| Time |
Nick |
Message |
| 05:01 |
pinesol |
News from qatests: Failed Running perl live tests <http://testing.evergreen-ils.org/~live/test.49.html#2019-06-28T04:58:00,775679270-0400 -0> |
| 06:52 |
|
agoben joined #evergreen |
| 07:09 |
|
rjackson_isl joined #evergreen |
| 07:16 |
|
JBoyer joined #evergreen |
| 07:17 |
bshum |
Well between the last qatest success and the start of the failures, the only patches to go into master were from LP#1823041 |
| 07:17 |
bshum |
So that's always a fun place to start looking :) |
| 07:19 |
csharp |
those appear to all be in the ng layer :-/ |
| 07:28 |
* csharp |
figured it out |
| 07:29 |
* JBoyer |
is on pins and needles |
| 07:29 |
csharp |
the failing test is basically, wget the offline interface and check for anything missing, in this case by grepping the string "404" |
| 07:29 |
csharp |
wget --no-check-certificate -m https://localhost/eg/staff/offline-interface 2>&1 |grep -B 2 404|grep https|grep -v robots.txt|wc -l |
| 07:29 |
bshum |
Yeah I just reading that in the test file. Hmm. |
| 07:29 |
csharp |
if there's any line that matches, it will be counted by wc -l and anything higher than 0 means failure |
| 07:30 |
csharp |
this is the line it's finding: --2019-06-28 07:27:46-- https://localhost/eg2/en-US/runtime.baaae695dca404f5a78b.js |
| 07:30 |
csharp |
so, um, bad test |
| 07:30 |
JBoyer |
csharp++ |
| 07:30 |
bshum |
csharp++ |
| 07:31 |
JBoyer |
That would explain why it worked fine for me when I was trying it locally, different filename. |
| 07:31 |
csharp |
right - it's just chance that this would have ever appeared, ever |
| 07:32 |
JBoyer |
looks like 404 -> ' 404 ' will take care of it. |
| 07:32 |
csharp |
ok - glad you know that because I don't know what bad output looks like :-) |
| 07:34 |
JBoyer |
If you just run the plain wget part you can see that the http status has a space to either side, so unless we end up with a file size that's 404 bytes, that should fix this test. |
| 07:35 |
csharp |
JBoyer: if you want to create a patch, I'll sign off/commit |
| 07:35 |
JBoyer |
++ |
| 07:44 |
|
bos20k joined #evergreen |
| 07:50 |
JBoyer |
Finally: user/jboyer/fix-offline-test |
| 07:51 |
JBoyer |
Had to re-setup some things because this new machine is apparently missing some things... |
| 07:53 |
csharp |
getting mine reset too since the livecheck creates data |
| 07:59 |
csharp |
sigh - the negative balances test is failing now too |
| 07:59 |
JBoyer |
Hmm. :/ |
| 08:00 |
csharp |
your fix works and I'm going to push it to master so we can move on from that, but I expect tests to still break |
| 08:01 |
csharp |
Script already running with lockfile /tmp/09-lp1198465_neg_balances.t-LOCK at /home/opensrf/Evergreen/Open-ILS/src/perlmods/blib/lib/OpenILS/Utils/Cronscript.pm line 151. |
| 08:01 |
csharp |
^^ that's what I'm seeing when I run it |
| 08:01 |
csharp |
oh - that's my fault |
| 08:01 |
csharp |
nevermind :-/ |
| 08:02 |
JBoyer |
Past run interrupted? |
| 08:02 |
csharp |
I canceled out of tests |
| 08:02 |
csharp |
yeah |
| 08:02 |
JBoyer |
+1 |
| 08:02 |
JBoyer |
I will say there are more warnings that I'd prefer when running them, but everything did work for me. |
| 08:02 |
csharp |
yeah, same here |
| 08:11 |
csharp |
ok, I fixed my issue and re-ran and all looks good - gonna go ahead and commit to master without worrying about a bug report |
| 08:12 |
csharp |
done |
| 08:12 |
csharp |
JBoyer++ |
| 08:12 |
JBoyer |
csharp++ |
| 08:15 |
pinesol |
