00:15:31 <shachaf> man, how 'bout that #haskell, huh
00:23:07 <zzo38> I wrote a official fan-game/fan-story policy for my "xyzabcde" game series. It is found at: http://zzo38computer.org/xyzabcde/fangame.txt Do you think it is OK?
00:32:20 -!- LKoen has quit (Quit: “It’s only logical. First you learn to talk, then you learn to think. Too bad it’s not the other way round.”).
00:34:16 <shachaf> Is there any language that does non-global GC effectively?
00:34:25 <shachaf> I think Erlang is the closest I know of.
00:50:56 <shachaf> As Saul Gorn said: "Tomorrow, which isn’t even here yet, will never be the day after tomorrow again."
01:21:01 -!- tromp has joined.
01:23:42 -!- tromp_ has joined.
01:25:35 -!- tromp has quit (Ping timeout: 252 seconds).
01:27:19 -!- FreeFull has quit.
01:28:09 -!- tromp_ has quit (Ping timeout: 252 seconds).
02:05:59 <zzo38> I don't know if LLVM does, either, but you can try to look maybe
02:46:40 <tswett[m]> So I'm retraining GPT-2 on #esoteric logs. :D
02:46:49 <tswett[m]> It's producing a lot of messages from "shashmaf".
02:47:36 * tswett[m] sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/aEBBWmuCSCXCSseeazmVRhta >
02:49:07 <tswett[m]> I was using too high a learning rate anyway. The starting point is a model that knows English really well, but the aggressive training caused it to forget all that.
02:49:49 <zzo38> And there is a timestamp with five pieces, too, it look like
03:11:42 -!- tromp has joined.
03:13:52 <tswett[m]> One of the samples consisted entirely of a single humongous timestamp.
03:14:20 <tswett[m]> I'll go ahead and paste the entire sample.
03:14:22 * tswett[m] sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/EgJPeXoVAePQqlqemMKxktil >
03:16:07 -!- tromp has quit (Ping timeout: 258 seconds).
03:42:58 <tswett[m]> "Language-bound variables are treated as if they were bound to variable length. To test whether a variable-length variable has some fixed-length state, let x be the bound variable of x, and let y be a variable of variable-length state, such that it has a variable-length binding that is one of the two variable-length bindings of the variable-length variable."
03:44:43 <kmc> shachaf: what did #haskell do now
03:46:19 <shachaf> I think it was just the usual monad nonsense.
03:47:48 <shachaf> speaking of nonsense, is rust trying to take every bad idea from npm/nodejs
03:48:11 <kmc> what'd they do now
03:48:14 <zzo38> Which bad ideas is that?
03:49:04 <shachaf> I just saw some Cargo packages and they seemed inspired by the npm ideal of "do one thing and do it poorly".
03:49:36 <shachaf> And also written by JavaScript people I think.
03:49:45 <zzo38> I should think it depends on the package. There are a lot of packages isn't so good.
03:49:47 <kmc> they seem hellbent on ruining a great systems language
03:50:03 <kmc> and this is true, 90% of everything is crap (being generous)
03:50:59 <zzo38> Yes, so that is going to do, regardless of the programming language in use. Just avoid to use the packages that is no good. The existence of it should not require you to use them if it is not applicable for your program
03:51:01 <kmc> shachaf: did you see that the rust-lang.org design was so disastrous that the entire website team resigned / was deposed after massive negative feedback from the community
03:51:13 <kmc> imo bring back my bullet points
03:54:12 <kmc> and there's going to be some kind of postmortem about the whole cock-up
03:54:14 <kmc> which is delayed
03:59:37 <shachaf> fortunately all sorts of people are working on systems languages now
04:04:36 <kmc> any cool ones i should know about?
04:12:59 <shachaf> what do you want from a systems language
04:18:31 <tswett[m]> Hey, it finally generated something that looks like a chat log. :D
04:19:19 <tswett[m]> Ethan_Gestapo has a conversation with Koenig and GangstaR about writing about dying.
04:20:29 <tswett[m]> Here it is. https://pastebin.com/H2dVTRuF
04:21:55 <tswett[m]> You know what's particularly weird? It generated chat logs—but not in the style of the chat logs I'm training on.
