00:00:02 I assume you have access to a dictionary, google, etc. 00:00:09 ehird, I did google it. 00:00:10 Therefore I do not dumb down my language. 00:00:16 as I mentioned above. 00:00:42 ehird, I prefer to make my message accessible to the target of the message. 00:01:19 ehird, out of pure interest, do you speak like that to other people of your own age. 00:01:27 AnMaster: You don't do a very good job of it; I can't understand you half the time. 00:02:08 And I could probably count the amount of times someone of my own age has been willing to voluntarily listen to me on a few hands. 00:03:37 ehird, 1) you are exaggerating 2) ais can though, same for Deewiant, so I suspect the issue is very much that you lack patience. 00:03:55 ehird, yes I can see why that would happen. 00:04:04 very easily. 00:04:21 (1) Please feel free to tell that to others of my age. (2) "I prefer to make my message accessible to the target of the message." if I cannot understand it, it is not accessible to me. You can understand "whoosh" to, you just have to be patient enough to look it up. 00:05:09 about 2, maybe I have not been unsuccessful in your case. I have tried however. 00:05:21 Not been unsuccessful? 00:05:24 You mean been successful? 00:05:29 indeed 00:05:30 typoed 00:05:36 either remove "not" or "un" 00:34:42 double negation :D 00:35:04 that sort of morphosyntactic error is BEAUTIFUL 00:39:40 psygnisfive, strange tate 00:39:42 taste* 00:40:02 well 00:40:13 its cool because they happen all over the place in very particular circumstances 00:40:31 people will double negate or misnegate things in a very precise way 00:50:23 psygnisfive, yeah and? 00:50:40 and its interesting 00:50:48 because there are all sorts of scope issues 00:50:49 psygnisfive, it happens when you rewrite part of the line. It isn't strange 00:50:56 and it's boring. 00:51:00 not really! 00:51:41 psygnisfive, when did you last write an esolang interpreter instead of discuss linguistics? 00:51:51 or compiler 00:51:53 or similar 00:52:07 i dont write compilers, but i wrote an interpret some time in the past. :P 00:52:24 listen, its not my fault all this esolang stuff is circling the same few ideas! >| 00:52:24 psygnisfive, recently? 00:52:33 psygnisfive, they aren't 00:52:36 they are too 00:52:45 now that is interesting 00:53:26 AnMaster: nobody cares that he's offtopic 00:53:28 That english says " too" meaning you disagree. 00:53:42 to me it parse as " as well" 00:53:49 actually 00:53:54 its not " too" 00:53:55 or "in addition to what you said" 00:53:58 its "V too" 00:54:02 psygnisfive, whatever. 00:54:08 well, its big difference 00:54:14 actually 00:54:17 psygnisfive, not to a non-expert. 00:54:21 its not even "v too" 00:54:28 its "T too" 00:54:40 since its an adverbial that after the T element 00:54:47 psygnisfive, actually it is " too" 00:54:48 "I have too gone to the store! :|" 00:54:49 even 00:55:00 yes, anmaster 00:55:03 possibly negation there 00:55:04 is a T element 00:55:19 psygnisfive, you mean a form of "be" in other word 00:55:25 yes 00:55:30 actually "did not! did too!" 00:55:32 is possible 00:55:32 "too" seems to be the positive version of "not" 00:55:36 so 00:55:39 since the distribution is apparently identical 00:55:39 not just be 00:55:40 any verb 00:55:42 even 00:55:44 what? 00:55:47 any T element 00:55:48 not any verb 00:55:57 psygnisfive, I have no clue whatsoever that means. 00:56:02 eh 00:56:04 its complicated 00:56:05 but 00:56:13 theres a syntatic position within a sentence called T 00:56:25 psygnisfive, make sure this doesn't become a tl;dr 00:56:33 right about here: Subject T Negation ... 00:56:36 a tl;dr for linguistics is 5 lines. 00:56:39 ;P 00:56:55 psygnisfive, mhm 00:57:11 some auxiliary verbs raise into T, sometimes there are things already in T, etc. 00:57:40 so for instance "will" is already in T: "I will not be going to the store" but remove "will" and you get "I am not going to the store" 00:58:12 psygnisfive, anyway. INTERCAL and Befunge are very different. So are Brainfuck and Underload. And what about Slashes and RUBE? 00:58:20 there are a lot of variation. 00:58:36 circling the same ideas. :P 00:58:37 Those are all old, AnMaster. 00:58:49 ehird, slashes is rather new isn't it? 00:58:54 maybe I misremember 00:58:59 but the other ones are old yes. 00:59:07 psygnisfive, what about Feather? 00:59:18 psygnisfive, oklotalk? 00:59:22 hold on, let me look at what these are like so i can properly bitch about them 00:59:26 oklotalk doesnt even exist 00:59:31 its fictional! 00:59:37 psygnisfive, yes it does. Not on the wiki though. 00:59:37 which is admittedly a novel concept 00:59:38 BUT 00:59:55 psygnisfive, and then there is OISC 01:00:05 esolangs.org isnt loading for me 01:00:22 psygnisfive, neither Feather or oklotalk are on the wiki. Trust me. 01:00:31 psygnisfive, you have to read channel logs. 01:00:35 im looking for rube and that other one right now 01:00:39 slashes 01:01:08 psygnisfive, also there was an idea about a language based on time traveling trains. 01:01:30 I don't think it was ever well speced. 01:01:34 But rather novel 01:01:43 sounds pretty boring. :P 01:01:56 psygnisfive, a lot less than linguistics at least. 01:02:03 much more so! 01:02:13 psygnisfive, and there is Photon, Not well speced either 01:02:27 but based on physics simulation of light reflecting. 01:02:32 oh slashes 01:02:44 ihope's regex like thing 01:02:44 psygnisfive, what about Thue? 01:02:53 thue is older than computers! 01:03:00 maybe 01:03:18 thue is like... a meta-language 01:03:20 almost 01:03:28 psygnisfive, IRP! 01:03:32 IRP? 01:03:35 yes IRP 01:03:41 that is more your style. 01:03:49 fuzzy human logic and so on ;P 01:04:04 i dont care about fuzzy human logic 01:04:12 psygnisfive, linguistics too 01:04:14 kind of 01:04:19 im just interested in novel ways of conceptualizing computational tasks 01:04:27 psygnisfive, then IRP is for you 01:05:44 i think i understand slashes 01:06:01 psygnisfive, IRP is more your style 01:06:01 it seems to be a time reversal of the normal chat s/// operation 01:06:10 ill look at irp 01:06:21 psygnisfive, err s/// isn't a "chat operation" as such 01:06:32 it is based on Perl, sed and so on 01:06:36 right 01:06:41 but like 01:06:45 s/// is totally a chat operation 01:06:46 AnMaster: stfu with "IRP is for you!" and "LINGUISTICS IS FOR FAGS" 01:06:51 it just runs in the users mind ;P 01:07:05 ehird: i agree. there are lots of breeders who like linguistics. 01:07:07 psygnisfive, " s/// is totally a chat operation" <-- how do you mean 01:07:15 is that what you call straight people 01:07:16 breeders 01:07:33 AnMaster: stfu with "IRP is for you!" and "LINGUISTICS IS FOR FAGS" <-- is psygnisfive the latter? I see. I wasn't aware. 01:07:37 in that we use s/a/b/ in chats to mean "replace that last instance when i said a with b" 01:07:49 anmaster: im a homofag. 01:07:53 AnMaster: err he practically cybersexes oklopol all the time 01:07:58 how can you not notice 01:08:02 OKLOPOL MAKES ME OK 01:08:03 GOD 01:08:07 YOU SEE HOW HE DRESSES 01:08:10 LINGVOJ VEJNAS! 01:08:12 STOP 01:08:19 AnMaster: STOP! STOP WITH THE SEX! 01:08:22 psygnisfive, I also use //s/// to. The more complex sed syntax. 01:08:22 OH GOD ESPERANTO 01:08:23 IT'S ILLEGAL 01:08:30 ehird, it is irrelevant 01:08:32 /s///? 01:08:33 to this 01:08:33 what? 01:08:35 nonsense 01:08:38 psygnisfive, no 01:08:38 get out of here with that 01:08:38 KAJ? 01:08:43 i know anmaster 01:08:47 i typed //s/// but irc is lame 01:09:00 psygnisfive, /foo/s/bar/quux/ would only match on lines where "foo" is found. 01:09:08 but would replace bar with quux on those 01:09:10 interesting 01:09:11 anyway 01:09:13 the first one on those lines 01:09:14 so in slashes 01:09:28 psygnisfive, big question is. Is Slashes TC or not!? 01:09:29 /a/b/c replaces a with b in c 01:09:34 is this correct? 01:09:43 where c is any source code 01:10:09 think so. Don't remember 01:10:22 interesting. 01:10:41 itd be interesting to see if that can be made comprehensibly useful 01:10:54 psygnisfive, TC or not is much more interesting 01:11:01 nah 01:11:06 non-TC languages can still be useful 01:11:08 psygnisfive, what 01:11:16 and also interesting 01:11:29 yes, but tc or not is *more* interesting. 01:11:29 AnMaster: read up on total fp 01:11:31 im willing to bet that slashes is TC 01:11:34 tc is almost entirely needless 01:11:53 There is no page titled "total fp". You can create this page. 01:11:59 No page title matches 01:12:15 Page text matches 01:12:15 1. SMETANA (3,831 bytes) 01:12:15 32: ...t of its steps (''n''! possibilities) - this is a total of ''n''*''n''! possible configurations. 01:12:15 38: If we take ''n'' to be finite, then the total number of possible configurations is bounded as w... 01:12:18 [...] 01:12:26 ehird, link to the relevant one 01:12:31 AnMaster: not an esolang. 01:12:47 Did you mean: total fta Top 2 results shown 01:12:51 AnMaster: http://lambda-the-ultimate.org/node/2003 01:12:59 AnMaster: & http://www.jucs.org/jucs_10_7/total_functional_programming/jucs_10_07_0751_0768_turner.pdf 01:13:23 thanks for the latter link especially 01:13:31 yes 01:13:37 it's a readable paper so you shouldn't have too many issues 01:13:43 ... 01:13:53 ehird, papers are easier to read than web pages in general IMO 01:14:00 readable as in language, foo 01:14:05 ehird, yes... 01:14:07 I meant that 01:14:18 I didn't mean typographically 01:15:21 anmaster: oh. irp. yes. 01:15:21 :P 01:15:58 ehird, very interesting. Will bookmark it for further reading tomorrow when I'm less tired. 01:16:10 ehird, I'd go as far as "extremely interesting" 01:16:29 yes, unfortunately it has issues so I'm not going to be programming with a total FP language any time soon 01:16:30 ehird, anyone tried implementing this in haskell yet? 01:16:36 but further research is very much warranted 01:16:40 AnMaster: there are no implementations afaik 01:16:46 you couldn't build it on top 01:16:56 you know 01:17:10 i like haskell's no-arbitrary-number-of-arguments model 01:17:12 i really do. 01:17:15 ehird, yes I can from the abstract see that there are some stuff that will be hard. I don't yet know if the paper solves the need to operate on codata in it. 01:17:26 but very often you need that for at least a part of the program 01:17:27 no, codata is essential 01:17:43 right 01:17:45 tho i imagine you could actually define some functions in haskell that can take an arbitrary number of args, with a delimiter 01:17:47 hmm 01:17:52 psygnisfive: yes. 01:17:54 with typeclasses 01:18:00 oh i dont even mean like tht 01:18:01 i like haskell's no-arbitrary-number-of-arguments model <-- Erlang has that too. 01:18:02 see Oleg; in fact, always see Oleg 01:18:07 i just mean like 01:18:11 you could define a function 01:18:15 f 1 2 3 END 01:18:19 yes, you do it with typeclasses 01:18:38 yes, im sure you COULD :P 01:18:42 but i mean in a different way 01:18:43 like say 01:18:47 impossible 01:19:20 eh.. i dont know! 01:19:24 give me a second to try 01:19:48 whats the normal conditional in haskell? 01:19:53 if 01:19:59 if .. then .. else .. end ? 01:20:04 no end 01:20:09 ok 01:24:37 hm. no couldnt do what i was thinking of 01:26:30 hm well... 01:36:27 ehird 01:36:39 how do i return a partial function in haskell? 01:36:47 my definition isnt working 01:36:47 :| 01:45:41 grr 01:45:44 so ehird 01:45:57 heres my not-working-in-haskell definition of sum 01:46:42 sum f 0 = f 0 01:46:42 sum f x = sum (\y -> y + f x) 01:47:50 and you call it 01:48:00 sum id 5 4 3 2 1 0 01:49:56 sum id 5 4 3 0 => sum (\y -> y + 5) 4 3 0 => sum (\y -> y + 9) 3 0 => sum (\y -> y + 12) 0 => 12 01:50:44 it works in theory. :p 01:51:40 but i keep getting "too few arguments" errors 02:08:12 -!