00:00:18 -!- oerjan has joined. 00:00:27 {)}:i{.!!-}:d{.*.256=!*}:s{.p}:o](0\~@ 00:00:30 And I thought Perl was bad. 00:02:32 What's that, awk? 00:03:06 Golfscript. 00:03:06 -!- FireFly has quit (Quit: swatted to death). 00:03:17 Aaah. 00:04:01 So, I want to work on a basic Bayesian filter type program, just to learn the basics of creating one. My goal is to separate between English and Non-English with text that is in Arabic letters. Perhaps also to have multiple languages, how could I start? I was reading some articles I found on Google but none really told me how to begin 00:07:59 * pikhq is willing to bet that this compiler is just too stupid for its own good. 00:12:23 if i wanted to scare people, i would do that. i would also pick a nick like "Guest17101" 00:12:35 at any rate, I see cpressey [...] 00:12:45 and then pretend to have MPD 00:13:34 oerjan: pikhq: any ideas? 00:14:59 i'd imagine counting the occurrences of the 100 most common words in english would go a long way. 00:15:31 although more information might discern shorter texts, maybe 00:16:24 hm idea: design sentences in a non-english language where every word is actually a legal english word 00:17:10 but not actually used as one (i.e. not a transparent borrowing) 00:17:49 sounds boring 00:19:20 hm doing this in norwegian is not as easy as i thought 00:19:30 night 00:19:44 have to avoid the endings with -er. or maybe... 00:20:41 oerjan, your challenge is to do that with Japanese. 00:20:43 I wish you luck. 00:20:54 FUNNY GUY 00:21:09 hur säger man lycka till på norska nu då? 00:21:15 lykke til 00:21:54 obviously the reason that won't work is that all japanese sentences contain transparent english borrowings. 00:21:58 how disappointingly obvious 00:22:14 heh 00:22:36 oerjan, find one valid Japanese word that is also an valid English word (spelling included) 00:22:48 it's just swedish with a different orthology 00:23:02 Lymia: ok there _might_ be another reason then :P 00:23:20 I'm so going to have to rewatch this movie sober later 00:23:39 pikhq: you don't know a japanese character which accidentally looks like an english word, do you? 00:23:46 *happen to 00:23:53 いいえ 00:24:00 -!- Vorpal has quit (Ping timeout: 250 seconds). 00:25:05 well if iiiiz was an english word and you had bad handwriting... 00:25:39 oerjan: No idea. 00:25:43 Lymia: tsunami 00:25:52 :P 00:26:17 Well, I do believe I've pinpointed the source of my woes. 00:26:20 offset (Move x : Loop a y : xs) = Loop (a+x) (Move x : y ++ [Move (x*(-1))]) : (offset $ Move x : xs) 00:26:25 *ouch ouch ouch* 00:26:31 pikhq, nope. 00:26:32 Can you see what's so painful about it? 00:26:35 The Japanese don't spell it like that. 00:26:41 lol, the error in that is so blatantly obvious 00:27:02 olsner: And remember, I was finding the *fixed point* of that function. 00:27:32 pikhq, doing it wrong as always :/ 00:27:44 pikhq, please turn that into English. 00:27:48 Or at least math. 00:28:00 pikhq: um i don't see why that should infinitely recurse at least 00:28:09 oerjan: That's only one line from the function. 00:28:15 Lymia: ehm, fixed points are math aren't they? 00:28:30 oerjan: Aaand it's not infinitely recursing. 00:28:41 just wrong? 00:29:12 oerjan: It's just reconstructing a list is all. 00:29:16 oerjan: See the "++". 00:29:27 olsner, translate that function into math 00:29:39 That list is, incidentally, 2 megabytes before it hits my parser. 00:29:46 Lymia: not my job, I'm just here telling you it could be 00:30:53 Suddenly, it executes in 5 seconds instead of 2 minutes. 00:30:56 i'm assuming this is brainfuck with relative positioning built into the representation 00:31:26 Yeah, relative positioning is built into my intermediate representation. 