Evergreen ILS Website

IRC log for #evergreen, 2022-06-17

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat

All times shown according to the server's local time.

Time Nick Message
06:02 pinesol News from qatests: Testing Success <http://testing.evergreen-ils.org/~live>
07:05 kmlussier joined #evergreen
07:05 kmlussier bshum++
07:05 kmlussier parts++
07:14 collum joined #evergreen
07:15 _collum joined #evergreen
07:16 JBoyer kmlussier++
07:16 JBoyer :D
07:22 kmlussier Excellent! I have karma for one more year! Thanks JBoyer!
07:31 rjackson_isl_hom joined #evergreen
08:08 kmlussier I'll just leave this here - https://docs.google.com/spreadsheets/d/1U_9sQ9snIg​F4TM9U39dD4XkjCUbbRxDHxjcqTP6gjlw/edit?usp=sharing and I'll see you all again around this time next year. Happy Friday!
08:16 RFrasur joined #evergreen
08:23 rhamby just in case anyone missed the registration link https://hopin.com/events/evergreen-internationa​l-hackfest-2022?code=R7fHd20cOcsFkNlh7hoGEZQbG
08:34 RFrasur There was mention yesterday about an issue with B&T's FTP.  Anyone know if that's an issue still?
08:41 RFrasur Hmm, it appears to be.
08:42 mmorgan joined #evergreen
08:43 miker should I be concerned that my "eeevil" karma is higher than my "miker" karma?
08:46 mmorgan kmlussier++
08:47 mdriscoll joined #evergreen
08:51 miker berick: I plan to start poking at the redisrf code, but wanted to follow up on the part that confused me most yesterday.  you were saying the router would need to listen on multiple sockets, but I still can't see why.  in xmpp, we need multiple sockets because the jid is the analog of a redis stream or list, but is /also/ the endpoint id of an xmpp session.  in redis, the router would just have one "incoming" connection to the "local" redis server and
08:51 miker it would xgroupread or blpop from all the lists/streams together, in one command. it would need to connect (perhaps persistently, perhaps transiently) to "foreign" redis servers, to throw messages at "foreign" lists/streams that have a listener (or drones directly!) behind them, but that's all during the processing loop for the message in hand and is all outgoing.
08:57 miker (also, as a side note, because stream entries are each an order set of key/value pairs, using streams rather than lists would let us unwrap some of the message envelop to be natively redis. IOW, we could reduce a /lot/ of the de/re-serialization by putting things like request id, thread trace, originating sender coordinates, etc, outside the message blob)
09:02 Dyrcona joined #evergreen
09:04 Dyrcona @karma
09:04 pinesol Dyrcona: Highest karma: "bshum" (2), "kmlussier" (2), "evergreen" (1), "community" (1), and "parts" (1).  Lowest karma: "evergreen" (1), "community" (1), "parts" (1), "bshum" (2), and "kmlussier" (2).
09:04 Dyrcona bshum++ kmlussier++
09:07 mantis1 joined #evergreen
09:11 Dyrcona Good morning, #evergreen!
09:12 Stompro joined #evergreen
09:13 mmorgan Good morning!
09:13 Stompro joined #evergreen
09:22 rjackson_isl_hom joined #evergreen
09:40 eby miker: / berick : going to test more next week but curious with your mention of setting workstation prefs as org unit settings. Does it act as a fall back default or as a forced preference?
09:49 phasefx berick: do you want to talk SIP2M sometime today?  I don't have an agenda, but we could determine/sanity-check next steps
09:51 berick eby: it acts as fall back default.  ignored in cases where a workstation setting value is already saved.
09:52 berick best way to create is to save the value you want at a workstation, then copy the JSON from the value into the newly created org unit setting.
09:52 eby Perfect thanks!
09:52 eby berick++
09:52 mmorgan eby: But you could also remove the ws setting type (and ws settings) if you want to force it.
09:52 berick mmorgan: yes, good call
09:52 berick if you want to force it, remove the WS setting entirely
09:53 mmorgan We have not done that, but have made wide use of the fallback!
09:53 mmorgan berick++
09:53 berick phasefx: yes, happy to talk today
09:54 phasefx berick++ I'll refresh my memory this morning and you can poke me whenever you'd like after lunch?
