00:02:20 <esolangs> [[Num]] https://esolangs.org/w/index.php?diff=83099&oldid=83049 * AndrewBayly * (-3751) /* Main */
00:02:46 <esolangs> [[Num]] https://esolangs.org/w/index.php?diff=83100&oldid=83099 * AndrewBayly * (+1) /* Main */
00:03:45 <fizzie> Oh, right, actually the UTC midnight was just now. Forgot about BST.
00:04:13 <b_jonas> oh... I didn't pay attention to that, yes
00:04:41 <int-e> ozone is fighting it out with chanserv, now there's a sight
00:07:47 <fizzie> Weird. The version of the bot I built on buster segfaults, but the one built on my daily-driver desktop (sid) doesn't.
00:07:55 <fizzie> I built it with Bazel, that's not supposed to happen. ;)
00:08:15 <fizzie> At least it's reproducible, so hopefully I can figure it out.
00:08:26 <int-e> "that was not supposed to happen" is a great candidate for famous last words
00:10:11 -!- DHeadshot_ has quit (Read error: Connection reset by peer).
00:11:47 <fizzie> It dies in std::thread::detach(). The system's almost entirely a single-threaded event loop, but there's one thread spawned to handle making blocking getaddrinfo calls, because I thought bundling an external DNS resolver like c-ares would be overkill.
00:14:04 -!- moony has changed nick to example.
00:15:13 <fizzie> https://0x0.st/-eHB.txt looks p. mysterious. What's that `call 0x0` doing there?
00:16:30 <esolangs> [[User:Icecream17/Arbitrary]] N https://esolangs.org/w/index.php?oldid=83101 * Icecream17 * (+512) Created page with "'''Icecream17/Arbitrary''' is ridiculous The following is subject to change {| class="wikitable" |+ Static programs |- ! Bits !! Hexadecimal source code !! What the source c..."
00:16:50 -!- example has changed nick to nickname.
00:17:06 -!- nickname has changed nick to example.
00:20:11 <fizzie> The context for the call is `std::thread(<function pointer>, <argument>).detach();`, which I at least naively thought would just mean "start an independent thread running that function".
00:20:23 -!- chibi has joined.
00:22:26 <chibi> Oh, Libera does ## instead of #?
00:22:45 <fizzie> By my interpretation, we *should* have been ##esoteric on freenode as well.
00:23:09 <keegan> yeah, Freenode had a bunch of channels grandfathered into single-# names
00:23:22 <keegan> although you could argue that #esoteric was the official channel of the open source esolang wiki
00:23:33 <keegan> just like how #trains is the official channel of trainbot
00:24:04 <int-e> keegan: https://esolangs.org/wiki/Esolang_talk:Community_portal#Freenode_and_the_future has some thoughts in that direction
00:24:10 <fizzie> I seem to recall someone kind of hinted we should think about migrating to ##esoteric on freenode, but we never bothered.
00:24:35 <fizzie> On the Libera policies page, I saw there's this new category of "community channels", between project channels and informal channels.
00:24:48 <b_jonas> fizzie: no, that's what we say the official policy is because there's a rare case when people when squat a #-channel and staff takes it over for the project with the relevant name for searchability, and we want to have something to back it up. they don't actually mind #-channels when they don't cause trouble like that, but they can't say that officially.
00:25:17 <b_jonas> any long-established #-channel that doesn't steal namespace is fine
00:25:37 <fizzie> As far as Libera goes, I think a "community channel" would be the reasonable designation for this one as the official community channel of esolangs.org, but `#esolangs` would IMO be a more reasonable name than `#esoteric`.
00:25:40 <int-e> b_jonas: I think they care when they have visibility, like #math which was basically forced to become ##math
00:25:41 <fizzie> (Despite the history.)
00:25:57 <example> yea, #esolangs would make more sense at this point
00:26:06 <b_jonas> int-e: yes, #math is stealing prime namespace, it's a channel name that people will type in even without outside references
00:26:15 <int-e> #esoteric, yeah, is probably small enough nobody cares. Also the attitude has probably shifted quite a bit over the years.
