| Time |
Nick |
Message |
| 01:16 |
|
sarabee joined #evergreen |
| 02:39 |
|
sarabee joined #evergreen |
| 06:40 |
|
rlefaive joined #evergreen |
| 07:58 |
|
jboyer-isl joined #evergreen |
| 08:30 |
|
rlefaive joined #evergreen |
| 08:32 |
|
mrpeters joined #evergreen |
| 08:37 |
|
maryj joined #evergreen |
| 08:38 |
|
mmorgan joined #evergreen |
| 08:41 |
jeff |
rangi: interesting. thanks for sharing! |
| 08:51 |
|
Dyrcona joined #evergreen |
| 09:29 |
Dyrcona |
Whee! Spend the first 20 minutes of my day looking through about 5 minutes-worth of acq logs.... |
| 09:29 |
* Dyrcona |
does not enjoy log diving. |
| 09:31 |
* mmorgan |
likes log diving. |
| 09:31 |
mmorgan |
Well, not the diving so much as finding the buried treasure. |
| 09:32 |
Dyrcona |
Well, I didn't find anything useful, just lots of looking up permissions and retrieving data. |
| 09:32 |
Dyrcona |
The user reported a "stuck" upload. |
| 09:32 |
* Dyrcona |
doesn't even know what that means, really. |
| 09:33 |
Dyrcona |
But, I advised just reloading the view and then started looking into the logs. |
| 09:33 |
Dyrcona |
By the time the user got back to me, turns out the upload had finished, but the client just didn't know. |
| 09:33 |
Dyrcona |
So, I stopped looking through the logs. |
| 09:34 |
Dyrcona |
All I know about acq, I learned from tickets, and it often takes me days to learn anything useful. |
| 09:44 |
mmorgan |
I've looked at the acq tables and logs, but am not too familiar with the workflow. It's hard to understand what's going wrong without that familiarity. Lots of moving parts to Acquisitions. |
| 09:52 |
kmlussier |
Dyrcona: Typically, when there's a stuck upload in acq, it usually finished behind the scenes. But it's then a pain for the user to find the PO where the lineitems were uploaded. |
| 09:52 |
kmlussier |
Dyrcona: We've found that missing permissions can sometimes cause a "stuck" upload. |
| 09:54 |
berick |
related: https://bugs.launchpad.net/evergreen/+bug/1514085 |
| 09:54 |
pinesol_green |
Launchpad bug 1514085 in Evergreen "Feature Request: Make Vandelay Asynchronous/Stateless" [Undecided,New] |
| 09:55 |
|
yboston joined #evergreen |
| 10:01 |
|
jwoodard joined #evergreen |
| 10:02 |
kmlussier |
Dyrcona: The CREATE_BIB_IMPORT_QUEUE is one permission we found causes that particular problem. |
| 10:07 |
tsbere |
kmlussier: I believe the problem came up after several others had uploaded fine. |
| 10:07 |
tsbere |
So I doubt permissions were at fault |
| 10:08 |
tsbere |
and I believe the resolution was "reload the interface and find it did, in fact, complete" |
| 10:10 |
kmlussier |
OK - just trying to add data points from previous experience. The resolution is pretty much the same when the permissions are the issue |
| 10:31 |
jeff |
kmlussier++ |
| 10:31 |
jeff |
datapoints++ |
| 10:38 |
|
rlefaive joined #evergreen |
| 10:48 |
|
RoganH joined #evergreen |
| 10:59 |
remingtron |
RoganH: you have the power to add jobs to the EG job web page, right? |
| 10:59 |
RoganH |
remingtron: yep |
| 10:59 |
remingtron |
RoganH: I sent one to the general list recently, and I saw another one come thru also. Could you add them? |
| 11:00 |
RoganH |
remingtron: sure I missed those, but I will do so today |
| 11:00 |
remingtron |
RoganH: many thanks! |
| 11:01 |
Dyrcona |
https://www.youtube.com/watch?v=0jnv0ejmLlY |
| 11:01 |
Dyrcona |
Your daily earworm courtesy of remingtron and RoganH. |
| 11:02 |
berick |
Dyrcona: haha, you never miss a chance to dis web apps |
| 11:02 |
RoganH |
Dyrcona++ |
| 11:02 |
Dyrcona |
berick: To be fair, I'm not sure you could make the tab flash in a desktop app, either. Probably depends on the toolkit. |
| 11:03 |
Dyrcona |
But don't mind me, in a ...mood... today. |
| 11:04 |
berick |
we just have the browser play The End (the doors) after it completes. |
| 11:04 |
Dyrcona |
berick++ # perfect! |
| 11:06 |
Dyrcona |
"Lost in a romance...wilderness of pain..." |
| 11:14 |
Dyrcona |
The End is probably perfect for my mood. |
| 11:14 |
* Dyrcona |
decided to play it. |
| 11:19 |
Dyrcona |
I think it would help if batch operations in general were asynchronous. |
| 11:19 |
Dyrcona |
Notifying the user that the work had finished in ways that don't interrupt their current work flow prove interesting. |