- pikhq has quit ("Lightning, no modem circuit protection."). 02:19:05 no? 02:19:07 :( 03:06:28 -!- Sgeo has joined. 03:11:35 -!- pikhq has joined. 03:50:23 -!- EgoBot has joined. 03:50:29 !help 03:50:30 Supported commands: bf_txtgen help 1l 2l adjust axo bch befunge bf bf16 bf32 bf8 dimensifuck glass glypho kipple lambda lazyk linguine malbolge pbrain qbf rail rhotor sadol sceql trigger udage01 unlambda whirl 03:50:48 Not complete, based on an entirely new codebase, source available at http://codu.org/projects/egobot/ , enjoy. 03:51:21 !bf +++++++++++++++[>+++>+>+++++++>+++++<<<<-]>>>>---.<----.+++++++..+++.<<-.------------.>>++++++++.--------.+++.------.--------.<<+.>-----. 03:51:42 See, it's so incomplete it doesn't even work when I run it the second time, only when I run it the first time X-P 03:51:49 -!- EgoBot has quit (Remote closed the connection). 03:52:35 -!- EgoBot has joined. 03:52:39 !bf +++++++++++++++[>+++>+>+++++++>+++++<<<<-]>>>>---.<----.+++++++..+++.<<-.------------.>>++++++++.--------.+++.------.--------.<<+.>-----. 03:52:39 Hello, world! 03:52:46 See, it actually does work ;) 03:53:24 lol 03:53:52 !bf +++++++++++++++[>+++>+>+++++++>+++++<<<<-]>>>>---.<----.+++++++..+++.<<-.------------.>>++++++++.--------.+++.------.--------.<<+.>-----. 03:53:52 Hello, world! 03:53:55 Sorry, paranoid :P 03:53:59 :P 03:54:09 is ego bot yours? 03:55:47 Egobot! Oh, how we missed thee and thy crashiness! 03:56:47 psygnisfive: Egobot is Gregor's work, yeah. 03:56:56 Wait, you rewrote Egobot? 03:57:11 grEGOrBOT eh? 03:57:12 !bf_txtgen Egobot 2.0! 03:57:15 138 ++++++++++[>+++++++>+++>+++++>+<<<<-]>-.++++++++++++++++++++++++++++++++++.++++++++.-------------.+++++++++++++.+++++.>++.>.----.++.<+.>>. [560] 03:57:34 psygnisfive: In fact, yes. 03:57:46 Doesn't do PMs any more? 03:57:55 !bf ++++++++++[>+++++++>+++>+++++>+<<<<-]>-.++++++++++++++++++++++++++++++++++.++++++++.-------------.+++++++++++++.+++++.>++.>.----.++.<+.>>. 03:57:55 Egobot 2.0! 03:57:56 It should. 03:58:03 Well, it doesn't. 03:58:25 !bf ++++++++++[>++++++++++>+>><<<<-]>---.>.<+. 03:58:26 a 03:58:33 Indeed :P 03:59:09 Hmm. It's got Dimensifuck support, too? Has it always, or am I just slow? 03:59:38 Erm. 03:59:42 Rephrase that. 03:59:45 I just did a quick sweep over the ones in old EgoBot, so probably it doesn't. 03:59:56 Which is to say, it may be in the menu but not actually working ;) 04:00:12 !dimensifuck 0+v.+^ 04:00:47 !test 04:00:47 Nick: GregorR 04:00:49 Hm 04:01:04 Guess not. That ought to be spamming stuff. 04:01:23 Does it ever spam a newline? :P 04:01:40 That is in [0..255], isn't it? 04:02:02 The same code in Brainfuck is +[.+], BTW. 04:02:37 Ah 04:02:43 !bf ++++++++++[>++++++++++>+>><<<<-]>---.>.<+. 04:02:43 a 04:02:50 Damnsicles ... 04:02:53 !dimensifuck ++++++++++. 04:03:17 Damnnabbit. 04:03:17 malbolge? really? 04:03:49 !bf ++++++++++[>++++++++++>+>><<<<-]>---.>.<+. 04:03:49 a 04:04:10 while read LN is valid bash, right? X_X 04:04:24 don't think so 04:04:33 you need a do; done 04:04:53 OK, obviously the rest of the loop comes after that X_X 04:04:57 I meant as a loop opener 04:05:28 !bf ++++++++++[>++++++++++>+>><<<<-]>---.>.<+. 04:05:28 a 04:06:12 !bf +++++[>++++++<]>++. 04:08:45 Oh, found the issue. Weird one. 04:09:04 Oh? 04:09:38 !bf +++++[>++>+++>++++[>++++>++++<<-]<-]>>-->>----.>.<--------.>-. 04:10:06 I said found, not fixed :P 04:10:24 OK, /now/ it's found and fixed. 04:10:35 Apparently read LN doesn't work right if there is no newline at the end of input. 04:10:36 (weird) 04:12:05 !bf ++++++++++[>++++++++++>+>><<<<-]>---.>.<+. 04:12:05 a 04:12:14 !bf +++++[>++>+++>++++[>++++>++++<<-]<<<-]>.>--.>>+.++++.>-------.<-. 04:12:15 -!- EgoBot has quit. 04:12:17 hahaha 04:12:18 -!- EgoBot has joined. 04:12:37 ............ wtf? 04:12:51 !bf +++++[>++>+++>++++[>++++>++++<<-]<<<-]>.>--.>>+.++++.>-------.<-. 04:12:53 -!- EgoBot has quit (Client Quit). 04:12:55 -!- EgoBot has joined. 04:13:04 It should be nigh-on impossible to cause that ... 04:13:12 hint: it's not a crash 04:13:23 Presumably that outputs \r\nQUIT or something. 04:13:26 But it checks for that. 04:14:19 !bf +++++[>++>+++>++++++>++++[>++++>++++<<-]<<<<-]>.>--.>>>--.-----.------.++++++++.<<++.>>---------.>+++++.<+++++. 04:15:42 What, exactly, *does* it print? 04:15:47 hang on 04:16:20 pikhq: \r\nQUIT 04:16:35 pikhq: But my script reads the output by-line, so that makes no sense X_X 04:17:22 Huh. 04:17:28 !bf +++++[>++>+++>++++++>++++[>++++>++++<<-]<<<<-]>.>--.>>>--.-----.------.++++++++.<<++.>>>+++.<------.--.>++.---.<++++++.>++.+++++.<-------.>----.<+++++. 04:17:29 -!- EgoBot has changed nick to SECURITYBUG. 04:17:38 :P 04:17:58 It's like SQL injection with more fun! 04:18:07 And less actual potential harm to anything. 04:18:21 !bf +++++[>++>+++>++++[>++++>++++<<-]<<<-]>.>--.>>+.++++.>-------.<-. 04:18:22 -!- SECURITYBUG has quit. 04:18:25 :) 04:18:33 hmm... now that's odd 04:18:49 shouldn't it have rejoined by now? 04:19:11 -!- EgoBot has joined. 04:19:15 there we go 04:19:19 !bf_txtgen \r\nQUIT HALDO 04:19:22 155 ++++++++++[>+++>+++++++++>+++++++++++>+<<<<-]>>++.>++++.<.>----.-----------------------------.<-------.>--------.<-.<++.>>-.-------.<--------.>+++.<+++.>>. [902] 04:19:27 Well that's clearly not going to work. 04:19:33 !bf +++++[>++>+++>++++++>++++[>++++>++++<<-]<<<<-]>.>--.>>>--.-----.------.++++++++.<<++.>>---------.>+++++.<+++++. 04:19:47 fixed? aw 04:19:52 :) 04:20:05 !bf +++++[>++>+++>++++[>++++>++++<<-]<<<-]>.>--.>>+.++++.>-------.<-. 04:20:06 -!- EgoBot has quit (Client Quit). 04:20:09 -!- EgoBot has joined. 04:20:10 Not fixed. 04:20:11 What. 04:20:11 The 04:20:12 FUCK 04:20:28 There is no logic by which this should happen. 04:20:43 Does it have infinite loop protection? 04:21:08 In terms of slowing the program down and limiting its CPU and clock time, yes. 04:21:10 !bf ++++++++++[>+++>+++++++++>+++++++++++>+<<<<-]>>++.>++++.<.>----.-----------------------------.<-------.>--------.<-.<++.>>-.-------.<--------.>+++.<+++.>>. 04:21:11 rnQUIT HALDO 04:21:44 Hrm. 04:22:44 !bf +++++[>++>+++>++++[>++++>++++<<-]<<<-]>.>--.>>+.++++.>-------.<-. 04:22:44 Down for maintenance! 04:22:46 -!- EgoBot has quit (Client Quit). 04:22:49 -!- EgoBot has joined. 04:22:50 There, fixed ;) 04:23:22 -!- bsmntbombdood has joined. 04:25:49 What the heck, it happened with your BF program, but not mine ... what weird thing is yours doing that mine isn't? 04:26:26 OH 04:26:30 I see what happened! 04:26:33 And I have a fix. 04:26:38 -!- EgoBot has quit (Remote closed the connection). 04:27:53 aw 04:28:07 also question, does anyone have a BF debugger? 04:28:07 Hmm. We've got EgoBot. Now I guess we're going back in time? 04:28:22 Should be only a matter of time before we're talking about LOLCODE again. :( 04:28:24 -!- adu has joined. 04:29:23 coppro: BTW, it's supposed to be \r\n, not \n\r ;) 04:29:30 oh 04:29:39 that might be why it worked lol 04:29:40 coppro: Also btw, apparently FreeNode accepts \r as a line delimiter. Who knew. 04:30:00 That /is/ why it worked. I didn't scrub for random \r's, because why would I? But apparently FreeNode thinks \r is a line delimiter by itself :P 04:30:27 hi 04:30:37 -!- EgoBot has joined. 04:30:39 'lo 04:30:55 I'll be damned. 04:31:17 !bf +++++[>++>+++>++++[>++++>++++<<-]<<<-]>.>--.>>+.++++.>-------.<-. 04:31:27 Voila, nothing :P 04:31:36 'Lo. 04:32:00 lo? 04:32:11 is that bf? 04:33:08 how would you write "true" and "false" in bf? 04:33:08 adu: That was a piece of BF that SOME shitty person used to break my bot :P 04:33:17 o 04:33:18 As in the string "true"? 04:33:34 !bf_txtgen true 04:33:34 Or the program? 04:33:38 62 +++++++++++++[>+++++++++>++++++++>+><<<<-]>-.--.+++.>---.>---. [218] 04:33:39 as in a 0 return status and a nonzero return status 04:33:56 Ah, that you can't do, or at least can't do consistently, as there's no well-defined way to exit with a particular status code. 04:34:04 :( 04:34:15 Alas, another nail in the BF-UNIX coffin :P 04:34:36 Unless you get PSOX out. 04:34:45 what if the 0th cell was the return status and >=1 are stdout? 04:35:05 adu: Then that wouldn't be BF :) 04:35:07 (one of these days, I'm going to make an implementation *and* a PEBBLE library for that...) 04:35:10 adu, PSOX can do it 04:35:15 Like pikhq said 04:35:37 adu: Then "true" would be the null string and "false" would be "+". 04:35:39 (I have PSOX on highlight. If you don't want me noticing, censor it. MUAHAHAHA) 04:35:59 pikhq: that makes sense :) 04:36:39 pikhq: what if the (-1)th cell was return status? 04:36:46 adu, PSOX can return statuses 04:36:57 whats PSOX? 04:37:21 It's a layer that goes between an esolang interpreter and stdio 04:37:21 False would be "<+", then... 04:37:52 !info 04:37:54 Sgeo: i see 04:38:00 D-8 04:38:18 !info 04:38:18 EgoBot is a bot for running programs in esoteric programming languages. If you'd like to add support for your language to EgoBot, get the mercurial source at https://codu.org/projects/egobot/hg/ 04:38:24 Well that was useful :P 04:38:57 http://trac2.assembla.com/psox/browser/trunk 04:39:36 Let's see if I can write true 04:39:42 Now, I expect to be getting some emails with hg bundles for the numerous languages that have been invented since I last touched EgoBot. If I don't see them, I WILL BE ANGRY :P 04:40:12 PEBBLE it is. 04:40:46 .+++++++.>+.<+++.>-..<.,,,[-].++.-.-.++++++++++. 04:40:48 There. 04:40:53 That should be TRUE in PSOX 04:41:17 .+++++++.>+.<+++.>-..<.,,,[-].++.-..+++++++++. 04:41:22 is FALSE 04:42:09 adu, feel free to assassinate me 04:43:25 * adu gives Sgeo a hug :) 04:43:37 Sgeo: is there a link to the PSOX docs? 04:44:03 coppro, http://trac2.assembla.com/psox/browser/trunk/spec 04:44:18 thanks 04:45:57 yw 04:49:51 pikhq: Any comments on multibot? :P 04:50:06 GregorR: Multibot? 04:50:11 Uh... Egobot, I assume? 04:50:25 pikhq: Oh, I assumed you had checked out egobot's source and so seen that it's just a set of scripts for multibot :P 04:50:43 Egobot is one of our finest traditions; glad to see it return. :p 04:50:45 Ah. 04:50:53 multibot is another "why do I keep writing IRC bots" IRC bot. 04:51:19 It's all scripted. Whenver it receives a message, it looks for an appropriate script to run by the command, channel, etc. 04:52:13 (Which is why it was so easy to reimplement EgoBot :P ) 04:56:24 -!- ab5tract has joined. 05:05:53 -!- GreaseMonkey has joined. 05:05:56 -!- adu has quit. 05:12:39 -!- psygnisfive has quit (Read error: 110 (Connection timed out)). 05:14:29 -!- adu has joined. 05:27:28 -!- sebbu2 has joined. 05:44:12 -!- sebbu has quit (Read error: 113 (No route to host)). 05:44:13 -!- sebbu2 has changed nick to sebbu. 06:03:43 !help 06:03:44 Supported commands: bf_txtgen help info 1l 2l adjust axo bch befunge bf bf16 bf32 bf8 dimensifuck glass glypho kipple lambda lazyk linguine malbolge pbrain qbf rail rhotor sadol sceql trigger udage01 unlambda whirl 06:03:49 Hey look, it's still running :P 06:09:23 -!- Sgeo has quit ("Leaving"). 06:22:15 help me rhonda 06:22:19 help help me rhonda 06:24:39 -!- pikhq has quit ("FOOBAR"). 06:37:41 !bf_txtgen help help me rhonda 06:37:45 141 +++++++++++++++[>++>+++++++>+++++++>+<<<<-]>>-.---.>+++.++++.<<++.>+++.---.+++++++.>.<<.>+.--------.<.>>++.<+++.>---.-.----------.---.>-----. [423] 06:37:53 !bf +++++++++++++++[>++>+++++++>+++++++>+<<<<-]>>-.---.>+++.++++.<<++.>+++.---.+++++++.