just keep clicking. The first print preview pops up, then I close that, then the next one pops up and shows affix IDs for some of the items.  Not all.
16:03 phasefx cool deal
16:05 csharp berick: thanks - I'll give it a shot
16:05 Stompro phasefx, maybe multiple instances of the print full list are stepping on each other, trying to update the same info?  Maybe we just disable the print full list after the first click until it is done?
16:07 phasefx Stompro: for me, it didn't seem to be concurrent pop-ups that showed the bug, but open close open close
16:08 Stompro phasefx, have you already taken down your dev system?
16:09 Stompro phasefx, could I try there quick?
16:09 phasefx Stompro: sure, it's back up, but with an extra commit beyond master
16:09 phasefx in the pull list grid UI, no less :) but shouldn't affect the bug
16:10 Dyrcona Stompro: I saw the same thing as phasefx. I could open it, look at it, close, then open it again.
16:12 Stompro phasefx, yep, I see it too on your test system.  Hmm, I'll check to see what I missed when creating the commit.
16:15 Dyrcona I'm only seeing it with affixes of -1. phasefx did you see it with others?
16:15 phasefx Dyrcona: yes, the REF one from concerto
16:16 Dyrcona OK. Could be that the list I was looking at had no actual affixes.
16:16 Dyrcona I was using production data with a small library.
16:16 phasefx I created a pull list in concerto by hitting up 3 bibs with copy level holds, and specifically modifying the item for the 2nd hold to have a prefix and suffix
16:17 Dyrcona Good way to test it. I just logged in and brought up the pull list. :)
16:17 phasefx I also played with location affixes too, because I forgot those were just for label printing
16:19 * phasefx is poking at lp1779316 and wanted to see what changes were made recently
16:19 Dyrcona I wonder if fleshing the affixes would be better than going back and retrieving them? We seem to do a lot of pcrud lookups in the web staff client for that sort of thing.
16:20 Dyrcona Lp1779316
16:20 Dyrcona Hey! Where's pinesol?
16:21 berick @who silenced pinesol
16:21 pinesol ericar silenced pinesol.
16:22 Dyrcona Oh, I forget that pinesol is out of order in my list of people in the room.
16:22 phasefx bug 1779316
16:22 pinesol Launchpad bug 1779316 in Evergreen "Web Client: Pull List Call Number include Prefix / Suffix" [Undecided,New] https://launchpad.net/bugs/1779316 - Assigned to Jason Etheridge (phasefx)
16:22 Dyrcona I thought Lp was supposed to work, too.
16:25 csharp still seeing actor drones get maxxed.  I upped them from 72 to 96, but they maxxed out on one server pretty quickly (I guess it's like adding highway lanes not improving traffic) - still seeing a single brick get pegged while others are underutilized - I guess cross-server xmpp is the only/best solution?
16:27 berick yeesh.  seen any patterns on what they are doing?
16:27 berick i wonder if we just have an interface or two that are spewing an unreasonable number of requests
16:27 Dyrcona The auth drones are C. One could actually attach a debugger to them.
16:28 berick there are places where a large grid will render w/ one or more API calls per row getting fired
16:28 berick Dyrcona: open-ils.actor
16:28 Dyrcona Oh... Time to call it a day, I think. :)
16:29 bshum Dyrcona: LP does work, but only if you have a space between ;)
16:29 bshum LP 1779316
16:29 pinesol Launchpad bug 1779316 in Evergreen "Web Client: Pull List Call Number include Prefix / Suffix" [Undecided,New] https://launchpad.net/bugs/1779316 - Assigned to Jason Etheridge (phasefx)
16:29 berick csharp: would be useful if you could check activity log for one of the explody bricks, fitler by open-ils.actor, and see if there are chunks of calls coming from a single client (e.g. by authtoken)
16:30 berick if you get a huge humber of calls (especially if they're the same api call) in a very short period of time, there may be a UI we can fix
16:33 csharp berick: will do
16:36 csharp berick: at first blush I'm seeing a lot of open-ils.actor.settings.retrieve.atomic calls
16:37 csharp also open-ils.actor open-ils.actor.settings.apply.user_or_ws
16:38 csharp open-ils.actor.ou_setting.ancestor_default.batch
16:39 csharp also open-ils.actor.user.itemsout.notices
16:39 csharp lots of each of those, but not so many that any one of them is obviously the cause
16:42 csharp but yeah, after trawling through about 20 minutes worth of activity logs, looks like it's mostly settings checks
16:45 berick but no clear cases of, say, 50 calls from a single client within a few seconds?
16:49 csharp fwiw, pcrud is no longer maxxing since I upped those limits, so we are better overall
16:50 csharp and only one brick is maxxing with actor at the moment
16:57 sandbergja__ joined #evergreen
17:00 sandbergja__ Hey all -- I've got Evergreen running on localhost, using docker.  But Firefox doesn't like weird ports like 7682 under a self-signed certificate. Does anybody know if there's a way I can adjust about:config so that Firefox isn't so picky about that?
17:00 sandbergja__ FWIW, I can use Chromium just fine.
17:02 pinesol News from qatests: Failed Running pgTAP tests <http://testing.evergreen-ils.org/~li​ve/test.42.html#2019-03-20T16:59:00,937625555-0400 -0>
17:03 sandbergja__ I tried setting network.security.ports.banned.override to 7682, but that didn't help :-(
17:05 mmorgan left #evergreen
17:13 berick sandbergja__: try navigating to https://HOST:7682/osrf-websocket-translator
17:13 berick it should prompt you to click through the cert warning
17:14 berick on FF even though it's the same cert from the same host, you have to click through per port.
17:15 berick or use nginx so you can share the port
17:16 sandbergja berick++
17:16 sandbergja thanks!
17:30 Bmagic Has anyone put a capcha on the patron "request a library card" form?
20:30 sandbergja joined #evergreen
20:35 Christineb joined #evergreen

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