| Time |
Nick |
Message |
| 00:34 |
pinesol_green |
[evergreen|Kyle Huckins] LP#1534787 Patron Message Center port - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=a6f1a4f> |
| 03:08 |
|
abowling joined #evergreen |
| 05:02 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
| 07:03 |
bshum |
jetlag-- |
| 07:03 |
bshum |
breakfast++ |
| 07:17 |
|
rjackson_isl joined #evergreen |
| 07:30 |
|
agoben joined #evergreen |
| 08:06 |
|
rlefaive joined #evergreen |
| 08:13 |
|
collum joined #evergreen |
| 08:37 |
|
jvwoolf joined #evergreen |
| 08:37 |
|
mmorgan joined #evergreen |
| 08:38 |
JBoyer |
It is a truly sublime feeling when the burn-it-all-down-and-begin-anew config reset switch on a brand new device *does* *not* *work*. It is the universe giving you the opportunity to reflect on the decisions that have led you to this point. Revel in it. |
| 08:50 |
|
bos20k joined #evergreen |
| 08:52 |
|
jvwoolf left #evergreen |
| 08:57 |
|
kmlussier joined #evergreen |
| 08:59 |
|
drigney joined #evergreen |
| 08:59 |
|
barbara joined #evergreen |
| 08:59 |
|
miker joined #evergreen |
| 08:59 |
|
akilsdonk joined #evergreen |
| 08:59 |
|
phasefx_ joined #evergreen |
| 08:59 |
|
rhamby joined #evergreen |
| 08:59 |
|
graced joined #evergreen |
| 08:59 |
|
abneiman joined #evergreen |
| 08:59 |
|
sallyf joined #evergreen |
| 08:59 |
|
Shae joined #evergreen |
| 09:01 |
|
sarabee joined #evergreen |
| 09:07 |
|
jvwoolf joined #evergreen |
| 09:23 |
|
yboston joined #evergreen |
| 09:32 |
|
Dyrcona joined #evergreen |
| 09:40 |
|
maryj joined #evergreen |
| 09:43 |
csharp |
berick: you asked me Friday afternoon if I was still seeing hold targeter issues - I sent you a @later on Saturday - have you seen it? (I can re-send) |
| 09:43 |
csharp |
or I can just paste it here :-) |
| 09:43 |
csharp |
< csharp> @later tell berick there are several errors per hour like those in this full threadtrace: http://pastebin.com/CcaMp6tH caused by "too many copies per bib" (I think) |
| 09:44 |
csharp |
we've received no complaints from the wild that correspond to those, but they're there |
| 09:51 |
Bmagic |
hi |
| 09:51 |
* Bmagic |
yawns |
| 09:51 |
csharp |
@coffee Bmagic |
| 09:51 |
* pinesol_green |
brews and pours a cup of Ethiopia Yirgacheffe Koke, and sends it sliding down the bar to Bmagic |
| 09:52 |
* Bmagic |
sits up at attention |
| 09:52 |
Bmagic |
Ethiopia knows what they are doing |
| 09:53 |
Dyrcona |
Hmm.. I just setup opensrf on a test VM. Services are up and running, but when I ran srfsh, I got this "Unable to bootstrap client for requests" |
| 09:56 |
Dyrcona |
log says it can't authenticate with jabber... |
| 09:57 |
* csharp |
suspects the ejabberd nightmare he experienced a couple of weeks ago |
| 09:57 |
Dyrcona |
Nope. Simpler than that. |
| 09:57 |
Dyrcona |
I was put the password in for the router user when I wanted the opensrf user's password, in .srfsh.xml. |
| 09:57 |
csharp |
ah |
| 09:58 |
Dyrcona |
And I got 4 for the opensrf.math test, so that works. :) |
| 09:59 |
Dyrcona |
Next, I get to transpose our Apache 2.2 config to Apache 2.4. |
| 10:02 |
* Dyrcona |
is preparing to upgrade to Debian 8. |
| 10:02 |
berick |
csharp: thanks! taking a look |
| 10:07 |
dbs |
Bmagic++ |
| 10:11 |
* csharp |
