| Time |
Nick |
Message |
| 00:16 |
|
Rogan joined #evergreen |
| 01:17 |
|
NIngalls joined #evergreen |
| 02:13 |
|
NIngalls joined #evergreen |
| 07:05 |
|
sleary joined #evergreen |
| 07:32 |
|
bgillap__ joined #evergreen |
| 08:00 |
|
BDorsey joined #evergreen |
| 08:28 |
|
dguarrac joined #evergreen |
| 08:30 |
|
redavis joined #evergreen |
| 08:39 |
|
mmorgan joined #evergreen |
| 08:48 |
|
sleary joined #evergreen |
| 08:50 |
|
mantis1 joined #evergreen |
| 09:07 |
|
terranm joined #evergreen |
| 09:32 |
|
Dyrcona joined #evergreen |
| 09:37 |
Dyrcona |
I was looking for the list of things not to do in Postgres that jeff(?) shared recently, and I found this gem on the Pg wiki: https://wiki.postgresql.org/wiki/Don't_Do_This#Don.27t_use_table_inheritance |
| 09:59 |
|
sleary joined #evergreen |
| 10:07 |
mmorgan |
claiming 1386 |
| 10:11 |
Stompro |
Baker & Taylor just let me know that they made progress on their cover art issues. It isn't 100% though... we just found some new ones, so they have finally cleared the cache, but I'm not sure they solved the problem. |
| 10:13 |
|
redavis joined #evergreen |
| 10:16 |
|
kmlussier joined #evergreen |
| 10:18 |
Dyrcona |
Grrr..... |
| 10:23 |
kmlussier |
Stompro: We've been pushing them on this issue. In a conversation we had a couple of weeks ago, they said they thought they could solve it on their end. I'm following up with them today because we are still seeing problems with the small selection of ones they thought they resolved. |
| 10:26 |
Dyrcona |
I would really appreciate some knowledgable eyes on https://pastebin.com/C3sCfC0e . My gut tells me I'm missing something really simple. |
| 10:27 |
|
mantis1 left #evergreen |
| 10:31 |
Dyrcona |
Also, on the above example, the db server does not log any errors... Oh wait a minute....I'm looking at the wrong database, I bet. |
| 10:32 |
Dyrcona |
No, I am looking at the correct database instance. |
| 10:34 |
berick |
Dyrcona: https://gist.github.com/berick/2a9cfb83548a7e8e47cb2ecd1e8437ea |
| 10:35 |
berick |
"-not" maybe? |
| 10:36 |
Dyrcona |
Yeah, I had "-not" originally, and I just put it back still nothing. |
| 10:36 |
berick |
well, a new error |
| 10:37 |
Dyrcona |
I wonder why I'm not getting errors logged on my db? |
| 10:37 |
berick |
the would be cstore errors |
| 10:37 |
berick |
never hit the db |
| 10:37 |
Dyrcona |
OK. I'll check the other logs... |
| 10:38 |
pinesol |
News from commits: LP1991294 - Stamping upgrade script <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=9966cc19577d585bcdb157ccdc32d72f716ea94a> |
| 10:38 |
pinesol |
News from commits: LP1991294 Add release note <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=03229b70cddd35aef5574947493884d9a7d4e5db> |
| 10:38 |
pinesol |
News from commits: LP1991294 Add Added Content tab ou settings and perm to seed data <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=92444f2f8bce0dc1a32cacdda3ac8cafebc161bc> |
| 10:38 |
pinesol |
News from commits: LP1991294 Staff NoveList Select Added Content tab <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=0f3e0f70423508aff782f087f676c1773d1c7922> |
| 10:39 |
Dyrcona |
It just tells me sever query error see logs for details, but there are no useful details. |
| 10:39 |
berick |
Dyrcona: the /openils/bin/test_json_query will give the details |
| 10:40 |
berick |
though i'm surprised it's not also in the logs |
| 10:42 |
Stompro |
kmlussier, how are you finding the problems... manually through staff reports? I'm considering creating a script to compare requests from the method that works for all our ISBNs to get a better list of problems. |
| 10:42 |
Dyrcona |
Do you mean the Everreen stderr logs or the Postgresql logs? |
| 10:43 |
berick |
i would expect json query issues to be logged with the cstore logs |
