| Time |
Nick |
Message |
| 04:22 |
|
tsbere_ joined #evergreen |
| 05:02 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
| 06:40 |
|
rlefaive joined #evergreen |
| 07:15 |
|
rjackson_isl joined #evergreen |
| 07:23 |
graced |
good morning #evergreen |
| 07:25 |
|
Dyrcona joined #evergreen |
| 07:26 |
Dyrcona |
You know what's funny? |
| 07:26 |
Dyrcona |
All of the crontab entries on our system do . ~/.bashrc. |
| 07:26 |
Dyrcona |
There's a block .bashrc to exit immediately if the environment is not interactive. |
| 07:26 |
Dyrcona |
The crontab environment is not interactive. |
| 07:27 |
Dyrcona |
I recall at MVLC, tsbere removed that block from the .bashrc files. |
| 07:29 |
Dyrcona |
And, then there's a redirect to /dev/null, so if there's any output, we don't get it. |
| 07:34 |
csharp |
Dyrcona: yuck |
| 07:50 |
|
rlefaive joined #evergreen |
| 08:09 |
jeff |
double yuck. |
| 08:12 |
jeff |
so, the various unaccent-related indices on actor.usr do break at pg_restore time due to 1) unqualified call to unaccent() within evergreen.unaccent_and_squash() AND changing the call to public.unaccent() isn't sufficient because the tsearch dictionary isn't in the search path, either. |
| 08:13 |
jeff |
instead of RETURN evergreen.lowercase(unaccent(regexp_replace(...))) we need RETURN evergreen.lowercase(public.unaccent('public.unaccent', regexp_replace(...))) |
| 08:14 |
jeff |
OR we could set an explicit search_path on the evergreen.unaccent_and_squash function, but we don't do that on any other functions at this time. |
| 08:14 |
Dyrcona |
jeff: I have not encountered that, but I am not actually have restored a database with the latest unaccent_and_squash, either. |
| 08:14 |
Dyrcona |
s/am/may/ |
| 08:15 |
Dyrcona |
I quit. |
| 08:15 |
jeff |
and technically we're lying about evergreen.unaccent_and_squash() being IMMUTABLE, but that's another story all together. |
| 08:18 |
jeff |
at least in 9.4, explicit search paths on functions do not appear to be exposed anywhere other than pg_catalog.pg_proc.proconfig |
| 08:19 |
jeff |
Dyrcona: good news is that the handful of authority-related indexes that used to fail no longer fail due to changes made for other mostly-unrelated reasons. :-) |
| 08:20 |
Dyrcona |
I've not had that many restore-related issues, honestly. |
| 08:20 |
* jeff |
nods |
| 08:20 |
jeff |
it used to be just those two or so authority indexes, and some people never even noticed them if they weren't heavy authority users. |
| 08:21 |
jeff |
those are fixed because the indexes no longer use a function ("mostly-unrelated reasons" above), and the unaccent indexes on actor.usr are 2.11 and up. |
| 08:22 |
Dyrcona |
Well, there was some issue with Z39.50 configurations that happened sometimes with -j > 1. |
| 08:22 |
jeff |
i think the upgrade script for this would be to CREATE OR REPLACE the evergreen.unaccent_and_squash() function, then create the indexes if they don't exist. if they exist, i don't think they need to be dropped and re-created. |
| 08:22 |
Dyrcona |
But not all the time. I never got errors about authority indexes. Are you sure you didn't add those, yourself? |
| 08:23 |
Dyrcona |
jeff: Sounds reasonable. |
| 08:23 |
bshum |
I'm pretty sure we fixed those z3950 issues last year |
| 08:23 |
Dyrcona |
Yes, we did. |
| 08:23 |
* bshum |
hasn't restored a database in forever |
| 08:25 |
jeff |
Dyrcona: you or tsbere shared a script a while back that you used as a standard post-restore to create the authority indexes that failed due to search path issues. |
| 08:25 |
jeff |
Dyrcona: i actually thought that you both had at one time or other. |
| 08:25 |
jeff |
another. |
| 08:26 |
Dyrcona |