>.<<.>+.--------.<.>>++.<+++.>---.-.----------.---.>-----. 06:37:53 help help me rhonda 06:42:22 !bf_txtgen aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa 06:42:27 196 ++++++++++++++[>+++++++>+++++++>+++++++>+++++++<<<<-]>-....>-.>-................>-.....<...<.....>...>......---------------------------------------------------------------------------------------. [358] 06:42:37 Hah 06:42:41 Goooo bf_txtgen :P 06:42:43 wow that is retarded 06:42:52 It uses a genetic algorithm, blame calamari. 06:43:10 (Also I cut it off at 1000 generations) 06:43:13 i won't be satisfied until the 3-bit encoding of brainfuck instructions is shorter than the original string 07:00:07 bf_txtgen also has a user-defined number of cells it always uses; by default 4. 07:05:39 lame 07:06:08 it should be allowed to use at least O(n) cells 07:11:04 -!- adu has quit. 07:16:35 -!- FireFly has joined. 07:16:49 fizzie: If you want tuned text generation, you download textgen.java yourself, !bf_txtgen is just for testing and fun :P 07:18:13 Yes, but it should tune-a-fish itself automatically; that's what computers are for, after all. 07:20:11 fizzie: I await your hg bundle with fixes to make it brilliant :P 07:23:53 Eh; it should also rewrite itself, that's also what computers are for. 07:26:33 Admittedly it would feel pretty silly when a brainfuck text generator took over the world. 07:35:51 -!- ab5tract has quit. 07:55:07 -!- FireFly has quit ("Later"). 07:59:59 -!- clog has quit (ended). 08:00:00 -!- clog has joined. 08:51:27 -!- tombom has joined. 09:04:08 -!- Slereah has joined. 09:58:13 -!- lament has quit (Read error: 110 (Connection timed out)). 10:04:37 -!- ais523 has joined. 10:48:31 -!- GreaseMonkey has quit ("YES -> thor-ainor.it <- THIS IS *DELICIOUS*!"). 11:24:54 hi ais523 11:25:13 hi 11:25:25 sorry, can't talk much, I'm in an assessed project open day atm 11:25:39 even though it's worth hardly any marks, basically the mark on this just determines whether my final mark rounds up or down 11:25:49 right. 11:57:04 -!- KingOfKarlsruhe has joined. 13:03:10 Re. Ulrich Drepper, hadn't seen this one before: http://sourceware.org/bugzilla/show_bug.cgi?id=956 13:05:46 -!- ais523 has quit (Read error: 110 (Connection timed out)). 13:07:41 -!- KingOfKarlsruhe has quit (Read error: 113 (No route to host)). 13:12:55 Deewiant, ouch 13:13:32 "Live with it." :-D 13:22:37 -!- ais523 has joined. 13:23:17 who was the guy who constantly closed a bug where the rpm database could get destroyed if you did some reasonably routine stuff as "WONTFIX" and eventually got fired over it 13:38:35 -!- ais523 has quit (Read error: 104 (Connection reset by peer)). 14:22:13 -!- ais523 has joined. 14:33:51 Deewiant: Wow, that's defiantly useless :P 14:33:59 Deewiant: I especially like "there is no bug" 14:44:35 * GregorR imagines Ulrich Drepper doing the Jedi mind trick "there is no bug!" 14:46:31 You must eat a bug to prove your loyalty 14:47:15 -!- MizardX has quit ("bbl"). 14:54:07 -!- Judofyr has joined. 15:11:58 http://blogs.msdn.com/vbteam/archive/2009/04/01/using-negative-sleeps-to-improve-responsiveness-in-vb-web-apps.aspx 15:12:06 it seems Microsoft had an April Fool's joke too 15:26:40 -!- ais523 has quit (Remote closed the connection). 15:29:37 -!- Slereah has quit ("Leaving"). 15:32:25 -!- Judofyr has quit (Remote closed the connection). 15:36:25 -!- coppro has quit (Remote closed the connection). 15:46:41 -!- BeholdMyGlory has joined. 15:50:30 -!- FireyFly has joined. 15:57:04 -!- FireyFly has changed nick to FireFly. 16:01:35 -!- Gracenotes has quit (Remote closed the connection). 16:35:15 -!- oerjan has joined. 16:38:45 !unlambda ````.Y.a.y.!i 16:38:46 Yay! 16:45:18 17:46:42 sum f 0 = f 0 16:45:18 17:46:42 sum f x = sum (\y -> y + f x) 16:46:08 not well typed 16:48:20 problem is, without type classes there is no way to assign a type to that which doesn't require sum f x to have the same type as sum f 16:48:49 which means they both must take the same number of arguments => infinity 16:48:57 (not allowed) 16:49:48 fix (\f s xs -> case xs of [] -> s; y:ys -> f (s+y) ys) 0 16:50:19 what? 16:50:37 That just came to mind 16:51:04 but psygnisfive doesn't use any lists 16:51:22 I know, I wonder what that sum function is summing :-) 16:51:43 17:49:56 sum id 5 4 3 0 => sum (\y -> y + 5) 4 3 0 => sum (\y -> y + 9) 3 0 => sum (\y -> y + 12) 0 => 12 16:51:51 Ah, its arguments 16:52:09 But yeah, that can't work without a type class 16:52:25 Hmm, or even with one? 16:53:16 it cannot work for every Num in any case, but specific types should be doable... 16:53:30 -!- Sgeo has joined. 16:54:07 Oh right, it takes two parameters, of course it can work 16:54:12 (because you cannot dispatch a type class on another one) 16:54:20 Yep 16:55:14 it might very well work with just the numbers too, printf manages to have only the arguments 16:55:46 but i recall it goes via a helper function that uses an accumulating list 16:56:02 Possibly, I haven't looked at it in much detail 16:56:28 well naturally, you could define sum = sum' id 16:56:36 Yep 16:57:38 (if you are careful about how you do it, that exact phrasing would probably trigger the monomorphism restriction) 16:58:04 Just give it a type signature :-P 17:05:58 20:28:22 Should be only a matter of time before we're talking about LOLCODE again. :( 17:06:11 i thought LOLCODE was after EgoBot vanished... 17:06:18 maybe not entirely. 17:10:00 23:26:33 Admittedly it would feel pretty silly when a brainfuck text generator took over the world. 17:10:24 whatever AI takes over the world _will_ be silly. 17:13:19 it will put a smile on your face. 17:13:29 what's left of it. 17:17:21 -!- Judofyr has joined. 17:25:02 -!- Sgeo_ has joined. 17:28:02 psygnisfive: oh, and another thing about that sum is that that 0 has the exact same type as the other arguments, so it won't help decide when to stop taking arguments at all - haskell is not dependently typed so _values_ don't influence _types_ 17:32:04 -!- FireFly has quit ("Later"). 17:38:03 -!- Sgeo_ has quit (Read error: 104 (Connection reset by peer)). 17:39:58 -!- Sgeo has quit (Read error: 110 (Connection timed out)). 17:41:11 -!- Gracenotes has joined. 17:52:05 -!- MizardX has joined. 17:52:10 -!- tombom has quit ("Peace and Protection 4.22.2"). 18:05:58 -!- pikhq has joined. 18:07:02 -!- MizardX has quit (Read error: 54 (Connection reset by peer)). 18:07:13 -!- MizardX has joined. 18:10:55 -!- Dewio has joined. 18:23:12 -!- Dewi has quit (Read error: 101 (Network is unreachable)). 18:43:04 hi oerjan 18:43:23 hi AnMaster 18:50:39 since when is egobot back up 18:50:50 since yesterday 18:51:03 it's apparently a new implementation 18:52:48 oh 18:53:02 does the new one do the suspend process crazy stuff too? 18:53:16 hm good question 18:53:17 !help 18:53:18 Supported commands: bf_txtgen help info 1l 2l adjust axo bch befunge bf bf16 bf32 bf8 dimensifuck glass glypho kipple lambda lazyk linguine malbolge pbrain qbf rail rhotor sadol sceql trigger udage01 unlambda whirl 18:53:28 !ps 18:53:37 !befunge 'a,@ 18:53:55 !befunge 'a,@,"b" 18:53:58 meh? 18:54:06 !befunge "a",@ 18:54:06 a 18:54:08 !daemon 18:54:20 !befunge <"a",@,"B"< 18:54:21 B 18:54:21 No daemons. :( 18:54:25 :( 18:54:30 !befunge <'a,@,"B"< 18:54:30 B 18:54:32 hm 18:54:35 !befunge 'a,@,"B"< 18:54:38 * oerjan hugs fungot 18:54:38 so 18:54:39 oerjan: " it was a golden crown." ( bruno had very fnord provided one, which fitted him exactly, by cutting out the centre of the crown. 18:54:48 clearly befunge-93 18:54:48 At least _you_ are user-extendable 18:54:49 Gregor a while back wrote a different bot. EgoBot is now a slight bit of patching on that. 18:54:51 I think. 18:54:57 !help 18:54:58 Supported commands: bf_txtgen help info 1l 2l adjust axo bch befunge bf bf16 bf32 bf8 dimensifuck glass glypho kipple lambda lazyk linguine malbolge pbrain qbf rail rhotor sadol sceql trigger udage01 unlambda whirl 18:55:00 Erm. 18:55:03 !info 18:55:03 EgoBot is a bot for running programs in esoteric programming languages. If you'd like to add support for your language to EgoBot, check out the source via mercurial at https://codu.org/projects/egobot/hg/ 18:55:12 Go forth, patch EgoBot! 18:56:34 hm that fungot quote looks like it's one piece 18:56:35 oerjan: once more: and, till then, i shalt say, in about three seconds. but it was also a worn and sad one, and told a tale ( or so i seemed to see a little further. ' principal fnord there are any, have the fnord so fnord a series of somersaults. however, it was, spread over the hearth-rug, all fnord and fnord 18:56:35 !befunge #@ #, "a"$# r "b",@ 18:56:36 b 18:56:39 hm 18:56:52 so it doesn't reflect on 93 instructions 18:56:54 err 18:56:56 98 ones 18:57:02 yet doesn't implement 98? 18:57:07 !befunge # <1.@.2 18:57:07 2 18:57:07 this sounds like crap 18:57:18 Deewiant, it did wrap around, I checked. 18:57:23 above 18:57:27 AnMaster: I was checking # over space. 18:57:28 EgoBot is wrapping a different Befunge interpreter. 18:57:30 ah 18:57:39 It'd be easy to get it to wrap a better one. 18:57:45 AnMaster: And also # over edge, accidentally. 18:57:52 hah 18:58:00 UNDEF: # over edge does not skip rightmost column in file 18:58:02 Deewiant, well I suspect it is 93 not 98 18:58:13 Yes, if it's 93 that's expected. 18:59:04 !befunge #@ #, #, "a"$# q "b",,@ 18:59:05 b 18:59:05 !befunge #2#.#@>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>1.@ 18:59:14 !befunge #2#.#@>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>1.@ 18:59:18 :-( 18:59:19 err 18:59:20 what 18:59:27 !befunge #2#.#@>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> 18:59:37 BAD: something very FUBAR 18:59:50 !befunge #2#.#@1.@ 18:59:51 1 18:59:58 !befunge #2#.#@>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>1.@ 18:59:58 1 19:00:08 Not that 19:00:14 !befunge #2#.#@>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>1.@ 19:00:15 1 19:00:16 Oh, oops 19:00:17 My bad 19:00:18 !befunge #2#.#@>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>1.@ 19:00:18 1 19:00:22 Deewiant, wha 19:00:23 what* 19:00:30 !befunge #2#.#@>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>># 1.@ 19:00:46 .. 19:00:47 UNDEF: # over column 80 does not skip column 0 19:00:50 !befunge #2#.#@>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>># 1.@ 19:00:50 1 19:00:52 Or loading of data is broken 19:00:53 !befunge #2#.#@>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>># 1.@ 19:00:58 yes 19:01:02 something happens there 19:01:15 AnMaster: In mine, the # is 80th 19:01:44 https://codu.org/projects/egobot/hg/index.cgi/file/50ebc27d8665/multibot_cmds/interps/befunge/bef.c#l1 19:01:45 My guess it loads only 80x25 as expected, and then the # over edge hits column 0. 19:01:49 +is 19:01:54 3 bef.c - The Original Befunge-93 Interpreter/Debugger in ANSI C 19:01:54 4 v2.21 Sep 20 2004 Chris Pressey, Cat's-Eye Technologies 19:02:40 With b97 support! 19:02:49 what?! 19:03:09 I think that's the only extant piece of code that can do anything with Befunge-97 19:03:17 I may remember incorrectly but I think so. 19:03:23 int use_b97directives = 0; /* flag : use b97-esque directives? */ 19:03:28 right 19:03:38 could we reverse engineer bf-97 from it 19:03:49 Parts of it, certainly. 