00:01:53 -!- BlueProtoman has joined. 00:02:11 `welcome BlueProtoman 00:02:15 BlueProtoman: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.) 00:02:19 Hello, oerjan. 00:10:28 imo thue is perfect for transducers 00:14:21 -!- DHeadshot has quit (Read error: Connection reset by peer). 00:26:24 Anyone know why my Brainfuck loops aren't working? I think it's the way I'm storing the []'s in a table, I don't think I'm doing it correctly. Code's coming up/ 00:26:41 -!- myndzi has quit (Read error: Connection reset by peer). 00:26:50 http://ideone.com/PT6Rp 00:27:11 -!- myndzi has joined. 00:28:40 It looks like it's doing more than checking loops 00:28:53 is brace_table a global variable? 00:28:57 No. 00:29:08 BlueProtoman: don't you need to save backwards jumps as well? 00:29:16 Lemme explain how this works. 00:29:36 I plan on implementing several Brainfuck dialects in my program. To do this, each dialect comes as a class that derives from AbstractBF. 00:29:51 So the Brainfuck class has a std::unordered_map brace_table. 00:30:18 Each entry is a pair of braces; the ['s position is the key and the ]'s position is the element. 00:30:48 When a [ is found, its position is recorded in brace_table, with its element initially being -1. 00:31:43 While all this is going on, I'm pushing [ positions onto a std::stack. 00:32:14 When a ] is found, its corresponding key is found with find(), and then that element is set to the ]'s position. 00:32:27 If any element is equal to -1, something went wrong. 00:32:51 And thus we don't run the program. Problem is, I seem to be looping to the beginning of the program whenever I hit a [. 00:33:17 BlueProtoman: well the bug is probably not in the part you pasted, then. 00:33:22 how does your program deal with jumping from a [ to a ] when the value is zero? 00:33:33 ais523: When the element is 0, you mean? 00:33:38 yes 00:34:10 It doesn't. I should fix that by testing whether tempbracestack is empty or not. 00:34:19 But that doesn't seem to be my problem. 00:34:33 BlueProtoman: um i think he means when the _tape_ value is zero. 00:35:01 oerjan: What do you mean the tape value? I don't quite follow. 00:35:30 BlueProtoman: for the standard brainfuck with a tape of cells 00:35:35 or array 00:35:46 oerjan: Oh, right. Lemme show you. 00:35:55 I handle all BF instructions as function objects. 00:36:11 But first I use a regex to filter out other characters. 00:36:42 Here's my entire class. http://ideone.com/ZTTwA 00:36:49 Including the instructions. 00:39:16 -!- Mefi has joined. 00:41:50 Any thoughts? 00:43:36 -!- oerjan has quit (Quit: Lost terminal). 00:45:51 -!- oerjan has joined. 00:46:06 is instructions[program[currentinstruction++]](); 00:46:15 well defined behavior? 00:46:39 as in, is it guaranteed when the increment happens wrt. the function call? 00:46:53 oerjan: I don't think the timing of the increment wrt the call is guaranteed 00:47:03 although if that doesn't matter, it's well defined in other respects 00:47:17 i'm pretty sure it matters for [ and ] 00:47:19 -!- itidus20 has joined. 00:47:20 unspecified rather than undefined, as in IIRC you're guaranteed that it will definitely happen either before or after 00:47:45 and no, it doesn't matter, just go backwards to the first [ or ] of the group 00:47:49 -!- Mefi has left. 00:47:56 (this actually works!) 00:48:11 (I used the trick to obfuscate a BF interp for the IOCCC once, but it didn't win) 00:48:19 ais523: erm in the given code it matters, currentinstruction is used to look up the point to jump to 00:48:38 oerjan: with [[, there's no way that the second [ can perform a jump 00:48:40 if it has already been incremented, that could well give a zero 00:48:44 likewise, with ]], there's no way the second ] can perform a jump 00:48:57 ais523: erm are you even understanding my point? 