| 10:45 |
Dyrcona |
berick++ : I found this: "open-ils.cstore: Possible SQL injection attempt; direct order by is not allowed" |
| 10:47 |
Dyrcona |
I had the same order_by working in a cstoreeditor search_action_circulation yesterday. |
| 10:49 |
berick |
some variation between json_query and the .direct. api |
| 10:50 |
Dyrcona |
Well, I've tried several vatiations on the order by, and no dice. |
| 10:51 |
|
briank joined #evergreen |
| 10:52 |
berick |
"order_by":[{"class":"circ","field":"usr"}] |
| 10:53 |
kmlussier |
Stompro: They really aren't difficult to find. We pull up a carousel on a library's web site, and it looks like a jack-o-lanterns smile. |
| 10:54 |
kmlussier |
Stompro: If you create a script, I'm sure our staff would appreciate trying it out! :) |
| 10:57 |
Dyrcona |
berick: Thanks. I removed the order by and I sill have issues. It's complaining about +circ in the subquery where clause, so I added "=" and it says "=" is not a column. II I do "--=", it's an invalid operator...... |
| 11:00 |
Dyrcona |
I changed it to "target" => {"=" => {"+circ" => "id"}} and now, I get: Not an ARRAY reference at ./fix line 49. |
| 11:02 |
Dyrcona |
Oh! It's an arrayref of hashrefs... berick++ |
| 11:17 |
Dyrcona |
I wonder if I could do all of that with a cstoreeditor search? |
| 11:19 |
eeevil |
to confirm, it's def "-not" rather than "not", and there are several order-by vocabularies, but [{class:"circ",field:"usr",direction:"asc"},...] is the "best" |
| 11:19 |
eeevil |
however, there is also "-not-exists" |
| 11:20 |
eeevil |
which is probably what you want instead of the separate layers |
| 11:26 |
|
terranm65 joined #evergreen |
| 11:29 |
eeevil |
and, you /may/ be able do that via new_editor()->search_action_circulation(), but I'm not certain ... it's certainly technically possible, search and json_query use the same internal sql-building logic, but search might be explicitly restricted for security ... I just don't recall if the query frame for the outer is passed to a correlated subquery in search mode the way it is in json_query mode |
| 11:33 |
Dyrcona |
eeevil: Thanks for the suggestions. I've got it working. I'm retrieving the circs from the ids. I guess I don't really need the order by because I'm jamming it all in a hash indexed by the user id. Originally, I thought I'd loop through the results from the editor search_action-_circulation, but then the query became more complicated.... |
| 11:34 |
Dyrcona |
I've used order by and limits in searches, along with id_list, but I've not tried joins before. |
| 11:35 |
Dyrcona |
Now, to see if I can just rune the AutoRenew handler like I think I can..... |
| 11:38 |
|
sandbergja joined #evergreen |
| 11:39 |
|
terranm joined #evergreen |
| 11:40 |
sandbergja |
terranm and I are claiming 1387 |
| 11:40 |
Dyrcona |
sandbergja++ terranm++ # pushing code |
| 11:41 |
mmorgan |
sandbergja++ terranm++ |
| 11:44 |
Dyrcona |
Hey! It seems to be working. My original SQL query returns fewer rows each time I run it. |
| 12:04 |
|
jihpringle joined #evergreen |
| 12:09 |
pinesol |
News from commits: LP1812241: stamp upgrade script <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=198570e4beb8736c194fa31a4d0ed57d7e4a6c6b> |
| 12:09 |
pinesol |
News from commits: LP1812241 follow-up: add default link note to bootstrap opac, details view, and addit... <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=9aecb6a1b21581edbc7886c2f37174e46d06b296> |
| 12:09 |
pinesol |
News from commits: LP1812241: add YAOUS for default URI note text <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=ebbe85d30119f5da289d23a9e854df6e1ab495de> |
| 12:12 |
Dyrcona |