jeff: Maybe. I know that it would recreate the missing z39.50 table. If it did authority indexes, I don't really remember, but now that you mention it, it has a familiar ring. |
| 08:27 |
jeff |
fun fact: the unaccent indexes only fail if there's data in the relevant columns being indexed, so a schema-only restore throws no errors and if you're dealing with a mostly-empty database you might not notice the one on actor.usr.second_given_name failing, because the admin user doesn't have a middle name. :-) |
| 08:45 |
|
mmorgan joined #evergreen |
| 08:47 |
dbs |
hmm do our --load-all-sample user names include any with accents? |
| 08:47 |
* dbs |
goes and looks |
| 08:47 |
|
bos20k joined #evergreen |
| 08:49 |
dbs |
yeah I don't think so, from a quick scan. that's probably not great |
| 08:51 |
Dyrcona |
I messed with a few of the names after loading that to test that branch. |
| 08:52 |
dbs |
right but we should have names and addresses as part of the stock set to avoid surprises down the road |
| 08:57 |
jeff |
and there are pg_tap tests for the unaccent function, but i agree -- names with accents would be useful to add to the sample data. |
| 08:57 |
* dbs |
will see about putting together a branch to add some baseball players to the stock set and figure out a way to tie it into web staff client so it gets attention |
| 08:57 |
* dbs |
is sick today so has some time to run repeated live test builds :) |
| 09:08 |
|
jvwoolf joined #evergreen |
| 09:08 |
jeff |
@weather --wind ktvc |
| 09:08 |
pinesol_green |
jeff: Cherry Capital, MI :: Mostly Cloudy :: 50F/10C | Wind: N 20mph (31mph gusts) | Tuesday: Partly cloudy and windy. High 53F. Winds SW at 20 to 30 mph. Winds could occasionally gust over 40 mph. Tuesday Night: Partly cloudy. Slight chance of a shower throughout the evening. Low 38F. Winds SW at 15 to 25 mph. | Updated: 15m ago |
| 09:09 |
jeff |
"occasionally" seems to be "every few minutes" |
| 09:11 |
csharp |
@weather 30345 |
| 09:11 |
pinesol_green |
csharp: Atlanta, GA :: Mostly Cloudy :: 57F/14C | Tuesday: Cloudy skies with periods of light rain this afternoon. Thunder possible. High 68F. Winds S at 10 to 15 mph. Chance of rain 70%. Tuesday Night: Periods of rain. Thunder possible. Low 47F. Winds WSW at 10 to 15 mph. Chance of rain 100%. |
| 09:12 |
csharp |
@band add Thunder possible |
| 09:12 |
pinesol_green |
csharp: Zoia knows how to make fusilli. |
| 09:12 |
csharp |
we keep not having the @band plugin |
| 09:13 |
jeff |
thunder was all last night. |
| 09:14 |
|
yboston joined #evergreen |
| 09:14 |
bshum |
It's probably something we could add. But I got annoyed with figuring out all the necessary dependencies for some of them. |
| 09:16 |
csharp |
bshum: don't mind me :-) |
| 09:17 |
csharp |
that is to say, please don't go to any effort for the purpose of fulfilling my desire to be silly :-P |
| 09:18 |
bshum |
Pfft, please..... That's the whole point of having the bot as far as I'm concerned! :) |
| 09:18 |
bshum |
Well that and lazy LP link grabbing |
| 09:18 |
bshum |
And git logs |
| 09:19 |
|
kmlussier joined #evergreen |
| 09:22 |
csharp |
:-) |
| 09:23 |
bshum |
But probably not today |
| 09:23 |
bshum |
:) |
| 09:26 |
Dyrcona |
If not now, when? :) |
| 10:04 |
Dyrcona |
And, speaking of restores....I just started one on a development db server... |
| 10:13 |
mmorgan |
I have a staff member who is intermittently seeing this error when opening an existing item in the unified editor: 'Error in volume_copy_creator.js, g.doc_id not valid' |
| 10:13 |
mmorgan |
Any ideas? |
| 10:14 |
bshum |
Autogen? |
| 10:14 |
bshum |
Did you change any org units stuff lately? |
| 10:15 |
Dyrcona |
More likely a timing issue. The editor is loading before some xul library thing that it needs. |