just bounced back to Fedora 25 after about 6 months on Ubuntu, following his usual migration pattern |
| 10:13 |
dbs |
csharp++ # good to keep your experiences broad! |
| 10:14 |
dbs |
maybe if I can make it to the hackfest (can't remember my travel details offhand) I'll work on shoring up Fedora 25 OpenSRF/Evergreen support |
| 10:14 |
csharp |
dbs++ |
| 10:15 |
* dbs |
peeks at http://youmightnotneedjquery.com/ for fun |
| 10:15 |
csharp |
maybe I'll dust off my occasional EG/centos project |
| 10:15 |
dbs |
alas there is no http://youdonotneeddojo.com/ |
| 10:15 |
csharp |
ha! |
| 10:16 |
dbs |
or even http://dojo13xwtfbbq.com/ |
| 10:17 |
csharp |
http://icantbelieveyourestillusingdojo13.biz |
| 10:18 |
dbs |
heh |
| 10:18 |
Bmagic |
lol |
| 10:19 |
Bmagic |
dbs++ |
| 10:19 |
Bmagic |
csharp++ |
| 10:20 |
dbs |
hey, dojo 1.3 would be in Grade 2 now if it was human |
| 10:20 |
Bmagic |
Some of my best years were in Grade 2 |
| 10:42 |
Dyrcona |
heh. |
| 10:42 |
* Dyrcona |
liked 1975, but that was before grade 2. |
| 10:43 |
Dyrcona |
So, instead of messing with config files, I spent 40 minutes messing with git branches, but I think I had botched my 2.11 config branch and now it's fixed. |
| 10:46 |
Dyrcona |
And I see something else I should update...I recently changed the zurl for simple2zoom. |
| 10:47 |
Dyrcona |
Question is: Do I use "localhost" or the FQDN? |
| 10:48 |
JBoyer |
Dyrcona, I believe the request from that machine and is relayed over the Z port, so localhost will work, unless you have reason to care that it's the FQDN. (I know this because I set up the FQDN but the Z server at the time couldn't resolve it in DNS and so that was a fun afternoon...( |
| 10:48 |
dbs |
localhost keeps it easy because you can use HTTP |
| 10:49 |
dbs |
HTTPS requires a newer version of the yaz utilities than are sometimes installed by default, which causes crashies |
| 10:49 |
dbs |
JBoyer++ |
| 10:49 |
Dyrcona |
dbs++ |
| 10:50 |
Dyrcona |
JBoyer++ |
| 10:50 |
Dyrcona |
I usually use localhost on my test VMs. |
| 10:50 |
Dyrcona |
I used the FQDN of the local machine in production. |
| 10:51 |
Dyrcona |
So, think I'll switch to localhost. |
| 10:53 |
Dyrcona |
Fun thing: I have 3 config branches: 2.10, 2.11, and master, so that's a couple of git checkout; git cherry-pick to keep them all in sync. |
| 10:53 |
berick |
csharp: fix pushed to same branch: http://git.evergreen-ils.org/?p=working/Evergreen.git;a=commitdiff;h=2771cbe124221842271c0b06c9b517fe6673256b |
| 10:58 |
* berick |
updates lp |
| 10:58 |
csharp |
rock on, I'll give it a go |
| 11:00 |
berick |
csharp++ |
| 11:06 |
|
khuckins joined #evergreen |
| 11:21 |
* Dyrcona |
likes it when patches just apply. |
| 11:22 |
|
Christineb joined #evergreen |
| 11:25 |
Dyrcona |
Hm... I think there might be an error in the default eg_vhost.conf for apache 2.4, but I need to verify it. I'm in a branch where I haven't modified it, yet. |
| 11:26 |
Dyrcona |
It's in something that isn't used much, though. |
| 11:42 |
|
lualaba joined #evergreen |
| 11:43 |
lualaba |
hello team. i need to import records over z3950. via yaz client everything is ok in DB i already define attributes. LIBRARY/opac/extras/sru?version=1.1&operation=searchRetrieve&query=hemingway&maximumRecords=0 returning xml, but from staff client 0 records any idea? |
| 11:50 |
|
bmills joined #evergreen |