Use of uninitialized value $auto_remaining in numeric lt (<) at /usr/local/share/perl/5.34.0/OpenILS/Application/Trigger/Reactor/Circ/AutoRenew.pm line 81. # I'll have to look into that one. Probably hit some that won't renew or something. |
| 12:13 |
Dyrcona |
AutorenewNotify events are being created. |
| 12:15 |
|
sleary joined #evergreen |
| 12:34 |
Dyrcona |
I'm going to paste the program somewhere others can get it. You might find it useful. |
| 12:34 |
|
terranm25 joined #evergreen |
| 12:55 |
Dyrcona |
Hmm.. I wonder if I could "select *" from circ and then use circ->from_bare_hash to make the ojbects? It would save a few thousand db queries..... |
| 12:56 |
Dyrcona |
I'm going to experiment with that. |
| 13:06 |
Dyrcona |
Well, that's interesting: Nov 9 13:03:43 egtest kernel: [794612.251537] opensrf-c[15727]: segfault at 55eadc21e790 ip 00007f6b9fd05482 sp 00007ffc063ac9c0 error 6 in libopensrf.so.2.1.1[7f6b9fcf7000+1f000] |
| 13:06 |
Dyrcona |
N |
| 13:08 |
Dyrcona |
That comes right after running the query. |
| 13:11 |
Dyrcona |
Well, this other vm really doesn't like running that query with *. |
| 13:24 |
Dyrcona |
The JSON tutorial says that {"circ":"*"} or {"circ":null} will use the default selection list, i.e. all fields. This 3.7.4 VM really doesn't like that. I get segfaults. Is it just too much data, I wonder? |
| 13:28 |
kmlussier |
sandbergja++ terranm++ |
| 13:41 |
|
redavis joined #evergreen |
| 13:47 |
|
sleary joined #evergreen |
| 14:01 |
|
BDorsey_ joined #evergreen |
| 14:16 |
|
terranm joined #evergreen |
| 14:24 |
|
Dyrcona joined #evergreen |
| 14:25 |
Dyrcona |
berick | eeevil: In case you have any ideas: https://pastebin.com/5Z9M9jNZ |
| 14:28 |
|
kworstell-isl joined #evergreen |
| 14:31 |
Dyrcona |
Oh, cool. Looks like a busted libc on Ubuntu and a few other distributions, maybe. |
| 14:33 |
* Dyrcona |
tries re-making OpenSRF and Evergreen. Maybe a linker issue. |
| 14:37 |
Dyrcona |
Nope. Something is definitely busted somewhere. |
| 14:39 |
Dyrcona |
I can also say that it's nor Redis' fault. One vm is using Redis and the other isn't. |
| 14:54 |
Dyrcona |
So, after doing the make clean, make, make install tango, it's now error 6 in libopensrf.so. Removing the order by doesn't make a difference. It looks like the problem is the {"select":{"circ":"*"}} because going to selecting just id and usr stops the segfaults. |
| 14:54 |
Dyrcona |
That's a pity because I think that version would be better if it didn't crash. |
| 15:02 |
|
sandbergja joined #evergreen |
| 15:28 |
|
terranm joined #evergreen |
| 15:31 |
Dyrcona |
And, nobody cares.... I should try it with produciton data on Debian to see if it is consistent. |
| 15:32 |
Dyrcona |
I find it funny that trying to past the query to IRC crashed my client, too. That's just some bad code, I tell ya. |
| 15:32 |
Dyrcona |
Why can't I type "paste?" |
| 15:34 |
terranm |
I care. I just don't know what you're talking about :) |
| 15:37 |
mmorgan |
:) |
| 15:41 |
|
jihpringle joined #evergreen |
| 15:42 |
|
kmlussier1 joined #evergreen |
| 15:44 |
|
mmorgan1 joined #evergreen |
| 15:46 |
Dyrcona |
:) |
| 15:46 |
Dyrcona |
Well, a certain valid query can crash opensrf, and I don't know why. |
| 15:47 |
Dyrcona |
Also, it looks like auto_renewal_remaining can be null, so the autorenew code should be adjusted to take that into account or auto_renewal_remaining should be not null with a default of zero..... |
| 15:47 |
* Dyrcona |
sighs. |
| 15:48 |
Dyrcona |
And, it looks like my query for the autorenewal fixer should take stop_fines into account. I don't want to renew LOST or CLAIMSRETURNED. I think I'll just use "stop_fines":null for now. |
| 15:52 |
Dyrcona |
renewal_remaining can also be null. Looks like if the circ rules doesn't allow renewals, these are null, but the autorenew code assumes that they'll have a number in the fields. |
