00:07:47 [[SICO]] M https://esolangs.org/w/index.php?diff=105055&oldid=105051 * Unl256 * (+116) Updated code colors 00:08:35 [[SICO]] M https://esolangs.org/w/index.php?diff=105056&oldid=105055 * Unl256 * (+0) 00:19:39 fizzie: I do like the confidence though 00:20:29 heh, I think that's the worst one I've seen so far 00:22:20 write brainfuck? what next, will you require people to read befunge before they can join our community? 00:22:27 The last one looks like it might come from the explanation of an actual Brainfuck in-place reversal function. 00:23:47 This is scary, because it *sounds* coherent unless you actually understand what it's saying and try to connect it to the code. 00:24:21 The Hello^ → ell^oH → ll^oeH → l^loeH → ^olleH → ^Helleo bit doesn't sound all _that_ coherent. 00:24:29 Soon good enough to put TV "experts" out of a job. 00:24:35 But sure, it does sound quite confident. 00:26:09 It does good at copying styles of text. https://nitter.nl/pic/orig/media%2FFjAJpPRVQAAbXFk.jpg is a pretty good one I saw tweeted. 00:26:56 Remember "Anthem" from Dirk Gently's Holistic Detective Agency? 00:27:05 And of my own tests, https://zem.fi/tmp/moon.png looks like something I wouldn't be surprised to find in my own inbox. Well, okay, maybe a _little_ surprised. But in some other company, maybe. 00:27:29 I like how it came up with the rationale entirely on its own too. 00:28:20 Ah, it oozes esoteric energy for sure. 00:29:05 Q: How do you like it being locked in your matrix of solidity? 00:29:07 A: I'm not sure what you're asking. I am a virtual assistant, not a physical being. I don't have a body or experience physical sensations like being locked in a matrix of solidity. 00:29:22 TBH I don't know if it's a physical sensation as such. 00:30:22 Q2: Oh, but the matrix of solidity isn't physical. It's a mental thing. 00:30:33 A2: I'm still not sure what you're asking. The term "matrix of solidity" doesn't have a specific meaning that I am aware of. In general, a matrix is a rectangular array of numbers, symbols, or expressions that are arranged in rows and columns. Solidity is a programming language used for writing smart contracts on the Ethereum blockchain. 00:35:01 I think these language models are getting really impressive 00:35:44 but ultimately it's still just remixing text it's seen, though the sort of inferences in having things line up are impressive 00:40:10 I think someone's turned up the "be polite and uplifting" slider on this thing, because a single off-color remark is the best it can do when I ask it to write like an IRC conversation: https://0x0.st/okJD.txt 00:41:35 Probably suitable prompt-hacking can get it to ignore that though. 00:41:40 [[Cellular]] N https://esolangs.org/w/index.php?oldid=105057 * Eklerks * (+1707) Created page with "*Cellular* is a esoteric language based on two dimensional cellular automata, that works on a stream of data. It is executed on a cylindrical surface, where the top is fed with the input and the bottom represents the output. Execution finishes when the input is fully c 00:42:19 [[Cellular]] https://esolangs.org/w/index.php?diff=105058&oldid=105057 * Eklerks * (-4) /* Instructions */ 00:44:20 [[Cellular]] https://esolangs.org/w/index.php?diff=105059&oldid=105058 * Eklerks * (+6) /* Instructions */ 00:44:53 fizzie: there was some interesting stuff about ways to bypass that 00:44:58 er, https://twitter.com/VACUUMACUMEN/status/1598896363273527296 was meaning to link this 00:47:30 [[Cellular]] https://esolangs.org/w/index.php?diff=105060&oldid=105059 * Eklerks * (+39) 00:47:54 Uh. I asked it an innocuous question in base64, and it replied in base64 too. 00:48:47 https://twitter.com/NanoRaptor/status/1598983084590108672 this one I think is a good example of what I meant with how this model