| 11:57 |
|
brahmina joined #evergreen |
| 12:17 |
|
jihpringle joined #evergreen |
| 12:43 |
Dyrcona |
I've had to do chown opensrf /var/lock/apache2 on a couple of Debian installs lately, particularly on Jessie. Maybe we should update the README? |
| 12:58 |
bshum |
+1 |
| 12:59 |
* bshum |
hasn't played with Debian lately, but wouldn't be surprised if that sort of thing changed with newer apache versions |
| 12:59 |
Dyrcona |
Well, that's not so useful: apache2: Syntax error on line 219 of /...ry |
| 13:00 |
Dyrcona |
bshum: That happened to me on Debian 7 and Debian 8. |
| 13:01 |
Dyrcona |
systemctl-- for truncating error messages to the width of the terminal rather than wrapping. |
| 13:03 |
bmills |
Dyrcona: we had to do the same when we upgraded our debian bits from wheezy —> jessie. |
| 13:03 |
Dyrcona |
So, apparently, we need a full path for Include /etc/apache2/eg_vhost.conf. |
| 13:04 |
Dyrcona |
At least, that appears to be what it's all complaining about. |
| 13:04 |
Dyrcona |
systemctl-- Again, for truncating error messages to the width of the terminal. I apparently need a 4K screen to get the full message. |
| 13:04 |
Dyrcona |
1080p just isn't big enough. |
| 13:05 |
csharp |
you can right arrow if you're in journalctl |
| 13:09 |
Dyrcona |
Yeah, journalctl was a bit better than the systemctl equivalent. |
| 13:09 |
Dyrcona |
systemctl-- just because. |
| 13:09 |
Dyrcona |
:) |
| 13:09 |
Dyrcona |
Anyway, turns out I don't need the path. I need to not typo file names in copy commands. |
| 13:09 |
Dyrcona |
I typoed the destination file name, so the copy worked. |
| 13:10 |
Dyrcona |
So, looks like my Apache 2.2-2.4 conversion worked. |
| 13:11 |
Dyrcona |
Or not... :( |
| 13:11 |
Dyrcona |
It failed to start, but this time systemctl didn't say anything...Guess what's coming.... |
| 13:12 |
Dyrcona |
systemctl-- :) |
| 13:14 |
Dyrcona |
Hah! That's funny! syslog says apache2 couldn't log errors to syslog.... No such file or directory: AH00089: Couldn't start ErrorLog process 'logger -p local7.error'. |
| 13:15 |
Dyrcona |
So, I'll just go back to files to get this working. |
| 13:16 |
dbs |
journalctl++ |
| 13:17 |
Dyrcona |
And now, apache is running. |
| 13:17 |
csharp |
has anyone been working on a systemd service for opensrf? |
| 13:18 |
Dyrcona |
Wouldn't surprise me if options to logger have changed since Debian 7. |
| 13:22 |
Dyrcona |
And, a search worked. This is promising. :) |
| 13:22 |
|
mllewellyn joined #evergreen |
| 13:29 |
Dyrcona |
All right, I don't know why logger complained like that. |
| 13:30 |
Dyrcona |
Maybe I need the full path? I'll try that after I test a z39.50 search. |
| 13:34 |
|
pgardella joined #evergreen |
| 13:35 |
Dyrcona |
Well, I guess the little machine is struggling. It timed out after 30 seconds. |
| 13:39 |
Dyrcona |
This doesn't look right: "searches":{"author":{"term":"coelho"},"keyword":{"term":"eg."}}, Looks like part of the index name ends up as a search tem. |
| 13:39 |
Dyrcona |
s/tem/term/ |
| 13:41 |
Dyrcona |
That comes from this Z search: find @attr 1=1003 coelho |
| 13:41 |
pgardella |
Good afternoon, everyone! I'm trying to restore a backup of the evergreen database onto another server for testing, and have run into some errors I've not hit before. Namely, functions not existing: |
