| Time |
Nick |
Message |
| 08:50 |
|
mmorgan joined #evergreen |
| 09:27 |
|
Dyrcona joined #evergreen |
| 09:32 |
goood |
@later tell Dyrcona re your duplicate query yesterday, you'll probably want one CTE (first_one) to find "a" using min() group by oclc, and another to find all that you can join on "oclc matches" and "ids don't match". might be able to just do it with a window in one select, actually... |
| 09:32 |
pinesol |
goood: The operation succeeded. |
| 09:34 |
|
Dyrcona joined #evergreen |
| 09:36 |
Dyrcona |
goood: Thanks. I was leaning toward a CTE. I'll give that a try today. Not sure that I'll need a window function, I might use an array_agg to get an array of the matching ids. |
| 09:37 |
goood |
we have a "remove this value from an array" function in the evergreen schema -- you could us least() or min() to yank the "first one" out, if you want one row per tcn |
| 09:41 |
goood |
well... I thought we did, but now I can't find it :( |
| 09:41 |
Dyrcona |
Well, I'm thinking I'll get a row like this in the end: min(id), array[other ids...] |
| 09:51 |
goood |
(oh, duh, just stock array_remove() ... anyway: select oclcnum, min(id), array_remove(array_accum(id), min(id)) from ... group by oclcnum; ) |
| 09:56 |
Dyrcona |
OK. something like that might work, but we dropped array_accum(). |
| 09:58 |
Dyrcona |
If I redesign the query like that, I won't need array_remove.. If I group by oclc number, I do a having to get rows with more than 1 array entry. |
| 10:04 |
|
mmorgan1 joined #evergreen |
| 10:06 |
Dyrcona |
Oh. interesting. I do need array_remove. I get results like this without it: ocolc 99997540 | {925827,925827} |
| 10:10 |
Dyrcona |
goood++ |
| 10:11 |
goood |
(yeah, s/accum/agg/) |
| 10:11 |
Dyrcona |
Hm.. I'm getting dups in the array still. |
| 10:11 |
Dyrcona |
ocolc 964630 | 597551 | {1481737,1481737} |
| 10:12 |
goood |
ah, there's uniq() |
| 10:12 |
Dyrcona |
ok. I was just going to check the documentation. |
| 10:12 |
goood |
select oclcnum, min(id), uniq(array_remove(array_agg(id), min(id))) from ... group by oclcnum; ) |
| 10:13 |
Dyrcona |
OK. uniq() is one of ours. :) |
| 10:14 |
goood |
well, it comes from intarry, IIRC |
| 10:14 |
goood |
an extension, but one we depend on |
| 10:15 |
Dyrcona |
I might not need the min and array_remove if I use uniq. Trying it again that way. |
| 10:16 |
Dyrcona |
Looks like I need to typecast: ERROR: function uniq(bigint[]) does not exist |
| 10:16 |
goood |
ah, yeah. unless you have more than 2B bibs, you should be ok :) |
| 10:18 |
eby |
@karma oclc |
| 10:18 |
pinesol |
eby: oclc has neutral karma. |
| 10:18 |
goood |
and, obv, "cardinality(blah) > 1" for the having, as you say. |
| 10:18 |
goood |
oclc-- |
| 10:18 |
goood |
@karma oclc |
| 10:18 |
pinesol |
goood: Karma for "oclc" has been increased 0 times and decreased 1 time for a total karma of -1. |
| 10:18 |
goood |
good, it's working! |
| 10:20 |
Dyrcona |
This is my having: "HAVING array_length(uniq(array_agg(CAST(record AS INTEGER))), 1) > 1;" |
| 10:22 |
Dyrcona |
And, I still need the array_remove because I'm getting results where the min repeats at the end of the array. Hm.... |
| 10:22 |
goood |
either should work fine, just slightly different return value semantics: https://www.postgresql.org/docs/current/functions-array.html |
| 10:22 |
Dyrcona |
{86604,598884,86604} |
| 10:23 |
goood |
so uniq isn't removing one of them? |
| 10:23 |
goood |
oh, nm |
| 10:23 |
Dyrcona |
Apparently. But it's not all cases, just some. |
| 10:23 |
goood |
mmm... don't like that |
| 10:24 |
Dyrcona |
Most rows look fine: {4665613,4670200,2171049,4667630} as an example. |
| 10:24 |
goood |
well, it's all just implementation details now! *disappears in a puff of smoke* |