00:32:04 well that particular rule looks correct to me... 00:32:34 *Correct*, yes. 00:32:43 or maybe you don't want to have that second offset since you apply fixed point anyway 00:32:44 *Insanely slow*, also yes. 00:35:04 well that's the only thing i can think of 00:35:41 The point is not that this optimisation functioned incorrectly. 00:35:48 The point is that it took 2 minutes to execute. 00:36:03 or maybe you don't want to have that second offset since you apply fixed point anyway 00:36:07 the point is blah-di-blah-di-blah? 00:36:30 i can see how that might give a blowup 00:36:41 oerjan: Uh, wouldn't that screw up the correctness of the function? 00:36:57 oh? 00:37:55 are you, or are you not, reapplying offset on the result of offset until a fixed point? 00:38:06 if so then it shouldn't 00:38:18 Oh, duh, I see what you're saying. 00:39:30 Does not seem to have helped, though. 00:39:50 i note you are not applying offset _inside_ the loop 00:39:53 *Loop 00:40:11 I am, however, doing "y ++ [Move (x*(-1))]". 00:40:28 oh you mean the problem is ++ slowness? 00:40:36 Pretty sure. 00:41:12 that would surprise me, unless you put a whole lot of Moves there... 00:41:24 appending a single element isn't slow i think 00:41:29 *To the end*. 00:41:38 I thought prepending was faster than appending. 00:41:45 ++ reconstructs the entire list and replaces the tail. 00:42:00 pikhq: it's a _little_ slower, but it shouldn't be _huge_ 00:42:15 : is O(1). ++ is O(n). Any further questions? 00:42:29 well... 00:42:50 O(1) is O(n). Any further questions? 00:42:51 well you said the list was long 00:43:11 (lost kingdom?) 00:44:31 -!- pikhq_ has joined. 00:44:44 5 minutes and it's still running. 00:44:57 And yes, it's LostKng. 00:45:10 i still don't understand why it should make an enormous difference on a list which you are traversing anyhow 00:46:33 -!- pikhq has quit (Ping timeout: 252 seconds). 00:47:09 pikhq_: have you used profiling? 00:47:46 (note: i never have, but i know ghc has a profiler) 00:48:21 I have been. 00:51:47 another thing: that x*(-1) might be lazily evaluated, since this is too complicated for ghc to notice it is needed? 00:52:32 The fields in the BF type are strict. 00:53:03 hm... i guess. 00:54:34 oh well. my hunch then is to give offset the initial and final offsets as arguments 00:55:11 and recurse immediately instead of building the whole list several times 00:57:46 argh that confused my brain 00:58:53 well maybe just the latter, then: 00:59:24 offset (Move x : Loop a y : xs) = Loop (a+x) (offset $ Move x : y ++ [Move (x*(-1))]) : (offset $ Move x : xs) 01:00:09 oh wait if you used my first idea it would split up: 01:01:33 offset m n (Move x : xs) = offset (m+x) (n-x) xs 01:03:27 offset m n (Loop a y : xs) = Loop (a+m) (offset m n y) : offset m n xs argh i'm not sure that's correct 01:03:54 clearly adding Move's makes things simpler to follow 01:04:45 offset m n xs should be equivalent to the old offset (Move m : xs ++ [Move n]) 01:09:15 offset m n (Loop a y : xs) = Loop (a+m) (offset m (-m) y) : offset m n xs 01:09:42 and 01:09:51 offset m n (Move x : xs) = offset (m+x) n xs 01:10:52 hm maybe that actually _is_ clearer, since you get rid of all the Move's 01:11:29 m and n should probably be strict arguments 01:11:47 pikhq_: ^ 01:12:34 offset _ n [] = [Move n] 01:13:49 hmm, this Move/Loop thing, I recognize it 01:14:17 i realize this last one is redundant at the program top level, but i think it's cleaner this way 01:14:38 olsner: well it's thinly disguised brainfuck 01:16:46 also: even I knew about that magic roundabout, but it was via looking up the children's show 01:17:01 hey even _i_ knew about it. i think it was featured on wikipedia or something. 01:20:53 `addquote You make a fist, shake it at the sky, and shout "why, GNU, why?!" -- that is the standard reportig practice. 