04:22:05 <tswett[m]> It added brackets around the timestamps. The training data doesn't have those.
04:22:25 <tswett[m]> Presumably, GPT-2's original corpus has some chat logs similar to that.
04:23:08 <tswett[m]> So my fine-tuned model has learned that it should generate chat logs, but it hasn't learned that it shouldn't put brackets around the timestamps.
04:24:17 <tswett[m]> Also, the timestamps happen to all have 28 as the hour.
04:29:41 <kmc> I am the queen of the 28th hour
04:31:49 -!- nfd has joined.
04:35:37 -!- nfd9001 has quit (Ping timeout: 250 seconds).
04:41:30 -!- adu has quit (Quit: adu).
04:44:17 <tswett[m]> We've never had a person here named esowiki, right?
04:44:30 <tswett[m]> Because it just generated a long monologue by someone named esowiki.
04:45:00 -!- S_Gautam has joined.
04:45:09 <shachaf> esowiki was a bot, and also a person pretending to be that bot.
04:45:22 <shachaf> Did it generate a long, stream-of-consciousness by tswett[m]?
04:45:46 <tswett[m]> I don't think I delivered very many of my famous monologues in 2018, though.
04:45:50 <tswett[m]> That's the training data I'm using.
04:52:20 -!- sebbu3 has joined.
04:56:06 -!- sebbu has quit (Ping timeout: 258 seconds).
04:59:46 -!- tromp has joined.
05:04:07 -!- tromp has quit (Ping timeout: 252 seconds).
05:08:03 -!- Vorpal has quit (Quit: ZNC - http://znc.sourceforge.net).
05:11:47 <shachaf> did you see my ridiculous c11 fmt macro
05:14:13 -!- nfd9001 has joined.
05:16:13 -!- nfd has quit (Ping timeout: 252 seconds).
05:20:38 <tswett[m]> Now the model has created somebody named "int_jonas".
05:21:56 <tswett[m]> 11:50:50 <int_jonas> Phantom: Phantom isn't Phantom, Phantom was Phantom's thought, Phantom is Phantom's thought
05:22:12 <tswett[m]> In this sample, every message has "Phantom" once every two or three words.
05:22:23 <tswett[m]> Okay, enough of this nonsense, I'm going to bed.
05:35:39 -!- Melvar has quit (Quit: thunderstorm).
05:47:35 <moony> oh is esowiki deat?
06:14:26 <HackEso> 1016) <Bike> http://www.amazon.com/Someone-Cuttlefish-Shapeshifter-Erotica-ebook/dp/B0087PTMW2 i hope you know this is going to /fuck up/ my amazon recs
06:38:38 -!- sprocklem has quit (Read error: Connection reset by peer).
06:40:13 -!- sprocklem has joined.
06:46:21 -!- tromp has joined.
06:47:42 -!- tromp_ has joined.
06:50:49 -!- tromp has quit (Ping timeout: 252 seconds).
06:51:55 -!- tromp_ has quit (Ping timeout: 252 seconds).
07:13:10 <fizzie> That might have been a reference to the bot.
07:13:39 <fizzie> Which is temporarily dead.
07:20:30 <fizzie> I've managed to diagnose it to the fact that it's statically linked; for whatevere reason that makes it segfault in std::thread::detach(). When I remove features = ["fully_static_link"] from the BUILD file it starts to work again.
07:20:36 <fizzie> Unfortunately that was a necessary workaround for the fact that GCC on the machine it's running on isn't new enough to compile the code. If they'd just go on and release Debian 10...
07:20:40 <fizzie> In the meanwhile, I'm not sure how to work around the workaround.
07:21:18 <shachaf> static linking is the best
07:21:28 <shachaf> if your program doesn't work with static linking it's broken hth
07:21:53 <fizzie> Well, I mean, glibc doesn't even support static linking for programs that call getaddrinfo.
07:22:08 <fizzie> external/fi_zem_bracket/event/socket.cc:258: warning: Using 'getaddrinfo' in statically linked applications requires at runtime the shared libraries from the glibc version used for linking
07:22:51 <fizzie> I don't think the breakage is necessary related to that, though, since it segfaults the same way even on the same machine it's built on nowadays.