19:04:00 I don't think it's worth the effort :-P 19:04:21 if (use_b97directives && (x == 0) && ((cur == dc) || ((accept_pound) && (cur == '#')))) 19:04:22 hm 19:04:54 Yeah, Befunge-97 had source directives like #foo at the start 19:05:02 Or something like that anyway 19:05:06 Can't remember what kinds of things they did 19:05:37 Befunge-97 also supports a 19:05:37 broad set of interpreter directives (very roughly analogous to assembler 19:05:37 directives or C preprocessor directives) to help the programmer take 19:05:38 advantage of the larger grid. 19:05:48 haha 19:06:21 Deewiant, it is very confusing the bit that parses these directives 19:06:36 Lines starting with =, sorry, not #. 19:06:53 Deewiant, # is involved too 19:06:54 it seems 19:07:12 Evidently; I don't know / can't recall what that's about. 19:08:08 http://www.bedroomlan.org/hacks/signature-befunge has an example of a Befunge-97 proggy 19:09:02 To me that if looks like it's saying; if at start of line, and we either have the directive-introducing character ("dc") or we have a # and the accept-pound setting is on, then do something. 19:09:17 So presumably there's a setting that makes it support #foo-directives too. 19:10:17 http://catseye.tc/projects/befunge93/eg/pi2.bf is Bef97; may be -98 as well. 19:10:26 but 19:10:26 (No directives.) 19:10:29 what about an actual spec 19:10:39 Ask Chris if he still has one. 19:10:49 I did, never got a reply 19:10:59 Somebody on the Language::Befunge mailing list has one he translated to French but lost the original 19:11:57 Deewiant, can't we get Slereah (spelling?) to back-translate it 19:12:04 if the French one still exists 19:12:19 I'm sure we could get many to back-translate it if you want it and he still has it :-P 19:12:53 I found a reference to Befunge-96, seems it had concurrency. Can't find anything else about it. 19:13:05 Yep, -96 is even more obscure than -97. 19:13:49 Deewiant, hm: http://www.mail-archive.com/perl6-all%40perl.org/msg42839.html 19:14:04 Yep, that's the guy I was talking about. 19:14:27 'I have translation of Befunge-93 spec with appendix D which mentions directive '='. I translated it myself in 1998, but lost original :-)' 19:14:33 what the hell is .pasm 19:14:45 Parrot source 19:14:50 oh ok 19:15:09 http://www.bedroomlan.org/hacks/soup is another 19:17:25 I recall having seen =I and =L somewhere but I don't know where. Oh well. 19:17:36 Deewiant, minifunge? 19:17:48 Heh, that could be. 19:18:33 "In Befunge-93, the data types supported were 32-bit integers and bytes, but Befunge-97 supports only machine-width integers. Befunge-97 also supports a broad set of interpreter directives (very roughly analogous to assembler directives or C preprocessor directives) to help the programmer take advantage of the larger grid. 19:18:33 Several free implementations of both old Befunge-93 and newer Befunge-97 exist, some explicitly for PCs and at least one that is highly portable (based on Perl). Fairly good documentation and some example programs are available on the web." 19:18:34 hm 19:18:40 no download links though 19:18:44 http://cgibin.erols.com/ziring/cgi-bin/cep/cep.pl?_key=Befunge 19:19:36 http://search.cpan.org/~jquelin/Language-Befunge-2.06/Befunge/IP.pm says "Language::Befunge::IP - an Instruction Pointer for a Befunge-97 program." but seems to be all 98 to me 19:21:41 https://codu.org/projects/egobot/hg/index.cgi/file/50ebc27d8665/multibot_cmds/PRIVMSG/tr_21.cmd#l1 19:21:46 AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA 19:22:13 What :-P 19:22:46 it is written in shell and C. I would never have thought anyone else was that crazy 19:22:54 I mean of course I have written irc bots in shell 19:23:01 (envbot being the most advanced one) 19:23:18 -!- KingOfKarlsruhe has joined. 19:23:21 Remember, Gregor is insane. 19:23:23 ;) 19:23:35 meh. Aren't we all more or less. 19:23:39 Since we are here. 19:23:43 True enough. 19:25:16 err 19:25:35 https://codu.org/projects/egobot/hg/index.cgi/file/50ebc27d8665/README#l1 19:25:38 seems like not 19:25:40 anyway 19:25:42 he could use cfunge quite easily for befunge-98 in it. It has a sandbox mode. 19:26:45 It's supposed to be the IRC equivalent of CGI :P 19:27:03 GregorR, do you sandbox the interpreters at some OS level as well or 19:27:41 EgoBot runs in a chroot, but the interps aren't really sandboxed from the bot itself. 19:28:00 (They have some ulimits, but that's nothing really P ) 19:28:01 *:P 19:28:04 GregorR, if it is running on a POSIX system you could drop a cfunge binary in for funge-98 (if you want a sandboxed one). Or CCBI I guess if you want to rely on chroot. 19:28:13 GregorR, what ulimits. 19:28:34 cfunge allocates optimistically, expecting a large program. 19:28:51 Few files, few forks, small files. It allows some 32MB of memory, so it shouldn't be an issue. 19:29:00 It's running on a Xen VM running... Linux, I think? 19:29:04 GregorR, as in ulimit -v or ulimit -m 19:29:59 Y'know what, I set it as ulimit -m, but you've just reminded me that that's wrong :P 19:30:01 GregorR, cfunge needs ulimit -v $(( 1024 * 23 + 512 )) to even get as far as mmap()ing the input file. 19:30:17 yes it mmap()s the input file for fast loading. 19:30:26 that is on x86_64 though 19:30:48 Server is on x86_64 19:31:01 I guess a build that disable NCRS and TERM would need less. Since it wouldn't need to load libncurses.so 19:31:09 What am I using now for befunge? :P 19:31:25 and if you are ok using gettimeofday() instead of clock_gettime() you could drop librt and libpthreads too 19:31:43 GregorR, a 93 interpreter. 19:32:10 which cfunge doesn't aim to provide. (it does have a semi-working compatiblity mode for 93, I plan to drop it as it doesn't fully support 93 with it) 19:32:10 Well, I can't look at that now, or even tonight. If somebody updates it and sends me a bundle I can import that easy, otherwise it'll be tomorrow at least. 19:32:24 GregorR, hm what about build system 19:32:28 for the programs 19:32:32 cfunge builds with cmake. 19:32:48 I can install cmake in the chroot, that's not a big deal. 19:32:58 I could add a MINIMAL switch which disables NCRS and TERM (would be useless for irc usage anyway) 19:33:19 I guess. 19:33:44 Don't worry about it. 19:33:48 CCBI has a fully working 93 compat mode as of 2-3 hours ago. 19:33:53 Deewiant, ! 19:33:57 but 19:34:00 does it have a sandbox mode 19:34:05 Not yet. 19:34:09 right 19:34:22 GregorR, also CCBI needs even more ram than cfunge in my tests. 19:34:32 93 compat mode is a sandbox mode though, right? ;-P 19:34:38 Deewiant, hah! 19:34:43 but 98 + sandbox 19:34:43 Maybe it's a bit too sandboxy 19:34:50 It steals half the sand 19:35:42 * GregorR awaits an hg bundle :P 19:35:48 Feel free to duke it out yourselves. 19:35:56 I don't really care that much :-P 19:36:27 GregorR, I don't use hg enough to know that. But I can provide a good old patch. I'm restructuring build system to help this work smoothly in your code. 19:36:33 One advantage of CCBI2 currently is that it has Unefunge, though 19:36:37 Which lends itself to IRC 19:37:18 GregorR, what about multi-line input for the interpreters 19:37:36 Just run Unefunge-98 19:37:42 AnMaster: Patch is fine. If you do want to make a bundle, it's as simple as: hg add ; hg commit -m 'Your commit message'; hg bundle > bundle.file 19:37:43 AnMaster: hg pull to get stuff from the hg repository. Make your changes, then hg bundle. 19:37:48 Erm. 19:37:52 GregorR, I see 19:37:53 What GregorR said. 19:37:53 :p 19:38:04 I'm more used to bzr indeed 19:38:07 AnMaster: It can interp URLs. !befunge http://foo 19:38:30 Wahh, that reminds me that I never finished the HTTP client in fungot. 19:38:30 fizzie: ' that was mean!' alice cried eagerly. ' you'll never guess! _i_ couldn't.' the leg of mutton before alice, who looked at it with great curiosity. 19:38:37 Yes, it is indeed mean. 19:38:44 GregorR, right. 19:51:18 * oerjan finds himself trapped inside tvtropes again 19:51:23 soon done (on the cfunge build system side) 19:53:24 interesting, cfunge with 23 MB RAM: Refuses to start. With 24 MB RAM: Can run all of mycology just fine. 19:54:25 without ncurses 20 MB works fine 19:54:43 $ (ulimit -v $(( 1024 * 19 )); ./cfunge ../../trunk/mycology/mycology.b98 ) 19:54:43 FATAL: Failed to process file "../../trunk/mycology/mycology.b98": Invalid argument 19:54:45 what 19:54:56 the file exists. Just odd failure message at 19 19:55:01 errno messup I guess. 19:55:02 :( 19:55:06 do you check file size in advance, or is the initial allocation entirely constant? 19:55:08 Hooray for errno 19:55:45 oerjan, well mycology fits into the static area of funge space. Which is the most commonly used area, located close to 0,0 19:55:54 so in advance in that case 19:56:27 Mycology fits because you made the array barely large enough that it does :-P 19:56:29 oh but then the static area has constant size 19:56:30 Deewiant, it is a pain to debug too, since gdb can't do it. and calling setrlimit in main with those values doesn't do it either 19:56:35 oerjan, yes it does. 19:56:46 Deewiant, Quite a bit larger. 19:56:52 512*1024 19:56:55 AnMaster: Rounded up to the next power of 2. 19:56:57 yes 19:56:59 err 19:57:01 no 19:57:07 Ah, 512, I thought it was 256 19:57:12 mycology would fit in 256*1024 19:57:16 indeed 19:57:21 But still :-P 19:57:36 Deewiant, well, anything larger made memory usage absurdly high 19:57:39 I suspect that if Mycology were 400 lines long that'd be 512*512 or 256*512 19:58:20 Deewiant, probably, Since it is the largest app I know of. You can adjust the values in funge-space.c if you need a larger program 19:58:22 subject to some rules 19:58:30 such as it must be a multiple of 16 bytes. 19:58:39 when you consider sizeof() 19:59:32 hm 19:59:35 !help 19:59:36 Supported commands: bf_txtgen help info 1l 2l adjust axo bch befunge bf bf16 bf32 bf8 dimensifuck glass glypho kipple lambda lazyk linguine malbolge pbrain qbf rail rhotor sadol sceql trigger udage01 unlambda whirl 20:00:00 GregorR, no perl? I should add a possibility to turn the PERL fingerprint off then I guess. 20:00:18 AnMaster: Perl isn't esoteric. 20:00:21 true 20:00:32 so the fingerprint would be useless in that chroot 20:00:36 unless it has perl 20:03:36 maybe 20:08:45 -!- tombom has joined. 20:10:00 without clock_gettime() (which on linux need librt, which in turn needs libpthreads) cfunge manages just fine with ulimit -v 15 even for mycology 20:10:01 Deewiant, ^ 20:10:35 You're well on the way to writing an embedded Befunge-98 interp :-P 20:12:12 -!- FireyFly has joined. 20:12:23 * oerjan swats FireyFly -----### 20:12:24 Deewiant, haha 20:12:30 oerjan, why 20:12:41 Bugs must be swatted 20:12:41 too long since last time 20:14:41 -!- FireyFly has changed nick to FireFly. 20:15:05 Deewiant, anyway: I already had checks for ncurses (optional dependency) and clock_gettime() (fallback on gettimeofday()). Just not config options to disable them even if they were supported by the system. 20:20:52 hm 20:20:58 maybe a libm free variant Deewiant 20:21:14 Embedded++ 20:21:20 core doesn't (as far as I remember) need libm anywhere. Some fingerprints do. 20:21:26 Deewiant, agreed. 20:25:37 AnMaster: Uhhh, wtf? It should not be possible for your interp to read or write arbitrary files, or run code that isn't in the isn't-this-harmless language of befunge :P 20:25:58 GregorR, I do have such an option: -S for sandbox 20:26:15 GregorR, however, it doesn't disable ncurses, since that is console IO only. 20:26:22 but now I have such an option. 20:26:31 Probably :P 20:27:11 GregorR, -S disables amongst other things: i and o, hides most (but not all) environment variables in y, fingerprints like SOCK, FILE and DIRF, ...) 