01:20:56 431) You make a fist, shake it at the sky, and shout "why, GNU, why?!" -- that is the standard reportig practice. 01:22:24 Autotup: on a scale of 1 to "Fuck you", how bad of an idea is it? 01:22:29 oerjan: Hmm. So, rewriting offset to make it actually keep track of all the state, and do its recursion itself... I like that. I like that a lot. 01:22:32 * oerjan wants to see the rest of that scale 01:22:47 yay 01:30:06 And there is absolutely, positively no need to find the fixed point of that function. Yay. 01:34:57 And now the assembler takes longer than my compiler. 01:36:01 Significantly longer, in fact. 01:37:33 -!- iconmaster has quit (Quit: Pardon me, but I have to go die in NetHack again.). 01:40:29 Shame it exhibits incorrect behavior. 01:45:49 -!- davve123 has joined. 01:47:34 -!- Wiz126 has joined. 01:51:27 -!- BeholdMyGlory has quit (Remote host closed the connection). 01:51:57 -!- ralc has quit (Quit: Leaving). 01:58:14 -!- Patashu has joined. 02:02:37 14:52:38 lets see, p[16] will always end up at 1 at the end I think 02:02:37 14:52:42 and p[10] at 0 02:02:45 unless p[10] was 0 to start with 02:05:03 -!- Aune has quit (Quit: Hath Deprated). 02:24:38 So. I am very, very confused by the behavior of this. 02:25:29 mhm? 02:26:00 ... Never mind, it seems to have... Fixed itself? 02:28:05 -!- pikhq has joined. 02:28:16 http://andlinux.sourceforge.net/screenshots/screenshot.png 02:28:18 Made of win. 02:30:31 Oooor not. 02:31:00 -!- pikhq_ has quit (Ping timeout: 255 seconds). 02:49:15 Now to just figure out why this is emitting superfluous "Set 0 0"s... 02:50:18 superfluous "Set 0 0"s that are superfluous 02:50:58 The thing is, the compiler is outputting Set 0 0 immediately after Set 0 0... 02:51:22 Technically not a *wrong* compilation, just a sign of something weird happening. 02:56:55 Lymia, oooh 02:57:09 "32-bit versions only 02:57:10 Oh, I see. Minor issue in my DCE. 02:57:10 " 02:57:32 Sgeo_, I'd love to set up a.... freakish Frankenstein system once they get it working on 64-bit 02:58:30 How much rage do you expect from Windows fanboys/Linux zealots 02:58:42 Might as well as port it to Mac OS too, so the Mac fanboys can have their fun too. 02:59:03 DCE? 03:00:36 pikhq: DCE? 03:09:54 -!- pikhq_ has joined. 03:10:13 oerjan: Dead code eliminator. 03:11:29 -!- pikhq has quit (Ping timeout: 260 seconds). 03:12:55 1.2 second build. 03:17:45 And the output's better, too! 03:17:57 Towers of hanoi now runs in 0.195 seconds. 03:18:38 And is 38756 bytes instead of 39432... 03:19:00 -!- lameNOT has joined. 03:20:20 wow a fox outside the window 03:21:10 And "only" uses 56M to compile. 03:21:34 Which I'd bet is most courtesy of the list structure being so big. 03:22:21 -!- lameNOT has changed nick to lament. 03:23:55 pikhq_: do you use the unbox-strict-fields option? 03:24:06 oerjan: Not presently. 03:25:01 That... Doesn't actually help much. 03:25:20 it only works for fixed size nonrecursive fields, obviously 03:26:35 hm i don't think it even works for Integer fields (since they are unbounded) 03:27:59 i suppose nothing will help much if you are doing optimizations which requires having the whole structure in memory at once 03:28:04 *require 03:29:05 well, nothing based on lists 03:29:51 I'm using Ints, anyways. 03:30:23 ah 03:30:52 definitely should work for those 03:33:05 Welp, my code generator is using half the time now. 03:33:16 And 44% of the RAM. 03:37:23 Thanks, oerjan, for the order-of-magnitude increase. 03:38:22 yay 03:39:34 Perhaps I should swap State for ST in the code gen... 03:41:04 http://sprunge.us/ZCaP Anyways, bam, there it is. 03:47:16 should there even be any consecutive Move's left after the offset pass... 03:47:36 No. 03:47:49 You'll note that I do not take the fixed point of the optimisations anymore. 