| 11:19 |
Dyrcona |
s/flow/could/ |
| 11:31 |
* Dyrcona |
should do another angular tutorial or at least more angular exercises. Use it or lose it as they say. |
| 11:34 |
|
vlewis joined #evergreen |
| 11:39 |
* berick |
sees FF finally supports websockets in shared workers |
| 11:39 |
berick |
good time to start testing that again for FF and to see if chrome fixed a bug from way back.. |
| 11:40 |
RoganH |
remingtron: weird, I never did find your's in my email but I found it via MarkMail, it'll be up in a moment |
| 11:44 |
remingtron |
RoganH: strange, glad MarkMail has it |
| 11:46 |
RoganH |
remingtron: published. and if there are email spirits I think they take pleasure in seeing how much they can mess with me sometimes. Or maybe I just have too much email and things go weird. /shrug |
| 11:50 |
remingtron |
RoganH: thanks. I think I've seen this issue before, where a list message doesn't come thru in someone's client, but there's proof it was sent (via MarkMail or another person's client) |
| 11:52 |
|
maryj joined #evergreen |
| 11:52 |
remingtron |
RoganH: I'm seeing some formatting issues with my job posting, mainly a few unexpected line breaks. |
| 11:52 |
|
bmills joined #evergreen |
| 11:53 |
|
bmills left #evergreen |
| 11:54 |
|
bmills joined #evergreen |
| 11:54 |
RoganH |
remingtron: yeah, those come through with posting sometimes. I'll double check it. I took a bunch out already but a few snuck through I guess. |
| 11:55 |
remingtron |
RoganH: thanks! |
| 11:56 |
remingtron |
RoganH: also, Grand Rapids Public LIbrary has stray capital "I" |
| 11:57 |
* remingtron |
is in nitpicking mode |
| 11:58 |
Stompro |
Does anyone know off the top of their head where the MARK Federateded search fields that are shown are configured? We had UPC added to ours, but the search isn't working, so I'm wondering if the config is wrong. |
| 12:00 |
RoganH |
remingtron: nit picking can be useful |
| 12:08 |
gmcharlt |
Stompro: Z39.50 source attributes |
| 12:08 |
Stompro |
gmcharlt, thanks. |
| 12:28 |
Stompro |
gmcharlt, How does that relate to the "Evergreen Catalog" search option in MARC Federated search, since there is no Z39.50 server entry for the local catalog? |
| 12:29 |
berick |
Stompro: that bit is hard-coded in the Z39 search API. |
| 12:30 |
berick |
Stompro: see open-ils.search.biblio.zstyle |
| 12:30 |
gmcharlt |
yep, around line 185 of Open-ILS/src/perlmods/lib/OpenILS/Application/Search/Z3950.pm |
| 12:38 |
Stompro |
Hmm, that looks pretty simple to add UPC to. Thanks for showing me the way berick++ gmcharlt++ |
| 13:24 |
|
jihpringle joined #evergreen |
| 13:42 |
|
bmills joined #evergreen |
| 13:52 |
|
sandbergja joined #evergreen |
| 15:01 |
jboyer-isl |
berick, bshum: I’m looking at web staff installation, is this basically how it goes: make -f …/Makefile.install *-developer; cd …/Open-ILS/web/js/ui/default/staff; npm install; bower install; grunt all ? |
| 15:04 |
berick |
jboyer-isl: by George, I think you have it. |
| 15:04 |
Dyrcona |
jboyer-isl: Yes. I do that before make install and after configure of course. |
| 15:04 |
Dyrcona |
berick++ # Better answer. :) |
| 15:04 |
jboyer-isl |
I needed both answers, I wasn’t certain re: order of ops. |
| 15:04 |
jboyer-isl |
Thanks much! |
| 15:04 |
jboyer-isl |
berick++ |
| 15:04 |
jboyer-isl |
Dyrcona++ |
| 15:06 |
Dyrcona |
BTW: I have found that it works better if you sudo -i before doing the make -f makefile.install *-developer. |
| 15:06 |
Dyrcona |
if you just sudo that command, you get a /tmp in your /home that is owned by root. |
| 15:07 |
Dyrcona |
You'll want to chown that back to your current user or one of the node.js commands will fail. |
| 15:08 |
Dyrcona |
That may just be a quirk of Ubuntu, but I would expect it to happen on Debian, too. |
| 15:08 |
jboyer-isl |
Dyrcona: well, in my case it’ll be a /tmp in the home dir for a user that is never used, but I’ll keep an eye out for it. Thanks |
| 15:09 |
jboyer-isl |
I’m definitely going to be finding the edges of the process since I’ve never done it before, heh. |
| 15:09 |
Dyrcona |
jboyer-isl: What I'm saying is the user that you're building Evergreen as will get that /tmp owned by root, then one of npm install, bower install, or grunt all will fail. |