20:31:29 Befunge (or rather cfunge) is terrifyingly powerful :P 20:33:12 Bah, cfunge supports hardly any fingerprints :-P 20:36:19 10:54:49 Gregor a while back wrote a different bot. EgoBot is now a slight bit of patching on that. 20:36:21 Boo 20:39:08 ? 20:39:46 I liked the old, shitty code. 20:40:00 Then you'll love the new, shitty code. 20:40:02 You have a knack for shitty 20:40:15 GregorR: It's too modular and not process-suspending enough 20:40:17 GregorR, cfunge is powerful but all the dangerous bits can be disabled. 20:40:22 Deewiant: I am to shit. 20:40:24 *aim 20:40:32 I thought *too 20:40:53 Deewiant: Well yeah, I aim too, shit. 20:41:02 without libm dependency cfunge manages to run mycology in 13 MB 20:41:05 * ehird downloads a debian installer cd 20:41:07 of ram 20:41:19 Fuck yeah Debian. 20:41:23 Or sth 20:41:32 People use Debian installer CDs? :P 20:41:39 Oh, or do you mean a netinst CD? 20:41:42 GregorR: Yes 20:41:49 How else am I meant to install? 20:42:04 disabling t should save a bit too 20:42:04 Netinst is THE way to install Debian. All other methods are EVIL. 20:42:13 So's your mom. 20:42:17 Anyway I'm just doing this inside a VM 20:42:25 To see if the installer can do the crazy LVM2 shit I need 20:42:34 ehird: Yeah, and she installs Debian with the DVDs, so what does that tell you! 20:42:49 Hagjaj! 20:42:53 * oerjan expects a face mention just about now, or would that be too obvious 20:42:57 What's the fastest way to a lover's heart? A beautiful and stylish designer watch! 20:42:57 http://foleyohev.cn 20:42:58 At Diam0nd Reps we make it easy to get a Rolex, Cartier, Bvlgari or any brand name that you think of. As long as it is considered a high class watch, you will find it in our one of a kind store! 20:43:01 http://foleyohev.cn 20:43:03 Don't believe me? Click here to enter Diam0nd Reps right now, and see it with your very own eyes! 20:43:05 oerjan: no, but your face would be 20:43:23 good, good 20:43:29 ....... do you often paste spam into #esoteric ? 20:43:56 Annoyingly often 20:44:05 it must be - irc script spam! 20:44:19 *+virus 20:44:20 Deewiant, this stripped down cfunge and without t, compiled with -O3 is 0.005s faster for "all fingerprints disabled, clean env, stdout to /dev/null" run of cfunge than the same but with libm, librt, and libncurses deps 20:44:27 AnMaster: < 0.01 = 0 20:44:39 Deewiant: You're really on the offensive again me :P 20:44:41 *against 20:44:46 ehird: :-P 20:44:48 Deewiant, not when total time is 0m0.019s vs 0m0.24s 20:44:51 err 20:44:53 Deewiant, not when total time is 0m0.019s vs 0m0.024s 20:44:54 I meant 20:45:01 AnMaster: Yes, it still is. 20:45:02 Deewiant: Wanna know who else went on the offensive against me? 20:45:04 Deewiant: Hitler. 20:45:09 AnMaster: *yawn* 20:45:23 GregorR: OPTAMAZATION 20:45:29 GregorR, I'm working on a build for the bot yes. 20:45:29 ehird: well everyone knows the finns sided with hitler 20:45:42 GregorR: Use rc/funge, it'll really annoy AnMaster. 20:45:49 that is unsafe 20:45:55 it doesn't (afair) have a sandbox mode 20:46:10 And we all know GregorR cannot use LD_PRELOAD or patch C programs 20:46:11 I'm sure most of the interps that EgoBot runs are horribly unsafe :P 20:46:19 AnMaster: At least check the results for statistical significance: "If your benchmarks are not significant at the 95% confidence level, we don't want to hear about it", as the FreeBSD devs say. 20:46:20 Or trust in #esoteric not to be dix 20:46:36 Deewiant, I did. I did average of 50 runs 20:46:37 for each 20:46:40 duh 20:46:41 AnMaster: irrelevant 20:46:54 ehird: I trust the people who are usually here, but not my evil enemies who happen to find out I have a bot here and come just to hax me. 20:47:02 GregorR: Like ehird? 20:47:02 standard variance was small, +/- 0.001s 20:47:05 That guy's a jerk. 20:47:15 Yeah, he's a douchenozzle. 20:47:19 :D 20:47:24 I'm a taco. 20:47:31 a hypertaco? 20:47:43 Hypertacodouchenozzle maximus. 20:47:48 The third. 20:47:50 Esquire. 20:48:08 i believe nouns in -e are usually neuter, so that should be maximum. 20:48:16 (in latin, that is) 20:48:21 TIME FOR SOME NETINSTALL 20:48:26 GregorR: Is using the graphical install a sin? 20:48:33 Because I want to use the graphical install 20:49:21 ehird: i believe the bible says _nothing_ about graphical installs whatsoever. shocking, i know. 20:49:23 GregorR: I need your priestly consultation. 20:49:29 oerjan: church of debian. 20:49:33 ah 20:49:36 ehird: Graphical install is tolerable so long as you use netinstall :P 20:49:47 GregorR: I bet you use a tiling window manager. 20:49:52 Ooh burn 20:49:56 I wish, there are no good ones. 20:50:06 That's because the concept is fundamentally flawed 20:50:09 Douchenozzle. 20:50:10 haskell burn 20:50:11 Ratpoison's nice. 20:50:15 When I had a tablet PC I really wanted to find one, but they're all ultra-keyboard-bound, which sort of defeats the purpose of a tablet PC. 20:50:28 Of course, Ratpoison's nice for keeping full-screen windows. ;) 20:50:35 lol@debian-installer has a prominent screenshot button 20:50:43 GregorR: Maybe you could get StumpWM to work for that? 20:50:49 pikhq: but that's not useful! take a web browser - w/ a large screen the lines will be too long to comfortably read 20:50:51 (Lisp scriptable WM FTW?) 20:50:59 most of the time you WANT windows smaller than your screen 20:51:34 I've had full-screen web browsing for ages. 20:51:57 Either you have a tiny screen, your eyes hate you, or the sites you use have text columns smaller than full-sized :) 20:52:26 My eyes probably hate me. 20:52:46 People have argued that fullscreen browsing will become less popular/necessary as resolutions increase, but in my experience web page size/complexity is increasing faster :P 20:54:02 [[Windows 7 leaps forward: "We were able to shave 400 milliseconds off the shutdown time by slightly trimming the WAV file shutdown music."]] 20:54:33 Hah 20:54:52 Even Apple wouldn't brag about that 20:55:30 Hey ho, hey ho, installing a-debian I go 20:55:38 Umm... wait. 20:55:43 It didn't ask me whether I wanted testing... 20:55:49 UPS LOL ^_^ 20:55:54 * ehird kills vm 20:56:09 hur im clever 20:56:20 Graphical expert install. That looks right. 20:56:39 ... codeword for "Please make the UI horrible." 20:57:06 Yeah. 20:57:20 The expert install is if you need to do rather screwy stuff. ;) 20:57:31 'cuz, you know, a button saying "Please give me testing instead" would just be too hard 20:58:15 http://cdimage.debian.org/cdimage/daily-builds/daily/arch-latest/i386/iso-cd/ 20:58:17 There it is. 20:59:55 Anyway, I still think Debian and Ubuntu should merge. 21:00:08 Take Debian, slap some of the Ubuntu polish on, bam. 21:00:44 pushed these changes. 21:00:44 I don't disagree, but then where would all the Canonical douchebaggery go? 21:00:54 GregorR: Which douchebaggery? 21:01:00 GregorR, now I'll try to add it to the bot. Where was the hg checkout url now again 21:01:01 As far as I can tell Canonical just funds Ubuntu 21:01:09 !info 21:01:10 EgoBot is a bot for running programs in esoteric programming languages. If you'd like to add support for your language to EgoBot, check out the source via mercurial at https://codu.org/projects/egobot/hg/ 21:01:20 ehird: They're known for never contributing anything back to any upstream ever. 21:01:32 GregorR, and. How do I test the bot. I understood I needed part of another checkout called multibot right? 21:01:33 GregorR: Right, that is a big issue. 21:01:39 AnMaster: RTFREADME 21:02:00 AnMaster: RTFNEW_LANGUAGE 21:02:16 AnMaster: I may move multibot over, but for the moment it's in https://codu.org/projects/stuff/hg/ 21:02:26 GregorR, right 21:02:26 AnMaster: And to run it you need socat, because I'm E_LAZY :) 21:02:39 GregorR, I have socat since ages. 21:02:43 it is very useful 21:02:44 for bash bots 21:02:49 Yup, same idea. 21:03:25 Stop! Debian testing time. 21:03:29 GregorR, can't pull from https://codu.org/projects/stuff/hg/ 21:03:31 ↑ i'm so hip 21:03:36 Can't patch this. 21:03:39 GregorR, abort: There is no Mercurial repository here (.hg not found)! 21:03:47 AnMaster: ... quay? >_O 21:03:48 the other one worked 21:03:49 hm 21:03:56 GregorR, what? 21:04:02 $ hg pull https://codu.org/projects/stuff/hg/ stuff 21:04:02 abort: There is no Mercurial repository here (.hg not found)! 21:04:08 I guess you meant "why" 21:04:13 AnMaster: clone, not pull. 21:04:17 oh right 21:04:18 sorry 21:04:25 typoed 21:04:34 I think just about anything that has gone from Ubuntu to upstream has been because of Debian hunting down the patches... 21:04:38 "Quay" is nolanguage for "what?" 21:04:45 ah 21:04:48 qué? 21:04:56 GregorR: I thought it was a creative misspelling of Spanish. 21:04:58 Exactly, only in nolanguage :P 21:05:18 pikhq: No, it's pronounced kway, not kei 21:05:27 Hmm. 21:05:38 GregorR, what naming scheme for befunge. Should the command/directory be befunge98 or cfunge? 21:05:38 No, it's pronounced neither kway nor kei :-P 21:05:38 Debian-Installer, la la la la. 21:05:41 Installing testing. 21:05:42 La la la la. 21:05:47 or should they not be the same 21:05:52 AnMaster: {be,une}funge98. 21:06:01 (as in command befunge98 and directory cfunge) 21:06:04 no unefunge in cfunge 21:06:06 ehird, cfunge doesn't do unefunge 21:06:11 Fix it. :) 21:06:13 Deewiant: It's pronounced between "keh" and "kei" sort of, but it's hard to write the Spanish spelling in English so shaddap X-P 21:06:14 ehird, not a bug 21:06:31 AnMaster: It should be cfunge. 21:06:32 GregorR: /ke/ 21:06:43 AnMaster: (The names that aren't consistent with that are olde :P ) 21:06:56 GregorR, Ah ok. 21:07:02 Qu'e? 21:07:05 There's no 'i' in there that anybody could discern, IMO. :-P 21:07:24 GregorR, one sec. What about license. 21:07:36 AnMaster: So long as it's F/OSS. 21:07:42 AnMaster: you can't copyright such a small glue file. 21:07:43 GregorR, right. cfunge is GPLv3 21:07:58 Oh, you mean the three-line glue file? :P 21:08:00 ehird, true. 21:08:05 and that wasn't about it 21:08:14 Good 21:08:25 I was wondering if there were issues with licenses on the interpreters. 21:08:35 like your "GPL is evil" style 21:08:37 AnMaster: RMS probably considers a bot linking. 21:08:42 Nah, I'm ambivalent. 21:08:43 See clisp fiasco... 21:08:49 I'd like to take this opportunity to mention that the debian installer is niiiiice. 21:08:51 Who's "GPL is evil" style? 21:09:10 GregorR, ehird 21:09:24 I have never said "GPL is evil", ever. 21:09:30 I just think it's stupid and misguided. 21:09:50 ehird, not that exact phrase. 21:09:55 I think it's a good idea and sensible for some reasonable subset of software. 21:10:16 Although admittedly my definition of that subset seems to have dwindled to near-zero over time for my own stuff X-P 21:10:16 I was making a summary of your MBs of anti-GPL rants for reasons of brevity. 21:10:41 I do not rant about the GPL every time it is mentioned. 21:11:34 Is it just me, or are Pastry, Kademlia and Tapestry all the same but for some minor details X_X 21:12:11 Is it just me, or is GregorR, his mom and his face all the same but for some minor details X_X 21:12:18 BUUUUUUUUUUUUUUUUUUUUUUUUUUURN 21:12:24 GregorR, format for USED_VERSION differs a lot I see. 21:12:28 which one do you recommend 21:12:36 something like http://rage.kuonet.org/~anmaster/cfunge/ 0.4.0+bzr:trunk:r760 21:12:38 maybe? 21:12:43 lol 21:12:48 AnMaster: That's also changed over time ... URL version 21:12:51 (I'm probably doing a 0.4.1 release later this week) 21:13:35 Gracenotes, whirl is missing a USED_VERSION file 21:14:07 AnMaster: GregorR 21:14:15 err 21:14:16 yeah 21:14:18 mistab 21:14:34 GregorR: Ew, Debian doesn't use sudo by default. 