00:49:05 so for each [, you can just give the instruction after it the same jump target as the [ itself 00:49:15 I'm understanding your point, but I think you're missing mine 00:49:32 which is that you have enough information to determine where to jump to even with an uncertainty of 1 in the IP 00:49:44 ais523: sheesh. you're being irrelevant to the code we're actually trying to debug. 00:50:03 oh, this is debugging? I thought you were trying to find a creative way to write a BF interp 00:50:52 -!- itidus21 has quit (Ping timeout: 244 seconds). 00:50:58 BlueProtoman: anyway, try putting the currentinstruction++ in a separate command afterwards. 00:50:59 oerjan: If instructions (which is a std::unordered_map>) returns nothing, it'll throw an exception, which I catch. 00:51:16 oh hm 00:51:25 or wait 00:52:08 BlueProtoman: that's not the problem. the instruction is found fine, but the code for [ needs currentinstruction to _not_ be incremented in between 00:52:11 afaiu 00:52:24 If it returns no function object, then we've hit a character that isn't a Brainfuck instruction, and was somehow missed by the regecx. 00:52:26 *regex. 00:52:27 well assuming my guess is right 00:53:04 OK, I'm getting somewhere. Still not finished, though. 00:53:06 Hold on. 00:53:25 BlueProtoman: my guess is that (1) the instruction for a [ is looked up, correctly (2) currentinstruction++ is incremented, too early (3) the code for [ breaks because of (2) 00:53:59 What's the third most relevant architecture beyond x86[_64] and ARM? PPC? MIPS? 00:54:07 Probably PPC. 00:54:10 oerjan: For some reason, I get �� as my output. 00:54:11 yeah 00:54:17 maybe some pic stuff 00:54:19 That is SUCH a distant third. 00:54:24 AVR? 00:54:30 Gregor: All current-gen consoles use it. 00:54:30 You know, for Arduinos. 00:54:35 pikhq: Yeah, that's true. 00:54:41 Arduinos are a tiny tiny tiny tiny fraction of all AVRs in use 00:54:45 -!- zzo38 has joined. 00:54:47 Hm. 00:54:52 PIC and AVR and their friends probably outnumber x86 and ARM by a large factor 00:55:09 BlueProtoman: what happens if you replace instructions[program[currentinstruction++]](); with instructions[program[currentinstruction]](); currentinstruction++; ? 00:55:11 I really mean general-purpose MMU-capable archs *shrugs* 00:55:17 Arduinos are expensive as hell 00:55:20 but what ebout the amount of software written for it? 00:55:21 Why are they that expensive anyways 00:55:31 are they still 50$? 00:55:34 Lumpio-: there are plenty of cheap clones 00:55:35 oerjan: I get garbage output, but no infinite loop. 00:55:38 i thought they were like $30 00:55:41 ouch 00:55:43 Looks like this; �� 00:55:47 BlueProtoman: ok scratch that then :( 00:55:58 I've come to see Arduinos as a representation of the new hipster electronics generation 00:55:58 It means I'm on the right track, though, doesn't it? 00:56:03 Them, and sparkfun 00:56:12 Expensive as hell but buy is it hip and cool 00:56:17 BlueProtoman: well it means it changed the behavior, at least. 00:56:20 s/buy/boy/ 00:56:27 Lumpio-: arduinos are reasonable development boards 00:56:37 * oerjan defers to real c++ experts, and goes for food -> 00:56:41 if you build a project around an arduino and leave it in there, that's kind of silly 00:56:48 Yeah, but that's what they do. 00:56:53 lumpio: haven't seen anything interesting sound synthesis wise from that stuff 00:56:57 but really you're whining about someone spending an extra $20 here and there 00:56:59 kinda disappointing 00:57:01 madbr: Arduino? 