00:26:32 <b_jonas> and it's a big channel with occasional drama
00:26:56 <fizzie> ISRT that also happened to ##java, which makes sense, because it's both obvious and having it as #java implies it's associated with whoever owns Java at the moment (Oracle?).
00:27:10 <b_jonas> #esoteric isn't. a few people from the other type of esoterica did find their way in, and some of them were probably genuine rather than sockpuppets trying to make fun of the welcome message
00:27:25 <int-e> b_jonas: I really think #math became ##math mainly to set an example. Any project claiming that name would have to deal with people who wanted to just discuss (talk *about*) math.
00:27:48 <example> pls don't set me. (lemme change my nick back)
00:27:54 -!- example has changed nick to moony.
00:28:57 <b_jonas> is there any kind of namespace continuity implied?
00:29:04 <b_jonas> I mean from freenode to libera
00:29:10 <b_jonas> or is it considered new domain?
00:29:16 <moony> uh, new domain afaik
00:31:37 <fizzie> In other news, IDGI. When I build it on that other system, the binary contains https://0x0.st/-eHS.txt as the implementation of std::thread::detach -- when I build it on this one, it just gets a PLT entry <std::thread::detach()@plt> referring to <std::thread::detach()@GLIBCXX_3.4.11>.
00:33:35 <int-e> so the dynamic linker messes up?
00:34:00 <int-e> ..wtf is that google thing doing in there
00:34:22 <fizzie> I think that's just a random symbol that happened to end up associated with address 0 for whatever reason.
00:35:28 <int-e> right, makes sense, sort of
00:35:49 <fizzie> But I don't know how it manages to get a "call 0" in there. Maybe I'll need to look at some intermediate files.
00:37:40 <fizzie> Oh, there *is* a difference in the BUILD files: the working one has commented out features = ["fully_static_link"].
00:37:54 <fizzie> That sounds odd enough that it's probably what's screwing it up.
00:38:27 <fizzie> Though I've no recollection of why that's there in the first place.
00:41:36 <fizzie> Let's give it a whirl.
00:41:48 -!- esolangs has joined.
00:42:05 <chibi> fizzie: Is #java a place for coffee discussions?
00:42:37 <fizzie> After two years or so of punting it until later, I can finally produce a new binary that works. (Well, assuming it does.)
00:44:30 <b_jonas> chibi: is Java coffee actually such a thing? because I'm quite sure I hadn't heard about that until the Java language people started to spread that origin story. I think they invented that to have a cool explanation for their name like Python.
00:45:25 <fizzie> Java the Indonesian island definitely predates Java the programming language, and is associated with coffee.
00:45:52 <fizzie> Wouldn't know any of the details, though, not being a coffee-ist.
00:46:27 <fizzie> Have I mentioned our team's hybrid coffee/tea break is called a "toffee break", but never has any toffee involved?
00:46:37 <b_jonas> the indonesian island does, yes
00:46:59 <chibi> Apparently so, though my knowledge of coffee extends to how to brew good coffee in a pot and that's about it
00:48:12 <chibi> I've never really had toffee, what's it taste like?
00:49:13 <int-e> fizzie: that 0x0 may be __gthrw_pthread_create, which is a /weak/ reference to pthread_create... which might stay 0 if you "forget" to link in pthreads?
00:49:38 <fizzie> It's a bit like fudge and caramel.
00:50:23 <int-e> (I'm looking at /usr/include/x86_64-linux-gnu/c++/10/bits/gthr-default.h)
00:50:59 <fizzie> int-e: Yeah, I think it must've been something about that fully_static_link thing, which I've reconstructed to having been an attempt to produce a binary that I could runs while I didn't have a system with exactly-matching library versions as the target machine big enough to run the build.
00:51:05 <fizzie> Guessing it just wasn't pthreads-compatible.
00:52:04 <chibi> Ooh, maybe that'd be nice to pair with some black coffee
00:52:37 <fizzie> https://github.com/tensorflow/tensorflow/issues/42057 sounds potentially related.
00:55:07 -!- Wezl has joined.
01:13:01 <esolangs> [[Special:Log/newusers]] create * CiaaiK * New user account
01:45:36 -!- reader445 has joined.
01:55:15 -!- Wezl has quit (Ping timeout: 240 seconds).