21:14:36 I vomited. 21:14:37 Do you look at your keyboard when you type, or something :-P 21:15:02 AnMaster: Lots of them are, I only started adding USED_VERSIONS files when I realized how stupid it was that I didn't have them X_X 21:15:17 GregorR, hm cfunge only supports out of tree builds. 21:15:18 AnMaster: Those files are all olde, imported from the old EgoBot rather than their upstream source. 21:15:27 My root password will be butt. 21:15:44 AnMaster: Then I suggest you make the Makefile do an out-of-tree build. Put it in a 'build' subdirectory if it allows in-subdirectory builds. 21:16:07 GregorR, a compile.sh would work better. Since I need to invoke cmake as mentioned to create the makefile. 21:16:15 well I could do a one-off Makefile 21:16:23 but it might break with future versions. 21:16:29 You can make a makefile that calls cmake... 21:16:37 Deewiant: WHAT? Unpossible. 21:16:39 Deewiant, a shell script for it is easier. 21:16:42 ... 21:16:43 WHAT? 21:16:47 :-D 21:16:50 How on earth? Make calls sh! 21:16:51 yes it is. 21:16:57 Are you on CRACK? 21:17:02 ehird, no. 21:17:05 Inconceivable! 21:17:07 less \ for one. 21:17:14 \\\\ 21:17:24 Your call is so complex that typing \s becomes an issue, AnMaster? 21:17:27 Your build system sucks. 21:17:45 stop trolling. 21:17:56 No, I'm seriously gobsmacked. 21:18:08 Indeed, "que" has no "i" sound in it, except when spoken by people whose native tongues always have "i" after "e". 21:18:11 If you aren't happy with it submit a patch. Or if GregorR isn't happy with it, I can convert it I guess. 21:18:12 Your build system is shit if making it work scriptedly requires so many lines that \ is a pain. 21:18:13 Such as English. 21:18:26 kerlo: "hello" has no i... 21:18:42 kerlo: ... English always has /i/ after /e/? 21:18:48 That's news to me. 21:19:08 No wonder I'm told I speak somewhat oddly. 21:19:18 Yeah. 21:19:24 I find that, sometimes, the easiest explanation is that the assertion is wrong. 21:19:39 Do you have a specific accent? 21:19:47 General American. 21:20:19 With personal quirks, from moving so @#%@# much. 21:20:30 I think the long "a" is usually /ei/, and no other phoneme (or whatever you call them) contains /e/. 21:20:54 kerlo: Not any? 21:21:01 /eni/ 21:21:02 ? 21:21:06 SHA-1 collisions in 2^52: http://eurocrypt2009rump.cr.yp.to/837a0a8086fa6ca714249409ddfae43d.pdf 21:21:08 That's not /e/. 21:21:16 You're right, it isn't. 21:21:36 Damned vowel shift. 21:21:51 What's ɛ 21:22:09 The main thing seperating us from Middle English. 21:22:19 Ah, open-mid versus close-mid 21:22:20 GregorR, what GCC version do you have 21:22:21 And the main thing making our spelling so weird. 21:22:37 AnMaster: Idonno, 4.whatever-installs-in-debootstrap-squeeze 21:22:53 GregorR, ok... 3.4 and later should all work 21:23:05 GregorR, what about cmake? cmake 2.6 or later is needed. 21:23:12 I can install that. 21:23:15 2.4 is a no-go. 21:23:33 AnMaster: You srsly think Debian has a gcc older than 4? 21:23:38 Now THAT's trolling. 21:23:39 ehird, no. 21:23:46 I didn't know he used Debian 21:23:48 cmake installed. 21:23:51 it might have been CentOS 21:23:52 ! 21:23:52 MinGW does 21:23:52 21:22 GregorR: AnMaster: Idonno, 4.whatever-installs-in-debootstrap-squeeze 21:23:57 he said that before you said 3.4 21:24:03 ehird, I didn't know *before* that reply. 21:24:04 also, centos is modern 21:24:19 People use CentOS? :P 21:24:25 Servers do. 21:24:39 They do when they're non-Debian-using idiots I guess. ^^ 21:25:58 GregorR, should I dump the whole source into the egobot tree or should I remove test and example programs and documentation (man page and such) first. 21:26:06 La la la, 6min32s remaining 21:26:10 Ladedededaaa 21:26:29 AnMaster: Just dump the whole source in, that should make it marginally easier to merge in later changes. 21:26:55 For a while, Debian Stable had a newer GCC version than Gentoo... 21:26:57 GregorR, a plain bzr export is 1.1 MB according to du -sh :) 826K if you use -bsh 21:27:06 (Gentoo recently jumped from 4.1 to 4.3) 21:27:24 AnMaster: WTFBBQ? That's a huge fekking program (assuming that doesn't include bzr control files?) 21:27:33 GregorR: His code is bloated with inline asm. 21:27:41 And other ridiculous optimizations. 21:27:43 no. That was just one file. 21:27:48 Just go for CCBI2 fi you want to keep your sanity. 21:27:50 *if 21:27:54 GregorR, no bzr control files. But it is a lot less if I strip all the doxygen comments. 21:28:01 ehird, you missed the bit about no sandbox. 21:28:01 You, after all, can get a D environment working 21:28:11 AnMaster: GregorR is so crap at coding D 21:28:15 no he isn't 21:28:19 he couldn't possibly disable a few fingerprints and comment out some instructions 21:28:20 but why would he want to do that 21:28:22 it'd be impossible 21:28:48 *yawn* 21:28:57 Yawn on the butt 21:29:22 ls 21:29:24 err 21:29:26 wrong window 21:29:52 rm -f horse_porn.avi 21:31:09 find porn -type f | grep -i horse | xargs mplayer -fs -shuffle 21:32:47 ok I got a makefile working for it. Just to irritate ehird. It is twice as long as the shell script. 21:33:00 Errr... 21:33:17 AnMaster: making a shell script a makefile is adding \s and prefixing with tabs, and adding a single rule 21:33:25 Ur doin it rong. 21:33:59 ehird, I'm doing it idiomatically! 21:35:09 ok if I really make the code dense it is just 3 lines longer than the shell script. But the shell script isn't dense. 21:36:54 http://pastebin.ca/1408848 http://pastebin.ca/1408850 21:36:59 decide which you prefer. 21:37:13 You've really got something to prove, haven't you? 21:37:16 GregorR, ^ 21:37:25 ehird, the make one actually looks nicer. 21:37:38 but I think the logic is easier to follow in the shell one. 21:37:49 Only if you're an imperative weenie. 21:37:55 .PHONEY: clean all 21:37:57 Fail. 21:38:00 It's .PHONY 21:38:08 ehird, thanks for the bug report. 21:38:21 Also, they idiomatically go before the definitions. 21:38:38 ehird, no? I always seem them at the end 21:38:47 * GregorR finds the Makefile much clearer. 21:38:49 AnMaster: See gnu make manual. 21:39:33 ehird, ok. The fact that iirc automess puts it at the end means nothing indeed. 21:40:04 -!- Leonidas has quit ("An ideal world is left as an exercise to the reader"). 21:40:05 AnMaster: BTW, Debian _has_ a rolling release system (you've dissed it on that point before) 21:40:07 sid/testing 21:40:17 ehird, that is rather unstable though 21:40:19 (er, that is, sid/unstable, /testing) 21:40:20 AnMaster: not testing 21:40:21 stable rolling release 21:40:28 ehird, testing is rolling release too? Hm ok 21:40:30 Testing is super-stable. 21:40:32 packages from sid go into testing when they're stable for a few weeks 21:40:40 AnMaster: then, testing freezes gradually every year or so 21:40:44 and is declared the stable release 21:40:45 and thaws 21:40:49 for more updates 21:40:55 I see. 21:40:58 a nice system 21:41:02 Fonsie be praised. 21:41:04 not *too* bad. 21:41:11 How is it at all bad? 21:41:19 It's rolling-release. 21:41:20 ehird, "testing freezes gradually every year or so" 21:41:36 AnMaster: That's just for a week or so. 21:41:40 To become the stable release. 21:41:43 that bit sounds a bit inconvinient. (sp?) 21:41:48 I'm sure you can do without minor updates for a week 21:41:49 ehird, a week or so, ok I guess. 21:41:54 ehird: Except when the release is lenny, then it's for a month or so X-P 21:42:04 GregorR: Mishaps happen :P 21:42:11 btw anyone remember the long wait between woody and etch 21:42:20 or whatever it was after woody 21:42:25 AnMaster: btw anyone remember the long wait between $debian_release and $other_debian_release 21:42:37 ehird, it was several years 21:42:39 GregorR calls debian stable debian obsolete :-) 21:42:46 I see what he means 21:43:00 Mind you, it has its place :P 21:43:05 It just happens to be obsolete. 21:43:24 (Some businesses basically count on all their systems being obsolete because then nothing ever has to change) 21:43:36 I use it for my router. 21:43:38 GregorR, you asked about size and bzr control files before. .bzr directory is 6.3 MB 21:43:41 "This system is obsolete." "Yeah, but it was obsolete two years ago and worked fine then." 21:43:57 Which is something that I honestly don't care about being recent, I just want it to work. 21:44:04 Exactly. 21:44:07 And, well, Debian stable Just Works(tm). 21:44:10 GregorR, but remember cfunge had over 750 revisions. 21:44:31 AnMaster: That is something I remember, being a piece of information I have never seen until this point. 21:44:40 um 21:44:46 oh you mean the version string 21:44:48 forgot about that 21:45:19 No, actually, I was being sarcastic, but thanks for pointing out that in fact I have seen that >_< 21:45:38 GregorR: I remember that you're pink. 21:46:33 That is roughly the skin color of "white" people, yes. 21:46:38 ehird, should I place the entry in the interps/Makefile for cfunge in dictionary order or at the end? 21:46:53 AnMaster: Logical order. 21:47:00 Dur. 21:47:08 ehird, um. "First"? 21:47:11 ;P 21:47:41 * ehird watches debian-installer go. 21:49:05 GregorR, can you pass arguments to the interpreter. 21:49:11 AnMaster: alphabetical. Recall that EgoBot is my code :P 21:49:21 GregorR: Your mom is your code. 21:49:24 AnMaster: Yes. interp_file "./interps/foo/bar -bleh -forp" 21:49:32 Oh, I thought you meant like 21:49:36 Cfunge's actual makefile 21:49:37 Not egobot's makefile 21:49:38 Durr. 21:49:44 GregorR, I should add a space to the file name ;P 21:49:52 "c funge" 21:50:00 Your mom should add a space to the file name. 21:50:02 "your face" 21:50:08 except it would probably break cmake. 21:50:08 AnMaster: Well then that won't work, but gee, it's almost as if I've intentionally set it up so there are no spaces. 21:50:17 GregorR, hah ok 21:50:34 GregorR, btw, do you support passing stdin from irc. Like: 21:50:47 ^bf ,[.,]!test 21:50:48 test 21:51:08 I could add that as a special case for e.g. bf, but not as a general case. 21:51:17 I don't support it generally mainly because I don't know how I want to handle that (and because it kills the program after 21:51:20 ah 21:51:23 30 seconds) 21:51:32 good :) 21:51:52 ^bf +[+-] 21:51:58 ...out of time! 21:52:12 hm I guess [] is valid. Just no one writes it. 21:52:14 ^bf +[] 21:52:20 ...out of time! 21:52:24 (of course) 21:52:37 Many people write it for infinite loops. 21:52:39 !bf +[] 21:52:51 ehird, true. But it isn't something you often see. 21:52:55 and he said 30 secs 21:52:56 !help 21:52:57 Supported commands: bf_txtgen help info 1l 2l adjust axo bch befunge bf bf16 bf32 bf8 dimensifuck glass glypho kipple lambda lazyk linguine malbolge pbrain qbf rail rhotor sadol sceql trigger udage01 unlambda whirl 21:53:03 hm concurrent 21:53:04 idea: 21:53:07 No 21:53:32 time a lot of programs so all finish at once with output 21:53:40 you have 30 secons for the input 21:53:48 then watch as it flood off when all is output at once 21:53:56 GregorR, or do you rate limit it properly? 21:54:06 I rate limit it very simply. 21:55:11 oh you want to disable TURT too. Will add that. 21:57:49 GregorR: you know, I never thought I'd use Debian on a desktop. Ever. Why must you crush my world? Dick :( 21:59:42 So why're you installing it 21:59:56 Deewiant: Because he's made me like it, with his evil mind tricksies. 21:59:59 :-( 22:00:06 Ah, okay. 22:00:15 Should be illegal to do such a thing. 22:00:37 3DSP BASE CPLI DATE FING FIXP FPDP FPSP FRTH HRTI INDV JSTR MODU NULL ORTH REFC REXP ROMA STRN SUBR TIME TOYS 22:00:44 those should be safe right Deewiant ? 