| 13:41 |
pastebot |
"pgardella" at 64.57.241.14 pasted "missing functions" (8 lines) at http://paste.evergreen-ils.org/45 |
| 13:42 |
pgardella |
this was with a -Fc backup of our live DB |
| 13:43 |
pgardella |
Anyone seen this before? |
| 13:43 |
Dyrcona |
pgardella: I think the database you're restoring into lacks the unaccent extension. |
| 13:44 |
|
rlefaive joined #evergreen |
| 13:45 |
Dyrcona |
Interestingly, I'm seeing the "keyword":{"term":"eg."} in production, too. This is on Evergreen 2.10.7. |
| 13:45 |
pgardella |
Hum OK. I create the DB with it, and then I think restore drops it and recreates it again. |
| 13:45 |
pgardella |
@Dyrcona I also get this: function metabib.compile_composite_attr(text) does not exist |
| 13:45 |
pinesol_green |
pgardella: It reads like a Nigerian 419 scam, but I think it is a sincere question sent to the wrong list. |
| 13:46 |
pgardella |
Oddly enough, this is for a library in Nigeria! |
| 13:46 |
Dyrcona |
:) |
| 13:46 |
Dyrcona |
Just type someone's name in IRC, if you put @ at the beginning of a line, the bot thinks it is a command for it to run. |
| 13:47 |
pgardella |
Ah. Too many systems all jumbled in my head |
| 13:50 |
Dyrcona |
So, you're just doing pg_dump -Fc database > somefile ? |
| 13:50 |
pgardella |
yes |
| 13:51 |
pgardella |
Then I was restoring with -C, which dropped the database I had prepped with the extensions. I just fixed that, so I'm not seeing the unaccent error anymore |
| 13:51 |
pgardella |
Now I'm just getting the evergreen ones: type "evergreen.query_int" does not exist |
| 13:52 |
pgardella |
function metabib.compile_composite_attr(text) does not exist |
| 13:52 |
csharp |
pgardella: sometimes it's a path issue (show search_path ; will let you see what the current path is) |
| 13:53 |
csharp |
mine shows evergreen, "$user", public |
| 13:53 |
Dyrcona |
Yeah, what csharp said. :) |
| 13:53 |
pgardella |
"$user",public |
| 13:53 |
Dyrcona |
Is your db user named evergreen? |
| 13:53 |
pgardella |
Yes |
| 13:54 |
pgardella |
and I'm running the restore as evergreen |
| 13:54 |
csharp |
pgardella: SET search_path TO evergreen, "$user", public; |
| 13:54 |
pgardella |
Sigh. I think I've found the answer to all of these already, I just need to get them all in one place at one time. |
| 13:54 |
csharp |
though if you're using psql as evergreen, "$user" should be evergreen too |
| 13:55 |
Dyrcona |
Yeah, that's why I asked. :) |
| 13:56 |
* csharp |
used to do most things as "postgres", which caused this sort of thing more often that it happens when using "evergreen" |
| 13:56 |
Dyrcona |
I've run into that, 'cause I've got databases where the user isn't evergreen. |
| 13:57 |
csharp |
which reminds me that we need to demote the evergreen user from superuser by default (pretty sure there's a bug on that) |
| 13:57 |
Dyrcona |
I've had to restore as superuser because I wanted evergreen to have lower permissions. |
| 13:57 |
Dyrcona |
:) |
| 13:57 |
csharp |
:-) |
| 13:57 |
csharp |
@dunno add pretty sure there's already a bug report about that. |
| 13:57 |
pinesol_green |
csharp: The operation succeeded. Dunno #52 added. |
| 13:58 |
pgardella |
Interesting. I set my search_path in my database "prep" script (which add extensions, etc), but pg_restore ignores that and sets its own |