03:47:54 aha 03:59:16 http://www.tuxradar.com/files/ubuntu_vs_7-3.png 03:59:25 Wtf does Windows use all that space for? 04:02:10 ... My root filesystem is 6.3G. 04:02:35 *I have a freaking full build system there*. 04:05:32 Does Windows use a DVD or CD? 04:05:42 DVD nowadays. 04:06:04 How the hell do you expand that DVD to such a huge amount of space? 04:06:27 A lot of the stuff on there probably compresses well, I suppose. 04:06:44 Is the problem due to the different system architectures, or what? 04:06:50 They should also compare with a full install of Debian. :P 04:07:04 How large is that on average? 04:07:14 Also, I'm curious. 04:07:24 A *full* install of Debian? Let me put it this way: it comes on 2 Bluray discs. 04:07:24 Is a default Windows installation, or Damn Small Linux more featureful? 04:07:29 pikhq_, hah. 04:08:36 I'm pretty sure DSL is more featureful. 04:09:38 Though perhaps you should look at Puppy. 04:09:44 As it's still maintained... 04:09:55 *Also* more featureful, BTW. 04:40:39 gah, i just woke up. 04:40:50 @note 04:40:50 Missing argument. Check @help for info. 04:42:23 @note Vorpal esotope-bfc does not calculate the fixed point of optimization passes due to the performance, so some cases might be missed by the optimization pass (at the first time). 04:42:23 usage: @vote 04:42:29 ..what? 04:42:36 lifthrasiir: @tell 04:42:46 ah, okay. 04:42:54 @tell Vorpal esotope-bfc does not calculate the fixed point of optimization passes due to the performance, so some cases might be missed by the optimization pass (at the first time). 04:42:54 Consider it noted. 04:43:05 Heh, "noted". 04:43:15 so lambdabot will tell him right after he joins again? 04:43:23 right after he speaks 04:44:04 -!- monqy_ has joined. 04:44:18 -!- monqy has quit (Disconnected by services). 04:44:27 -!- monqy_ has changed nick to monqy. 04:56:38 -!- hagb4rd has joined. 04:59:40 good day! 05:01:00 morning! 05:01:12 -!- pikhq has joined. 05:01:18 -!- pikhq_ has quit (Ping timeout: 255 seconds). 05:10:11 -!- augur has joined. 05:51:00 -!- myndzi has quit (Ping timeout: 240 seconds). 05:51:23 -!- myndzi has joined. 05:54:40 -!- monqy_ has joined. 05:55:34 -!- monqy has quit (Disconnected by services). 05:55:38 -!- monqy_ has changed nick to monqy. 06:26:30 lifthrasiir: Say, you're Korean, aren't you? 06:26:40 lifthrasiir: What're your thoughts on fan death? 06:27:03 (I have no fucking clue what that is in Korean) 06:28:21 I woner. 06:28:24 wonder* 06:28:42 What would happen in your average programming class if you submitted obfuscated or golfed code? 06:28:58 https://secure.wikimedia.org/wikipedia/ko/wiki/%EC%84%A0%ED%92%8D%EA%B8%B0_%EC%82%AC%EB%A7%9D_%EC%82%AC%EA%B3%A0 This is apparently the Wikipedia page... 06:29:07 I think your lecturer would give you the eye brow 06:29:12 Unless he's the kind of person who used to do it 06:29:36 And the title is entirely sensible with the hanja. 扇風機死亡事故. See, entirely sensible. 06:39:50 There is a moth in my room :( 06:41:31 i have this trick involving an empty glass and a sheet of paper 06:41:43 To kill moths 06:41:48 Damn thing keeps touching me 06:42:00 um i practice a catch and release policy _most_ of the time 06:42:41 well i guess it needs to settle down for a moment before you can do it 06:43:13 I'm going to just try to go to sleep, I think 06:43:30 (ants get flushed down, after i got the impression they learn to come back - and bring friends) 06:43:59 Also, the only light in my room is the computer 06:44:10 aha 06:44:18 I think it left my room 06:44:31 No, still here 06:44:36 i thought it was unhealthy to use a computer screen without ambient light... 06:44:52 How so? 06:45:06 harder on the eyes? 06:45:08 Can I touch the thing with gloved hands? 06:45:27 well if you're going to kill it anyhow... 06:45:40 I don't care if it dies or not 06:45:46 i've never heard that moths are particularly poisonous 06:45:57 but who knows what exists in other countries 06:45:59 -!