22:00:55 For what definition of 'safe' 22:01:02 Deewiant, for EgoBot safe 22:01:06 Deewiant: He's going to run a nuclear reactor on befunge 22:01:24 Deewiant, no file or socket IO to begin with, 22:01:25 He's going to run all the nuclear reactors on Befunge. 22:01:39 Deewiant, or other stuff you wouldn't want in egobot. 22:01:42 Here goes Debian! 22:01:48 TOYS generate an svg file with a fixed name. 22:01:54 so I turned that off. 22:01:58 TURT. 22:02:03 err 22:02:03 yes 22:02:04 TURT 22:02:10 was the one I turned off 22:02:15 OH GOD DEBIAN IS BREAKING 22:02:17 ;_; 22:02:20 Oh god, Debian has a text-only bootup. 22:02:21 Ahem. 22:02:25 * ehird pulls self together. 22:02:33 It just didn't recognize my virtual audio card. 22:02:34 I think. 22:02:48 text-only is nice 22:02:49 IMO 22:03:07 No MODE? 22:03:12 Ah right, you don't implement it 22:03:16 My system had a kernel failure, apparently :< 22:03:17 Deewiant, indeed. 22:03:18 Nom nom nom. 22:03:29 Deewiant, it isn't very useful either IMO. 22:03:45 Um, I can't change my system resolution. Wtf Debian? 22:03:52 A deque is noticeably handier than a stack 22:04:28 Oh god it comes with Epiphany 22:04:57 Deewiant, oh right THAT was why I didn't implement it. 22:05:06 And I've occasionally felt the want for switchmode 22:05:15 Hovermode is more a curiosity, I admit 22:05:28 Deewiant, the IP mode thingy I don't have anything much against. the deque bit though... 22:05:34 Although it's not completely useless either 22:10:36 * AnMaster reads on how to setup the bot. 22:11:29 err 22:12:09 GregorR, there is no makefile for multibot? 22:12:39 gcc multibot.c -o multibot -levent 22:12:40 :P 22:13:01 GregorR, and then just copy the binary into egobot ? 22:13:21 Yeah 22:13:38 GregorR, it doesn't like -Wall -Wextra ;P 22:13:48 -!- tombom has quit ("Peace and Protection 4.22.2"). 22:13:59 Wha? Does for me, I always compile it that way to make sure it's sensible ... 22:14:15 well 22:14:19 it isn't warning free 22:14:24 which might not be the same thing 22:14:25 Oh noes 22:14:25 ;P 22:15:33 nothing seriously bad though 22:15:49 some unused parameters, comparsion between signed and unsigned 22:15:51 http://pastebin.ca/1408899 22:16:14 Grr, virtualbox guest additions needs linux 2.6.27 22:17:26 GregorR, "Use: multibot ", but where do I set the network 22:17:59 You don't, I assume. 22:18:01 AnMaster: You use socat 22:18:03 Or, you use socat. 22:18:05 Snap. 22:18:06 oh right 22:19:26 GregorR, got an example command for it? socat have so many ways to run apps in 22:19:31 no idea which egobot wants 22:19:56 socat TCP4:irc.freenode.net:6667 EXEC:'./multibot MrMultiBot esoteric' 22:20:02 thanks 22:20:09 GregorR: wow, debian now have a Really Please Break This distro 22:20:11 "experimental" 22:20:19 ehird: Yup :P 22:20:26 ehird: sid was too stable. 22:20:30 GregorR: LOL WAT? 22:21:36 err 22:21:41 what do they call it 22:21:46 this ultra-sid 22:22:07 AnMaster: experimenta 22:22:07 l 22:22:13 ah 22:23:32 yay it works (tested on a local test ircd) 22:23:49 (because I'm too lazy to setup anything more secure than "separate user" 22:23:51 ) 22:24:48 GregorR, shouldn't you setup ignore for all those *.bin or something 22:24:54 hg st is unreadable ;P 22:25:22 *eh* 22:25:28 what 22:25:39 not that I know how hg does ignore 22:26:17 -!- BeholdMyGlory has quit (Remote closed the connection). 22:26:47 AnMaster: .hgignore 22:27:18 Deewiant, one per dir, or one at top of tree 22:27:31 (the latter is how .bzrignore is done) 22:27:32 I always had just the latter 22:27:38 The former sounds SVN-y :-P 22:28:16 Deewiant, yes. What is the syntax. I mean if I do ./foo in .bzrignore it means "only when relative tree root" 22:28:22 as opposed to any "foo" 22:28:27 AnMaster: man hgignore. 22:28:32 right 22:28:36 I tried hg help ignore 22:28:40 and such 22:30:29 Is there a depressant that targets the area of the brain that converts words into keystrokes? 22:30:47 I suffer from typing-words-that-I-mean-to-type-later-on-but-not-now-itis. 22:32:11 -!- ehird_ has joined. 22:32:17 Hello bruthas. 22:32:19 From Debian. 22:33:31 GregorR, making sure the .hgignore works now. Not perfect but a good start. 22:33:48 *eh* 22:34:02 GregorR, http://pastebin.ca/1408922 22:34:06 is that useful to you? 22:34:11 and why "*eh*"? 22:34:15 I never use hgignore. I find that when people count on it working, it doesn't, and they don't count on it working when it's not there :P 22:34:50 GregorR, I'm doing this commit thingy now. How do I set proper author/commiter/whatever for hg? 22:35:28 echo -e '[ui]\nusername = Your Name ' > ~/.hgrc 22:35:42 right 22:36:02 seems I already have that file 22:36:16 contains "[extensions]\nhgext.transplant=\nhgext.gpg=" 22:36:23 where \n is replaced of course 22:36:33 no idea why 22:36:40 * AnMaster adds the ui thing to it 22:37:23 A green. 22:37:48 GregorR, are you happy with one in the format of: username = Arvid Norlander 22:37:59 AnMaster: Those pesky hg-cloning spambots. 22:38:10 ehird, web interface... 22:38:10 duh 22:38:22 AnMaster: Because spam bots can't remove nospam, and replace at and dot. 22:38:30 That's just too advanced for them, they've been working on it for years. 22:38:35 ehird, they can. But I have yet to see the TLD bit. 22:38:52 AnMaster: So I assume you don't run a spam filter. 22:38:53 there are tricks like that AIS uses too 22:39:02 Yes. They're stupid too. 22:39:05 ehird, I do. But I prefer to get as little as possible anwyay. 22:39:54 I encrypt my email with AES. 22:40:00 The address that is. 22:40:54 GregorR, http://rage.kuonet.org/~anmaster/cfunge.bundle 22:40:56 Time to install a Weasel of Ice. 22:41:21 Oh, I have it already. 22:41:23 GregorR, I hope that bundle is correct. file says "data". 22:41:56 since it says "compressed" in hg help bundle I would have expected something like "gziped data" or such from file 22:42:01 or bzip2ed 22:42:02 or whatever 22:42:38 GregorR, that one doesn't include the .hgignore file 22:43:15 GregorR, tell me if you have any issues with it. I'm around for about 40 more minutes before going to bed. 22:43:24 GregorR: can you make debian use gksudo instead of the gksu? 22:43:28 *-the 22:46:55 ah 22:47:04 it is HG10 followed by bzip2 file 22:48:29 BitTorrent is cute. 22:48:48 "I'm bored, so I'm going to upload this 2-gigabyte game 21 times, 'kay?" 22:49:20 Hmm, I wonder if I should say something on-topic at some point. 22:50:05 kerlo, write a befunge-93 interpreter (no IO but meh) in subleq? 22:52:18 GregorR, there? 22:54:24 -!- ehird_ has quit ("Ex-Chat"). 22:57:42 That wouldn't be saying something on-topic. 22:58:06 kerlo, true, would be doing on topic 23:00:29 hg is a lot easier to use than git IMO. Things are named pretty much like you expect them. Some differences to bzr, svn, darcs and so on of course. But they are still pretty logical. No odd "add files already tracked" for example (yes there is the -a option for git, but that isn't what you would have expected). Easy to figure out errors (not that errors are as common as with git either). Only think th 23:00:30 at is slightly irritating is hg merge after hg pull instead of auto merging if you have no local changes compared to the branch you are pulling from. 23:01:46 I think I'll agree with ehird and say "familiarity is not intuitiveness" 23:02:00 Or, in this case, s/intuitiveness/logicality/ 23:02:01 IMO you shouldn't need to read manuals for version control systems in general. Once you learned one DVCS it should be a simple matter of reading a list of "what is the command called in this one". About half a screen at most. 23:02:16 Deewiant: Stop trying to make AnMaster not find every oppertunity to dis git; it's an instinct of his, part of the "Be a Complete Moron" package. 23:02:21 What you're saying is that all VCSs should be identical 23:02:29 Except for their command names 23:02:37 The customer may choose any VCS he likes, so long as it is bzr. 23:02:45 Deewiant, Well no. options can differ of course. 23:02:54 But point is, user interface should be compatible. 23:02:57 more or less 23:03:15 VCS Newspeak. 23:03:16 That's like saying that you should only need to learn one language, all others should be compatible 23:03:38 it shouldn't be like the difference between emacs and vi. it should be like the difference between vi and vim, or emacs or xemacs, or nano and pico 23:03:43 You can't have a compatible user interface if you do things fundamentally differently 23:04:04 Deewiant, then provide a compatibility wrapper. 23:04:16 AnMaster: http://hg-git.github.com/ 23:04:18 in fact I'm going to check bzr-git out and use that. 23:04:19 There's your newspeak. 23:04:20 Sure, I'll wrap Haskell for VB compatibility 23:04:27 ehird, I know about hg-git. 23:04:28 But you're an idiot btw. 23:04:37 AnMaster: I'm sure you do. It was just released today. 23:04:46 And more or less only reddit knows about it. 23:04:49 I'm sure you knew. 23:04:56 ehird, um... I heard about it in #mercurial 23:04:57 ... 23:05:01 I am in there you know 23:05:09 Last I heard you disliked hg. 23:05:19 ehird, correction: I *prefer* bzr 23:05:24 but hg isn't too bad. 23:05:49 Regardless, you're saying that we should not be able to use a superior UI because it's not one you've used before. 23:06:05 In response, I say, fuck you; if everyone followed that, we'd be using a UI identical to DOS. 23:06:06 ehird, You have a flaw in that argument. 23:06:10 and it is "superior" 23:06:15 git's UI isn't 23:06:21 Troll. 23:06:37 ehird, so are you when you are claming bzr's UI isn't better then. 23:06:58 Deewiant: do you think he'll go into another tirade if I say "whoosh"? 23:07:11 sure it is subjective... 23:07:11 An opinion on an UI is one thing; claiming that a software should provide a compatible UI for users of other software is quite another. 23:07:36 Deewiant, I'm saying that confusing UI and steep learning curve are not features. But bugs. 23:07:40 ehird: No, we'd be using a UI identical to UNIX or something older. 23:07:50 AnMaster: So you don't want emacs to exist? 23:07:55 yes I think darcs fail here by calling it "recording" instead of "commit" for no good reason. 23:07:56 After all, it has a steep learning curve for notepad users. 23:08:00 ehird, there is viper-mode 23:08:04 for vi users 23:08:05 and 23:08:06 Emacs should provide a Notepad compatible UI. 23:08:13 And if it doesn't, that's a bug. 23:08:14 (remember: DOS's UI was intended as an imitation of the Bourne shell. It did so poorly. ;)) 23:08:17 ehird, almost. emacs in X is close. 23:08:18 viper-mode and vimpulse both suck for vim users 23:08:19 In fact, everyone should use notepad UIs. 23:08:20 At least IMO. 23:08:26 As different UIs are bad. 23:08:29 ehird: Notepad should offer an Emacs-compatible UI, you mean. 23:08:31 Meanwhile, you're still an idiot. → 23:08:33 ehird, I used BBEdit on classic Mac OS before I used emacs. 23:08:35 go figure. 23:08:52 AnMaster: Confusing and steep are subjective issues and hence can't be bugs. 23:08:53 it wasn't too large difference. 23:09:16 Certainly, if practically everybody agrees that something is confusing then something is probably wrong. 23:09:17 Deewiant, Compared to main one currently. Which is svn. 23:09:20 But that isn't the case. 23:09:22 it used to be cvs 23:09:35 do you think there is a reason for svn using similar command names to cvs? 23:09:41 not perfectly similar 23:09:43 The goal of git is not to replace svn. 23:09:44 but close enough 23:10:04 I thought the goal of git was to replace BitKeeper. ;) 23:10:20 yes it was. Sad thing is it spread outside kernel development. 