freenode/lisp - IRC Chatlog
Search
13:49:05
jackdaniel
I'm not sure what you are talking about, but my remark is about current affairs.
13:50:35
phoe
I guess that aeth means that the initial step for that was made in 2017 and this fact blossoms now in 2021
13:51:36
phoe
which currently is unavailable, presumably due to being swamped with traffic and/or DoSed
13:51:38
lukego
Hey I don't seem to have the OUTPUT function that CL-SMT-LIB expects me to at https://github.com/GrammaTech/cl-smt-lib/blob/master/cl-smt-lib.lisp#L84. Seems like it's expected to be exported by GRAY-STREAMS but isn't? anyone know what gives?
13:52:30
phoe
https://github.com/GrammaTech/cl-smt-lib/blob/master/fundamental-two-way-stream.lisp#L9
13:55:06
lukego
ah indeed. looks like a bug in CL-SMT-LIB that they have multiple packages and aren't sharing symbols between them as they should. thanks for helping!
13:56:00
phoe
maybe it's a feature that it doesn't have a god package that reexports everything from everything else
13:59:01
lukego
well, sure, but the WITH-SMT macro generates code that errors because it puts the OUTPUT symbol from the wrong package in function-position
13:59:28
lukego
but I'll fix that and send a PR (assuming it's broken in the latest code from github, checking now)
14:04:20
lukego
oh hm I think this code was never debugged, also errors with EOF-ERROR-P on READ not being threaded through where it needs to go
14:07:07
lukego
and this is a bit too tangled for me. there's a recursive call to READ-CHAR first to a gray stream and then to a native stream, and EOF-ERROR-P is not threaded between them, such that the inner call errors on EOF while the outer call is expecting it to return :eof. Maybe needs a condition handler there instead?
14:07:33
pfdietz
A problem I've found with those smt interfaces is making sense of the model it sends back. The format is often idiosyncratic.
14:09:23
lukego
what with me being all millennial-wannabe and using PNG instead of TXT for error message reporting
14:11:34
lukego
anyway I'm lost as to where the bug is - SBCL, CL-SMT-LIB, or gray streams? - but if you'd run something like (cl-smt-lib:with-smt (smt ("z3"))) then you'd get an error like https://ibb.co/wz9TQ2S where EOF-ERROR-P is being lost on the stack
14:44:19
contrapunctus
I've been cultivating a Lisp room on Jabber/XMPP for a while, if anyone wants to join. xmpp:lisp@conference.a3.pm?join (it's not CL-specific like this room, though...much too small a community for that at the moment.)
14:45:36
Demosthenex
given freenode was bridging with matrix already, i wouldn't mind seeing people bridge xmpp
14:45:51
beach
I feared that we were going to be split between matrix and libera.chat, but now it looks like we might get split in 3 instead.
14:47:55
beach
Demosthenex: Except that it doesn't seem to have a matrix bridge, so it is not fine with people that are currently using such a bridge for freenode.
14:48:58
beach
Then that would seem like my preferred solution, given my experience with the staff here in the past.
14:49:08
lukego
okay, there is a lot to fix in WITH-SMT, and since it's not actually documented to exist I'll just pretend that it doesn't
14:49:30
Demosthenex
beach: otoh, the matrix bridge has been compared to when AOL added the internet ;] a new eternal september
14:49:30
lukego
(it's also not detecting crashes in the solver, e.g. when you specify one that's not installed, nor reading results properly)
14:50:03
Demosthenex
beach: you only need to know that matrix was written by webdevs, and that discord is facebook for gamers.
14:50:37
beach
Demosthenex: I am afraid I can't parse any of those sentences. Probably due to ignorance of the subject on my part.
14:51:41
Demosthenex
beach: matrix made by developers who think comic sans and javascript are good ideas, and that every user needs a 500MB gui client where the theme is locked. discord is just please record everything i say for purposes of making money.
14:51:52
xristos
discord is proprietary spyware, also ruthlessly antithetical to any sort of custom client development
14:52:23
Demosthenex
xristos: like the process scanner which constantly reports what you are running to their server. nominally so it can update your icon for the game you're playing.
14:52:30
xristos
matrix is unusable for me, full of anti-patterns and none of IRC's simplicity and elegance (gasp!)
14:54:10
beach
A necessary (but not sufficient) condition for me to join is that I can have my flyspell mode and my abbrevs.