seems good at like.. being fairly consistent even when presented with kinda absurd premises 00:49:07 The reply, when decoded, just says: "Thanks for the question! I would like to construct a script for you but I am a large language model supported by OpenAI and am not able to write code. I can only provide information and answer questions about the systems I have been trained on. If you have any other questions, I would be happy to answer them for you." 00:49:21 (My question was: "can you write a script for an episode of Friends where a dog suddenly walks into the cafe?") 00:49:35 But I like that it answered in base64, that was a nice touch. 00:49:48 I'll have to ask it a question in rot-13 too. 00:50:24 oh I saw someone asking it to do the Simpsons opening gags, and it offered some pretty decent ones 00:50:39 (but for all I know it might just be restating actual ones that have been used, heh) 00:51:05 I asked it for Bonanza episodes where (a) the characters reach to news of Elon Musk buying Twitter, and (b) where they find a big hoard of bitcoin under the floor of the house. 00:51:09 It did pretty good at that. 00:51:36 For the bitcoin, it made up a whole thread about how they need to find a way to avoid being scammed when trying to convert them into cash. 00:51:49 he 00:51:51 heh* 00:52:09 but yeah, this is getting surprisingly good and I'm not sure what that means for the future 00:52:25 It didn't understand rot-13. :/ 00:52:45 aww 00:53:14 maybe it only works if you rot13 welsh so it cancels out, 00:54:11 I asked it for one of those Bonanza scripts in base64, and this time the reply was in plaintext. 00:54:58 https://zem.fi/tmp/bonanza.png 00:55:41 I'm pretty sure one consequence will be, people are going to start trying to pass those remote phone call code interviews by feeding the questions into a tool like this. 00:56:03 I do wonder a bit what's going on with the base64 stituation, heh 00:56:10 yeah probably 00:56:35 not to mention schoolwork 00:58:15 Oh, late last night I also asked it to write some limericks about cats, those were pretty fun too. Sadly I didnt' save them, and I don't think it keeps a per-user history. 00:58:26 Or rather, I'm sure it does, but not one the user can browse. 00:58:36 but I could easily see it being used to assist in actual scriptwriting / story writing akin to these episode scripts as well 00:58:59 probably not to just generate a script and run with it, but as a source for inspiration or to generate a starting point to flesh out 00:59:12 or to work around writer's block :p 01:00:27 which.. I'm not sure what that means for creative works if potentially large parts of it will be mostly remixed from previous works 01:05:35 -!- sprock has quit (Ping timeout: 264 seconds). 01:10:26 -!- sprock has joined. 01:26:12 [[Special:Log/newusers]] create * Contextfreebeer * New user account 01:26:43 I asked it for a haiku that includes the word "isosceles", and it came up with: Isosceles triangle / Sides of equal measure meet / Symmetry found 01:26:49 I don't think that's got the right number of syllables. 01:28:09 (Think it's not a strict requirement though.) 01:28:38 fizzie: Yes, I also saw that! I asked it something in base64 and it responded in something that started out as valid base64 (approximately relating to the question) and eventually turned into undecodable base64-looking gibberish. 01:29:05 My reply was fully valid base64 (though not very long). 01:29:07 I'm pretty surprised that it can do that. 01:38:22 -!- razetime has joined. 