| 15:53 |
Dyrcona |
Hmmm. how best to say "is not null" in JSON? |
| 15:57 |
Dyrcona |
"Any other operator used with "null" turns into IS NOT NULL." Handy, that. |
| 15:57 |
Bmagic |
Anyone have any suggestions for resolving SSO looping? With nginx reverse proxy in front of Apache. We've got it where Evergreen is opening the SSO page (MS auth) - and then auth happens, and a redirect back to the OPAC, but then, Evergreen doesn't seem to respect the auth and redirects again |
| 16:06 |
Dyrcona |
mmorgan++ for Lp 1835953 |
| 16:06 |
pinesol |
Launchpad bug 1835953 in Evergreen "Autorenewals should not be attempted on circs where auto_renewal_remaining is NULL" [Medium,Confirmed] https://launchpad.net/bugs/1835953 |
| 16:28 |
jeff |
general handy Chrome tip: the usual Help -> About Google Chrome interface will almost always trigger an automatic update if available, so it’s not a great way to see what version of Chrome you’re on without potentially changing the version of Chrome. The URL chrome://version does not appear to have that side effect. |
| 16:31 |
Dyrcona |
Meh.... test_json says my JSON has a syntax error that I can't see. |
| 16:32 |
Dyrcona |
"Expected quotation mark to begin hash key; didn't find it" That's because there's no hash key at position 595 and I don't think a comma is required there either. |
| 16:37 |
Dyrcona |
The braces are all balanced, too. That's one thing Emacs is really good at ... balancing parenthesis-like characters. |
| 16:40 |
Dyrcona |
Ok. I think it was a stray comma from the Perl that I converted. |
| 16:40 |
Dyrcona |
Too much editing of this thing.... |
| 16:44 |
Dyrcona |
Norhing like last minute changes to ruin your evening and your day off.... (I'm thinking of tomorrow.) |
| 16:50 |
Dyrcona |
Awesome sauce. I think I have something to run tonight that will do what I need. |
| 17:06 |
|
kmlussier1 left #evergreen |
| 17:08 |
|
mmorgan1 left #evergreen |
| 17:09 |
|
terranm joined #evergreen |
| 17:09 |
terranm |
Claiming 1388 |
| 17:40 |
pinesol |
News from commits: Merge branch 'main' of git.evergreen-ils.org:Evergreen <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=3a4181dcf07c497dc6415d29d0eb54234a2c221e> |
| 17:40 |
pinesol |
News from commits: LP1899976 Stamp upgrade script <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=0508ed23f393cfe4c749eed78c7181bf09b38d82> |
| 17:40 |
pinesol |
News from commits: LP1899976 Adjust AutoRenew Delay <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=64ea028e6abfe1ebef2d685ddf23a1760716fdfa> |
| 17:40 |
pinesol |
News from commits: Lp 1899976: Adjust AutoRenew Delay <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=3c26b261dc18a86ef7210cd95f25402db061823a> |
| 17:41 |
terranm |
Bah! |
| 18:03 |
|
Rogan joined #evergreen |
| 18:10 |
pinesol |
News from commits: LP2040303 Record Match Sets: Buttons out of alignment <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=35f823507ea3b885d88bd77c34ddfe6bfda7fc2c> |
| 18:10 |
pinesol |
News from commits: LP2039606 Multiselect accessibility & button size <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=ac189f851081f9c3ae3d0337ff8daafcc5422fdc> |
| 18:10 |
pinesol |
News from commits: LP1812414 Keyboard shortcut to log out <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=ed11404dd3869a7a87cfb66137b88868e1529d9f> |
| 18:40 |
pinesol |
News from commits: LP2039310 Change Vandelay queue links to buttons <https://git.evergreen-ils.org/?p=Evergreen.git;a=commitdiff;h=1e43062874764d2a65c50d425b2a2630f798f663> |
| 19:47 |
|
jihpringle joined #evergreen |
| 19:58 |
|
jihpringle joined #evergreen |
| 20:30 |
|
sandbergja joined #evergreen |
| 21:56 |
|
sandbergja joined #evergreen |
| 22:18 |
|
Rogan joined #evergreen |