- hagb4rd has quit (Quit: hagb4rd). 06:46:38 oerjan: It's not a fear of being poisonous, just a common overreaction to insects. 06:46:55 (we don't even have any dangerous spiders in norway iafaik) 06:47:03 *-i 06:47:11 yeah right 06:47:30 There are certainly dangerous insects in the US, but not fucking moths. 06:47:58 we had this spider having babies in the bathroom one year, and my housemate (my landlady had rented to a girl this year) freaked out 06:48:14 i just collected them all with a clotheshanger 06:49:06 I'm just going to try to go to sleep 06:49:58 although i don't care much about most insects, you don't have to touch them with the glass and paper method 06:52:07 (1) take an empty glass and a piece of paper (2) catch the bug between the glass and some surface (3) slide the paper between the glass and surface (4) remove from surface carefully not letting it escape (5) do whatever you want 06:52:51 I'm just going to go to sleep, I think 06:53:37 GET AWAY FROM MY COMPUTER 06:53:38 * Sgeo_ cries 06:55:46 I'm afraid if I close my computer I'll get moth blood on it 06:56:25 you can wipe that off, i'm sure 06:57:16 * oerjan feels like a pot calling a kettle black in the rare spot where the pot isn't 06:57:56 hm? 06:58:36 Sgeo_: you are freaking out over a practical matter. just like i do all the time. but not with one i have trouble with. 07:01:25 If I leave my door open and the lights in my room off and the light in the hall on, will it fly out? 07:02:05 i've tried that before, with mixed results 07:04:27 -!- shachaf has quit (Ping timeout: 248 seconds). 07:13:02 -!- shachaf has joined. 07:14:24 Yet again, my battery isn't charging despite the laptop being plugged in 07:17:35 -!- augur has quit (Read error: Connection reset by peer). 07:20:15 -!- augur has joined. 07:26:57 -!- Sgeo has joined. 07:27:13 -!- Sgeo_ has quit (Ping timeout: 246 seconds). 07:33:58 -!- adayleryan has joined. 07:38:08 -!- zzo38 has joined. 07:42:40 -!- adayleryan has left ("Leaving"). 07:52:12 -!- cheater__ has changed nick to retaehc. 07:58:08 The damn thing attacked my mouth 08:00:06 oerjan, I think I injured it but not killed it 08:00:19 It's on the floor pathetically. Should I kill it, would that be merciful? 08:00:44 Your moral dillemae amuse me. 08:00:56 Almost as much as unusual plurals do. 08:01:17 -!- Vorpal has joined. 08:18:51 -!- monqy has quit (Quit: hello). 08:24:23 -!- Lymia has quit (Remote host closed the connection). 08:27:27 -!- retaehc has quit (Ping timeout: 240 seconds). 08:39:27 -!- lament has quit (Ping timeout: 240 seconds). 08:42:06 * oerjan guesses it's too late for mercy now 08:48:32 -!- Phantom_Hoover has joined. 08:49:31 It got under something and I can't reach it now 08:50:43 -!- zzo38 has quit (Remote host closed the connection). 08:56:23 How easy/difficult would it be to parse code of a language into an AST in Javascript? 08:56:51 Meh, guess I should just try it 09:12:13 -!- oerjan has quit (Quit: leaving). 09:18:27 Sgeo: Mozilla did it 09:18:56 I am not a group of people 09:19:09 I guess I should figure out what tokenizing is 09:22:54 -!- MigoMipo has joined. 09:27:24 -!- retaehc has joined. 09:31:48 -!- augur has quit (Remote host closed the connection). 09:44:42 -!- retaehc has changed nick to cheater_. 09:45:41 -!- cheater_ has changed nick to retaehc. 09:51:20 -!- augur has joined. 09:58:47 -!- Slereah has joined. 