| 15:09 |
Dyrcona |
IIRC, it is grunt all. |
| 15:09 |
Dyrcona |
I discovered this when building the 2.9-beta tarball. |
| 15:10 |
jboyer-isl |
Is it an issue re: the dir itself, or files within it? I may just pre-create it to get around that if it’s that simple. |
| 15:11 |
Dyrcona |
jboyer-isl: I believe it is just the dir itself, but you could chown -R to be sure. |
| 15:11 |
jboyer-isl |
(doing this by hand at least once might be a bit more instructive that worrying about conf management at the moment...) |
| 15:11 |
Dyrcona |
my setup process on vms typically invovles a sudo -i before I install prerequisites. |
| 15:12 |
Dyrcona |
Yeah, I did it by hand a couple of times before I added it to my setup scripts. |
| 15:12 |
Dyrcona |
Took two or three tries to get it right. |
| 15:13 |
Dyrcona |
You can also go back and do it after make install, but you'll need to run make install again with the same STAFF_CLIENT_VERSION. |
| 15:14 |
Dyrcona |
That was how I did it the first few times, before I realized the proper order of things. |
| 15:16 |
jeffdavis |
Has anyone else had trouble with locally hosted cover art (i.e. local images not displaying) in 2.8+? |
| 15:21 |
jeff |
jeffdavis: i don't think i've heard any reports. do you have any more details? |
| 15:23 |
jeff |
Dyrcona: the sudo behavior that you describe is pretty typical of how sudo works. I'm fairly certain that none of our documentation recommends the use of sudo. |
| 15:23 |
mmorgan |
jeffdavis: I'm not aware of problems, but we only have a few. Can you remind me where those are stored on the server so I can find a record number to test? |
| 15:23 |
Dyrcona |
jeff: I didn't think it did, but I know we administrators sometimes take shortcuts. |
| 15:24 |
jboyer-isl |
jeff: That’s true, but this is the only situation I’m aware of that it matters. |
| 15:24 |
jeff |
mmorgan: /openils/var/www/opac/extras/ac/jacket/medium/r/RECORDNUMBER by default |
| 15:24 |
Dyrcona |
And, it was something that I remembered being an issue for me, so thought I'd throw it out there. |
| 15:24 |
mmorgan |
jeff++ |
| 15:24 |
mmorgan |
I'll take a look. |
| 15:24 |
jeff |
"medium" may be one of small, medium, large |
| 15:25 |
jeff |
and you could still have isbn-based images |
| 15:25 |
jeff |
which wouldn't be in the r/ subdir |
| 15:26 |
tsbere |
jeffdavis: I don't see any major issues with the code or default apache configs. |
| 15:26 |
tsbere |
jeffdavis: Though, thinking about it, were you loading jacket images by ISBN before? |
| 15:26 |
jeff |
tsbere: have recent changes broken backward compatability with that? |
| 15:27 |
tsbere |
jeff: I was thinking more "If TPAC is loading based on record ID, and their customs are all ISBN-based, the customs won't show up any more" |
| 15:27 |
tsbere |
Still there, but not referenced |
| 15:27 |
jeff |
ah, right. |
| 15:27 |
jeffdavis |
tsbere: We load images by ISBN for most sites. We have one site that has locally hosted images. On 2.6 the local images were loaded based on record ID; on 2.8 they're not showing up, and the filename appears to be ISBN based. |
| 15:28 |
jeffdavis |
i.e. in AddedContent.pm, around line 86, $r->filename is /openils/var/web/opac/extras/ac/jacket/small/{ISBN}, not /openils/var/web/opac/extras/ac/jacket/small/r/{RECORDID} |
| 15:29 |
tsbere |
jeffdavis: And the file exists and has the correct ownership? |
| 15:29 |
mmorgan |
jeffdavis: We're on 2.8.2 and ours seem to be working fine: http://evergreen.noblenet.org/eg/opac/record/3578725 |
| 15:30 |
jeffdavis |
tsbere: well, jacket/small/r/{RECORDID} exists with correct ownership/perms; jacket/small/{ISBN} does not exist locally |
| 15:30 |
jeffdavis |
mmorgan: interesting, thanks for checking! |
| 15:30 |
tsbere |
jeffdavis: If you are loading off of ISBN and the jacket is on the record ID that would be the opposite of what I thought. |
| 15:31 |
tsbere |
Still the general idea of "the custom jacket is rigged for one lookup method, but you are using the other" though |
| 15:31 |
mmorgan |
Note that we haven't added any since we've been on 2.8, though. |
| 15:35 |
jeff |
jeffdavis: check to ensure that the size you're trying to load exists -- you said that jacket/small/r/{RECORDID} exists, but are you trying to load a medium or large url? |