[evergreen|Jason Boyer] Correct False Positive on Offline Test - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=a150021> |
| 08:41 |
|
Dyrcona joined #evergreen |
| 08:50 |
|
aabbee joined #evergreen |
| 09:01 |
|
jvwoolf1 joined #evergreen |
| 09:05 |
Dyrcona |
git++ |
| 09:08 |
dbwells |
csharp++ JBoyer++ |
| 09:13 |
Dyrcona |
csharp++ JBoyer++ tests++ fragility-- |
| 09:14 |
Dyrcona |
sed++ |
| 09:15 |
Dyrcona |
I got the haproxy config issue sorted out for Ubuntu. Gonna test it on Debian 8 & 9 a bit later. |
| 09:15 |
csharp |
tests might be a good way for the new devs group to get involved |
| 09:16 |
Dyrcona |
csharp: I have had that thought. Writing tests it a good way to learn the back end, and maybe that should be a conference presentation. |
| 09:17 |
|
yboston joined #evergreen |
| 09:18 |
csharp |
yeah definitely |
| 09:18 |
|
mmorgan joined #evergreen |
| 10:20 |
|
sandbergja joined #evergreen |
| 11:00 |
|
khuckins joined #evergreen |
| 11:11 |
Dyrcona |
@dunno search autogen.sh |
| 11:11 |
pinesol |
Dyrcona: No matching dunnos were found. |
| 11:11 |
Dyrcona |
@dunno add Have you run autogen.sh? |
| 11:11 |
pinesol |
Dyrcona: The operation succeeded. Dunno #63 added. |
| 11:12 |
Dyrcona |
Yeah, things were a little wonky just now, until I remembered to run autogen.sh. |
| 11:12 |
Dyrcona |
@dunno |
| 11:12 |
pinesol |
Dyrcona: Go away, or I'll replace you with a very small shell script! |
| 11:17 |
|
mmorgan joined #evergreen |
| 11:31 |
pinesol |
[evergreen|Cesar Velez] LP#1831781: add eg-help-popover Angular component - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=ab82f5c> |
| 11:31 |
pinesol |
[evergreen|Galen Charlton] LP#1831781: tweaks to eg-help-popover - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=a1a03fa> |
| 11:36 |
sandbergja |
csharp++ |
| 11:36 |
sandbergja |
JBoyer++ |
| 11:36 |
sandbergja |
bshum++ |
| 11:36 |
sandbergja |
Thanks for fixing that test! |
| 12:00 |
|
jamesrf joined #evergreen |
| 12:04 |
|
jihpringle joined #evergreen |
| 12:09 |
|
JBoyer joined #evergreen |
| 12:09 |
|
jihpringle79 joined #evergreen |
| 12:18 |
|
yboston joined #evergreen |
| 12:26 |
|
jihpringle joined #evergreen |
| 12:42 |
|
yboston joined #evergreen |
| 12:43 |
|
khuckins_ joined #evergreen |
| 12:45 |
|
jamesrf joined #evergreen |
| 13:06 |
Dyrcona |
@dunno |
| 13:06 |
pinesol |
Dyrcona: zarro boogs found |
| 13:18 |
jonadab |
Heh. |
| 13:22 |
Dyrcona |
So, I'm not sure if this haproxy or something else, but I keep getting blank pages when testing the web staff client with a Debian host. I have to reload just about every page. I'm also using Chromium 75 and an Incognito window. |
| 13:23 |
jonadab |
We get that in production when we have network performance issues. |
| 13:23 |
Dyrcona |
By "Debian host," I mean that I see the same behavior with Jessie (8) and Stretch (9). |
| 13:23 |
Dyrcona |
Well, this is a vm running on my laptop. The network should actually be RAM... |
| 13:24 |
Dyrcona |
Could be haproxy slowing things down. I'm going to logout and try with regular Chrome windows. |
| 13:26 |
Dyrcona |
We've been getting reports of blank pages in production, and I've done a few things that seem to have helped, but chasing this stuff down is like chasing rainbows. |
| 13:26 |
berick |
Dyrcona: beware if you are using a dummy cert, it can lead to page load funkiness |
| 13:27 |
Dyrcona |
berick: Yes, a self-signed cert is another factor. I do not like the notion of "dummy cert." |
| 13:27 |