| 10:15 |
mmorgan |
No, no changes to org units. |
| 10:16 |
Dyrcona |
Which reminds me why developers should use 486 DX4s while frontline staff use Pentium IIs. :) |
| 10:16 |
Dyrcona |
But, that was a while ago. |
| 10:17 |
csharp |
mmorgan: check Admin -> For developers... -> JavaScript Console to see if there's more detail |
| 10:18 |
bshum |
bug 787561 |
| 10:18 |
pinesol_green |
Launchpad bug 787561 in Evergreen "sporadic g.doc_id is undefined error in volume/copy creator" [Undecided,Triaged] https://launchpad.net/bugs/787561 |
| 10:18 |
csharp |
ah |
| 10:18 |
csharp |
bshum++ # launchpad-fu |
| 10:18 |
bshum |
So sporadic. Timing issue like Dyrcona says |
| 10:19 |
mmorgan |
bshum++ |
| 10:19 |
kmlussier |
That bug still occurs? For some reason, I thought there had been a fix for it. |
| 10:19 |
bshum |
I tried tsbere's fix then and it broke stuff apparently. According to my notes |
| 10:19 |
bshum |
I think it went away and we stopped chasing it after |
| 10:19 |
bshum |
(Or we just quietly ignored it) |
| 10:19 |
Dyrcona |
bshum: Sounds about right. |
| 10:20 |
* csharp |
cranks up time machine to visit 2012 and ask younger bshum what was up |
| 10:20 |
* Dyrcona |
wants in on this time machine thing. |
| 10:20 |
bshum |
Younger bshum had potential. |
| 10:20 |
bshum |
I miss him too |
| 10:20 |
berick |
but the parachute pants were too much |
| 10:20 |
csharp |
berick++ |
| 10:21 |
bshum |
I did finally buy jeans last year. First time in a decade or so. |
| 10:21 |
csharp |
berick: the mention of parachute pants just set the sick beat of Freakazoid going in my head :-) |
| 10:22 |
berick |
hah |
| 10:22 |
csharp |
Z O IDS |
| 10:24 |
mmorgan |
I get the impression the error is a newish thing for this user, but will check back. It certainly is sporadic, and they can click through and continue their work. |
| 10:29 |
|
abowling joined #evergreen |
| 10:29 |
bshum |
Honestly, it doesn't surprise me if the bug still manifests since 2.2 with new XUL. Nothing substantial has changed with XUL client in years |
| 10:31 |
Dyrcona |
The user isn't using a 485 are they? ;) |
| 10:31 |
Dyrcona |
oops. |
| 10:31 |
kmlussier |
bshum: Yeah, my surprise was based on a faulty memory that there was a tested fix there. I should add a fixedinwebby tag to that LP bug. |
| 10:31 |
Dyrcona |
Spoiled my own joke. |
| 10:31 |
Dyrcona |
It was tested. It mostly sorta worked. :) |
| 10:32 |
bshum |
Hehe |
| 10:32 |
Dyrcona |
Apparently, it might have broken other things. |
| 10:32 |
bshum |
Like most things |
| 10:33 |
* Dyrcona |
is trying to figure out how to print arbitrary text, in this case new queries, to a file from psql. |
| 10:33 |
Dyrcona |
I swear there is an obvious way to do this, but the docs are not helping me. |
| 10:35 |
dbs |
\o output_file\nSELECT 'arbitrary text' FROM actor.usr LIMIT 1; ? |
| 10:38 |
Dyrcona |
dbs: Yeah, that isn't what I want, but may have to do. |
| 10:38 |
Dyrcona |
I know about \o and \qecho. |
| 10:38 |
dbs |
There are probably better ways but quick'n'dirty :) |
| 10:39 |
Dyrcona |
I must be remembering sybase, 'cause I swear there was print statement. |
| 10:39 |
Dyrcona |
\p isn't what I want, either. |
| 10:40 |
Dyrcona |
I basically want \echo to got standard out and query output, too. |
| 10:40 |
Dyrcona |
I want some arbitrary text in a file, basically an undo query. |
| 10:40 |
Dyrcona |
Guess I'll have to do it with Perl or store some values in a table. |
| 10:42 |
Dyrcona |
Yeah, I can just throw the ids into a new table. Hopefully, I'll remember what it was for when the time comes. :) |
| 10:57 |
|
JBoyer joined #evergreen |
| 11:11 |
|
brahmina joined #evergreen |
| 11:28 |
|
khuckins joined #evergreen |
| 11:37 |