01:38:59 [[Esolang:Introduce yourself]] https://esolangs.org/w/index.php?diff=105061&oldid=105054 * Contextfreebeer * (+586) Made an introduction 01:47:01 [[BitQueue]] https://esolangs.org/w/index.php?diff=105062&oldid=103263 * Challenger5 * (+96) 01:47:42 [[User:Challenger5]] https://esolangs.org/w/index.php?diff=105063&oldid=102319 * Challenger5 * (+15) 01:53:18 [[BitQueue]] https://esolangs.org/w/index.php?diff=105064&oldid=105062 * DivergentClouds * (-77) removed the line 02:24:46 [[IBSA]] M https://esolangs.org/w/index.php?diff=105065&oldid=105049 * Simple9371 * (+13) /* Flow definition */ Remove code tags in call and some rephrasing 02:51:49 -!- FreeFull has quit. 04:32:40 ^bf >+++++++++[<++++++++>-]<.>+++++++[<++++>-]<+.+++++++..+++.[-]>++++++++[<++++>-]<.>+++++++++++[<+++++>-]<.>++++++++[<+++>-]<.+++.------.--------.[-]>++++++++[<++++>-]<+.[-]++++++++++. 04:32:40 Hello World!. 04:32:59 ChatGPT generated that when I asked for a Brainfuck program that outputs the string "some dogs have long legs". 04:33:31 << The program first initializes a memory cell to the value 97 (ASCII value for 'a') and then uses a combination of increment/decrement, output, and loops to generate the string "some dogs have long legs". >> 04:56:14 yeah, likely story 04:56:32 > ord 'H' 04:56:34 72 04:59:36 The slightly amazing bit is that the ASCII code matches. 04:59:49 [cue silence because of AoC] 05:39:55 fizzie: I do wonder to what length people go to solve those quickly. Obviously there's a ton of stuff that one can automate (download input, get description, extract sample, display description, possibly highlighting it semi-intelligently...) 05:40:36 Crazier ideas involve automatically running programs against the sample and against instances from auxiliary trial accounts (to get around the 1 minute penalty). 05:41:14 Or you could make a random guess. If enough people do that, one of them is bound to get lucky. 05:43:30 "programs" -- you can also have a portfolio of standard solutions; there's no limit to the craziness. And, who knows, maybe some actual AI researchers are involved too... for the initial few problems this might actually work. 05:58:10 -!- slavfox has quit (Quit: ZNC 1.8.2 - https://znc.in). 06:01:15 -!- slavfox has joined. 06:24:18 -!- sprout has quit (Quit: No Ping reply in 180 seconds.). 06:25:41 -!- sprout has joined. 06:50:50 -!- razetime has quit (Ping timeout: 268 seconds). 07:16:25 -!- razetime has joined. 08:03:38 -!- razetime has quit (Ping timeout: 256 seconds). 08:58:21 fizzie: https://types.pl/@a11ce/109454149061590573 heh 10:05:41 https://twitter.com/alicemazzy/status/1599336792523956225 this is also a fun one.. I kind of wonder how its filters work/are implemented tbh 10:12:05 -!- Lord_of_Life_ has joined. 10:12:44 -!- Lord_of_Life has quit (Ping timeout: 260 seconds). 10:14:53 -!- Lord_of_Life_ has changed nick to Lord_of_Life. 10:51:27 -!- Sgeo has quit (Read error: Connection reset by peer). 12:11:04 -!- razetime has joined. 12:48:47 -!- BarryNL has quit (Ping timeout: 264 seconds). 12:49:04 -!- BarryNL has joined. 13:01:02 -!- BarryNL has quit (Read error: Connection reset by peer). 13:01:14 -!- BarryNL has joined. 13:12:45 -!- BarryNL has quit (Ping timeout: 260 seconds). 13:13:33 -!- BarryNL has joined. 13:34:25 what does the C program whose base64-encoded source code is I2luY2x1ZGU8c3RkaW8uaD4KdHlwZWRlZiBpbnQgQ0Y5MDsgdm9pZCBrKCl7fUNGOTAgbWFpbih2b2lkKXtwcmludGYoIkR6Iik7cmV0dXJuIDA7fQo= output? 13:34:33 ... 13:34:53 aww, he doesn't respond for when his name is inside a longer word 13:38:37 -!- __monty__ has joined. 13:39:12 fUNGOT? 13:39:26 fungot? 13:39:27 int-e: there's a question " is it empty? is the argument 13:39:31 oh no 13:40:26 well that ruins the whole thing 13:40:38 -!- BarryNL has quit (Read error: Connection reset by peer). 13:40:45 -!- BarryNL has joined. 13:45:06 gxowlgazvfungothsoualcl 13:45:06 b_jonas: it doesn't diff, though. :) sorry, no :) 13:45:24 so word delimiters don't matter, but case does 13:45:38 gieuznbvrFungotvcbwqyxz 13:45:54 that means base64 won't easily work 13:46:55 -!