09:55 berick +1
09:55 Stompro eby, I started a bug on documenting the workstation settings/org unit stuff that may be useful to you - https://bugs.launchpad.net/evergreen/+bug/1843454
09:55 pinesol Launchpad bug 1843454 in Evergreen "Docs: Default/Read Only Print Template Options" [Undecided,New]
09:57 mmorgan Stompro++
09:57 berick Stompro++
09:58 jvwoolf joined #evergreen
09:59 Dyrcona @decide local git or virtual machine git
09:59 pinesol Dyrcona: go with local git
10:07 * Dyrcona wonders if we should add some instrumentation to pingest.pl (and possibly drop the .pl extension).
10:16 Dyrcona Hmm. I guess rebasing from master to rel_3_7 is not a good idea.
10:18 miker Dyrcona: do I feel my ears burning? ;) (if you're looking at what I think you're looking at, if you can test closer to master I think you'll have more success)
10:18 Dyrcona Remote syntax in git can be annoying: orign/rel_3_7 for one command and origin[space]rel_3_7 for another.
10:19 Dyrcona miker: I am. I plan to start testing with rel_3_7 today and an upgrade to rel_3_9 next week if that's OK.
10:20 mdriscoll joined #evergreen
10:20 Dyrcona I suppose the different syntax is what they call commitish (in the / case) and remote[space]branch in the latter.
10:22 Dyrcona I should copy the database that I want to use so it won't get obliterated this weekend.
10:23 * Dyrcona is tempted to test on Pg 14 since there was an update this morning... ;)
10:43 miker Dyrcona: as far as order and timing goes, I'm just glad someone's looking at it, but just to note that deadlock+symspell branch was developed after 3.7 (recall, nocase browse) so starting with vanilla 3.8 or 3.9 (/not/ patched with this and then upgraded) may be less of a headache. I get that "with your data" testing isn't particularly feasible at scale for versions past what you're actually on, though.
10:46 Dyrcona miker: Right. I previously made a 3.7 backport and only had had to change the db upgrade script. Looks the same still, but I'll see what happens. If it just blows up, I'll move to upgrading to 3.9 over the weekend.
10:47 Dyrcona I'm going to compare the implementation of metabib.reingest_metabib_field_entries in the two db upgrades. I think we can delete the one from the WWW upgrade.
10:51 Dyrcona Yeah, only difference is the check for reification.
10:52 Dyrcona Barring any other surprises, I think it will work on 3.7 with just the same modifications as before.
11:01 Dyrcona Downgrading on a vm is also fun...
11:07 miker right, if we take both commits, we just need the latest one
11:07 jihpringle joined #evergreen
11:07 miker downgrading? never 'eard of 'im.
11:11 csharp_ downgrader? I never even met her!
11:11 Dyrcona :)
11:11 * csharp_ can't hackfest today - tied up in holds issue
11:11 Dyrcona :(
11:11 csharp_ frakking age protection
11:11 Dyrcona csharp_++
11:11 csharp_ it's for the birds
11:12 csharp_ a workaround for some nimrod in the early PINES planning meetings who said he was never going to buy new items again because other libraries could buy them for them
11:12 * Dyrcona is too lazy to build a new vm, so rm -rf /openils/*, and reinstall OpenSRF and older version of Evergreen. Also had to rm node, but I wrote a script for that some years ago. :)
11:12 csharp_ libraries said "oh yeah?" - thus age protection
11:13 berick hold my beer
11:13 Dyrcona Age hold protection is a decent feature IMHO. It just doesn't always do what the users expect.
11:14 * Dyrcona takes a short walk while the database is copied...
11:14 mmorgan csharp_: Trying to understand the issue. Age protection expires at some point ...
11:14 csharp_ it's fine, I'm just mad that it's failing for some reason
11:14 csharp_ some stupid HUMAN did something somewhere that broke stuff
11:14 * csharp_ may be the human in question :-)
11:15 berick spiderman <-> spiderman
11:15 mmorgan Failing to protect?
11:15 berick irc_memes--
11:15 mmorgan Or being over-protective?
11:16 Dyrcona @blame Humans
11:16 pinesol Dyrcona: Humans HAXORED Dyrcona's SERVERZ!!!!
11:16 csharp_ berick++
11:16 miker csharp_: I don't recall hearing that age protection story, but I don't doubt it. basically why public bookbags took so long to happen :)
11:16 mmorgan Darn humans :-/
11:17 csharp_ my solution to the problem with a consortium like PINES is to have a consortial agreement to purchase or rent pop materials with a formula in place for fair distribution
11:17 csharp_ during the first 12 months of life, those items are consortium owned
11:18 csharp_ after which the copies are fairly distributed to the libraries to be "owned" by them outright (unless they're rentals)
11:19 csharp_ miker: yeah -  I think this was in the 1998-1999 timeframe if my understanding is correct
11:19 csharp_ the director in question is long dead, I believe, but his ghost haunts us all
11:20 csharp_ @blame add $who - what a bunch a bastards!