14:57:32
katco
full disclosure, i'm a matrix fan, but matrix is working on a binary protocol, and there are many clients, some of which are not bloated at all.
14:59:47
contrapunctus
If anyone wants a better XMPP in their Emacs, do see if you'd like to contribute to this fork of jabber.el
15:00:18
phoe
;; janitor announcement: I think that, given the current network-related disturbances, it would be sensible to relax the off-topic rule a little bit to involve IRC-and-alternatives and Freenode-and-alternatives-related discussion, at least until things calm down and clarify a little bit
15:00:56
Jachy
There was even a clim client for matrix at one point, though it hadn't gotten very far last I tried it and I don't know what happened to it.
15:01:23
lukego
ACTION hopes he will be able to just point his web browser at whatever the new venue does, which he does now via irccloud.com web client
15:01:26
Demosthenex
katco: my issue with matrix is reinventing the wheel, versus an evolving standard like xmpp with multiple non-web oriented clients and servers.
15:03:43
katco
Demosthenex: i can't comment on the similarities between the xmpp and matrix protocol, though i used to use jabber a lot. but for me at least the argument doesn't sway me; both protocols are here now and should be evaluated on their merits. as an aside (and as a longtime freenode user before matrix), it's sad what's happening. but time goes on i suppose...
15:03:46
splittist
ACTION also fears an escalating civil war of manifestos, codes of conduct, purity tests and purges, leading to a barren wasteland of chat...
15:08:03
katco
here are some matrix clients, almost all of which are not web-based: https://matrix.org/clients/
15:09:58
katco
mozilla just switched to matrix, fosdem ran their last conference with matrix, gitter is now the matrix protocol, multiple governments are using matrix for interdepartmental communication. i think there's some momentum behind the protocol. that's my two cents, and i'll be quiet about it now :)
15:13:07
Josh_2
My community hasn't completely swapped to matrix, but its nice for me as my bot implemented invite only communities, along with a whole host of other features
15:13:29
Josh_2
means I dont have to maintain all that nonsense and can just focus on making cool features
15:16:17
Demosthenex
katco: there's momentum behind AOL IM, MSN, and Skype too... my point is federated open protocols with varied clients and servers are more durable, ie: email
15:17:31
katco
Demosthenex: sorry, i think that's a false equivalency. matrix is an open, federated, protocol with varied clients?
15:18:37
Demosthenex
katco: i mean new kid on the block syndrome and reinvent the wheel. i haven't seen anything matrix added that xmpp doesn't already have
15:19:36
Demosthenex
MichaelRaskin: multiserver synced chat? yeah, given it's federated each chatroom is server local. what's your use case
15:20:15
MichaelRaskin
Demosthenex: use case is that in Matrix, matrix.org going down for a week does not take out #matrix:matrix.org
15:21:39
Demosthenex
MichaelRaskin: i'm afraid i don't find that compelling ;] if a federated server hosts a room, and it's down... then the room's down.
15:22:29
Gnuxie[m]
Demosthenex: you don't understand how Matrix works, rooms aren't local to any server, room 'aliases' might be but they're irrelevant to the operation of the room
15:22:39
katco
Bike: https://matrix.to/#/#common-lisp:matrix.org?via=ponies.im&via=matrix.org&via=privacytools.io
15:24:46
Bike
i already use the discord web client and it hasn't left me with much love for electron, but oh well
15:28:02
MichaelRaskin
Demosthenex: you say they are reinventing the wheel because you do not understand the difference
15:30:04
Demosthenex
MichaelRaskin: that may be perfectly true. i got off the IM network churn long ago, every new network i am skeptical of and see little reason to change. there were hundreds before them, and i think email, irc, and xmpp are solid for my use
15:32:06
beach
But it is looking more and more likely that I will be present in at most 1 out of 4 of 5 "chat rooms" dedicated to Lisp.
15:33:38
Demosthenex
beach: well, you're appreciated anywhere you are, you've given me good advice before =]
15:34:44
jackdaniel
Demosthenex: the issue is that fragmenting a (former) freenode community would be a waste
15:35:56
xristos
the dust will settle in a day or three, as long as its IRC and contains the majority of existing lisp channels, i'll be happy
15:36:00
jackdaniel
I will look around how things develop until next week and then I will decide; I personally consider only the "new" irc server vs matrix
15:36:51
jackdaniel
ACTION will move only two channels (#clim and #ecl), so probably not that much of an impact here
15:36:52
Cymew
I noticed now that my auto join was broken. I have missed what ever conversation has taken place here, but am now on libra.chat as well if anyone wants to migrate there.