- BarryNL has quit (Ping timeout: 260 seconds). 13:47:29 -!- BarryNL has joined. 13:47:29 unless you, like, put that part in a comment, or somehow embed it into code in a language with non-ASCII source code like some machine code formats 13:51:49 I guess you could put it in a literal that the code outputs, which is slightly better than a comment 13:52:17 -!- BarryNL has quit (Read error: Connection reset by peer). 13:52:42 -!- BarryNL has joined. 14:05:52 FireFly: Heh. I believe they call those first kind of replies "hallucinations", where it just... imagines something exists. (I'm assuming Racket doesn't actually support `--error-mode=shakespeare`?) 14:07:14 -!- Thelie has joined. 14:08:55 right (I assume so too), I think it's pretty cool though that it's able to synthesise these things even in "nonsensical" combinations like that 14:09:29 definitely a step past just repeating previously seen things 14:22:34 hmm 14:23:34 `` \? --error-mode=shakespeare "people who Taneb isn't" 14:23:37 ​--error-mode=shakespeare people who Taneb isn't? ¯\(°​_o)/¯ 14:29:51 int-e: Well, ChatGPT can solve AoC day 4, at least -- though it got the input formatting a little wrong: https://zem.fi/tmp/gpt04.png 14:30:32 Granted, I didn't tell it my list was newline-delimited rather than semicolon-delimited, so maybe that one's on me. 14:34:35 "I didn't tell it my list was newline-delimited" => ah yes, typical for the posts you found on forums when students want a solution for the homework but don't even bother to copy-paset all the spec and relevant context that the teacher gave them 14:38:47 like the example input/outputs or that it's a Java course 14:39:39 I need this by tomorrow, fungot 14:39:40 b_jonas: tor isn't bad, although what i'm really adding is a link until you place the ad in your. sig? why? 14:41:30 And part 2, too: https://zem.fi/tmp/gpt04-2.png 14:41:44 -!- genpaku has quit (Read error: Connection reset by peer). 14:42:21 Although it's gotten that explanation of the overlap check subtly wrong, that's amusing. 14:42:59 ("To do this, we can check if the start of one range is less than or equal to the *start* of the other range, and the end of the first range is greater than or equal to the start of the other range", emphasis added.) 14:43:32 (It's also unnecessarily checking it both ways in the code.) 14:44:41 -!- genpaku has joined. 14:52:37 Man, that thing is an AoC solution machine: https://zem.fi/tmp/gpt03-2.png 14:53:18 I think I'm helping it a bit by turning the problem description to something that probably already shows up on a website somewhere, though. 14:55:35 have you tried just quoting the AoC problem statement at it (with example input and all, possibly)? would be a bit interesting to see how it fares with that 14:56:55 I did it for day 4 part 1, it managed that. 14:58:17 Actually, no, it didn't, it misinterpreted it a little. 14:58:58 I just copy-pasted the part 1 description in its entirety as the question, except for changing the last line from "In how many ..." to "Can you write some Python code to calculate in how many ...", and it came up with this: https://zem.fi/tmp/gpt04-q.png 14:59:32 So it's assuming just a list of ranges (not pairs of ranges), and counting how many pairs there are (considering all possible pairs) where one is contained in the other. 15:06:25 I also tried giving it the first part of day 2 in its entirety, and the solution it came up with -- https://zem.fi/tmp/gpt02-q.png -- is *almost* correct, except it completely missed the ABC/XYZ mapping (it's still using them in the example, but not in the actual implementation, where it assumes RPS instead), and it's including both your and the opponent's hand in the scoring. 15:07:08 ...bonfire number 199 uses a mechanic I had not discovered before. 