07:23:11 <shachaf> getaddrinfo is kind of a scow api anyway
07:26:23 <fizzie> I don't know. But I don't want to start figuring out how to make Bazel build for that. Although apparently it's maybe been done.
07:26:56 <shachaf> is there any excuse for these things to be complicated instead of simple
07:27:48 -!- nfd has joined.
07:30:53 <fizzie> I guess it's hard to make complex things simple.
07:31:50 -!- nfd9001 has quit (Ping timeout: 272 seconds).
07:32:54 <shachaf> but why is it complex in the first place
07:35:53 <b_jonas> shachaf: Cargo packages have been doing that for a while, as in, CPAN-like dependency hell. you don't have to use those packages.
07:37:25 -!- tromp has joined.
07:38:47 <b_jonas> fizzie: you could install a newer gcc then. it doesn't have many dependencies.
07:40:05 -!- APic_ has joined.
07:41:57 -!- nfd has quit (*.net *.split).
07:41:59 -!- uplime has quit (*.net *.split).
07:45:46 -!- john_metcalf has quit (Ping timeout: 272 seconds).
07:47:44 -!- wmww has quit (*.net *.split).
07:47:46 -!- APic has quit (*.net *.split).
07:47:46 -!- olsner has quit (*.net *.split).
07:52:43 -!- wmww has joined.
07:52:43 -!- olsner has joined.
07:53:19 -!- nfd has joined.
07:53:19 -!- uplime has joined.
08:10:47 -!- Phantom_Hoover has joined.
08:23:23 -!- b_jonas has quit (Quit: leaving).
08:39:33 <fizzie> That might be plausible. It's taking a long time to build though.
09:02:04 -!- tromp has quit (Remote host closed the connection).
09:13:47 -!- tromp has joined.
09:16:36 <fizzie> (Maybe I should've done --disable-bootstrap.)
09:30:06 -!- AnotherTest has joined.
09:34:18 -!- wob_jonas has joined.
09:35:03 <wob_jonas> fizzie: yes; what can speed it down a lot is if you configure it with --enable-languages=c,c++,fortran or the like, so that it doesn't try to compile the java and golang and ada and objective c and whatever else it has now standard libraries
09:45:06 <wob_jonas> and yes, you shouldn't disable bootstrap unless you're cross-compiling
09:45:24 <wob_jonas> onto some unusual architecture where you can't native compile gcc
10:03:00 -!- APic_ has changed nick to apic.
10:03:04 -!- apic has changed nick to APic.
10:12:56 <wob_jonas> fizzie: thank you for spending time to maintain those tools though, they're very useful
10:20:28 -!- Lord_of_Life has quit (Ping timeout: 246 seconds).
10:22:14 <fizzie> I did --enable-languages=c,c++, which should be enough. Now I'll be out for the day, but in the evening I'll hopefully be able to bring things back up.
10:24:14 -!- Lord_of_Life has joined.
10:34:54 -!- oklopol has joined.
10:39:55 -!- atslash has quit (Quit: Leaving).
11:01:09 -!- S_Gautam has quit (Quit: Connection closed for inactivity).
11:41:43 -!- S_Gautam has joined.
11:42:30 -!- AnotherTest has quit (Ping timeout: 258 seconds).
12:29:25 -!- sebbu3 has changed nick to sebbu.
12:38:38 -!- Melvar has joined.
12:39:51 -!- arseniiv has joined.
13:17:57 -!- AnotherTest has joined.
13:20:36 -!- rain1 has joined.
13:22:03 -!- rain2 has quit (Ping timeout: 252 seconds).
13:23:51 -!- AnotherTest has quit (Ping timeout: 250 seconds).
13:26:27 -!- ^[ has quit (Ping timeout: 250 seconds).
13:28:34 -!- ^[ has joined.
14:20:09 -!- rodgort has joined.
14:39:49 -!- oklopol has quit (Ping timeout: 246 seconds).
14:43:31 -!- AnotherTest has joined.
14:51:07 -!- AnotherTest has quit (Ping timeout: 248 seconds).
15:30:40 -!- LKoen has joined.
15:48:43 -!- wob_jonas has quit (Remote host closed the connection).