dbs |
berick: am I reading ubuntu-auto-installer correctly, in that the the web staff client tests don't get run? |
| 11:39 |
berick |
dbs: looking at the 16.04 installer, I see a 'grunt test' in there |
| 11:39 |
berick |
ditto 14.04 |
| 11:40 |
dbs |
ah, not reading correctly! I was looking at the live tests section. thanks! |
| 11:41 |
berick |
cool, no prob |
| 11:52 |
|
mmorgan1 joined #evergreen |
| 12:04 |
|
mmorgan joined #evergreen |
| 12:06 |
|
mmorgan2 joined #evergreen |
| 12:09 |
|
mmorgan joined #evergreen |
| 12:17 |
|
mmorgan left #evergreen |
| 12:27 |
|
mmorgan joined #evergreen |
| 12:33 |
|
sandbergja joined #evergreen |
| 13:27 |
|
jvwoolf joined #evergreen |
| 13:29 |
|
JBoyer-alt joined #evergreen |
| 13:36 |
|
JBoyer-alt joined #evergreen |
| 13:37 |
|
JBoyer-alt-bin joined #evergreen |
| 14:08 |
|
kaffenkj joined #evergreen |
| 14:21 |
Stompro |
Hello, if I want to set the value of a $cgi->param "$cgi->param('status')" would that go in the apache eg.conf, in the virtualhost section as a "SetEnv status 0,1,2,3" ? Or is that only for $ENV{}? I'm trying to set the list of statues and shelving locations for the item_age_browse supercat feed. |
| 14:22 |
Dyrcona |
Stompro: $cgi gets parameters from the query string and/or post variables. |
| 14:22 |
Stompro |
Dyrcona, ahh, much easier.. thanks. |
| 14:22 |
Dyrcona |
So, you put status on the end of the URL, like ?status or &status. |
| 14:23 |
Dyrcona |
Usually with an = followed by a value. |
| 14:25 |
|
khuckins_ joined #evergreen |
| 14:38 |
miker |
Stompro: and to set more than one status, add the cgi param more than once ... ?status=0&status=1&status=2 |
| 14:38 |
Stompro |
miker++ I was just in the process of asking that, awsome. |
| 14:39 |
miker |
Stompro: http://search.cpan.org/~leejo/CGI-4.35/lib/CGI.pod for future reference |
| 14:41 |
Dyrcona |
Speaking of CGI, I was researching memory leaks with Apache and mod_perl yesterday, and the most common answer was, "Use fast-cgi, instead." |
| 14:41 |
miker |
Dyrcona: orly? |
| 14:42 |
Dyrcona |
Well, that or debug with Devel::Leak. |
| 14:42 |
Dyrcona |
But, yeah serverfault/stackoverflow, a number of "answers" were to switch to fcgi. |
| 14:42 |
miker |
Dyrcona: as in CGI::Fast? |
| 14:43 |
Dyrcona |
Not sure, there are a couple of fast-cgi modules for Apache. |
| 14:43 |
Dyrcona |
And not just for Perl. |
| 14:43 |
miker |
well, but at that level, it's at odds with mod_perl, no? |
| 14:45 |
Dyrcona |
Exactly. The answer is to basically abandon mod_perl, which is no answer at all. |
| 14:45 |
miker |
what I mean is, CGI.pm is just a wrapper around html construction and URL parsing ... and knows when it lives inside mod_perl. it's not really CGI-the-architecture |
| 14:45 |
miker |
heh |
| 14:45 |
miker |
ok, we're saying the same thing, then :) |
| 14:45 |
Dyrcona |
Right, they're talking about fast-cgi for Apache and not the CGI.pm. |
| 14:45 |
Dyrcona |
'Cause it's cgi, but ... fast. ;) |
| 14:46 |
Dyrcona |
No answer...Kind of like a recent mailing list post I made. |
| 14:47 |
Dyrcona |
Anyway, the "answer" of going to cgi amused me. |
| 15:13 |
|
mmorgan1 joined #evergreen |
| 15:26 |
* phasefx |
wants to make CGI great again (grins and runs away) |
| 15:28 |
jeffdavis |
I can't seem to get apache2-websockets to work in my Ubuntu Trusty environment. After going through the install steps in the OpenSRF install doc, `/etc/init.d/apache2-websockets start` gives no errors but it doesn't actually seem to start. |
| 15:29 |
jeffdavis |
`strace /etc/init.d/apache2-websockets start` shows a bunch of "No such file or directory" errors. |
| 15:29 |
jeffdavis |
but I'm not sure if those are real errors or if it's just iterating through alternate paths for things like env and rm |