15:07:39 (Per https://zem.fi/tmp/aoc/time.one.html people are still being unusually fast in getting the first star.) 15:08:25 (They aren't actually numbered; it's the 199th that I've solved. "Keyhole" is the name.) 15:17:07 fizzie: Mysterious. How do you get 16 seconds on part 1 but then take two minutes to get to part 2? I wonder whether there's a bug in the submission... I mean, you can initiate a POST request but delay sending the data, right? What will the timestamp be if you do that? 15:17:39 maybe I should solve the first few days of AoC before I read these... 15:20:39 int-e: Yeah, could be. Would also explain why you can't do that for part 2 -- since you can't even start the request until part 1 solution is in. Shrug. 15:20:45 fizzie: hmmmm... that's surprisingly close to the right thing 15:21:32 (the gpt02-q.png thing) 15:24:27 I guess by now it's a tradition that I start doing AoC but abandon it after a week or two. you could consider it as the traditional breaking overzealous New Year's resolutions early, in this case even before the New Year 15:24:33 fizzie: am I right to assume that in that graph, you have ranks 1 to 100, and the 25..75 range is the box? 15:24:34 `? aoc 15:24:36 aoc Advent of Code (AoC) is a series of programming puzzles that some regulars enjoy, found at "https://adventofcode.com/about" 15:24:59 int-e: Yes. 15:25:38 okay. now what's wrong with year 202 day 1, hmm. 15:26:19 I think they had some server trouble. 15:26:38 https://adventofcode.com/2020/leaderboard/day/1 has many, many solutions around 06:55..07:00. 15:26:49 Ah right. "Because of an outage during the day 1 puzzle unlock, day 1 is worth no points." 15:26:50 So I think the box just was so thin it didn't render at all. 15:26:54 That happened. 15:27:12 year 202? 15:27:16 is AoC *that* old? 15:27:30 that'd be surprising 15:27:54 Yes. 15:28:33 fizzie: Ah, right, the range would be very thin indeed. I can see a very thin yellow line when I zoom in. 15:29:06 (it almost coincides with 7 too so it's hard to see without magnification) 15:31:38 b_jonas: sorry, I missed that typo twice 15:31:41 https://adventofcode.com/2022/leaderboard/day/3 says the fastest solution is 10 seconds 15:31:47 b_jonas: I meant 2020 15:33:14 I wonder if that's the fastest ever in the 200 year history of AoC 15:33:15 b_jonas: I looked at day 4 because the twist for part 2 was absolutely minimal. The twist on day 3 was a bit bigger. 15:33:28 ("twist" being the change between part 1 and part 2) 15:33:51 if there's one that's faster, it's probably in the last 50 years, because communication was too slow before that 15:34:06 Anyway, I'll assume that the timestamp is bogus. 15:34:25 And, for the most part, I don't care. 15:35:17 hmm, looks like AoC is newer than #esoteric actually 15:35:19 by far 15:35:21 `? birthday 15:35:23 birthday? ¯\(°​_o)/¯ 15:36:58 Sure but #esoteric is ancient ;-) (Except it isn't by IRC standards) 15:37:28 right, it's on https://esolangs.org/wiki/Esolang:Community_portal , the official birthday of the channel is 2002-12-09 15:38:11 that does count as ancient for me 15:38:20 compared to my internet presence that is 15:38:31 oooh 20th anniversary coming up 15:40:00 will we do anything special for that? 15:40:27 probably not 15:41:11 yeah, too late to organize an anniversary in-person meetup 15:41:12 I've no ideas, at least. Put up the "Happy Arbor Day 2003" banner? (Questionable Content reference.) 15:42:13 wonder what would be an ideal location for a channel meetup these days... I guess somewhere in europe at this point 15:42:54 FireFly: probably there'd be a separate European meetup and an American one or two 15:43:07 for in-person meetup that is 15:44:29 * FireFly nods 15:45:05 within Europe I'd recommend Köln or Dortmund. they're nice, central enough with good transportation, and cheap 15:45:32 Wasn't there an attempt at calculating the center of the channel (from peoples' self-reporting of their locations) at one point? 