00:57:09 these are largely people with jobs who make more than $20 an hour 00:57:23 though I guess arduino isn't very good for audio anyways 00:57:38 Do the AVRS used in arduinoes even have hardware multiplication 00:57:53 maybe they just want to make something cool, not spend all day soldering chips by hand to appease some crusty old electronics beardo 00:58:07 basically you sound like more of a hipster than them 00:58:15 "i was into electronics before it was cool and accessible" 00:58:38 Any tips for my Brainfuck interpreter? I no longer get infinite loops with my output, but I do get garbage output (and I know that the program doesn't run completely). http://ideone.com/4dXq8 00:59:00 this is like the people who whine about any software not written in C 00:59:22 omg think of all the inefficiency, i'm using literally megabytes of memory, or 0.025% of the total available on my system 00:59:44 BlueProtoman, you are outside of the program, probably after it 00:59:46 Lumpio-: to answer your question, no the ATmega328 and friends do not have hardware multiply 00:59:51 ook 00:59:59 rszeno: What makes you say that? 01:00:16 kmc: ...actually I wasn't into electronics before it was cool and accessible .__. 01:00:19 output 01:00:26 Lumpio-: so you're a *poseur* hipster :) 01:00:28 I've barely even played with MCUs yet 01:00:54 look like reading a random place in memory 01:00:58 I do have ideas and plans to build a new computer, probably using an existing CPU architecture for practical purposes; if I could build a MMIX which is cost-effective I would do that; GCC already can compile for MMIX. However I will also need a GPU or DSP; can you recommend one which has Free software and can do audio and video? 01:01:25 rszeno: I don't see how that's possible, since I have a check to make sure the program stops running when currentinstruction > program.size(). 01:01:56 Gregor: PPC is used in game consoles and a lot of embedded stuff, MIPS also used in embedded stuff plus Chinese government is pushing it, s390 / s390x still quite relevant in enterprise 01:02:16 PPC might be the most common synthesizable core for FPGAs 01:02:17 i guess you change beaviour when you play with indexes 01:02:35 Why is Chinese government pushing it? 01:02:42 Wow, s390 still gets used? 01:02:57 zzo38: Because x86 is dominated by untrustworthy capitalist swine 01:02:58 Maybe I should use an iterator instead? 01:03:03 s 01:03:32 They're making their own processors so they can have hardware spyware on them 01:03:38 Gregor: IBM doesn't believe in letting things die. 01:03:50 yeah that's kind of the point of buying enterprise stuff 01:04:15 Lazy people use enterpricey stuff with long-term support. 01:04:18 imo is a bad idea to use oop and streams for this but probably is only my opinion, :) 01:04:20 True haxors live on the bleeding edge 01:04:25 haha 01:05:03 Can you suggest the CPU, GPU, DSP to use? 01:05:09 hmm actually the #1 (publicly known) supercomputer in the world is based on SPARC64 01:05:25 -!- ais523 has quit. 01:05:35 Can we make a Checkout compiler to any GPU? 01:06:39 is it possible to use a fast gpu from anything else than PC, ARM or PPC? 01:07:01 Does zzo38 even need anything besides a text mode? 01:07:34 For what I am making, it needs user-definable modes. The BIOS can include the text mode built-in 01:08:04 And it need to support composite video out, and preferably also component video out. 01:08:15 -!- derdon has quit (Remote host closed the connection). 01:08:33 What kind of resolution are we talking about 01:08:44 NTSC 01:08:51 Not PAL? :( 01:09:08 I could make a PAL version as well, but at first just NTSC 01:09:08 At any rate, that kind of resolution is not really hard to generate. 01:09:30 Make your own little GPU out of an FPGA or something. 01:09:40 lumpio: sounds hard no? 01:09:55 Nah, NTSC is easier to generate and practically all modern TVs support it 01:09:59 Not only will an FPGA might be slow, but the ones I know of are encrypted and lack Free software 01:10:00 madbr: What does? 01:10:21 Slow? 01:10:23 pffrt 01:10:38 designing a nice gpu that does something else than just regurgitate the content of the VRAM 01:10:40 FPGA might be slightly slow, possibly 01:10:46 And what do you mean encrypted, are you talking about the dev tools? 