02:00:00 -!- DHeadshot has quit (Remote host closed the connection).
02:12:44 <esolangs> [[Finites at Fredy's]] N https://esolangs.org/w/index.php?oldid=83102 * Salpynx * (+4732) This was _supposed_ to be a parody, but has become more of an uninspiring, but accurate, review. "Fair use" either way.
02:28:35 -!- chibi has quit (Read error: Connection reset by peer).
02:32:14 -!- chibi has joined.
03:52:23 <esolangs> [[Esolang:Community portal]] https://esolangs.org/w/index.php?diff=83103&oldid=82807 * New Army * (-63) The link doesn't seem to be dead.
03:54:24 <esolangs> [[Esolang talk:Community portal]] https://esolangs.org/w/index.php?diff=83104&oldid=83098 * New Army * (+187) /* LifeWiki links */
04:20:31 -!- arseniiv has joined.
04:42:32 -!- reader445 has quit (Quit: Connection closed).
04:45:51 -!- ski_ has joined.
05:15:52 <arseniiv> hi let this day be nice to everyone!
05:19:51 <shachaf> Confectionery is neatomatic 4,000.
05:33:12 -!- ski_ has changed nick to ski.
06:09:45 -!- jinn has joined.
06:09:57 <jinn> hello is this the compilier design channel from freenode?
06:10:18 -!- jinn has left.
06:10:41 -!- chibi has quit (*.net *.split).
06:10:42 -!- HackEso has quit (*.net *.split).
06:10:42 -!- lambdabot has quit (*.net *.split).
06:13:08 -!- lambdabot has joined.
06:15:49 -!- chibi has joined.
06:19:17 -!- esolangs has joined.
06:19:17 -!- Melvar has joined.
06:20:49 -!- chibi has joined.
06:26:31 -!- jinn has joined.
07:12:28 -!- dbohdan has quit (Remote host closed the connection).
07:12:38 -!- iovoid has quit (Remote host closed the connection).
07:12:44 -!- dbohdan has joined.
07:13:25 -!- dbohdan has quit (Changing host).
07:13:25 -!- dbohdan has joined.
07:14:35 -!- lambdabot has quit (Remote host closed the connection).
07:15:06 -!- lambdabot has joined.
07:16:32 -!- iovoid has joined.
08:29:16 <esolangs> [[Spider solitaire]] M https://esolangs.org/w/index.php?diff=83105&oldid=83032 * ColorfulGalaxy (disambiguation) * (+31)
08:45:49 <esolangs> [[Esolang talk:Community portal]] https://esolangs.org/w/index.php?diff=83106&oldid=83104 * Int-e * (+189) /* LifeWiki links */
09:22:26 <esolangs> [[Esolang talk:Community portal]] https://esolangs.org/w/index.php?diff=83107&oldid=83106 * ColorfulGalaxy (disambiguation) * (+237) /* LifeWiki links */
11:26:04 -!- imode has quit (Ping timeout: 272 seconds).
11:36:09 <fizzie> Actually, we've never had +t (or problem with "unauthorized" topic changes either), that was just a default setting. Let's get rid of that.
11:36:14 -!- ChanServ has set channel mode: +o fizzie.
11:36:16 -!- fizzie has set channel mode: -t.
11:36:20 -!- fizzie has set channel mode: -o FireFly.
11:36:24 -!- fizzie has set channel mode: -o fizzie.
11:36:29 <fizzie> (Tab completion strikes again.)
11:54:18 <int-e> Oh, -t is allowed? It failed in #esoteric yesterday, somehow...
11:59:05 <fizzie> You need to "chanserv set #chan mlock" from the default (+nt-lk) to something that either doesn't mention t, or has -t. I just set it to +n-lk.
12:06:09 <fizzie> Trying to manage all these channels in one client is a bit of a headache. Only one of my lurking set (#perl) has entirely moved. A couple others just set up forks.