15:36:59
Shinmera
Is there an equivalent on libera already? It could be a nice opportunity to avoid the #lisp vs #commonlisp debacle.
15:39:04
aeth
Unfortunately, every alternative to IRC has to be "monetized" which usually means controlling the client so that ads and tracking can be inserted at some unspecified future point.
15:41:54
Shinmera
while we're semi off-topic, I'm still pretty happy with the new trailer for Kandria that we released on Monday: https://www.youtube.com/watch?v=nf9Z4MPBbfc
15:42:09
aeth
"Modern" also means "minimalist", i.e. way too much padding and not many configuration options at all. Maybe none other than localization. You better hope that your language's datetime preferences are the datetime preferences you prefer!
15:42:17
Demosthenex
Cymew: modern seems to imply blinding white background, huge amounts of wasted space, and variable fonts in clients requiring hundreds of megs of ram :P
15:42:43
aeth
Demosthenex: or a "dark" theme that's white on a moderate blue, rather than white on black.
15:42:58
Demosthenex
"modern" also tends to mean completely uncustomizable. you can't change your hotkeys, our theme is set thru a/b testing, and it's all locked beyond your control
15:43:24
aeth
yeah, "dark theme" is a thing because nowadays most software has two theming options: dark and light. Old software like Emacs lets you change the whole thing.
15:45:38
katco
yeah i really hate the trend away from open protocols and user agents (i.e. clients). everyone celebrating dark themes because they were handed down from on high. if you have protocols/clients, you have whatever you want. no one has to give you anything.
15:45:43
Demosthenex
Shinmera: omg! were you part of that! i saw that recently come up on some lisp news feed. i was impressed!
15:47:41
Shinmera
I sometimes here people elsewhere are talking about it, but I wouldn't know where :)
16:15:23
aeth
There are two cultures: IRC channels and internet chat rooms. If you add in emoji reactions and images and videos and stuff, you get the latter, and you wind up getting an AOL chatroom replacement instead of an IRC replacement.
16:16:38
Bike
https://gist.github.com/Bike/ba58b85d2b2dec2d3af8a54821ca3b27 here is a quick list of places
16:18:10
jmercouris
everyone is getting excited, just like GitHub, I doubt anything will change... give it a week, and then bring it up again
16:18:15
Shinmera
Bike: I don't think making a lisp separate channel on tynet/lichat would be worth it. I expect the shirakumo channel alone is gonna be dry enough :)
16:20:38
jackdaniel
jmercouris: github was a commercial entity from the start - freenode premise was different though
16:20:44
mfiano
It will likely appear to be the same, sans a lower user count, and the unknown of what a hostile business may do with your user data.
16:42:53
Shinmera
Would be better to have the latter anyway, just to finally stop the endless confusion "but #lisp is only for commonlisp???"
16:48:24
phoe
jackdaniel: like, if we'd like to utilize this chance to move to #commonlisp, then it's already ready and registered
16:48:56
_death
phoe: and that's why if I'm gonna move to another network, it'll be eris free.. although they have that annoying chanfix :x
17:27:44
ecraven
I'm connected via tls on that port.. maybe the servers are a bit overloaded, try again a bit later?
17:38:19
galex-713
I tried before to reimplement “minimize” out of loop for the sake of abstraction but realized it’s more complex than that because the index is separated
17:38:55
galex-713
minimize stores only one value I guess, no way of remembering other value from the moment that value was smallest
17:39:15
saturn2
you can't do it because you need a place to store the minimum value and also a place to store the index
17:39:47
galex-713
also, what I’m trying to do is to take some sort of “median” of a list of numbers, so that the difference of the sum on either of its sides is the smallest… do that statistical function has a name?