| 10:24 |
Dyrcona |
uniq(array_agg(CAST(record AS INTEGER))) as "Records |
| 10:34 |
Dyrcona |
picky... uniq(array_remove(array_agg(CAST(record AS INTEGER)), CAST(min(record) AS INTEGER))) |
| 10:41 |
Dyrcona |
I removed a join on bre where not bre.deleted. That was a mistake. :) |
| 10:42 |
eby |
https://mitchellh.com/writing/ghostty-leaving-github |
| 10:46 |
Dyrcona |
Yeah, and Getnoo pulled their GitHub mirrors because of the "use AI or get out" attitude. The GitHumb CEO (Microsoft employee) has basically said as much in the past few months. I'll see if I can find some references. |
| 10:47 |
Dyrcona |
Consequently, our planned move to GitHub is on hold, pending review, etc. Partly because of this, partly because we're volunteers and there's really only 3 of us directly involved and we've all got other stuff to do. |
| 10:48 |
Dyrcona |
Here you go: https://www.businessinsider.com/github-ceo-developers-embrace-ai-or-get-out-2025-8 |
| 10:49 |
eby |
got usage based copilot to sell |
| 10:56 |
Dyrcona |
https://img-9gag-fun.9cache.com/photo/aKGoNAO_700bwp.webp |
| 11:14 |
Dyrcona |
eby: There is also this: https://mitchellh.com/writing/my-ai-adoption-journey . So, it's about the outages and not copilot for Ghostty. |
| 11:14 |
eby |
yeah i meant the push to embrace from MS |
| 11:15 |
Dyrcona |
Yeah. |
| 11:16 |
eby |
got a few projects I use that have moved off github due to the outages |
| 11:19 |
Dyrcona |
I guess I don't use GH enough to notice personally. I am more affected by Launchpad outages. And, that's another thing. I'm considering ditching Ubuntu completely because every release seems to be worse for my use case than the previous one. |
| 11:20 |
Dyrcona |
I think Ubntu 14.04 was the last really good release. They have all been worse in some way since. |
| 11:21 |
eby |
yeah i'm working on ditching as well |
| 11:23 |
eby |
updated an old kiosk install the other day and was constant 'we see you added a ppa and pinned the package to it but here is a snap package instead' |
| 11:24 |
Dyrcona |
Snaps are one thing. Generally bugginess of things I rely on (GNU Emacs) is another. |
| 11:25 |
Dyrcona |
They (Canonical, Ubuntu community) can't seem to get GNU Emacs to work properly with Wayland. I don't have these issues with Arch where they don't patch GNU Emacs. |
| 12:00 |
|
jihpringle joined #evergreen |
| 12:00 |
|
mmorgan1 left #evergreen |
| 12:27 |
Bmagic |
#startmeeting Let's have a wonderful meeting - wanna see where the logs go |
| 12:27 |
pinesol |
Bmagic: Meeting started at 2026-04-29T12:27-0400 |
| 12:27 |
pinesol |
Bmagic: Current chairs: Bmagic |
| 12:27 |
pinesol |
Bmagic: Useful commands: #action #info #idea #link #topic #motion #vote #close #endmeeting |
| 12:27 |
pinesol |
Bmagic: See also: https://hcoop-meetbot.readthedocs.io/en/stable/ |
| 12:27 |
pinesol |
Bmagic: Participants should now identify themselves with '#here' or with an alias like '#here FirstLast' |
| 12:27 |
Bmagic |
#endmeeting |
| 12:29 |
Bmagic |
#endmeeting |
| 12:29 |
pinesol |
Bmagic: Meeting ended at 2026-04-29T12:29-0400 |
| 12:29 |
pinesol |
Bmagic: Raw log: https://evergreen-ils.org/meetings/evergreen/2026/evergreen.2026-04-29.1227.log.json |
| 12:29 |
pinesol |
Bmagic: Formatted log: https://evergreen-ils.org/meetings/evergreen/2026/evergreen.2026-04-29.1227.log.html |
| 12:29 |
pinesol |
Bmagic: Minutes: https://evergreen-ils.org/meetings/evergreen/2026/evergreen.2026-04-29.1227.html |
| 12:30 |
Bmagic |
working |
| 12:31 |
Dyrcona |
Bmagic++ |
| 12:31 |
Bmagic |
I found this interesting: https://evergreen-ils.org/irc_logs/ |
| 12:31 |
Bmagic |
must be pre-irc-bot |
| 12:32 |
Bmagic |
somewhere around 2015, we stopped keeping the logs there, and instead, using the ilbot to do the job |