12:13:49 <esolangs> [[User:Batata]] https://esolangs.org/w/index.php?diff=83108&oldid=83067 * Batata * (-9)
12:14:51 <esolangs> [[User:Batata]] https://esolangs.org/w/index.php?diff=83109&oldid=83108 * Batata * (+12)
12:15:19 <esolangs> [[User:Batata]] https://esolangs.org/w/index.php?diff=83110&oldid=83109 * Batata * (-3)
12:16:51 <int-e> fizzie: did I mention that I set up a separate shell account for tracking Libera channels :P
12:24:31 <esolangs> [[User:Batata]] https://esolangs.org/w/index.php?diff=83111&oldid=83110 * Batata * (+43)
12:25:55 <esolangs> [[User:Batata]] https://esolangs.org/w/index.php?diff=83112&oldid=83111 * Batata * (+14)
12:26:03 <esolangs> [[User:Batata]] https://esolangs.org/w/index.php?diff=83113&oldid=83112 * Batata * (-2)
12:41:09 <jinn> elliot and ion are u here in the new channel?
12:42:21 <int-e> jinn: try /who ##esoteric
13:08:46 -!- Wezl has joined.
13:19:53 -!- Trieste has joined.
13:20:18 -!- moony has quit (Quit: Ping timeout (120 seconds)).
13:21:15 -!- Wezl has quit (Ping timeout: 265 seconds).
13:21:18 -!- iovoid has quit (Quit: iovoid has quit!).
13:23:00 -!- Bowserinator has quit (Remote host closed the connection).
13:25:00 -!- Wezl has joined.
13:33:53 -!- moony5 has joined.
13:34:13 -!- moony5 has changed nick to moony.
13:34:50 -!- iovoid has joined.
13:40:31 -!- Bowserinator has joined.
13:56:36 -!- moony has changed nick to cd.
14:14:37 -!- u0_a61 has joined.
14:14:37 -!- Wezl has quit (Read error: Connection reset by peer).
14:19:53 -!- u0_a61 has changed nick to Wezl.
14:41:04 -!- cd has changed nick to moony.
14:50:33 -!- Bowserinator has quit (Quit: Blame iczero something happened).
14:50:42 -!- iovoid has quit (Quit: iovoid has quit!).
14:53:16 -!- moony has quit (Quit: Ping timeout (120 seconds)).
15:04:43 -!- moony has joined.
15:05:09 -!- Bowserinator has joined.
15:05:25 -!- iovoid has joined.
15:30:52 -!- imode has joined.
15:39:04 <esolangs> [[Esolang talk:Community portal]] https://esolangs.org/w/index.php?diff=83114&oldid=83107 * Taneb * (+360)
15:41:15 -!- moony has changed nick to moon.
15:47:02 -!- imode has quit (Quit: WeeChat 3.1).
15:47:17 -!- imode has joined.
15:48:07 -!- river has joined.
15:48:20 -!- moon has changed nick to Guest17817.
15:48:30 <river> can we get an esoteric cloak?
15:48:35 <imode> goodbye, freenode, hello libera.
15:49:43 -!- Guest17817 has changed nick to moony.
15:55:42 <int-e> river: You'd need to find a Libera staffer who appreciates the joke.
15:56:08 <int-e> Sad as it is, fizzies boring esolang community idea is more likely to fly.
15:56:38 <int-e> (or esolangs, whatever)
15:57:50 <int-e> otoh who needs cloaks :P
15:58:14 <river> people without cool domain names
16:06:46 -!- Bowserinator has quit (Ping timeout: 264 seconds).
16:08:40 -!- Bowserinator has joined.
16:14:53 <moony> get a cool domain name
16:18:01 -!- imode has changed nick to FuckAndrewLee.
16:18:48 -!- FuckAndrewLee has changed nick to AndrewLee.
16:18:52 -!- AndrewLee has changed nick to imode.
16:19:14 -!- moony has changed nick to cd.
16:33:25 <fizzie> int-e: As a compromise solution, let's register a community with the boring name ("esolangs"), but include `#esoteric` in the list of claimed channel names, on the basis of historical precedent and discoverability.
16:33:46 <fizzie> In either case, I'd kind of give it a bit of time first to see how this Libera thing evolves, before starting a discussion with staffers. I feel like they might have better things to do at the moment, anyway.