01:11:14 madbr: A RAMDAC is all a true programmer needs! 01:11:16 ¬u¬ 01:11:23 Lumpio-: Yes. Also the program to load into the FPGA is encrypted for all the ones I know of 01:11:30 What kind of stuff did you have in mind >madbr 01:11:35 Tiles? Sprites? 01:11:37 ...3D? 01:11:44 zzo38, http://www.milkymist.org/mmone.html ? 01:11:52 3d is kinda hard so I was thinking more like tiles and sprites :D 01:12:26 Tiles and sprites shouldn't be /that/ hard 01:12:26 tiles or other static layers aren't too hard actually, it's sprites that are hard 01:12:54 At least with the kind of power we have these days 01:13:15 well, with the kind of power we have you can just use a framebuffer 01:13:42 Then again I've never done any video output more advanced than http://qp.virkkunen.net/private/jbxmAAqned 01:13:57 was thinking something like a snes/genesis style chip 01:14:00 ...I should really get back to that thing and add a stable clock source, the jitter is awful 01:14:15 that generates a scanline worth of data each scanline 01:15:14 one nice idea I had was to precalculate which sprite is on top of each pixel 01:15:22 No. Oh no. Why am I considering looking at Eiffel again 01:15:24 and then just load the pixel that's on top 01:15:42 so that you have more or less "infinite fillrate" 01:15:50 ie can put lots of large objects 01:16:02 haven't figured out how to do lots of small sprites though :/ 01:16:19 Sprite systems usually have fixed sizes for sprites anyways 01:16:25 (Or a couple of choices) 01:16:33 And to make large objects you use a lot of small sprites. 01:18:08 well, yeah most classic sprite systems use transparent colors 01:18:23 BlueProtoman: is +1 correct in currentinstruction = brace_table.find(currentinstruction)->first +1; ? 01:18:34 which means you have to read every potential pixel to see if it uses the transparent color 01:18:48 which puts an absolute limit on fillrate 01:18:52 oerjan: It should be. That puts me at the instruction past the ]. 01:19:29 BlueProtoman: past the [ you mean? but you have a currentintstruction++ to be performed after that, you know... 01:19:33 (ie if a pixel is hidden by another pixel, you can't have it for free) 01:19:41 *currentinstruction++ 01:19:50 oerjan: Wait. I don't understand. Mind elaborating on what I'm doing wrong? 01:20:24 BlueProtoman: does that +1 in the code for ] take into account that currentinstruction is automatically incremented after each instruction run? 01:20:27 essentially I think it would make more sense to store the transparency mask separately 01:20:45 oerjan: It might not, actually. 01:20:59 madbr: But you still have to read each pixel of the mask 01:21:04 Do software developers usually write code with only half an idea of what they're doing? 01:21:16 lumpio: but the mask can be 1bpp 01:21:27 So? 01:21:29 BlueProtoman: all the time, i hear (but then i'm not a software developer) 01:21:31 BlueProtoman: yes, at least me 01:21:37 which means you're reading 32 pixels at the same time if your memory bus is 32 bits 01:21:44 Oh, OK. 01:21:50 OR 01:21:57 the mask can be stored as segments 01:22:03 Where does that mask come from in the first place 01:22:08 Does each sprite have one? 01:22:31 "segment 1 starts at pixel a and finishes at pixel b, segment 2 starts at pixel c and finishes at pixel d" 01:22:53 lumpio: yeah 01:23:29 but yeah if you're using a segment list you still have to read the segment list 01:23:40 BlueProtoman: btw did i mention at the start that i thought the brace_table should contain entried for the ]'s as well? then you don't need an expensive find afaiu 01:23:42 so there's still a limit on the number of sprites you can put onscreen 01:23:47 *entries 01:23:48 Is EiffelStudio good? 01:24:00 Are there good alternatives if I want to write proprietary programs in Eiffel? 