| 13:58 |
Dyrcona |
So, I'm going to bug my Z search issue, though I might fire up VM with master to see if it happens there, first. |
| 13:59 |
pgardella |
logging in as evergreen, I see the search_path as evergreen, public |
| 13:59 |
Dyrcona |
pgardella: With -C it definitely will. |
| 13:59 |
pgardella |
without -C |
| 13:59 |
pgardella |
this time |
| 13:59 |
Dyrcona |
Ok. |
| 13:59 |
Dyrcona |
I've not run into this lately. |
| 14:00 |
pgardella |
This is my full restore command: pg_restore -U evergreen -W -d evergreen -h localhost evergreen.dump |
| 14:01 |
csharp |
pgardella: may not help anymore with this particular issue, but I usually add a "-v" to pg_restore and run it within 'script' so it captures the output so I can see any errors |
| 14:02 |
csharp |
pgardella: that *should* work fine |
| 14:03 |
Dyrcona |
I usually do -C and -d postgres, but yeah. |
| 14:03 |
csharp |
it shouldn't matter, but the only things I do differently are PGUSER=evergreen (though that should be essentially the same as -U evergreen) and -j<number of concurrent jobs> |
| 14:03 |
pgardella |
The errors print to the screen |
| 14:03 |
Dyrcona |
yeah, I us -j 12 or so. :) |
| 14:04 |
pgardella |
:Dyrcona So your restore adds the extensions correctly? |
| 14:04 |
pgardella |
You'd think I never used IRC before ;) |
| 14:04 |
Dyrcona |
Well, they're in the database that's being restored more or less. |
| 14:04 |
csharp |
pgardella: I usually run the create_database_extensions.sql script in the Pg directory in the Evergreen source code |
| 14:05 |
Dyrcona |
I used to, but haven't done that lately. |
| 14:05 |
pgardella |
That's what's was in my prep script |
| 14:05 |
csharp |
gotcha |
| 14:05 |
Dyrcona |
I just straight up restore the database. |
| 14:05 |
pgardella |
I'll try it again with -C and -d postgres this time |
| 14:06 |
Dyrcona |
My dump command is pg_dump -Fc evergreen > /path/file |
| 14:07 |
Dyrcona |
This was my most recent restore: pg_restore -U evergreen -h localhost -C -c -d postgres -j 12 eg-backup.Tuesday-20170110.dmp |
| 14:08 |
csharp |
time PGUSER=evergreen pg_dump -F d -j 48 -v -f pines-prod-2.11.1-2017-01-14.dmp evergreen is mine |
| 14:08 |
Dyrcona |
I think -c is redundant with -C, but there it is. |
| 14:08 |
pgardella |
it's not. That I found. You have to add -c if you want it to drop exisitng data first |
| 14:09 |
Dyrcona |
I'd use more cores, but rebuilding metabib.real_full_rec.index_vector index takes most of the time. :) |
| 14:09 |
csharp |
PGUSER=evergreen pg_restore -v -j48 -d evergreen /var/lib/postgresql/pines-prod-no-auditor-2017-01-09.dmp |
| 14:09 |
Dyrcona |
pgardella: You are probably right. I haven't looked at the man page lately. |
| 14:09 |
csharp |
Dyrcona: true - it's super exciting for the first few minutes, then there's the process that takes forever |
| 14:10 |
csharp |
like... pingest.pl! :-) |
| 14:10 |
Dyrcona |
Yeah. |
| 14:10 |
pgardella |
Hey, this time it looks like it's working! |
| 14:10 |
csharp |
pgardella++ |
| 14:10 |
berick |
csharp: fyi more speed improvements https://bugs.launchpad.net/evergreen/+bug/1596595/comments/20 |
| 14:10 |
pinesol_green |
Launchpad bug 1596595 in Evergreen "Hold targeter features and refactoring" [Wishlist,New] |
| 14:10 |
csharp |
berick++ # codin' |
| 14:14 |
Dyrcona |