16:35:23 <b_jonas> fizzie: I would mention that it's easier to register a channel and community on OFTC
16:36:40 <b_jonas> or more like, you don't have to register communities
16:36:43 <b_jonas> you can just register a channel
16:36:47 <b_jonas> mind you, that works on freenode too
16:37:06 <b_jonas> the whole group thing is just an official path to resolve namespace conflicts in single-flat-land
16:37:37 <b_jonas> OFTC doesn't bother with that
16:39:48 <fizzie> Yeah, I know; I don't think the difficulty is an issue, though. And I think there's something nice about having a claim that's not just "we were here first", though I don't have a rational reason for that feeling.
16:49:43 -!- jix has joined.
17:13:15 -!- cd has changed nick to moony.
17:14:14 <b_jonas> fizzie: does fungot have an command that a bot can use to request that fungot ignore it?
17:15:05 <b_jonas> or maybe a channel that the bot can join to request this?
17:16:54 -!- Bowserinator has quit (Changing host).
17:16:54 -!- Bowserinator has joined.
17:17:16 <fizzie> Unfortunately not. There's a single regular expression (matched against the message prefix) that defines the ignore list.
17:17:37 <fizzie> Not even persisted, I just re-apply it every time I restart it.
17:17:46 <fizzie> By grepping in my logs if it's fallen off the backscroll.
17:17:52 <fizzie> IOW, a highly productionized setup.
17:34:42 -!- kspalaiologos has joined.
17:48:57 -!- nakilon has joined.
18:30:36 -!- imode has changed nick to FuckAndrewLee.
18:33:16 -!- FuckAndrewLee has changed nick to imode.
18:45:44 <b_jonas> fizzie: hmm... I might have to replace his invocation keyword then
18:46:45 <b_jonas> or just convince you to change that regex
18:46:53 -!- arseniiv has quit (Ping timeout: 264 seconds).
18:47:59 -!- arseniiv has joined.
18:50:50 -!- velik has joined.
19:19:31 <fizzie> I'm happy to change that regex whenever. If it was written in a more reasonable language, I could also consider adding a feature to request being ignored, but as it stands it might be quite long before I'd get to that.
19:19:47 <fizzie> I should probably clean up that list anyway, it's got a lot of legacy cruft. It's currently:
19:19:50 <fizzie> ^(EgoBot|HackEgo|HackEso|toBogE|Sparkbot|optbot|lambdabot|oonbotti|metasepia|ruddy|preflex|evalj|idris-bot|passwordBOT|jconn|applybot|blsqbot|fnordbot|termbot|otherbot|j-bot|esowiki|bfbot|egelbot)!
19:40:52 -!- imode has changed nick to kronk.
20:06:11 -!- kronk has changed nick to imode.
20:48:09 <int-e> the table on https://libera.chat/guides/extbans makes my eyes hurt a bit
20:48:35 <int-e> (alternating between grey on white and white on grey for code spans)
20:52:11 <fizzie> Hmm, it's all light-on-dark for me, just slightly different shades. But I've got a prefer-dark thing going on, maybe it styles for that.
21:02:50 -!- eight has joined.
21:05:58 <int-e> https://int-e.eu/~bf3/tmp/alternate.png
21:06:10 <int-e> oh, dark theme, hmm
21:13:15 <fizzie> Well, it's not *that* dissimilar in the dark theme either. https://zem.fi/tmp/extbans.png
21:13:45 <fizzie> Got the same thing going where the code span color is the same as the background of the *other* row.
21:16:50 <int-e> I get the idea and the logic but my brain hurts ;)
21:17:46 <Taneb> I've just dropped my account with freenode nickserv and removed the config from my bouncer. Feels like an end of an era
21:19:52 -!- eight has quit (Quit: leaving).
21:23:51 -!- mrkajetanp has joined.
21:37:15 <arseniiv> shachaf: all of a sudden found your answer on SO about Fix, Mu and Nu in recursion-schemes. Thanks for writing it, clarified a great deal
21:37:29 -!- kspalaiologos has quit (Quit: Leaving).
21:40:44 -!- shikhin has joined.