11:20 pinesol csharp_: Error: You must be registered to use this command. If you are already registered, you must either identify (using the identify command) or add a hostmask matching your current hostmask (using the "hostmask add" command).
11:20 csharp_ @blame add $who - what a bunch a bastards!
11:20 pinesol csharp_: The operation succeeded.  Blame #32 added.
11:21 mmorgan csharp_: So when the copies are distributed to the libraries, do they create brand new item records for them?
11:22 csharp_ mmorgan: never got to that level of detail in my fantasy :-)
11:22 csharp_ presumably they would already be cataloged and could just change owners
11:24 mmorgan Right. That being the case, I'm not seeing why age protection would be problematic.
11:25 csharp_ it would rid libraries of the need to think of *their* new items as having particular value
11:25 csharp_ pop materials pretty much never land on the shelf for the first 6 months of life anyway, so why complicate things?
11:26 csharp_ "at least they are never here because *our* patrons have them"
11:27 csharp_ plus, my assumption would be a potential cost savings by purchasing/renting at scale
11:27 csharp_ as Michael Scott would say win-win-win
11:28 csharp_ this is all in theory, btw - it's never gone beyond my crackpot scheming :-)
11:50 rhamby Isn't crackpot scheming a wird al cover band?
11:51 berick i really hope so
11:51 rjackson_isl_hom joined #evergreen
11:55 berick @ana crackpot scheming cover band
11:55 pinesol berick: Pong mock scratched vibrance
11:56 berick @band add Scratched Vibrance
11:56 pinesol berick: Band 'Scratched Vibrance' added to list
12:00 Dyrcona csharp_: Make it a floating collection and no one has to mess with ownership.
12:00 Dyrcona Also, good luck with ridding "libraries of the need to think of *their* new items as having particular value"
12:11 csharp_ @band add [ana [someone]]
12:11 pinesol csharp_: Band 'Im surly jar' added to list
12:16 Dyrcona That's my band: Surly Jar... :)
12:18 berick heh
12:29 jamin We definitely appreciate age protection... allowed us to get rid of a lot of manual "denewing" processes and circ rules.  Whereby members could hold on to stuff locally at first, then share.  (Without which we wouldn't have been able to start sharing at all.)
12:32 Dyrcona miker: I'm ready to start a timed pingest on my 3.7 production data without doing the symspell set up first. Last time I tried that it started out being relatively quick, then slowed down as more records were ingested.
12:33 Dyrcona And, I will use the --delay-symspell option.
12:53 jihpringle joined #evergreen
13:13 mdriscoll joined #evergreen
13:15 pinesol News from commits: Docs: fix display error <https://git.evergreen-ils.org/?p=E​vergreen.git;a=commitdiff;h=45a4a8​22ca16605111b225888c848134e64fb348>
13:16 JBoyer I'm also basically out of pocket / on the bench / choose your own favorite game / sport related euphemism for unavailable for any real hacking at the fest. Additionally, if anyone wanted to take the reins / read the agenda for the dev meeting this afternoon I would be happy to just listen in and pop in should I have something worth saying.
13:19 miker Dyrcona: huzzah! with that option it /should/ be as fast as before (insert-only into the unlogged table) and then at the end there'll be a pause while it shoves the changes in, in one go
13:22 Dyrcona miker: We'll see.
13:22 Dyrcona I also just tested a 3.7.3 to 3.9 db upgrade that I prepared for next week, and I got this error: psql:3.7.3-3.9-upgrade-db.sql:3509: ERROR:  cannot drop type search.search_result because other objects depend on it
13:22 Dyrcona DETAIL:  function search.query_parser_fts(integer,intege​r,text,integer[],integer[],integer,int​eger,integer,boolean,boolean,integer) depends on type search.search_result
13:22 Dyrcona I wonder if we have something old or custom  hanging around?
13:23 Dyrcona The suggestion in the error report was to use drop cascade.
13:24 Dyrcona Looks like we won't be needing it.
13:27 jeffdavis We have two versions of search.query_parser_fts in our environment and I've had to manually drop the second one for the upgrade scripts to work in testing.