| 15:32 |
bshum |
jeffdavis: Did you create the SSL certs in /etc/apache2/ssl ? |
| 15:32 |
bshum |
Usually things don't work for apache2-websockets without that, since the config file expects the SSL cert to be around |
| 15:33 |
bshum |
The creation of the SSL certs happens during the Evergreen install usually |
| 15:33 |
bshum |
For myself, I tend to skip starting websockets till after Evergreen is done and apache configured, and then I stop/start the whole shebang |
| 15:33 |
bshum |
I think the README makes note about the SSL cert issue, but it's easy to miss if it's your first time around |
| 15:34 |
bshum |
*OpenSRF README |
| 15:35 |
* bshum |
muses that one has to create the directory too for /etc/apache2/ssl |
| 15:35 |
Dyrcona |
See, I keep a tarball of /etc/apache2/ssl handy for the real servers/vms. |
| 15:36 |
Dyrcona |
tar -C / -zxf ssl.tgz |
| 15:36 |
bshum |
Well if you've got real certs, awesome :) |
| 15:36 |
* bshum |
has been lazy |
| 15:36 |
Dyrcona |
:) |
| 15:36 |
bshum |
No publicly resolvable hostnames for me anyways |
| 15:36 |
bshum |
Maybe someday again |
| 15:37 |
Dyrcona |
Right for test vms, it's a self-signed cert, though I've considered using my own CA. |
| 15:37 |
jeffdavis |
ah, looks like an issue with hostnames in apache2.conf, not ssl |
| 15:37 |
bshum |
I always wished I had spent more time with the CA options in our Windows environment |
| 15:37 |
Dyrcona |
You can make your own CA, just don't expect to be automatically added to nss-cacerts. |
| 15:38 |
bshum |
hostnames, eh |
| 15:38 |
bshum |
jeffdavis: GL with that part then :) |
| 15:43 |
jeffdavis |
Asking in IRC is a surprisingly effective way to make myself see what I'm doing wrong. :P |
| 15:43 |
Dyrcona |
jeffdavis: Ditto. |
| 15:44 |
bshum |
Been there plenty of times |
| 15:46 |
* Dyrcona |
double checks his passwords in opensrf_core.xml. |
| 15:46 |
Dyrcona |
That was a recent instance of that, I had a router password where I wanted the opensrf password. |
| 15:47 |
Dyrcona |
And, I should be ready to test OpenSRF. |
| 15:48 |
Dyrcona |
And, no surprises, it works. |
| 15:49 |
|
kmlussier joined #evergreen |
| 15:49 |
Dyrcona |
I installed prerequisites from master, but installed 2.4.2 and configured partly using the 2.5.0-rc instructions with the addition of setting max_stanza_size in ejabberd.yml. |
| 15:50 |
Dyrcona |
Why? Because Ubuntu 16.04..... |
| 15:50 |
Dyrcona |
Fun stuff... |
| 15:50 |
Dyrcona |
Now for something similar with Evergreen 2.10.... |
| 15:51 |
|
Jillianne joined #evergreen |
| 17:00 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
| 17:01 |
|
jvwoolf joined #evergreen |
| 17:06 |
|
jvwoolf1 joined #evergreen |
| 17:07 |
|
mmorgan1 left #evergreen |
| 17:25 |
|
frenchstudent joined #evergreen |
| 17:25 |
|
mllewellyn joined #evergreen |
| 17:25 |
frenchstudent |
hello |
| 17:26 |
frenchstudent |
My friend and I are working on evergreen software for our class ( library informatisation ) and we are searching for some reviews on it. |
| 17:30 |
berick |
frenchstudent: hi, I don't have any reviews to point you at. just wanted to say, welcome! |
| 17:33 |
abneiman |
frenchstudent: Marshall Breeding collects a lot of ILS reviews on his website, librarytechnology.org |
| 17:33 |
frenchstudent |
thanks !! |
| 17:38 |
frenchstudent |
it's perfect !! There is articles and reviews on the software !! Thank you really !!! We were blocked on this part !! |
| 17:39 |
berick |
abneiman++ |
| 17:40 |
abneiman |
:) you're welcome -- it's a very useful site |
| 20:22 |
csharp |
abneiman++ |
| 21:30 |
jeff |
oh, right. bug for unaccent search_path issues. |
| 21:53 |
|
Enjabain joined #evergreen |