23:10:34 Eh, better than BitKeeper spreading. 23:10:39 true 23:10:42 AnMaster: Yes, there are at least two good reasons: 1. they share paradigms, 2. SVN was intended as a free alternative of CVS 23:10:47 pikhq, I can't argue with that 23:10:54 (damned Larry McVoy is still trying to push that on Tcl... :() 23:11:04 Deewiant, um. CVS was, and is, free? 23:11:08 open source. 23:11:26 you must meant "visual source safe" or something? 23:11:37 I thought its license wasn't free but maybe I misremembered 23:11:48 s/a free/an/ then, same difference 23:12:10 LICENSE="GPL-2 LGPL-2" 23:12:13 at least nowdays 23:12:59 AnMaster: source code of CVS was submitted to FSF and has been hosted by FSF long ago. 23:13:15 mhm 23:13:21 iirc, since 1990 23:13:29 lifthrasiir, so what was Deewiant thinking about 23:13:30 Alright, misremembered then. 23:13:33 i don't know 23:14:58 GregorR, removing all extra docs/examples/tests except LICENSE and CREDITS from cfunge source reduced size to 708K 23:15:32 261621 src 23:15:45 well I include build system and license and such too 23:15:49 anyway i think both git and hg have its pros and cons, and they were indeed developed with each other's influences 23:16:04 278432 total 23:16:06 so i don't care that flame war as long as they are improved together. 23:16:09 with only what is needed to build it is 605K 23:16:25 100K for the build system? Sheesh. :-P 23:16:25 628K if you want the tools to generate new fingerprints and update fingerprint list 23:16:36 Sheesh. 23:16:42 Deewiant, 78K for the LICENSE file 23:16:58 29K for the build system 23:16:58 All that for 5x the speed and 0.1x the features ;-P 23:17:05 Deewiant, um what? 23:17:11 no. 23:17:24 Compared to CCBI. 23:17:25 I removed docs, examples and test cases 23:17:36 Deewiant, what is the ohcount stats for it 23:17:49 Total 56 7195 1015 12.4% 2111 10321 23:17:56 hey, python has no build system except for tiny setup.py! ;) 23:18:03 c 108 11503 4734 29.2% 2222 18459 23:18:05 lifthrasiir: ccbi.rf is 70 bytes 23:18:05 well 23:18:09 Total 113 12145 5037 29.3% 2372 19554 23:18:10 overall 23:18:27 lifthrasiir: That's my build system :-P 23:18:28 Deewiant, as you can see. I have a much higher comment ratio 23:18:35 Deewiant: heh, good. 23:18:44 Deewiant, the actual source code lines are not that much more. 23:18:46 AnMaster: 7195 vs. 12145 23:19:10 Bah. Stick it all on a single line. 23:19:16 Deewiant, that includes those maintenance shell scripts 23:19:17 Arguably comments shouldn't be removed entirely from the equation. 23:19:31 Deewiant, yes. Nor blank lines 23:19:40 though we are close there 23:20:00 My blank-lines ratio is much higher :-P 23:20:04 Deewiant, and your code is a lot denser than mine *seds away all the assert* 23:21:38 Deewiant, another thing. D has a larger standard library 23:21:50 so for example my DATE is more complex 23:21:57 needing to do the same calculations as yours 23:22:12 you do it like Tango.Date.ConvertToJulian 23:22:15 or something like that 23:22:15 iirc 23:22:33 while I had to implement it in terms of libc functions 23:22:36 Consider getting a date library? :-P 23:22:42 You don't /have/ to implement everything from scratch 23:23:00 Deewiant, I don't implement all from scratch. Considered using ncurses for TERM? 23:23:21 I use terminfo for term 23:23:32 Deewiant, anyway if I exclude lib the number is closer to your 23:23:34 c 95 8709 3539 28.9% 1594 13842 23:23:43 Which, I guess, implies curses anyway 23:24:06 Deewiant, lib contains hash library and string concat library. Oh and XML generating one for TURT. 23:24:20 suddenly it is a lot closer. 23:24:56 And I've still got twice the amount of fingerprints + unefunge + trefunge + befunge93 :-P 23:25:05 and I even have fewer modules than you do. You have to have the number. Since each *.c has a *.h file (roughly, there are a few *.h with no *.c) 23:25:31 Fewer modules I would expect 23:25:52 Deewiant, How much does the deque thing in MODE slow you down 23:26:14 I don't know, I can't remember a time when I didn't have a deque thing. 23:26:40 Deewiant, slower than plain stack I assume? 23:26:56 I. Don't. Know. 23:26:59 ok 23:28:43 Deewiant, where is the main bottle neck in ccbi btw? I assume you have profiled. 23:28:55 GC. 23:28:58 aha 23:29:07 Which probably means Funge-Space access. 23:29:41 yes I guess so. But even with no static funge-space I don't have a huge malloc() overhead... Oh wait I use memory pools. Right. 23:29:45 forgot that. 23:30:40 As we recently discovered with that 30-MB Gutenberg file, D's hash maps appear to suck a lot more than yours. 23:31:20 Deewiant, yes. And the hash library I use is pretty sucky IMO. I mean linked lists! Instead of rehashing with another algorithm and trying somewhere else. 23:31:35 only way to make it reasonable is to use a large initial size. 23:31:52 Something I can't affect. :-P 23:32:01 D's uses binary trees IIRC. 23:32:10 0x40000 23:32:15 I forgot what that unit meant 23:32:25 Deewiant: but you can make your own hash in D. i cannot. 23:32:26 -!- KingOfKarlsruhe has quit (Remote closed the connection). 23:32:27 Nice, just enough for all of Mycology. 23:32:27 Deewiant, the code of the hash library is a mess IMO. 23:32:31 hashtable* 23:32:32 Deewiant, is it? 23:32:35 lifthrasiir: It didn't actually help much. 23:32:39 AnMaster: Looks like it. 23:32:57 Deewiant, It uses CRC32. There would be collisions. 23:33:00 Yes, by far. 23:33:10 Mycology is less than 0x20000 23:33:29 AnMaster: Most of them are accessed only once. 23:33:33 Deewiant, anyway if you gave me an example of a real befunge program that was larger... 23:33:42 There is none. 23:33:42 than mycology 23:33:55 Feel free to write one. 23:33:58 Deewiant, maybe that is why mycology fits "just in" 23:33:59 :P 23:34:18 lifthrasiir: And actually, that same hash has since been added to the stdlib so my code probably slows it down :-P 23:34:25 Deewiant, actually ^ul in fungot can run way out of static space 23:34:25 AnMaster: " no, she never does, i'm fnord this is always done however fnord fnord: to fnord on capital, you know!' 23:34:36 but I would need 128 MB for the static space to fit it in 23:34:44 AnMaster: slowdown.b98's purpose is to pull stuff out of your static space :-P 23:34:52 Deewiant, and I manage that just fine :P 23:35:09 Sure, I didn't expect it to blow up or anything 23:35:20 Deewiant, just a lot slower of course 23:35:39 Deewiant: but you can make your own hash in D. i cannot. <-- why not 23:35:54 python can load C *.so 23:35:56 I was thinking that it'd even the diff between CCBI and cfunge but that didn't really work out 23:36:02 so you could write your own FASTBIGHASH 23:36:03 or such 23:36:11 AnMaster: technically i can make it, but it will be much slower. and i don't like C modules basically. 23:36:20 lifthrasiir, why will it be much slower. 23:36:30 I didn't realize that it spends a million iterations in the main area which just increases the gap due to cfunge's static area 23:36:33 pure python module, i mean 23:36:41 And of course I didn't realize the shrinking-bounds issue 23:36:46 lifthrasiir, iirc all calls, even for internal modules use the same calling convention. Which is far from optimal 23:36:54 well 23:36:57 the same set of them 23:36:59 there are a few 23:37:10 METH_O, VARARGS and a few more iirc 23:37:28 I'm sleeping -> 23:37:41 if i want i can write the better funge.space module in C, but that's not my way 23:37:48 I was thinking that it'd even the diff between CCBI and cfunge but that didn't really work out <-- :D 23:38:11 Deewiant, and then there is the slow hash library of yours 23:38:20 of course someone can replace funge.space with C module and that IS my goal, though ;) 23:38:27 lifthrasiir, true. efunge avoids loadable modules too. 23:38:31 but 23:38:34 ets is pretty ok 23:38:58 since it is intended as the raw table backend of the erlang mnesia database 23:39:05 but you can use just ets directly 23:39:09 which I do in efunge. 23:39:20 I used a dict initially. A LOT slower. 23:39:36 I'm talking about a 10 second difference there. 23:39:51 of course someone can replace funge.space with C module and that IS my goal, though ;) <-- hm? 23:40:03 you mean if someone else do it 23:40:08 or what 23:40:35 AnMaster: "i won't do it but someone can do if he/she really wants" 23:40:47 lifthrasiir, and then you would use it... I see 23:41:28 lifthrasiir, do you implement TURT yet 23:41:49 yet, but will implement before 0.5 final 23:42:03 err yes or no 23:42:26 you meant "no, but will implement before 0.5 final" right? 23:42:28 yes 23:42:37 right 23:43:24 lifthrasiir, python -O made no difference for pyfunge it seems 23:43:37 * AnMaster wonders what -O is supposed to do 23:43:49 I mean, optimise clearly 23:43:51 but how 23:44:42 AnMaster: i'm not sure but it is supposed to optimize bytecode iirc. 23:45:41 -!- oerjan has quit ("Good night"). 23:46:18 it doesn't make a big difference and sometimes made the program slow, for example pyfunge. 23:47:17 lifthrasiir, hm your test suite programs (just testing them) doesn't output newline at the end? 23:47:44 AnMaster: for convenient editing, yes. you have to append newline to *.expected files. 23:48:23 lifthrasiir, I would append newline to the test suite output to reduce the confusion of messing up my terminal when I'm testing cfunge on it 23:48:52 -!- coppro has joined. 23:48:55 ~/funges/interpreters/pyfunge/tests/befunge98/iterate-zero.b98 that sounds like one of mine 23:49:15 yes I have an iterate-zero.b98 too 23:49:26 which is different 23:50:04 i made it for completeness, and i'm just trying to make all tests for all commands 23:50:19 lifthrasiir, mycology... 23:51:04 surely mycology does good job, but it doesn't test obscure cases which made pyfunge incompliant sometimes. 23:51:33 lifthrasiir, sure your sysinfo one is correct? 23:52:07 what do you mean, befunge98/sysinfo.b98 test or "y" command implementation of pyfunge? 23:52:08 Grepping for Py_OptimizeFlag doesn't show that the -O flag would do much more than turn off __debug__ and discard asserts. 23:52:14 lifthrasiir, possibly both. 23:52:24 lifthrasiir: I get BAD: 18y and 19y (offset to greatest point) didn't push <91, 42> from it. 23:52:43 other than that no errors 23:52:47 what, i'll check for it 23:52:49 and I pass mycology correctly. 23:52:56 that could be my fault 23:53:13 I don't know. Too late to try to figure out that code 23:53:43 lifthrasiir, do you have any automated runner script for then 23:53:47 since there are so many 23:53:58 that is better with mycology. Just a few files. 23:54:11 pyfunge-test is for it, but not so useful for testing other interpreter than pyfunge 23:54:36 could you not just change the name of the interpreter executable in it 23:55:43 you can, but some tests contain different options which don't match with other one 23:55:52 eh? 23:56:34 *.options file, if present, gives additional option to pyfunge when tested 23:56:54 -!- iano has joined. 23:57:02 iirc there are two or three tests use it for testing custom division-by-zero behavior in befunge-93 mode. 23:58:06 right 23:58:21 i have to separate pyfunge-specific tests and other tests clearly, but that will be done later 23:58:37 its primary purpose is testing pyfunge right now ;) 23:58:39 lifthrasiir, "BAD: null character reflects" 23:58:47 where does it say in 93 that it shouldn't 23:59:45 not that I implement anything but an option to disable SGML spaces for 93