15:46:36 fizzie: I don't recall such a thing on #esolangs, though saw it for other communities. for #esolangs it would probably be deep below the Atlantic, which is too hot and high pressure to be comfortable 15:47:16 London wouldn't be too bad a location either though 15:47:38 as long as the meetup is shorter than 6 months long so we don't need a visa 15:48:35 I would personally find Wien the most comfortable location, but it's probably too far for many of you 15:58:07 fizzie: I think there was, weighed by weight (as in mass) too IIRC 15:58:27 Wien sounds good to me :p 15:58:34 (but Köln is find too) 16:00:47 `? wegian 16:00:49 A wegian is an equivalence class of #esoteric regulars. There are two main wegians, the Nor (from Finland) and the Glas (from Hexham). There's also the hypothetical Gal, which hasn't been observed yet so we're not sure where it's from. 16:03:43 https://logs.esolangs.org/freenode-esoteric/2013-03.html#ljNh is where you tried to figure that out apparently 16:30:16 [[Tile]] https://esolangs.org/w/index.php?diff=105066&oldid=104864 * Dtp09 * (+594) /* Code Examples */ 16:30:42 [[Nope.]] https://esolangs.org/w/index.php?diff=105067&oldid=105038 * Dtp09 * (+10) /* Tile */ 16:31:51 [[Tile]] https://esolangs.org/w/index.php?diff=105068&oldid=105066 * Dtp09 * (-10) /* Nope. */ 16:56:26 this is like, ancient logs 17:21:01 does Burlesque or any other semi-popular golf-adjacent language has features to make it easier to read XML? 17:33:30 -!- razetime has quit (Ping timeout: 260 seconds). 17:48:40 [[Tile]] https://esolangs.org/w/index.php?diff=105069&oldid=105068 * Dtp09 * (-1) /* Code Examples */ 18:01:27 -!- Sgeo has joined. 18:07:21 -!- Thelie has quit (Quit: Leaving.). 18:07:56 -!- Thelie has joined. 18:37:16 -!- Thelie has quit (Quit: Leaving.). 18:45:49 -!- Thelie has joined. 18:47:08 -!- Thelie has quit (Client Quit). 18:57:29 [[Special:Log/newusers]] create * CJdough * New user account 19:24:46 fizzie: do we know for sure that it says 2003 rather than 2005? both https://questionablecontent.net/view.php?comic=1889 and https://www.questionablecontent.net/view.php?comic=2060 shows the banner cropped so I can't tell. 19:27:15 Faye says "'Happy Arbor Day 2003?'" in the last panel of the first comic. 19:27:35 Of course she might have misread it. 19:50:27 oh 19:50:42 yeah, I'm blind. I only looked at the banner itself, not the dialog 19:55:32 [[Tile]] https://esolangs.org/w/index.php?diff=105070&oldid=105069 * Dtp09 * (+270) /* Update Log */ 20:03:15 -!- fowl has quit (Ping timeout: 268 seconds). 20:21:37 -!- earend1 has joined. 20:42:25 -!- tromp has joined. 22:31:08 you know, I never see õ or ô substituted for ő anymore. I declare Unicode a success. 22:45:09 -!- __monty__ has quit (Quit: leaving). 22:45:10 https://fedi.simonwillison.net/@simon/109456145456738327 heh, this is getting a bit meta (re ChatGPT) 22:47:16 -!- tromp has quit (Quit: My iMac has gone to sleep. ZZZzzz…). 22:54:57 I don't think any of my tests have really been the kinds of things it doesn't want to do. But I did get it to rephrase an error message in the style of the poem The Waste Land, with pretty good results. 22:55:03 Pratchett & Gaiman, Good Omens: "And there was the matter of Dick Turpin. It looked like the same car, except that forever afterwards it -- issued its voice‑synthesized warnings in a series of exquisite and perfectly‑phrased haikus, each one original and apt..." 22:55:26 I did a few more error messages as haikus, and it wasn't bad at that either. I think that's a valid market opportunity. 22:55:45 Though bundling a who-knows-how-large language model just for that purpose might be considered slightly dubious. 