So, I'm going to give up on my z39.50 searches that are timing out and restore yesterday's dump to this server. |
| 14:17 |
pgardella |
Spoke too soon. I'm back to the function unaccent missing. |
| 14:19 |
pgardella |
Trying it again with the -U evergreen |
| 14:19 |
Dyrcona |
Sounds like you have a script to drop the database, recreate it, and restore into that. I used to do that, but lately I've found a restore just works for me. |
| 14:20 |
pgardella |
Dyrcona: That's what I'm trying now. |
| 14:20 |
Dyrcona |
yeah. |
| 14:20 |
pgardella |
I just keep making rookie mistakes over and over |
| 14:23 |
dbs |
We all do! |
| 14:24 |
JBoyer |
Dyrcona, I've been running into the same thing with our dump/restores. Since it's only 2-3 indexes I just rebuild them once it's done, but I've not been able to successfully use -C since moving from 2.9 to 2.11. |
| 14:25 |
Dyrcona |
JBoyer: OK. I'll look forward to that. I'm restoring 2.10. |
| 14:25 |
Dyrcona |
JBoyer: Is it indexes you're missing or functions used by indexes? |
| 14:26 |
JBoyer |
I don't really have any suggestions to go on right now, as adding the schema to the index def (public if memory serves) didn't help. |
| 14:27 |
JBoyer |
I get the same message pgardella was talking about when it tries to restore the indexes. Since the func isn't found the indexes don't exist when the restore complets. |
| 14:27 |
Dyrcona |
Is the unaccent extension in the db create script? I think so, but it might have been missed. |
| 14:28 |
JBoyer |
I've been using a complete dump restore (with -C) for ages, that's breaking. The only time I've ever bothered with pre-creating the db is when I want to restore with a different name. |
| 14:31 |
pgardella |
Latest tests done: With using the examples restores that dyrcona and charp gave me, I get the missing uaccent error and the missing functions. Using the pre-create the DB, I just get those missing functions. (I'm on 2.10.5, by the way.) |
| 14:31 |
Dyrcona |
What Pg version? |
| 14:32 |
pgardella |
I'm going to 9.3 right now, from a 9.1 db |
| 14:32 |
Dyrcona |
OK. My restores are going to 9.5 from 9.2. |
| 14:33 |
pgardella |
I did try originally a 9.5, but with 2.10.5, it only lists 9.3, so I went back to try that one. |
| 14:34 |
Dyrcona |
pgardella: It works with 9.5, if you run a couple of upgrade scripts after. |
| 14:34 |
pgardella |
I figured it would, but I want to get to a clean restore before I try the 9.5 route again. |
| 14:35 |
Dyrcona |
0979.function.grp-tree-functions-via-cte.sql & 0982.function.perm_functions_without_connectby.sql |
| 14:35 |
Dyrcona |
Sure. Makes sense. |
| 14:35 |
Dyrcona |
I seem to recall having fun with restores on 9.3. |
| 14:35 |
|
teletype01 joined #evergreen |
| 14:38 |
Dyrcona |
Umm, that's nice. I stopped all services, but some cstores and storages are still talking to the database... |
| 14:48 |
abowling |
have an overdue fine that ran to 20.94, but the item price is 13.96. anyone run into "gotchas" on this that might help me troubleshoot? |
| 14:49 |
|
khuckins joined #evergreen |
| 14:49 |
|
JBoyer joined #evergreen |
| 14:49 |
pgardella |
Dyrcona: are those two scripts I should run? |
| 14:50 |
Dyrcona |
If you switch your db to Pg 9.5, yes. |
| 14:50 |
pgardella |
ok |
| 14:50 |
Dyrcona |
The restore seems to work fine, then I run those scripts after. |