| 12:33 |
Dyrcona |
It's a different log bot. |
| 12:50 |
goood |
Dyrcona: re GH, there's also the "oops, we forgot to be git and rewrote your history on main" issue... fun times, there |
| 12:53 |
goood |
I'm also not happy with wayland. setting aside "push a window to my local machine" (the litteral point of X), it doesn't know how to shade windows. like, that seems like a small xrandr feature, but it soooooo handy. |
| 12:56 |
|
jihpringle joined #evergreen |
| 13:42 |
csharp_ |
goood: I've not tried this yet, but your comment got me curious and I found https://unix.stackexchange.com/questions/776767/what-is-the-ssh-x-equivalent-for-wayland |
| 13:45 |
Dyrcona |
I haven't used X forwarding in quite a while. I used to use it quite a lot. I guess I don't miss that with Wayland. |
| 13:45 |
|
pinesol joined #evergreen |
| 13:47 |
Bmagic |
remote desktop in Linux is one of the area's that it lacks. I've had a hard time finding a Remote Desktop replacemnt. Some distro's support MS RDP out of the box (fedora), and even then, it's not great, because you have to have a desktop session already started for it to work. |
| 13:47 |
Bmagic |
if you remotely reboot your computer, you can't RDP back in |
| 13:48 |
Bmagic |
outside of that, VNC isn't acceptable, so I'm left with services like Rustdesk or Anydesk. Which are technically free, but the expierence is sucky |
| 13:51 |
jeff |
for some use cases, if you're using KVM/qemu/similar, SPICE is a likely option but client support is far from universal. There is no great option for a macos SPICE client, last I looked. |
| 14:03 |
Bmagic |
my use case is: leave a computer permanantly connected to the internet somewhere. One with a lot of resources. Then travel and remote into it (desktop) and use it for all of my work, and never my travel computer |
| 14:04 |
Bmagic |
none of the remote desktop solutions that I've messed with are as good as MS RDP unfortunately. It's like MS has a monopoly on the tech |
| 14:10 |
Dyrcona |
Well, I think MS invented RDP, no? |
| 14:11 |
Dyrcona |
Bmagic: Sounds like you want a UNIX server and X term laptop. :) |
| 14:11 |
Dyrcona |
Maybe a Chromebook? :P |
| 14:21 |
goood |
Bmagic: at it's core, X /is/ the remote desktop protocol. like, that's it's whole point. wayland it tossing all the babies out with the bathwater, IMNSHO :( |
| 14:23 |
Dyrcona |
goood: I think Wayland is trying to be something different from X, but I haven't looked into its origins too deeply. |
| 14:25 |
Dyrcona |
I will say that it feels like everyone is trying to shove their vision of "the future" on everyone else all the time (wrt Wayland, Snaps, AI, Systemd, CBDCs, well... everything). |
| 14:25 |
* Dyrcona |
doesn't want to get too political but it is difficult not to. |
| 14:28 |
Dyrcona |
Going back to my query for duplicates by control number, I've been asked to exclude records where the edition statement basically matches /(hardover|paperback) edition/, but I think that's only the case when either is present in both records, i.e. one is 'hardcover edition' and the other is 'paperback edition' because these are records that were improperly copied or something like that. |
| 14:28 |
goood |
meh, when something is all "we're replacing that dusty old thing (X) with our shiny new compositor" ... it kinda feels like it's trying to be X |
| 14:30 |
Dyrcona |
goood: Point taken, but I don't think Wayland strives to the network capabilities of X, and it's meant for the people who won't use that capability. As I said, thought, there's always this feeling that you need to switch from one to other, and it often feels like that decision is being made for you when the updates all support Wayland and support for X gets less and less. |