21:41:34 <arseniiv> hm is it obvious at first sight if μx. x + x and μx. μy. x + y are (naturally) isomorphic? Oh, nvm they are both empty
21:41:58 <arseniiv> (in the strict sense of course)
21:44:48 <int-e> arseniiv: you could use ux. 1+x+x and uxy. 1+x+y instead
21:44:59 <arseniiv> as for the general situation… The allows ⊥ and x1 + x2 inductively, but as μy. x + y allows only
21:47:18 <arseniiv> as for the general situation… The type A ≡ μx. x + x allows ⊥ and x1 + x2 for x1, x2 :: A inductively, but B ≡ μy. x + y allows only ⊥ and c + y for y :: B so for example (⊥ + ⊥) + ⊥ :: A and not :: B. Hehehe
21:49:41 <esolangs> [[Esolang talk:Community portal]] https://esolangs.org/w/index.php?diff=83115&oldid=83114 * Fizzie * (+1128) /* Freenode and the future */ Thoughts about networks.
21:50:21 <arseniiv> int-e: yep that’s an interesting elaboration on 1 + x which is precisely the direction I looked in. Now we have naturals with two successor kinds in the first case (and in case of conaturals, 2^N infinities) and in the second case…
21:51:22 <int-e> arseniiv: x is free in B, I'm confused
21:52:32 <arseniiv> I planned to write first about μy. x + y which I did but then I forgot I hadn’t written about the full thing
21:53:40 <arseniiv> (but still (⊥ + ⊥) + ⊥ shouldn’t inhabit μxy. x + y, if I’m correct)
21:55:48 <arseniiv> okay, μxy. 1 + x + y is of course just μx. N + x and that should give us… yep, again less possibilities: first we use just successors of the first kind and then just those other ones, no intermixing, so not naturally isomorphic with μx. 1 + x + x
21:56:08 <arseniiv> I wonder if these two types occur somewhere in practice
21:58:27 <arseniiv> for example conaturals are great for timeout values (like how many small-step evaluations we are allowed to take), even if encoded in a naïve way. That has some charm IMO
22:00:22 <arseniiv> also one could try something with 1 + (x, y) or (1 + x, 1 + y)
22:00:39 <int-e> arseniiv: let X = μx y. x + y, Y = μy. X + y; ⊥ inhabits X and Y, so ⊥ + ⊥ inhabits X, so (⊥ + ⊥) + ⊥ inhabits X.
22:04:52 <arseniiv> wait, that Y is something completely different
22:05:15 <int-e> Basically thinking of it as data A = A A B; data B = B A B vs. data C = C C C
22:05:50 <arseniiv> though would we even get to μx. x + x by applying this ΛX.Y repeatedly to X?
22:06:41 <arseniiv> int-e: but those datas are mutually recursive whereas X, Y aren’t
22:07:24 <int-e> arseniiv: X already does the full mutual recursion. Y just occurs in unfolding the outer fixed point.
22:09:53 <int-e> μx y. x + y = ⊥ | Y, Y = ⊥ | X + Y
22:10:05 <int-e> so I guess my datatypes aren't reflecting this accurately
22:11:04 <int-e> And there are two ways to get a bottom from μx y. ..., but we have no way of distinguishing them? Bottoms suck.
22:11:23 <int-e> μx y. 1 + x + y doesn't have that issue
22:12:36 <arseniiv> <int-e> And there are two ways to get a bottom from μx y. ..., but we have no way of distinguishing them? Bottoms suck. => though in the haskellization we should get different bottoms due to explicit wraps and unwraps, wouldn’t we?
22:13:05 <int-e> arseniiv: which is how I realized that my representation is wrong
22:14:05 <int-e> data A = A B; data B = B A B <-- it's more like this, and then you do get distinct _|_ and A _|_ before branching. Also, evidently, I'm doing a product instead of a sum.
22:14:34 -!- Wezl has quit (Ping timeout: 264 seconds).
22:14:34 <int-e> data B = B1 A | B2 B
22:15:07 <arseniiv> I agree 1 + x + y and not concerning ourselves with ⊥-containing terms is clearer
22:15:21 <arseniiv> yep I hadn’t even see you did a product at the first time
22:16:18 -!- river has quit (Quit: Leaving).