13:27 jeffdavis So yeah, probably an old version of that function sticking around on ancient EG installs.
13:30 Dyrcona jeffdavis: Thanks. That's what this looks like.
13:30 Dyrcona I'm using cascade on the type drop, and that appears to work.
13:32 Dyrcona It's fun running upgrades over multiple version. :)
13:32 smorrison joined #evergreen
14:18 RFrasur @mmorgan @Dyrcona @JBoyer @miker @rhamby @phasefx @csharp_ @berick (yes, picking @'s that I saw and know stuff) 1. What does "pre-fetch" holds mean?  and, 2. Is Evergreen's FTP hardcoded as passive?
14:18 pinesol RFrasur: Have you run autogen.sh?
14:19 RFrasur @pinesol - No.  I can't.
14:19 pinesol RFrasur: Beyond here be dragons.
14:20 berick RFrasur: if the number of holds exceeds the page size of the grid, you can chose to fetch only the visible holds or all the holds
14:20 berick the latter being prefetch
14:20 berick this can make other operations quicker after the fetching is done
14:20 berick helpful if you know you want the whole list (e.g. for printing)
14:20 berick less helpful if you have 1,000 holds
14:21 RFrasur Oh, so if you wanted to retarget all holds, you'd have that selected?
14:21 RFrasur If there were more than 100?
14:21 RFrasur (in holdings view, that is)
14:22 * csharp_ just discovered ef in psql and mind is blown
14:22 csharp_ been copying/pasting/screwing around with whitespace forever
14:24 rhamby *nods* \ef is super useful
14:24 berick RFrasur: i'd have to look.  sometimes the UI will fetch the hold IDs, but not the holds themselves, and perform actions just using the IDs
14:24 berick retarget only requires the hold id, for example
14:25 berick in either case, the functionality is the same, it's just a question of when (and if) you fetch all the data
14:25 RFrasur berick++
14:25 RFrasur Useful no matter.  Thank you.
14:27 miker RFrasur: re FTP, you mean for EDI? it's not hard coded afaict, but passive is recommended because firewalls
14:29 RFrasur miker, yes for EDI.  Just off the phone with B&T about EDI orders failing with their FTP.  They recently changed IPs cuz of this and that.  We use their host name rather than the IP anyway, but apparently only active FTP is working for them right now.  The person I talked with was under the impression that Evergreen FTP is set up as strictly passive.  He used the word "hardcoded."
14:30 miker well, it's configurable, but on the server. not through the staff interface (AFAIK)
14:31 RFrasur Def not through the interface...(now I'm going to go verify that I didn't miss a tick box somewhere).
14:31 RFrasur miker++
14:33 RFrasur newp.  No tick box (LP ticket? maybe in a second or 360).
14:35 miker yeah, it's definitely only controlled by an environment variable when the fetcher and pusher are run. if B&T are alone in needing active, then it will take a little work because each vendor account will need its own crontab line, I think
14:35 miker definitely good to LP. should be able to control that at the remote_account level, I think.
14:37 RFrasur They're working to fix it on B&T's end to allow for passive again (it was broken in their IP transfer somehow), but in the meantime....hmm.  Maybe we just wait.  So, a LP.
14:38 JBoyer I think it's probably hard-coded for you RFrasur because I don't think active FTP won't work with your network. (It has been A Bit, so I'm not 100% certain)
14:38 Dyrcona RFrasur: We've had issues with B&T for a couple of days. Stompro and I discussed it briefly yesterday.
14:38 Dyrcona We use PASV.
14:38 RFrasur JBoyer, I think you're correct in that.  Dyrcona, it was your discussion yesterday that prompted me to call them.
14:39 Dyrcona We set FTP_PASSIVE in the crontab.
14:40 RFrasur Sigh.  So, we wait for them to fix things.
14:40 Dyrcona Yeah. It's broken on their side.
14:40 RFrasur Well, Dyrcona++ JBoyer++ miker++ thank you.
14:41 Dyrcona There are firewall modules that proxy FTP so you can use active mode, but I don't have control of the CWMARS firewall.
14:41 JBoyer Yeah, Active FTP is an archeological curiosity in 2022, not a viable protocol for most hosts.
14:41 RFrasur Yeah, and our environment is kinda convoluted.
14:41 Dyrcona I agree with Stompro that it would be better if B&T supported SFTP or SCP.
14:42 Dyrcona JBoyer: In my experience, it is still 1995 at most library vendors as far as the Internet is concerned.