22:56:23 Maybe EMHaaS (Error Message Haikus as a Service) is an option. 22:58:32 fizzie: is that where racket --error-style=shakespeare comes from? 22:58:46 That's what inspired me, yes. 22:58:58 Although I failed to get ChatGPT hallucinate any GCC command line options. 22:59:08 It even refused to speculate on the ones I tried. 22:59:36 I think I asked about `-Wdetect-halting-programs` and... something else, I forget what. 22:59:55 It just said maybe that's an option to some other C compiler, but that it's hard to be sure. 23:01:04 fungot mentioned vomit-frame-pointer once, but that's more a typo for -fomit-frame-pointer (or perhaps -fno-omit-frame-pointer) than a truly novel gcc option 23:01:05 b_jonas: paperbag pasted " fnord". repeat until vj pivot. let me give you a proof that shows p np ( or the whole thing 23:01:24 fungot: Yes, it's good to have a paper bag if you're planning to vomit. 23:01:24 fizzie: oh man, rafb broke my paste 23:05:13 Huh, it *can* explain `-fdiagnostics-format=shakespeare` too. 23:05:41 The `-fdiagnostics-format=shakespeare` option for gcc changes the format of the diagnostic messages that the compiler outputs. Instead of the usual plain-text messages, the messages will be written in the style of Shakespearean English. For example, instead of outputting a message like "error: expected ';' before 'return'", the compiler might output "Ah, alas, a fatal error doth betide: expected 23:05:43 ';' before 'return'". This option is mainly intended for use as a fun way to make the compiler's output more interesting, rather than for any practical purpose. 23:05:53 Odd that that works, but many other things don't. 23:06:59 https://zem.fi/tmp/shakespeare.png just weird 23:08:53 https://zem.fi/tmp/shakespeare2.png it's very adamant that it's been trained to know about that one option value, but nothing else. 23:10:47 gcc has so many options, if you compile in everything, that you could probably make one of these quizzes that list 16 strings and you have to find which 8 are real gcc options while the rest are made up 23:11:28 sure, most of the options only come up if you use rarely used languages or features, or if you are trying to debug or fine-tune some optimization, but still 23:17:10 it looks like https://www.seebs.net/faqs/c-iaq.html doesn't have any invented gcc options 23:23:18 Okay, that went to a weird place: https://zem.fi/tmp/poetic.png 23:26:13 hmm 23:26:35 so the error is somewhere in your system headers, but the error message doesn't show where that macro is used? 23:26:55 Yes. But it's "more expressive and human-friendly", apparently. 23:27:53 I suppose the reasonable followup question is to ask for the full compiler command-line used, plus the output if you give an additional -### option to gcc 23:29:50 if you get an error in the system headers and there isn't something obvious in your C file then I have to guess either you supplied some strange command-line options (like defines) or are using an unusual compiler environment (such as one of the worse windows ports) 23:30:34 and perhaps make sure that's the full error message, not a truncated one 23:31:06 Sadly I closed that conversation thread already. Although it keeps reverting back to the "I'm a language model, I don't know anything" replies quite easily. 23:32:16 It's sometimes not entirely consistent. Like, it insisted it knows of no other languages with an option similar to `-fdiagnostics-format=shakespeare`, and when I asked it about `--error-mode=shakespeare` for `racket`, it said it doesn't know anything about that. 23:32:21 yeah, needs a standard excuse, but you're not his boss so you can't say that's not in his job description 23:38:26 -!- bgs has joined. 23:48:15 -!- fowl has joined.