| 14:51 |
pgardella |
I'll need to stick with 9.3 then, as the whole reason I'm working on this copy is to set up a master-master db server in Nigeria. So I need to have the same functions on both sides of the atlantic. |
| 14:51 |
Dyrcona |
OK. |
| 14:58 |
csharp |
abowling: so is the fine supposed to be capped at the item price? |
| 15:01 |
bshum |
Did someone alter the price of the item after it hit max fine? :) |
| 15:01 |
bshum |
Assuming it was supposed to cap at the item price |
| 15:02 |
bshum |
Either that or someone put a 50% ratio on something |
| 15:02 |
abowling |
bshum: that's what i was wondering, but i don't think so. |
| 15:02 |
Dyrcona |
There's also a setting (IIRC) that controls whether or not overdues are removed when an item goes lost. |
| 15:02 |
bshum |
Like they keyed in a rate of 1.5 or something |
| 15:02 |
abowling |
bshum: exactly what i just figured out too |
| 15:02 |
Dyrcona |
So, the percent was over 100? |
| 15:02 |
abowling |
cause indeed 20.94 is 150% of 13.96 |
| 15:02 |
bshum |
There are ways of setting limits based on percentages |
| 15:03 |
abowling |
csharp: according to their director, it is |
| 15:03 |
bshum |
Usually I would expect people to set it down, like 0.75 or something |
| 15:03 |
bshum |
But maybe they got it backwards |
| 15:03 |
* bshum |
never played with the percentages |
| 15:03 |
abowling |
csharp: but i'm looking to verify that |
| 15:07 |
|
genpaku joined #evergreen |
| 15:08 |
dbs |
master-master! dark magic |
| 15:11 |
|
mmorgan1 joined #evergreen |
| 15:19 |
abowling |
bshum++ jboyer++ for your FINE help! |
| 15:20 |
abowling |
i've got all the confirmation that this is a head scratcher |
| 15:29 |
Dyrcona |
Well, I've confirmed the z39.50 search "bug" on master. |
| 15:29 |
Dyrcona |
Going to Lp it. |
| 15:37 |
JBoyer |
Dyrcona++ |
| 15:37 |
Dyrcona |
Funny thing: I'm not sure it affects the results. I'm trying to determine that. |
| 15:38 |
csharp |
for google music subscribers, integrate play music with your desktop: https://www.googleplaymusicdesktopplayer.com/ |
| 15:38 |
csharp |
off-topic, I know, but I had to share |
| 15:38 |
khuckins |
csharp++ |
| 15:39 |
csharp |
found when troubleshooting Fedora + Firefox + Flash + Google problem |
| 15:39 |
teletype01 |
thanks csharp ! |
| 15:43 |
Dyrcona |
And, it does appear to affect search results. |
| 15:44 |
JBoyer |
If it didn't that'd practically be a yes to "Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?" |
| 15:45 |
dbs |
Flash-- |
| 15:46 |
csharp |
dbs: srsly - apparently Google is kinda testing a flash-free version of play music, but it's not consistently/widely available |
| 15:47 |
csharp |
https://bugzilla.redhat.com/show_bug.cgi?id=1405287 - and then I'll shut up :-) |
| 15:47 |
dbs |
csharp: I think you have to go Chrome if you're doing anything with Flash these days (and yeah I hatesessss the long-standing Flash dependency Google Music has) |
| 15:48 |
dbs |
At least Evergreen doesn't have a Flash dependency. |
| 15:48 |
csharp |
the key word is "YET" |
| 15:48 |
csharp |
:-D |
| 15:48 |
dbs |
csharp++ |
| 15:48 |
JBoyer |
dbs, I see that as the only safe way to use it period. The only way Flash gets on a machine I'm in charge of is if Chrome or Edge keeps it up to date. |
| 15:48 |
dbs |