| 15:36 |
tsbere |
jeff: From my understanding, the /r/ image exists, but the site is using ISBN-based images, not record-based. |
| 15:36 |
jeff |
jeffdavis: try loading a jacket image URL by hand. if the URL doesn't load, i'd suspect apache or perl mod issues. if the URL loads but doesn't appear in the TPAC I'd suspect templates being out of date, a bug, or conflict with other AC provider. |
| 15:38 |
|
jlitrell joined #evergreen |
| 15:39 |
jeff |
the design is that apache hands to the perl handler, and the perl handler looks for a file on disk. if present, it defers back to apache to serve the file. |
| 15:50 |
jeffdavis |
The problem seems to be that the Perl handler is looking for a file at path jacket/small/{ISBN} when it should be looking for a file at jacket/small/r/{RECORDID} |
| 15:50 |
tsbere |
jeffdavis: That would be due to using ISBN lookups and not record ID lookups |
| 15:50 |
jeffdavis |
I can load the /r/ image URL directly: https://testing2.catalogue.libraries.coop/opac/extras/ac/jacket/small/r/109936309 |
| 15:51 |
tsbere |
jeffdavis: You would have to tell TPAC to use record ID lookups instead |
| 15:51 |
jeffdavis |
tsbere: prior to 2.8 we could use ISBN lookups when grabbing remote content, but record ID when using local content, and the latter would override the former. |
| 15:52 |
tsbere |
jeffdavis: The record ID lookup will do a remote ISBN lookup anyway, pulling the ISBN from the record to do so. |
| 15:53 |
tsbere |
jeffdavis: I do not recall seeing "do a record ID lookup if there is no ISBN" but you might have that as a local customization? |
| 15:54 |
jeffdavis |
No, that's backwards. It used to do a local record ID lookup first, and then default to remote ISBN lookup (or whatever identifier you were using for remote content). |
| 15:54 |
tsbere |
jeffdavis: That is what record ID lookup *does*. It never requests the ISBN URL, though, because the backend module does the ISBN image lookup based on the record ID. |
| 15:55 |
tsbere |
Thus, if you put a /r/ jacket image in place it loads that, then it checks the cache, then otherwise it loads the MARC, looks up the ISBNs/UPCs/whatever, does a lookup, and returns what it gets |
| 15:56 |
tsbere |
AKA, unless you want to put override images in by ISBN you should probably tell the TPAC to load by record ID across the board. |
| 16:25 |
* mmorgan |
is trying to create an action trigger that sends an email message and also creates a usr_message for the patron's message center. Is it possible to do both with one trigger, or do I need 2 different triggers? |
| 16:27 |
berick |
mmorgan: you can do it in the same trigger |
| 16:27 |
berick |
mmorgan: see the "Message ..." fields |
| 16:29 |
mmorgan |
berick: Ok, thanks. Thought I could do it with one, but I can't get the message part to work. |
| 16:29 |
berick |
e.g. the "template" is what goes into the email. the "message template" is what goes into the user message. |
| 16:29 |
berick |
pretty sure all the message fields have to be set, too |
| 16:29 |
jeff |
"When an A/T definition has both message_usr_path and message_template defined, we gather the user, sending lib, template, and title and push those into the environment." |
| 16:29 |
jeff |
"After the user-defined reactor has successfully run, we pass the environment off to a special-purpose handler that generates output for actor.usr_message.message and, upon success of that, we insert the user message into the database." |
| 16:30 |
jeff |
quoting from commit 6e373d0 |
| 16:30 |
pinesol_green |
[evergreen|Mike Rylander] LP#1410369: Create user messages when configured to do so - <http://git.evergreen-ils.org/?p=Evergreen.git;a=commit;h=6e373d0> |
| 16:34 |
mmorgan |
jeff: Thanks for the reference. I must just have something out of kilter. Will keep poking. |
| 16:34 |
mmorgan |
berick++ jeff++ |
| 16:38 |
jeffdavis |
thanks tsbere and jeff, I think that got me looking where I needed to be looking |
| 16:45 |
|
sarabee joined #evergreen |
| 17:06 |
* mmorgan |
is calling it a week. Happy Thanksgiving! |
| 17:07 |
|
mmorgan left #evergreen |
| 21:51 |
|
bmills joined #evergreen |
| 22:52 |
|
dbwells joined #evergreen |