01:24:05 oerjan: It does. The key is the ['s, the element is the ]'s 01:24:17 BlueProtoman: no i mean, index _both_ ways 01:24:27 oerjan: I could use a multimap for that, I guess. 01:24:29 then it's just a simple lookup either way 01:24:47 Maybe. BRB, dinner 01:25:47 (I dunno if somebody's already mentioned this but is there a point in indexing loops beforehand?) 01:27:04 Lumpio-: in bf? well i guess you _could_ do it as you go, but it's a natural part of parsing... 01:27:23 if you do software rendering VGA style, you probably have to precompute the transparency mask of each sprite too 01:27:28 Also speaking of sound synthesis on MCUs 01:27:32 it's definitely more efficient than searching for the matching [] every time :P 01:27:37 http://qp.virkkunen.net/private/edjvvEcfiH 01:27:42 checking every pixel for transparency is way too slow 01:27:43 This is as far as I got before getting bored ¬u¬ 01:28:05 This thing doesn't have a hardware multiplier either so it can only handle 3 or 4 channels or so 01:28:12 Lumpio-: wait was your question about something completely different? :P 01:28:13 Also due to my lacking math-fu there's an awful lot of noise. 01:28:20 oerjan: Nope 01:28:56 what's the waveform? 01:29:25 you can do a nice soundtrack in 4 channels 01:29:26 Well you can at least find backward jumps by keeping a stack of open loops 01:29:33 ...provided that you have lots of sample data :D 01:29:40 But I guess forward jumps would still result in a scan 01:29:52 It's just a simple sine wave 01:29:58 With an envelope I drew in GIMP 01:30:06 After looking at a sample of a music box. 01:30:37 It would work better with music with less low notes. 01:30:50 But I couldn't be bothered to look for more MIDI files 01:31:06 what's the max data size? 01:31:12 umm 01:31:18 I think that thing had 8kB of memory 01:31:30 ah yeah 8kb is too tight for a MOD 01:32:02 Oh actual samples would be much easier to play, you could play way more tracks 01:32:15 But an envelope means at least one true multiply per sample 01:32:27 well, with samples you probably want volume 01:32:32 (True as in it cannot be optimized completely into a series of shifts and adds because both arguments are unknown) 01:32:50 Nah you can do decent music with a constant volume for each sample 01:32:52 Pre-multiplied 01:32:53 though some old DOS software cheated and used a LUT for the volume :D 01:33:08 was still a 16k LUT tho 01:33:20 heh 01:33:33 Well I could have LUTs for a couple of levels of volume I guess 01:33:34 lumpio: doesn't that multiply the amount of data you need very fast? 01:33:41 Assuming 8-bit output one LUT is only 256b 01:33:55 some HW synths totally optimize out the multiply too 01:34:02 How does that work 01:34:10 yamaha FM synths have no multipliers 01:34:18 I couldn't figure out how to do it without a true multiply 01:34:26 what they do is that they don't store sin(x) 01:34:33 they store log(sin(x)) 01:34:37 ooh 01:34:39 and exp(x) 01:34:40 :D 01:34:43 oooo 01:34:45 Mathematical! 01:34:47 I get it 01:34:58 You know that might work 01:35:04 hmm 01:35:20 Actually now that I think of it I think I've heard of that before 01:35:43 for better emulation of the yamaha fm chips, someone actually de-encapsulated one and read the bits out of the LUT with a microscope 01:38:11 kmc, imo will be a good start to jump to image processing and forget about audio for all the rest of his life, :) 01:38:36 kmc: yeah I remember seeing that 01:38:57 kmc: I think the opl3 has a real multiplier instead of the logsin + exp thing 01:40:08 Does an IDE which has a license saying you may only use it for "free" projects count as "free"? 