"Good news, here's a pull request replacing Dojo 1.3 with Flash" |
| 15:49 |
csharp |
haha |
| 15:49 |
dbs |
JBoyer: moi aussi |
| 15:49 |
JBoyer |
"And we call this new interface, Rooms." |
| 15:49 |
JBoyer |
Hah, I'm too harsh, it only looks vaguely flash-y. |
| 15:50 |
Dyrcona |
Flash is so 2005.... *rolls eyes* |
| 15:56 |
Dyrcona |
An, I misspoke. The search bug does NOT seem to actually affect the results, well not with a sample set of basically 1. :) |
| 15:56 |
Dyrcona |
So, nothing conclusive. |
| 16:04 |
|
genpaku joined #evergreen |
| 16:14 |
miker |
Dyrcona: what does your config.metabib_search_alias table contain? in particular, does it have an entry like this: |
| 16:14 |
miker |
alias | field_class | field |
| 16:14 |
miker |
------------------------+-------------+------- |
| 16:15 |
miker |
eg.keyword | keyword | |
| 16:17 |
Dyrcona |
miker: It does. |
| 16:19 |
miker |
and, I assume, an "eg.author" alias as well? assuming so, the problem is almost certainly in the SRU mapping, then |
| 16:20 |
Dyrcona |
Yeah, eg.author | author |
| 16:20 |
Dyrcona |
I saw this in production and in a stock master. |
| 16:21 |
Dyrcona |
I'm looking at the stock master db right now. |
| 16:28 |
miker |
Dyrcona: you have debug-level logging in apache, I'd be interested to see what lines that start with "SRU toEvergreen: " end in... |
| 16:30 |
Dyrcona |
miker: I don't but I'll activate it and try again. |
| 16:34 |
Dyrcona |
I'm not getting anything that looks like that in the logs. |
| 16:36 |
Dyrcona |
I do get this: GET /opac/extras/sru/CONS/holdings?version=1.2&operation=searchRetrieve&query=eg.author%20%3D%20mozart&startRecord=1&maximumRecords=0 |
| 16:36 |
Dyrcona |
When I did my SRU search I posted on the bug, I used version 1.1. |
| 16:39 |
Dyrcona |
This shows up in osrfsys.log: [2017-02-07 16:33:40] /usr/sbin/apache2 [INFO:1863:SuperCat.pm:1996:148650319118639] SRU search string [eg.author = mozart] converted to [eg.author:mozart site:CONS] |
| 16:41 |
Dyrcona |
I love how some parts of the system are using localtime and others are using UTC. |
| 16:41 |
Dyrcona |
I'm going to add some more info on a bug comment. |
| 16:59 |
Dyrcona |
So, some more digging and it looks like specifying a version=1.2 to SRU is the real issue. |
| 17:00 |
pinesol_green |
News from qatests: Test Success <http://testing.evergreen-ils.org/~live> |
| 17:08 |
|
mmorgan1 left #evergreen |
| 17:11 |
Dyrcona |
Well, that's it for today. |
| 17:30 |
|
mllewellyn left #evergreen |
| 17:34 |
|
jvwoolf left #evergreen |
| 18:05 |
csharp |
berick: early testing of your last fixes shows this error: ERROR: cannot pass more than 100 arguments to a function# |
| 18:06 |
csharp |
I can get the full output to you shortly |
| 18:11 |
csharp |
berick: http://pastebin.com/brVt8xrj |
| 18:12 |
berick |
doh! |
| 18:12 |
berick |
ok |
| 18:12 |
berick |
i can fix that, just use an array. |
| 18:13 |
berick |
csharp: if I don't patch until the morning, you all are still trucking along OK (minus the known issue)? |
| 18:14 |
csharp |
berick: yeah - all good after reverting the last two commits |
| 18:14 |
csharp |
berick++ |
| 18:14 |
berick |
csharp: ok, good, thanks for hanging in there |
| 18:26 |
|
jihpringle joined #evergreen |
| 18:44 |
|
bmills joined #evergreen |
| 18:49 |
|
teletype01 left #evergreen |