| 14:31 |
Dyrcona |
s/thought/though/... I make that typo all the time. |
| 14:32 |
* goood |
screams in systemd... |
| 14:33 |
* Dyrcona |
screams in initrc. |
| 14:36 |
Dyrcona |
ugh. I'm going to need a CTE or two for this new exclusion. |
| 14:39 |
Bmagic |
it doesn't seem like we can use anubis in front of our wordpress site, on account it's plugins are also doing a bunch of redirects. Infinite loops ensue |
| 14:42 |
Bmagic |
about the remote desktop: no, I don't want a unix server with X redirection. I litterally want to see a picture of the remote desktop and click on it's icons, and use it's programs. I don't* want to redirect those X windows over to my local computer. Because* I want to be able to close my session and pick it back up again from another computer. Same windows open and everything |
| 14:42 |
Bmagic |
litterally/literally |
| 14:43 |
Bmagic |
It's a pain to have to re-launch all of my work and pickup where I was. All of that needs to just stay open, for weeks on my remote computer |
| 14:43 |
Dyrcona |
Bmagic: I believe that was possibly at some point with X, but anyway.... RDP pm a Chromebook should work. |
| 14:44 |
Dyrcona |
HOME KEYS....jfjf... :( |
| 14:44 |
Dyrcona |
Modulo typos in my responses.... |
| 14:44 |
Dyrcona |
!-3s/pm/on/ |
| 14:45 |
Bmagic |
It's the RDP piece that doesn't work on a Linux desktop. Not well at least. I've got it working/running several times, but there are caveats every time. One example I gave above is the "you have to already have a session running" |
| 14:46 |
Dyrcona |
I assumed you're using a Window desktop. The few times I've used RDP, it has mostly been to connect from Linux to Windows. |
| 14:46 |
Bmagic |
yeah, the computer in this example, right now, is running Windows. I remote into it from various OSes, Linux, ChromeOS, Android, etc. |
| 14:47 |
Bmagic |
I'd love to swap it out for a different OS, but.... reasons above |
| 14:47 |
eby |
i have used the rustdesk like you mentioned but also like you mentioned i think login screen requires X11 so wayland is a no go |
| 14:48 |
Bmagic |
Wayland is one of the issues, but even with xorg, I can't get the same expierence I have now |
| 14:48 |
eby |
the wayland idle system is also a problem for me |
| 14:51 |
Bmagic |
back on the wordpress issue: I don't think we're gonna get Anubis reverse proxying on top of it |
| 14:52 |
Bmagic |
I removed nginx, and went straight apache and Anubis, but this issue now |
| 14:52 |
Dyrcona |
Bmagic: We didn't have it before. |
| 14:52 |
Bmagic |
right, I was hoping we could add it as part of the migration |
| 14:52 |
Bmagic |
I think* it works if I put nginx back in |
| 14:53 |
Dyrcona |
It's proxies all the way down. |
| 14:53 |
Bmagic |
I think it was jeff that was interested in dropping nginx, so I thought I would |
| 14:54 |
Dyrcona |
Well, the fewer moving pieces, the better, in general. |
| 14:54 |
Bmagic |
agreed |
| 14:54 |
Bmagic |
we could live without it, and see how it goes.... we already know it's going to get clobbered |
| 14:54 |
Dyrcona |
My opinion: Leave Anubis out until it looks like we need it. |
| 14:54 |
Bmagic |
maybe we could employ cloudflare |
| 14:56 |
Bmagic |
ok, I'll leave it out. That's the way it's setup at the moment, and it is working. So, Imma move over to some other work for now |
| 14:56 |
Bmagic |
anubis is* working for the other things: dokuwiki, irc-logs |
| 14:56 |
Dyrcona |
Speaking of moving on, I think I'm going to give up on refining this query. |
| 15:12 |
Bmagic |
Dyrcona++ |
| 15:26 |
|
Dyrcona left #evergreen |
| 15:26 |
|
Dyrcona joined #evergreen |
| 16:53 |
eby |
https://copy.fail/ |
| 19:54 |
|
eeevil joined #evergreen |