22:17:20 <int-e> There must be a bunch of papers about how this invalidates generic programming in Haskell ;)
22:19:37 <int-e> . o O ( The Haskell programmer's heaven is a bottom-less pit. )
22:19:43 <arseniiv> and it shows I wrote something incomprehensible about μxy. 1 + x + y being the same as μx. x + N :\
22:20:31 <arseniiv> int-e: I bet what many haskellers’ jokes are about is Just Nothing
22:20:45 <arseniiv> at least I can definitely say fix Just Nothing does what it says
22:20:52 <int-e> > fix error -- practical joke
22:20:54 <lambdabot> "*Exception: *Exception: *Exception: *Exception: *Exception: *Exception: *Ex...
22:21:36 <arseniiv> (it’s a bit unfair I came up with those two some half an hour earlier and just copied them here)
22:22:01 <arseniiv> yeah I remember fix error, I adviced it to several people
22:22:18 <arseniiv> I have it adviced to this time*
22:22:38 <ski> > fix show
22:22:39 <lambdabot> "\"\\\"\\\\\\\"\\\\\\\\\\\\\\\"\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"\\\\\\\\\\\\\...
22:22:58 <int-e> > iterate error "" -- kind of disappointing
22:22:59 <lambdabot> • Couldn't match expected type ‘[Expr]’
22:22:59 <lambdabot> with actual type ‘(a0 -> a0) -> a0’
22:23:54 <int-e> fix f = last $ iterate f undefined -- find the error
22:23:59 <ski> imho, escape characters ought not to be escaped, using themselves as indicator character
22:24:06 <arseniiv> I remember b_jonas saying something about rational escape syntax
22:24:45 <lambdabot> undefined = error "Prelude.undefined"
22:24:52 <int-e> (there's one error!)
22:25:05 <ski> GNU Screen gets this right. `^A' is the (default) escape character. to send a literal `^A' (e.g. to a nested session), you do `^Aa'. for two nested, you get `^Aaa'. linear instead of exponential
22:25:17 <ski> (iirc TMux does the same)
22:26:06 <int-e> . o O ( don't nest screens )
22:26:58 -!- mrkajetanp has quit (Quit: WeeChat 3.1).
22:27:09 <int-e> ^Aa is slightly less convenient at depth 1.
22:27:09 <ski> if `\' was escaped as `\/' (say), then `fix show' above would be `"\"\/"\//"\///"\////"..."'
22:27:24 <int-e> Also taking ^A away from readline is wrong.
22:27:30 <b_jonas> I might have said that I use \x5C all the time instead of \\
22:27:51 <b_jonas> and that I wish K&R set a simple letter escape to mean a backslash, like \s instead of \\
22:27:57 <b_jonas> so that we could have it standardized by now
22:28:05 <ski> (hard to avoid nesting Screens, when you have a mediate SSH (or `su', say))
22:28:07 <int-e> So... personally I'm going against the quoting efficiency with escape ^B^B
22:28:36 * ski normally uses `escape ^Zz'
22:28:44 <ski> (i don't suspend that often anyway)
22:29:10 <int-e> I suspend things a lot more than is healthy.
22:29:11 <ski> b_jonas : yes
22:29:27 <b_jonas> but as for screen, I use ^Qq , and in the rare case when I have two screens nested, one of them uses ^Qq and the other ^Aa
22:29:29 <ski> (using `escape ^Oo' as alternate)
22:29:40 <int-e> ^Z, less a file, forget about the background job... later that day, kill half a dozen pending jobs
22:30:09 <int-e> but there's also the banal ^Z; kill %1 trick that often works when ^C doesn't
22:31:10 <int-e> fungot: where are you and what's an idiosyncracy?
22:32:20 <b_jonas> I do admit that I decided Consumer Society has a less than ideal escape syntax, one that can cause leaning toothpick syndrome
22:32:34 <fizzie> Hmm, why didn't that work.
22:32:35 <ski> what's that ?
22:33:07 <ski> "Consumer Society" ?
22:33:11 <int-e> fizzie: well, no hackeso?
22:33:14 <fizzie> No, the `? freenode -- and I guess it's because HackEso is gone, but I'm not sure why.