14:42 RFrasur JBoyer, it was my momentary hope that we could switch to active until they got passive fixed so that the 8 POs languishing could go through. Which isn't too bad right now.
14:43 Dyrcona "We set it up in 1993, and it works. Why should we change it?"
14:43 RFrasur I did learn (?) that Follett and B&T are divested of one another though.  That's...something. (The impetus for the IP change)
14:44 RFrasur The guy on the phone was chatty.
14:45 pinesol News from commits: lp1942647 Provide Warning when deleting Term linked to Courses <https://git.evergreen-ils.org/?p=E​vergreen.git;a=commitdiff;h=b36ac4​1bd10b2a79849f1d1249442088ee77804b>
14:45 collum joined #evergreen
14:45 miker that seems like a short union ... but in pandemic time who knows? oh, 2016, much longer than I thought
14:46 RFrasur lol, miker, gettin' old over there? :D
14:46 RFrasur (it was short, but why miss an opportunity to make an old joke?)
14:46 miker I haven't had a birthday yet. it's only March 859th, 2020
14:47 RFrasur miker++
14:48 Dyrcona miker++
14:48 RFrasur I want to add to that joke (buh dum dum), but would have to reverse engineer to much with my OWN aging hardware.
14:48 Dyrcona RFrasur++
14:49 * Dyrcona sympathizes with Marvin the Android: I have a pain in all of the diodes down the left side of my body.
14:49 mmorgan Dyrcona++
14:49 RFrasur Dyrcona++
14:50 mdriscoll joined #evergreen
14:51 miker RFrasur++
14:59 shulabear joined #evergreen
15:09 csharp_ Dyrcona: did you have to rejigger JSON as text datatype stuff on PG 10? - our age protection not working was happening because "true" != true
15:10 csharp_ the setting value is literally "true" and the hold request permit function is comparing it with 'true'
15:10 csharp_ my dirty fix is trimming the quotes, but I feel like there hasta be a better way
15:11 csharp_ (I know the dev meeting is in progress but I can't attend - have to pick up my daughter from her day program)
15:17 miker csharp_: ah! that's a library setting editor bug! I think I fixed that recently. sec
15:21 csharp_ miker++
15:26 Dyrcona csharp_: We haven't noticed, but I'm 99% all of our JSON booleans are true without quotes.
15:27 tlittle joined #evergreen
15:29 jihpringle joined #evergreen
15:55 Bmagic JBoyer++ # Dev meeting
15:56 tlittle RFrasur Any chance you'd be willing to update the acq listserv that B&T FTP has some issues right now? I got a report today that we weren't getting invoices from them and didn't realize it was a widespread issue until I saw you mention it
15:56 RFrasur Yep, I'd be glad to.
15:57 tlittle Thank you!
15:59 tlittle RFrasur++
16:07 RFrasur tlittle, there are gaps in what I sent, but hopefully it's mostly comprehensible.
16:08 tlittle No, that's wonderful! I appreciate it. Good call on adding the bit about seeing the PUT errors--I'm going to back through our EDI messages for the past couple of days to see how many we've got affected
16:09 RFrasur And shout out to the "new" PO search, lol.  Very easy to pinpoint things.  It'd be even easier if our libraries used standard providers, but no need to get too crazy.
16:09 tlittle I have two reports that run everyday to let me know if there are errors on EDI invoices or with ORDERS/ORDRSP messages but I've been lazy this week because of the conference and hadn't been looking. Oops
16:10 RFrasur Eh.  Not the end of the world.  There's little to be done at this point anyway.  "Wait until [they] fix(es) it."
16:10 tlittle True facts
16:11 Dyrcona Here's the script that I use to stamp versions in files when making local branches, it could be useful for the release process: https://gist.github.com/Dyrcona/​933dd1a2e9035fdc48162dbcfcb51a2d
16:23 Stompro joined #evergreen
16:23 Guest5426 joined #evergreen
17:10 jvwoolf left #evergreen
17:17 mmorgan left #evergreen
17:18 Bmagic I'm logged into our WP Evergreen website for the first time. Who's in charge?
17:31 Bmagic I volunteered to take a stab at organizing our conference prestation in a searchable "catalog". My natural inclination is to integrate that into our website. But WP is old...
17:31 Bmagic presentations*
17:32 Bmagic my keyboard is dropping keystrokes... not sure what's going on but it's been doing that lately

| Channels | #evergreen index | Today | | Search | Google Search | Plain-Text | summary | Join Webchat