01:40:25 (Not asking about EiffelStudio in particular, although that inspired the question) 01:41:57 rszeno: huh? 01:42:54 i saw a chip ones was a simple one, i suspect a yamaha fm is more complicated 01:43:20 yamaha fm isn't that complicated 01:43:22 it's smart :D 01:44:02 meh now I have to try the log/exp thing out 01:44:35 essentially it has one oscillator and one enveloppe 01:44:44 and multiplexes the hell out of it :D 01:47:14 so most of the chip is taken up by registers actually 01:48:57 How do you generate multiple frequencies with a single oscillator? 01:50:31 armonics? 01:50:59 the 'oscillator' is a lookup table for sin(x) (or log(sin(x))) right? 01:51:05 yeah 01:51:10 lumpio: easy 01:51:14 so for each sample you can have several internal clock cycles 01:51:22 and latch the output into a different register each time 01:51:45 you store the current phase for each oscillator in a different register 01:51:51 and current frequency too 01:52:12 audio sample rate is way below any reasonable clock rate for logic, so this is a good tradeoff 01:52:35 for each oscillator, you read the phase, read the frequency, add together, store that back in the phase register 01:52:39 oh 01:52:43 So it's not actually just one oscillator 01:52:47 well 01:52:56 Just one look-up table with multiple oscillators built around it 01:53:02 ...that's what I did with my thing. 01:53:08 they need different phase registers 01:53:17 and different parameter registers of course 01:53:24 but all the processing logic is shared 01:54:23 I would like the DSP or GPU to be able to do tiles, sprites, and video playback. 3D rendering is not important. 01:54:23 like, read phase 1, read freq 1, add, write phase 1, process waveform 1, process volume 1, add to total register 01:54:37 then, read phase 2, read freq 2, add, write phase 2, process waveform 2, process volume 2, add to total register 01:54:49 etc, repeat for all the channels 01:55:04 then take the total register and write that to the DAC 01:55:36 zzo: I'm not sure hardware tiles are that useful :D 01:55:45 they'Re easy to do tho 01:56:01 video playback is different though 01:56:38 depends on if you'Re fine with paletted video essentially 01:56:43 madbr: It doesn't need the feature to do tiles built-in to the hardware! What I mean is, that it is capable of being programmed to do these things. 02:00:10 hm 02:01:33 by the time you have a superscalar pipelined architecture and if you don't have too much pixels to fill, you can do that in software 02:01:41 ie pentium 02:02:40 essentially you need enough fillrate to fill the screen at least once per frame 02:04:16 because then, aside from the transparency stuff (best SNES feature!), you can use smart code to fake any numbers of layers etc you like and limit overdraw 02:04:16 For video playback it is OK if it requires the CPU and the GPU or DSP to work together 02:04:35 is paletted (256 color?) video ok? 02:04:48 or it has to be some proper algo like mpeg? 02:04:57 No. It has to be proper video such as Theora or Dirac. 02:05:32 ok then you need a beefy cpu 02:06:16 might need hardware FFT even 02:06:55 you're jumping in the hundreads of mhz range 02:07:26 which means all the 2d stuff will be totally fine and you've got so much cpu that you can just software mix the audio 02:07:54 OK. The criteria are that it is not extremely expensive and does not require proprietary software to program it 02:07:56 unless you want high resolution on the 2d (640x480 or more?) 02:08:14 I told you, the resolution should be enough for NTSC. 02:08:17 oh 02:08:22 then 320x240 is enough 02:08:26 Just standard definition 4:3; no high definition is needed. 02:08:43 essentially you need an ARM system on a chip 02:09:06 Which one? ARM7? ARM11? 02:09:19 dunno 02:10:07 Other than the processors, hard drive, optical drive, connectors, etc, I intend to design the rest of the hardware myself such as what is connected and in what way, as well as most of the software too. 02:10:35 does decoding theora require floating point? 