22:33:31 <b_jonas> ski: it's an esolang I'm making, or one of a sibling pair of esolangs rather
22:33:36 <int-e> 08:10:45 --- Netsplit *.net <-> *.split quits: lambdabot, HackEso, chibi
22:33:46 <int-e> so gone for ... 16 hours
22:33:58 <ski> how does its escaping work ?
22:34:18 <int-e> cause a netsplit, don't come back --> successful escape
22:34:43 <ski> chibi is a Scheme bot ?
22:35:04 <b_jonas> ski: won't tell yet, the details are not public, but it's one that expands exponentially if you try to nest it
22:35:05 <fizzie> int-e: Oh, it had a temporary network problem, but one that caused it to be immediately "connection refused"ing, and systemd went all "Start request repeated too quickly" and gave up.
22:35:05 <ski> oh, apparently not
22:35:21 -!- HackEso has joined.
22:35:25 <ski> b_jonas : i'm sorry to hear :/
22:35:34 <fizzie> That was simple enough.
22:35:44 <int-e> fizzie: ah. I think lambdabot just sleeps for 3 minutes and tries again
22:35:50 <int-e> (well the loop that controls it)
22:37:09 <fizzie> I don't have fungot over on this side, though I guess I could. It wouldn't have the same shared "repository" (in terms of ^def and such), but OTOH those commands aren't really used much.
22:37:31 <b_jonas> fizzie: you don't have to, if you only have fungot on one side then you won't need to make it ignore the bridge
22:39:13 <b_jonas> fungot will work well that way because it's triggered by a keyword inside the line, pear tree style, unlike most bots that are triggered by a marking at the start of the message contents, so they should be present on all sides or have special code to accomodate the bridge
22:39:32 <b_jonas> yeah, there are caret commands where the trigger is at the start of a line
22:39:33 <lambdabot> int-e: Stop impersonating me, it's annoying!
22:40:05 <arseniiv> hm this gives me strange ideas
22:40:15 <b_jonas> yes, #esoteric tends to do that
22:40:49 <int-e> arseniiv: some sort of parse error
22:40:55 <int-e> > 2 + 2 maybe maybe?
22:40:56 <lambdabot> parse error (possibly incorrect indentation or mismatched brackets)
22:40:57 <arseniiv> hm should have made “maybe maybe” a comment
22:41:18 <fizzie> Bridges are kind of awkward, but I guess there's the benefit that we might see a cross-network botloop.
22:41:43 <b_jonas> fizzie: yes, that's why I asked about fungot's ignore facilities
22:41:51 <b_jonas> so the bot can ask fungot to ignore it
22:49:49 <fizzie> int-e: I'm sure that delay is something I could configure in the .service file. Probably by setting "RestartSec=...", because the log also says "Service RestartSec=100ms expired, scheduling restart" between the 6 attempts it made.
22:50:41 <int-e> fizzie: ah I don't know about that, I have a python script
22:51:45 <fizzie> I've been trying to be all modern.
22:52:48 <int-e> I have systemd units somewhere else, but nothing that actively connects over the network and has to try to be nice .
22:53:24 <fizzie> fungot I've got running in a screen that I just start completely manually. :)
22:59:10 <int-e> fizzie: why do it manually when you can do https://paste.debian.net/1198292/
23:00:08 <int-e> (the best part is where it sleeps for 0.3 seconds because `top` is so slow to start up)
23:01:09 <int-e> my lambdabot setup is *not* modern or nice
23:01:53 <b_jonas> none of my setups are modern
23:01:58 <b_jonas> I don't believe in modern stuff
23:02:34 <b_jonas> or at least I believe in only some modern stuff and not others
23:02:40 <int-e> fizzie: The thing is, I started out exactly as you described... set up screen, start the bot from inside screen
23:02:59 <int-e> fizzie: and over time I added more default screens, and then I got tired of doing it manually every reboot :)
23:03:29 <int-e> but the path of least resistance was to just keep the basic setup and automate it
23:03:55 <int-e> and by now the result is not safe for audiences outside of this channel
23:06:10 <int-e> and it's kind of useful for experimentation... if the socat needs tweaking I can ^C there and modify the command line
23:30:44 -!- arseniiv has quit (Ping timeout: 265 seconds).
23:31:36 -!- mrkajetanp has joined.