berick |
IIRC, chrome refuses to load the offline worker script when it doesn't like the cert, and that can prevent the page from loading intermittently |
| 13:27 |
Dyrcona |
Well, this is just normal pages, like check in and check out, etc. |
| 13:27 |
berick |
that's accompanied by JS errors though, so should be obvious |
| 13:28 |
berick |
Dyrcona: offline db caching happens on every page |
| 13:28 |
Dyrcona |
I have error loading shared worker in the console, so that must be it. |
| 13:28 |
berick |
yep |
| 13:28 |
Dyrcona |
chrome-- |
| 13:29 |
Dyrcona |
I'll try with Firefox, but I'm pretty sure that I've got the haproxy config working. |
| 13:30 |
Dyrcona |
Firefox's cert warning has gotten even scarier than it used to be. |
| 13:30 |
jonadab |
For developers, browsers really need to have a command-line argument that says, effectively, all certs are valid by fiat. |
| 13:31 |
jonadab |
It's not practical to do a Let's Encrypt cert for something that's not public facing. |
| 13:31 |
jonadab |
Like a test setup. |
| 13:31 |
berick |
and the process for making chrome trust a self-signed cert is a bit more cumbserome now as well. |
| 13:32 |
Dyrcona |
Well, I could use my let's encrypt wildcard cert. |
| 13:32 |
Dyrcona |
I used to use my own CA and add it to my browsers. |
| 13:33 |
jonadab |
Actually, I'd be happy if it were domain specific, like --ignore-cert-errors=mydomain.example.com |
| 13:33 |
Dyrcona |
Reason that I could use my wildcard cert is that I also use dnsmasq with /etc/hosts entries for my test vms, so they're accessible as {vmane}.sigio.{com,org,net} |
| 13:34 |
Dyrcona |
bleh... typos... :) |
| 13:34 |
Dyrcona |
I wish the browser just had a "I know what I'm doing" configuration option. |
| 13:34 |
jonadab |
^ Indeed. |
| 13:51 |
Dyrcona |
Reports question: Someone made a report template on our training server. Can I just dump the row to a file and then insert that into the production database server, or does some other magic have to happen for the template to be usable on production? |
| 13:52 |
rhamby |
Dyrcona: I've moved templates around that way (accounting for users and all that of course) |
| 13:52 |
|
khuckins joined #evergreen |
| 13:53 |
Dyrcona |
rhamby: Thanks! That was what I thought. I think the users and folder ids will be the same. Template id might already be used. |
| 14:41 |
|
khuckins_ joined #evergreen |
| 14:54 |
|
yboston joined #evergreen |
| 15:02 |
Dyrcona |
rhamby++ |
| 15:03 |
jeffdavis |
In 3.3, setting the default locale for server-side XUL and XHTML to something other than en-US is giving me internal server errors on some Dojo UIs in the web client. |
| 15:04 |
jeffdavis |
The reason is that some of the templates used in those UIs try to include "/opac/locale/${locale}/lang.dtd", but lang.dtd only exists in the en-US locale, not any others (at least in my environment). |
| 15:04 |
jeffdavis |
I'm not sure if this is unexpected behavior, perhaps a result of removing some XUL stuff. |
| 15:04 |
jeffdavis |
It is different from pre-3.3 though. |
| 15:08 |
Dyrcona |
Has any XUL been removed in 3.3? |
| 15:11 |
jeffdavis |
Maybe not removed, but some build steps were disabled by default (bug 1790923). I don't remember seeing this on a previous 3.2 server but local config may have been different. |
| 15:11 |
pinesol |
Launchpad bug 1790923 in Evergreen "disable XUL staff client by default" [Wishlist,Fix released] https://launchpad.net/bugs/1790923 |
| 15:25 |
Dyrcona |