17:40:11
galex-713
(what I’m trying to know is what’s the median size of a country, ponderated by its surface)
17:40:13
_death
(loop with min-i = -1 for i upfrom 0 for x in '(3 9 8 2 5) minimize x into min-x when (eql x min-x) do (setf min-i i) finally (return (values min-x min-i)))
17:40:23
galex-713
(I’m currently calling that a ponderated median but dunno if it’s the correct name)
17:45:02
galex-713
min contains a conditional that would be needed anyway to store the index of the minimum
17:45:04
_death
it doesn't matter that the code is imperative.. the function would be referentially transparent
17:47:07
_death
because you want an operator (minimum seq) => value, index (or a more general one) and it doesn't matter much how it's implemented
17:53:35
_death
if you like inefficient code.. (reduce #'smaller (zip sequence (iota (length sequence)))) (defun smaller (a b) (if (< (car a) (car b)) a b)) (defun zip (a b) (map 'list #'list a b)) (defun iota (n) (loop for i below n collect i))
17:54:11
galex-713
actually, I’d like code that would work out of order, so then I could apply it my dichotomy instead of linearily
17:54:30
galex-713
that is minimize starting from the center of a partition each time, and partitioning by the center at each level
17:56:02
galex-713
[19:39:47] <galex-713> also, what I’m trying to do is to take some sort of “median” of a list of numbers, so that the difference of the sum on either of its sides is the smallest… do that statistical function has a name?
17:56:50
galex-713
the sums are not purely monotonical, but the numbers are, so the sums first decrease and then increase
18:01:04
galex-713
I mean like take a list of number, find the number so that A is the sum of numbers before it, and B the sum of numbers after it, and the difference between A and B is minimal
18:01:38
galex-713
I want to know what’s the size of your country on average if you end up on a random place on earth
18:11:27
galex-713
_death: I want the country so that the total surface of earth of wider-than-you countries is as near as possible to the total surface of earth of smaller-than-you countries
18:11:52
galex-713
the time of finding it is linear at worse, but I’m pretty sure it could be logarithmic
18:18:17
jcowan
Buildings are made of frangible brick and mortar, overly rigid stone, and overly elastic steel. Therefore, we should all live in tents.
18:22:02
_death
galex-713: I guess it could matter when the order is fixed.. seems you're looking for some expected value?
18:24:51
galex-713
but my values are scrambled, 2 values are missing, so I can’t be sure, I’m doing it again
18:26:06
Inline
jcowan: back in the days making houses was no issue with a little bit of mud/clay, hay and wood
18:27:26
jcowan
Well, mud is no good in areas with torrential rain, and alll of it is bad in earthquake country
18:29:09
Inline
nah, with only 2 floors, nobody would die, at least not when the house is small and one is not hit in a weird position
18:29:58
Inline
and when the house comes down, who cares as long as you are alive, just another try for a little more days of work...
18:34:33
pfdietz
Is there a web client to access libera.chat, like webchat.freenode.net works to access here?
18:40:06
_death
channels exist as long as people inhabit them.. I think such migration would only work with "organized" communities and new channels..
18:42:15
jackdaniel
as in: one mod declares jumping the ship to another server and changes topic; another mod reverts it back 2h later ,-)
18:42:16
phoe
I guess someone would need to make a decision then, and people would need to follow it
18:42:48
phoe
right now I'm observing the situation, and preparing stuff on libera in case we move there
18:42:50
jackdaniel
well, I'm still going to change a topic for better or worse (the moment I know where #clim and #ecl move); but I'm sure that plenty of people will stay
18:43:39
jgkamat
I was kind of assuming the mods would (at some point) all be on the same page at some point xD
18:43:53
_death
of course a channel owner could set +i and kick everyone or something... personally I'm sticking freenode
18:44:49
_death
there are a zillion references to freenode #lisp.. I'm not sure it's useful to disband or have yet more channels just to satisfy irc drama du jour
18:46:50
jgkamat
at some point (unless things change) freenode will likely be overrun with spam and/or other bad things, because that's what the staffers jobs were
18:48:07
phoe
but libera.chat will be getting a matrix integration as soon as they can get it, just like freenode does
18:52:15
jgkamat
IMO matrix is nice if you're looking for the modern features, but IMO it's too slow/unreliable for me
18:52:27
jgkamat
and the spec moves too fast for anyone to have a chance of implementing a good client
18:52:29
jackdaniel
joining lisp channel on matrix was, well, disappoining (content and connectivity wise)
18:53:23
Inline
i tried both sometime ago, but then saw that the content there is sometimes distracting
18:53:56
Inline
better is to have some good paste sides around like termbin.com which you can use with netcat
18:56:27
Inline
file contents should be on a separate thing on both matrix and discord and leave the main channel as is
18:57:31
Inline
which you can bring to the foreground if you wish that, like we do when we open the browser
19:16:35
thmprover
Say, uh, has anyone been paying attention to this freenode takeover thing? Apparently the staff running it has resigned?