02:10:39 or dirac 02:11:05 I don't know. 02:11:59 if you need fast floating point that rules out arm11 I think 02:14:26 But I intend to design how the memory and that stuff is connected myself, in an attempt to prevent such things as DRM from being programmed in, as well as for other purposes too. Even if the ARM11 does not have fast floating point, whatever CPU is used if a GPU or DSP which is not tooo expensive and does not require proprietary software to use, can do some things the CPU doesn't, that might work too. 02:17:20 GPUs tend to be proprietary 02:17:49 I know that; but is there a DSP which would be sufficient for this purpose? 02:17:59 on the raspberry pi I think you can't even write to the GPU in hardware, you have to boot in linux and use precompiled drivers 02:18:06 depends 02:18:28 for 3d, software rendering is reasonable at low resolutions with beefy processors 02:18:36 if you can live without bilinar filtering :D 02:19:30 I don't need 3D rendering; although surely someone will program it to make some (perhaps not particularly good quality) 3D rendering if they want to. 02:19:36 sometimes SIMD stuff helps 02:21:25 -!- itidus22 has joined. 02:22:36 but yeah 3d hardware implementation details are jealously guarded these days as far as I can tell, a bit silly 02:25:18 -!- itidus20 has quit (Ping timeout: 244 seconds). 02:25:34 -!- david_werecat has quit (Ping timeout: 245 seconds). 02:26:08 -!- itidus22 has changed nick to itidus21. 02:44:15 -!- BlueProtoman has quit (Quit: Leaving). 03:05:51 -!- rszeno has left. 03:11:32 -!- oerjan has quit (Quit: Good night). 03:21:29 -!- rszeno has joined. 03:50:03 -!- itidus21 has left ("Leaving"). 04:10:55 -!- asiekierka has joined. 04:49:19 -!- jaba has joined. 04:50:44 -!- jaba has quit (Client Quit). 04:51:12 -!- rszeno has left. 05:09:10 -!- asiekierka has quit (Quit: Wychodzi). 05:30:18 -!- MoALTz_ has joined. 05:32:41 -!- MoALTz has quit (Ping timeout: 252 seconds). 05:34:54 -!- madbr has quit (Quit: Radiateur). 05:37:06 -!- MoALTz_ has quit (Ping timeout: 244 seconds). 05:41:57 -!- MoALTz has joined. 06:35:23 -!- rszeno has joined. 06:40:01 -!- rszeno has left. 06:48:20 -!- augur has changed nick to poststructuralis. 06:48:42 -!- poststructuralis has changed nick to augur. 07:02:28 -!- Taneb has joined. 07:03:43 Hello! 07:07:33 -!- Phantom_Hoover has joined. 07:19:46 -!- nooga has joined. 07:20:42 kmc: People's explanations of things in #haskell irritate me these days even though I can't pinpoint exactly why. 07:20:50 I should probably just pull a kmc and leave. 07:21:40 that's right 07:21:58 just stand up, put on shades, make some kind of witty parting remark 07:22:05 then casually flick a lit cigarette behind you as you leave 07:22:06 Like ddarius? 07:22:10 causing the whole place to go up in flames 07:22:14 00:22 ddarius was last seen on #haskell 80 days, 5 hours, 14 minutes and 58 seconds ago, saying: BMeph: No. 07:22:19 did ddarius have a witty parting remark 07:22:20 @seen kmc 07:22:20 Unknown command, try @list 07:22:38 00:22 kmc was last seen 3 seconds ago, saying: 07:22:49 oh heh 07:22:53 cause i just asked it in pm ;P 07:22:58 Sploisprot. 07:23:00 it was some boring quote from #git 07:23:01 sorry bro 07:23:08 brochaf 07:23:21 sorry, bro. sbarro. 07:23:43 You should've said the launch code to #haskell as you were leaving. 07:23:58 the launch code to #haskell is (repeat 0) 07:23:59 > repeat 0 07:24:00 [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,... 07:24:04 you have to type the whole thing, though 07:24:07 > last (repeat 0) 07:24:12 Terminated 07:25:04 I guess lambdabot doesn't compile with -O 07:25:13 > lаst (repeat 0) 07:25:14 0 07:28:56 > last $ map (flip showCReal pi) [1..] 07:29:00 mueval-core: Time limit exceeded 07:29:03 Aw. 07:32:49 gotta sleep 07:32:52 'night all 07:42:58 -!- Taneb has quit (Ping timeout: 245 seconds). 07:43:49 -!- azaq23 has quit (Quit: Leaving.). 07:59:20 good morning europe! 08:03:01 A bit late for a morning. 08:03:37