jeffdavis: You could always try the manual steps to build the xul client files and put them in place on the server. |
| 15:28 |
Dyrcona |
Looks like that isn't necessary. I appear to have the server side files in a master build from earlier today. |
| 15:31 |
Dyrcona |
Looks like the server-side files get installed in /openils/var/web/xul/legacy... Might want to check the symlinks. |
| 15:40 |
|
stephengwills left #evergreen |
| 15:43 |
jeffdavis |
bug 1757455 |
| 15:43 |
pinesol |
Launchpad bug 1757455 in Evergreen "Deprecate i18n building for XUL client" [Medium,Fix released] https://launchpad.net/bugs/1757455 |
| 15:47 |
jeffdavis |
commit 9023f96c removed the steps that put lang.dtd etc in opac/locale/{$locale} |
| 15:47 |
pinesol |
jeffdavis: [evergreen|Ben Shum] LP#1757455: followup to fix installing i18n - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=9023f96> |
| 15:52 |
Dyrcona |
Well, that's progably it, then. |
| 15:52 |
Dyrcona |
probably, even. |
| 16:04 |
pinesol |
[evergreen|Dan Wells] Docs: Revise 3.3.2 bugfix release notes - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=8d937f5> |
| 16:17 |
jeffdavis |
reported the above to LP, bug 1834692 |
| 16:17 |
pinesol |
Launchpad bug 1834692 in Evergreen "Some staff client UIs break with alternate default locale" [Undecided,New] https://launchpad.net/bugs/1834692 |
| 16:20 |
|
jihpringle joined #evergreen |
| 16:31 |
Dyrcona |
jeffdavis++ # It's Friday. |
| 16:35 |
Dyrcona |
jeffdavis: Can explain what you mean by "installed locale?" Does it have to be installed from O/S packages or do you mean one of the languages with Evergreen translations? |
| 16:36 |
jeffdavis |
The latter. In our case we build locale stuff from source; for installing from a tarball I think you could use any of the translations that come with EG (but I haven't tested that). |
| 16:37 |
jeffdavis |
"installed translations" would maybe be a clearer way for me to put it |
| 16:37 |
Dyrcona |
jeffdavis: Thanks. That is what I thought you meant, but yeah... |
| 16:38 |
* Dyrcona |
looks into how to install locale stuff from source.... |
| 16:40 |
jeffdavis |
cd build/i18n && make LOCALE=fr-CA install |
| 16:40 |
jeffdavis |
before the main `make install` step |
| 16:41 |
Dyrcona |
jeffdavis: Thanks! |
| 16:42 |
Dyrcona |
ooh.. looks like I need some additional packages. |
| 16:42 |
jeffdavis |
Thanks for looking at the bug! |
| 16:42 |
Dyrcona |
Yeah, I just want to confirm it. I'm not promising a fix. |
| 16:44 |
jeffdavis |
packages to install on Ubuntu 16.04: translate-toolkit, python-polib, maybe build-essential and/or bsdmainutils |
| 16:44 |
Dyrcona |
jeffdavis: I just hit the ubuntu-xenial-translator prereq option in Makefile.install |
| 16:45 |
Dyrcona |
Probably overkill as it is also doing -devloper again. |
| 16:47 |
Dyrcona |
I just realized the I'm doing this on my 3.2 vm, so it may not be affected.... |
| 16:48 |
Dyrcona |
I'll try it on master afterward. |
| 16:53 |
Dyrcona |
Oh,nice. I thought it was taking a long time, but my VPN connection got shut down. |
| 16:55 |
mmorgan |
Last minute question on a Friday. I have an action trigger event that 's pending and I just want to run that single event. There are lots of other pending events that I don't want to run. Is there a way to do that? It's the only event for this event_def, if that's a way to do it. |
| 16:56 |
Dyrcona |
mmorgan: you can run individual events from srfsh. |
| 16:57 |
Dyrcona |
Hang on a sec. |
| 16:58 |