09:59:08 w. t. f. 09:59:08 Vorpal: You have 1 new message. '/msg lambdabot @messages' to read it. 09:59:18 hm 09:59:59 lifthrasiir, ah okay, my program did go for fixed point. And yet it was faster than your. Heh. 10:00:07 (your program did some things better) 10:00:44 weird, make is removing generated files .c at the end of make all. Why 10:00:49 there is nothing in there to do that 10:01:23 %.c : %.pyx 10:01:23 $(CYTHON) $(CYTHONFLAGS) -a $^ 10:01:23 %.so : %.c 10:01:23 $(CC) -fpic $(CFLAGS) -I$(PYINCLUDES) -L$(PYLIBPATH) $(LDFLAGS) --shared -o $@ $^ $(PYLIB) 10:01:27 I have those two rules 10:01:38 yet it removes the .c at the end. WHY 10:02:10 pikhq, there? Maybe you can help with this. 10:03:20 oh found it 10:03:22 http://www.gnu.org/s/hello/manual/make/Chained-Rules.html 10:03:26 hm, how to avoid that 10:07:24 right, found it 10:25:44 -!- FireFly has joined. 10:40:34 -!- CakeProphet has quit (Quit: leaving). 10:41:39 -!- CakeProphet has joined. 10:42:04 -!- Vorpal has quit (Read error: Operation timed out). 10:42:51 -!- Vorpal has joined. 10:44:22 -!- CakeProphet has quit (Client Quit). 10:44:42 -!- CakeProphet has joined. 10:46:19 -!- CakeProphet has quit (Client Quit). 10:49:15 -!- CakeProphet has joined. 10:49:17 -!- CakeProphet has quit (Changing host). 10:49:17 -!- CakeProphet has joined. 10:49:20 !bf_txtgen test 10:49:26 61 +++++++++++++[>+>+++++++++>++++++++><<<<-]>>-.>---.<-.+.<---. [308] 10:51:08 -!- CakeProphet has quit (Client Quit). 10:51:35 -!- CakeProphet has joined. 10:54:00 -!- CakeProphet has quit (Client Quit). 10:54:24 -!- CakeProphet has joined. 10:56:10 Lists in LSL are immutable. I am going to cry. 10:56:33 Sgeo, what is LSL in this context 10:56:40 Linden Scripting Language 10:56:43 Sgeo: time to implement a state monad, of course. 10:58:29 Sgeo, never heard of that 11:05:48 lifthrasiir, interesting, in esotope-bfc it isn't safe to run all the optimiser passes more than once. Not sure which ones mess up here yet. 11:09:02 Vorpal, the language used in Second Life 11:09:14 I see 11:09:17 CakeProphet, is that even possible in a language without first-class functions? 11:09:54 I don't think so. There'd be no way to do bind. 11:10:15 unless you can eval strings, or have some other method to treat code as data. 11:10:47 CakeProphet, wouldn't eval be treating code as data rather than data as code? 11:10:53 err 11:10:56 the other way around I meant 11:11:17 hmmm 11:11:22 can go both ways depending on what you mean. 11:11:31 hm, okay 11:11:57 CakeProphet, for me it seems that eval treat data (a string for example) as code (by executing it) 11:12:05 I was thinking of the string as the code being stored as data. But yes, it's also data as code in that the string is data that is being re-interpreted as code. 11:12:16 but if you can treat data as code, you can go the other way around by making the bulk of the code your data 11:12:16 yeah 11:13:16 Sgeo: use Perl instead. 11:13:35 lol use perl in SL 11:13:44 actually bind would look pretty nice in Perl. 11:14:54 sub bind(&$) { my ($codeblock, $monad) = @_; ... #monad-specific implementation code goes here } 11:15:33 bind {$_ + 1} $monad_int 11:15:38 -!- BeholdMyGlory has joined. 11:16:32 though it might be better to swap the arguments of bind since the code block will generally be larger in size than the monad argument. 11:18:08 if you did some like. 11:18:21 bind ($[\&&]) {...} 11:18:35 then you could pass a coderef or a literal anonymous block. 11:18:54 ...for fun and profit. 11:20:35 CakeProphet, check CPAN, if there is no monad library for perl yet, make one 11:21:58 -!- NihilistDandy has joined. 11:23:54 Vorpal: surely it's in CPAN... 11:24:06 yep, there it is. 11:24:09 heh 11:24:10 Class::Monad 11:24:18 er Class::Monadic rather 11:24:22 CakeProphet, what about functors? 11:24:40 oh wait this is something different 11:24:45 monadic methods are "singleton" methods 11:24:49 whatever that means. 11:25:53 well Perl doesn't really have anything like typeclasses, I doubt I could make a Monad module since you'd have to implement the code differently for each monad. 11:26:07 but Monad helper functions? possibly. 