16:08:54 -!- AnotherTest has joined.
16:21:12 -!- FreeFull has joined.
16:24:31 -!- sleepnap has joined.
16:44:22 <int-e> ski: We have such entries for many regulars. I was wondering if this one says anything about a crusade against abstraction... or perhaps some other silly thing. But no such luck.
16:45:51 -!- b_jonas has joined.
16:48:27 -!- Sgeo__ has joined.
16:50:19 -!- AnotherTest has quit (Ping timeout: 252 seconds).
16:52:06 -!- AnotherTest has joined.
16:52:09 -!- Sgeo_ has quit (Ping timeout: 252 seconds).
16:56:07 <b_jonas> yet another new mulligan rule for M:tG? wtf
16:58:40 -!- AnotherTest has quit (Ping timeout: 272 seconds).
17:17:58 -!- sleepnap has quit (Ping timeout: 268 seconds).
17:28:38 -!- LKoen has quit (Remote host closed the connection).
17:34:45 -!- LKoen has joined.
18:30:54 -!- sleepnap has joined.
18:41:09 -!- S_Gautam has quit (Quit: Connection closed for inactivity).
18:44:05 <fizzie> Blah. Now it says "null context when constructing CivetServer. Possible problem binding to port."
18:44:09 <fizzie> Well, at least that's an error message, instead of a mysterious segfault.
18:45:08 -!- esowiki has joined.
18:45:17 <fizzie> Oh, there was just an instance running in gdb from previous testing.
19:04:17 -!- arseniiv_ has joined.
19:06:21 -!- arseniiv has quit (Ping timeout: 252 seconds).
19:13:23 -!- Sgeo_ has joined.
19:16:37 -!- Sgeo__ has quit (Ping timeout: 252 seconds).
19:17:58 -!- Sgeo__ has joined.
19:21:17 -!- Sgeo_ has quit (Ping timeout: 258 seconds).
20:06:22 -!- nfd9001 has joined.
20:10:18 -!- uplime^ has joined.
20:15:09 -!- nfd has quit (*.net *.split).
20:15:10 -!- uplime has quit (*.net *.split).
20:18:08 -!- lynn has quit (Ping timeout: 248 seconds).
20:19:05 -!- lynn_ has joined.
20:20:51 -!- AnotherTest has joined.
20:25:11 -!- AnotherTest has quit (Ping timeout: 252 seconds).
20:31:37 -!- sleepnap has quit (Quit: Leaving.).
20:32:43 -!- sleepnap has joined.
20:51:21 <fizzie> CivetWeb is an embedded web server thing, I use it for metrics (and in another binary, for serving the logs).
21:14:30 <kmc> civit poo coffee
21:16:54 -!- uplime^ has changed nick to uplime.
21:28:03 -!- aloril_ has joined.
21:32:02 -!- aloril has quit (Ping timeout: 268 seconds).
21:37:54 -!- LKoen has quit (Quit: “It’s only logical. First you learn to talk, then you learn to think. Too bad it’s not the other way round.”).
21:42:55 -!- FreeFull has quit (Ping timeout: 252 seconds).
21:43:53 -!- FreeFull has joined.
21:54:23 -!- arseniiv_ has changed nick to arseniiv.
22:19:44 -!- tromp has quit (Remote host closed the connection).
22:20:16 -!- Lord_of_Life_ has joined.
22:20:48 -!- Lord_of_Life has quit (Ping timeout: 248 seconds).
22:21:24 -!- Lord_of_Life_ has changed nick to Lord_of_Life.
22:24:51 -!- sleepnap has quit (Quit: Leaving.).
22:51:23 -!- b_jonas has quit (Quit: leaving).
22:53:57 -!- Phantom_Hoover has quit (Read error: Connection reset by peer).
23:49:25 <Sgeo__> I'm looking at a piece of code to convert opaque internal stuff into euler rotations, and I need to reverse it, and it should be easy, but my brain just hates me
23:50:34 <shachaf> euler rotations? i just met 'er hth
23:51:23 <Sgeo__> I don't want to need pen and paper to do math
23:52:21 -!- arseniiv has quit (Ping timeout: 252 seconds).