Dyrcona |
At the srfsh prompt: request open-ils.trigger open-ils.trigger.event.fire <event_id> |
| 16:58 |
mmorgan |
Oooh! Thanks Dyrcona! |
| 16:58 |
mmorgan |
Dyrcona++ |
| 16:59 |
Dyrcona |
I've used that on pending and collected events. |
| 16:59 |
Dyrcona |
I think it works regardless of the event state. |
| 17:01 |
pinesol |
News from qatests: Testing Success <http://testing.evergreen-ils.org/~live> |
| 17:03 |
|
jvwoolf1 left #evergreen |
| 17:06 |
mmorgan |
Dyrcona: Interesting, each time it's fired it bumps to the next state. Works for my purpose though! Thanks again. |
| 17:06 |
Dyrcona |
It should only need to be fired once. Something must be wrong with the data. |
| 17:09 |
mmorgan |
Ok, it bumps between valid and found. I'll need to revisit on Monday. Have a good weekend everybody! |
| 17:09 |
|
mmorgan left #evergreen |
| 17:11 |
Dyrcona |
jeffdavis: Did you also modify the PerlSetVar bits for Locale that are lower down in eg_vhost.conf? |
| 17:13 |
jeff |
This morning, I helped a group of children talk with an astronaut via amateur radio. Then, I helped tear down the antenna and wrap cables. Now, I'm shaking my fist at two different interfaces outside of our control that are being uncooperative. |
| 17:14 |
jeffdavis |
Dyrcona: do you mean OILSWebLocale and OILSWebDefaultLocale? No, they're all commented out on my test system. |
| 17:14 |
Dyrcona |
jeff: Sounds like a full day. :) |
| 17:14 |
jeff |
really, if one of these things was going to go well and one was going to go poorly, I'm okay with the way the chips fell. :-) |
| 17:14 |
Dyrcona |
And, we need more bib record in French. :) |
| 17:14 |
Dyrcona |
Yeah, better to NOT have problems with the kids and the astronaut. |
| 17:15 |
berick |
jeff: you should have asked the astronaut how to fix the interfaces |
| 17:15 |
jeff |
berick: AOS to LOS was approximately nine minutes. We didn't have a lot of time. :-) |
| 17:16 |
berick |
gotcha, keep the discussion focused on space lazers |
| 17:17 |
jeff |
Due to scheduling constraints on the ISS, there's no opportunity to do an actual test before the contact. Those few seconds waiting for them to respond after we started calling CQ were... tense. |
| 17:19 |
Dyrcona |
jeffdavis: In Firefox, I get a "this page isn't redirecting properly" message. |
| 17:19 |
berick |
kids (ney, all) must of loved that |
| 17:20 |
jeff |
Followed by "Whiskey eight tango charlie mike, this is november alpha one sierra sierra, how copy?" |
| 17:20 |
jeff |
And LOTS of silent celebration and high fives. |
| 17:21 |
Dyrcona |
jeffdavis: I uncommented the other variables and I'm getting too many redirects. I'll comment them out and try again. This still looks like the same problem to me. |
| 17:21 |
jeff |
Hour long event at https://youtu.be/4OnrVCbWW9E radio contact begins around 19:34, or https://youtu.be/4OnrVCbWW9E?t=1934 |
| 17:22 |
jeff |
It was a really awesome way to spend a morning. |
| 17:23 |
jeffdavis |
definitely very cool |
| 17:25 |
Dyrcona |
jeff: Definitely more interesting than my morning. |
| 17:26 |
Dyrcona |
jeffdavis: Some interfaces give me an internal server error, and others are redirecting me too many times, but the result is the same. |
| 17:31 |
Dyrcona |
Bug confirmed. |
| 17:32 |
* Dyrcona |
calls it a day. |
| 17:32 |
Bmagic |
jeff++ |
| 21:27 |
|
Dyrcona joined #evergreen |
| 21:47 |
|
Dyrcona joined #evergreen |
| 21:58 |
Dyrcona |
jonadab++ # For mentioning using letsencrypt certs for development. |