11:26:43 CakeProphet, hm. Surely you can do this somehow with perl. I mean, can't you execute pretty much anything at parse time 11:26:51 and thus do some sort of magic transform on the code 11:27:01 I seem to remember some ACME:: modules do that 11:27:18 CakeProphet, so I bet you can fake typeclasses in perl even 11:27:19 :D 11:28:11 well it's not like there's a type system or anything... 11:28:27 well, there is, but not a static one. 11:29:06 CakeProphet, hm what about untyped monads... 11:29:10 would that even work? 11:29:36 well it would work like bind/return but without typechecking, basically. 11:29:56 CakeProphet, sounds good enough for perl 11:30:14 and yeah, I do believe you can apply Perl to Perl code before compilation phase. 11:30:17 so, do notation is possible. 11:30:23 CakeProphet, now I feel like I want do do monads in C, probably using some sort of macros to type it 11:30:32 ...sounds horrid. 11:30:39 at least with Perl there would be some nice syntax. 11:31:30 CakeProphet, saw pikhq's C library with closures? Using lots of preprocessor + some GNU extensions 11:31:38 hm did he do continuations as well? I don't remember 11:31:39 I've seen examples that remove the need for semicolons in perl and just append non-empty lines with semicolons via regex. 11:31:53 I don't have the file for it here I think 11:31:58 ask pikhq when he is around 11:33:30 I guess a Monad module for Perl would just be a class that you could inherit from, that would standardize the names and overload some operators 11:34:25 ...I wonder, if you can overload operators and still do the anonymous {} notation that you can get if use & in the subroutine prototype. 11:34:46 $monad >= {...} 11:34:58 I doubt it. 11:35:16 try it? I have no idea, I'm not very good at perl. 11:35:29 no I think it has to be the original sub by its original name for the parser to know what to do. 11:36:45 to overload operators you have to do something like: use overload ('>=' => \&bind, '+' => \&empty_bind); 11:38:33 -!- wareya_ has joined. 11:38:54 but making a do syntax for Perl sounds like a fun time with regex. 11:40:20 turning a linear list of program lines into a single line of nested binds? I can't wait. 11:41:05 and it's Perl so I'm not allowed to use anything other than regex to do that. 11:41:55 -!- wareya has quit (Ping timeout: 260 seconds). 11:42:43 heh 11:45:31 -!- Phantom_Hoover has quit (Ping timeout: 246 seconds). 12:03:15 -!- BeholdMyGlory has quit (Remote host closed the connection). 12:03:31 Good idea/bad idea: Attempting to compile some code by hand without knowing what the compiler looks like exactly, just a vague idea 12:03:46 I want to get a feel for how it should work 12:16:15 DANGIT I DON'T WANNA WRITE MY OWN GARBAGE COLLECTION 12:19:03 Um, I'm not sure if I need GC... although the fact that I'm wondering that makes me rather nervous about my semantics 12:20:26 Oh, I figured out why. There is no heap. 12:20:31 How worried should I be about this 12:20:32 ? 12:28:18 Oh look, there are mutable structures in LSL. But they can only hold floats 12:40:09 If I detect that call/cc is not used, I can skip the CPS transform! 12:40:57 (inb4 halting) By "not used", I mean not mentioned in the source code 12:52:31 -!- malorie has joined. 13:09:44 -!- pikhq_ has joined. 13:11:28 -!- pikhq has quit (Ping timeout: 244 seconds). 13:47:52 -!- oerjan has joined. 13:54:15 well Perl doesn't really have anything like typeclasses, I doubt I could make a Monad module since you'd have to implement the code differently for each monad. 13:54:50 you might use the scala (iirc) method of letting monad elements be objects of a class that implements bind as a method for its first argument 13:55:07 with a different class for each monad, naturally 13:56:09 -!- NihilistDandy has quit (Quit: *yawn* Breakfast.). 13:56:30