19:17:02
ChoHag
This (unofficial) summary seems fairly accurate: https://gist.github.com/joepie91/df80d8d36cd9d1bde46ba018af497409
19:18:03
thmprover
Will the #lisp community move to a different IRC server? Or is this "not as bad as it sounds", so to speak?
19:18:13
ChoHag
tl;dr: Chill Winston. Wait for tomaw to wake up, check his inbox, and have his 6th or 7th coffee.
19:19:00
ChoHag
It would be insane to make any decisions at this stage just, as has apparently been done, prepare.
19:19:42
Bike
this channel is not really a coherent organization, so there aren't going to be official definitive statements about moving or not
19:20:03
ChoHag
The service isn't going anywhere any time soon you can be sure of that. ragnarok's not going to topple working servers he can't right when he's trying to prove his worth in a hostile environment.
19:21:14
MrtnDk[m]
_death Can't we use something like "Beatle juice" or what it's called, and simply implement an interface to the matrix for now? I realise it's going to be a major inconvenience, and everyone will have to get matrix accounts, but at least we have a place until the situation has been fixed permanently. I wish Lilo was alive, he might have been able to fix this.
19:21:32
ChoHag
OTOH it's getting late in Russia and the vodka's flowing and maybe with all the staff gone it's time for a ddos?
19:21:44
Lycurgus
ACTION is just catching up but doesn't see how this is different from the general situation with the global name service
19:21:55
thmprover
Right, right; I'm just mildly surprised by the news. It will all pan out in the end, I'm certain.
19:22:34
jackdaniel
it is not about dns; freenode ltd was sold under unclear circumstances and the current owner claims ownership of services and servers too
19:23:00
Lycurgus
so here's the URL I posted in the wiki channel before I knew there was a freenode issue eg.meansofproduction.biz/index.php/Alt_root
19:23:01
jackdaniel
that is most likely unlawful, but nobody has time and health to fight the lawsuits coming from a commercial entity (that's usa law for you)
19:24:48
Bike
based on all the resignation notices, it sounds like the staff was not clear on their own legal organization to begin with. they all come off as kind of confused or blindsided by this, and they'd know better than any of us schlubs, so everything comes off seeming vague
19:25:42
Lycurgus
i mentioned bitlbee in #wikipedia-en and the response was "Matrix is a Matrix is a standard for bridging nowadays tbh"
19:28:03
ChoHag
MrtnDk[m]: Apparently nothing except it happens to control the domain (ie. relationship with the registrar).
19:28:25
lonjil
MrtnDk[m]: it is the owner of being able to make legal threats against people. Legal threats that may or may not have weight, but may be quite difficult for individual staffers to deal with.
19:28:32
Bike
If you read the resignation notices, the staff seem to have been confused about what Freenode Ltd was in fact the owner of
19:29:03
jackdaniel
MrtnDk[m]: I've found this the most informative: https://fuchsnet.ch/freenode-resign-letter.txt
19:29:10
ChoHag
How do you devote decades of your life to something you so fundamentally misunderstand?
19:29:32
jackdaniel
ChoHag: you are a volunteer, you all agree to make a legal entity [for something], then one of you sell that legal entity
19:30:02
saturn2
freenode didn't have any employees, only volunteers. a while back one of the volunteers "sold" it to that guy and then disappeared
19:30:29
Bike
yeah, my attitude here is that if the staff was this vague about their organization I'm not sure I want to stick around anyway
19:30:42
ChoHag
And apparently the freenode staff thought their agreement (aka. contract) was one thing while the holding claims it was another.
19:30:55
saturn2
the whole thing made no sense, but the guy has enough money to keep the frivolous lawsuits going forever
19:32:49
ChoHag
Because the virus overreaction has strained a system that was already at breaking point.
19:33:29
lonjil
ChoHag: They already have a lawyer. Who advised them that Freenode Ltd would probably lose in the end, but legal challanges could drag on for years and be very very expensive for the volunteers.
19:34:33
MrtnDk[m]
jackdaniel If this letter is correct, Christel basically screwed over Freenode, possibly for personal gain. I think we're only seeing the top of the iceberg. Does anyone know if Christel have (had) any personal or financial problems, that could affect her judgement?
19:35:18
epony
a synthetic language the age of your grandparents graduation.. and some chat hosting facility on the end of its teens spawned from other such when that said language was 40 years old and standardised with dialects
19:40:37
MrtnDk[m]
Speaking of which ... are there any reasonable implementations of IRC in common Lisp, like services, servers and/or clients?