←2007-09-17 2007-09-18 2007-09-19→ ↑2007 ↑all
00:04:26 -!- ehird` has quit.
00:16:26 <lament> it's hard for compiler and interpreter to be the same word.
00:58:33 -!- sebbu2 has quit ("@+").
02:17:46 -!- edwardk has joined.
03:58:56 -!- edwardk has left (?).
04:22:59 <bsmntbombdood> CLOACAL IMMUREMENT
04:39:34 <oerjan> i generally try to avoid that.
04:40:06 <bsmntbombdood> good, that will probably extend your life
04:42:30 <bsmntbombdood> but it sure would make a good story, no?
04:43:28 <oerjan> no.
04:43:53 <bsmntbombdood> no to your no
06:19:03 -!- oerjan has quit ("leaving").
07:07:32 -!- rutlov has joined.
07:19:32 -!- rutlov has left (?).
07:23:42 <lament> stealing matlab is hard
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:13:47 -!- Guilt has joined.
08:59:47 -!- helios24 has quit (Read error: 104 (Connection reset by peer)).
09:01:34 -!- helios24 has joined.
09:34:53 -!- sebbu has joined.
09:59:38 -!- Guilt has quit ("CGI:IRC at http://freenode.bafsoft.ath.cx:14464/ (EOF)").
10:28:57 -!- ehird` has joined.
10:38:32 -!- helios24 has quit (Read error: 60 (Operation timed out)).
10:39:50 -!- SEO_DUDE55 has quit (Read error: 104 (Connection reset by peer)).
10:43:33 -!- SEO_DUDE55 has joined.
11:30:40 -!- SEO_DUDE55 has quit (Remote closed the connection).
11:41:39 -!- oerjan has joined.
11:41:51 -!- puzzlet_ has quit (Remote closed the connection).
11:41:56 -!- puzzlet has joined.
12:58:00 -!- SEO_DUDE101 has joined.
13:20:59 -!- jix has joined.
14:35:59 <oerjan> <CTCP>ACTION tests<CTCP>
14:36:13 * oerjan tests
14:36:25 <oerjan> ?
14:36:57 <oerjan> ah.
14:39:45 <ehird`> \1ACTION tests\1, you mean
14:39:51 <ehird`> ctcp :p
14:43:20 <oerjan> huh? the second was right wasn't it?
14:45:58 -!- jix has quit (Nick collision from services.).
14:46:10 -!- jix has joined.
14:48:31 <oklopol> yeah!
14:48:50 <oklopol> so right it almost peeked out from the left already
15:30:20 -!- helios24 has joined.
16:38:09 -!- oerjan has quit ("leaving").
17:04:43 -!- bsmnt_bot has joined.
17:16:50 -!- bsmnt_bot has quit (Read error: 145 (Connection timed out)).
18:37:19 -!- jix has quit (kornbluth.freenode.net irc.freenode.net).
18:37:19 -!- Overand has quit (kornbluth.freenode.net irc.freenode.net).
18:37:19 -!- Chton has quit (kornbluth.freenode.net irc.freenode.net).
18:37:19 -!- Eidolos has quit (kornbluth.freenode.net irc.freenode.net).
18:38:01 -!- jix has joined.
18:38:01 -!- Overand has joined.
18:38:01 -!- Chton has joined.
18:38:01 -!- Eidolos has joined.
18:53:22 * SimonRC has dinner.
19:42:13 -!- Figs has joined.
19:42:22 <Figs> ah...
19:42:24 <Figs> don't you just love operator abuse?
19:42:26 <Figs> http://rafb.net/p/YGBwxZ75.html
19:43:50 -!- sebbu2 has joined.
19:56:05 <oklopol> hehe clever
19:56:36 <Figs> it's pretty basic
19:56:42 <Figs> I could do a better job with more time
19:56:51 <Figs> but I was just fed up with syntax tree generators yesterday :P
19:57:12 <Figs> if I keep doing this shit, eventually I'll have a full lisp-interpreter in C++ .... >.>
19:57:18 * GregorR watches Figs turn C++ into LISP
19:57:18 <oklopol> closer to lisp than your earlier parenthesis one, was my point
19:57:24 <GregorR> Whoops, spoke to late :P
19:57:24 <oklopol> ...
19:57:30 <GregorR> *too
19:57:33 <Figs> :p
19:57:41 <oklopol> i recall us once before saying the same thing at almost the same second
19:57:49 <Figs> lol
19:58:02 <oklopol> actually
19:58:10 <oklopol> you almost said that at the same second too :P
19:58:30 <Figs> #define node(s) tree(s) isn't really needed... I was just feeling lazy :P
19:58:33 <GregorR> Now, Plof3 will be trivially able to be turned into LISP
19:58:54 <oklopol> ah you added that syntax thing
19:58:57 <Figs> I don't really know enough lisp to convert the whole thing
19:59:04 <GregorR> oklopol: Yup 8-D
19:59:09 <GregorR> Runtime-defined grammar! Weeeh
19:59:11 <oklopol> i don't like that.
19:59:17 <Figs> but supposedly I only need about 7 functions...?
19:59:18 <oklopol> you know what the reason is?
19:59:29 <Figs> (according to (paul graham))
19:59:35 <oklopol> I WAS GONNA DO THAT FOR OKLOTALK YOU MIND THIF
19:59:37 <oklopol> *THIEF
20:00:25 <oklopol> lisp needs if + define + lambda + a few operators
20:00:53 -!- ololobot has joined.
20:01:01 <oklopol> >>> (+ 1 1)
20:01:04 <oklopol> >>> sch (+ 1 1)
20:01:05 <ololobot> 2
20:02:03 <oklopol> GregorR: is grammar first-class?
20:02:23 <oklopol> can you pass it for the ultimate obfuscation
20:02:27 <oklopol> *pass it around
20:02:54 -!- sebbu has quit (Success).
20:03:07 -!- Figs has left (?).
20:17:05 -!- Tritonio has joined.
21:04:02 <GregorR> s
21:04:14 <GregorR> Why do I keep typing 's' :P
21:04:47 <GregorR> No, grammatical elements are not first class ... you could build a grammar which has constructs which themselves resolve to grammatical actions, and then pass around those constructs.
21:08:16 <GregorR> That is, grammatical elements are defined entirely internally and committed by operations in the stack code, but you could make an object which encapsulates that stackcode operation in a defined way.
21:10:45 <GregorR> (FYI, the stackcode itself is not directly accessible from user code, you need to define a grammar for it)
21:44:45 -!- jix has quit ("CommandQ").
22:00:09 <ehird`> who wants to help me write the ultimate bf compiler i've had in mind for a while now?
22:00:16 <ehird`> it'll do BF->C, and some really heavy optimization
22:03:07 <ehird`> hopefully it'll be the most complete BF compiler out there
22:03:46 -!- RedDak has joined.
22:04:46 <ehird`> =)
22:11:47 -!- sebbu2 has changed nick to sebbu.
22:16:56 <ehird`> <crickets chirp, nobody cares>
22:43:06 <bsmntbombdood> everyone wants to do an optimizing bf compiler
22:47:07 <ehird`> yes, except i've had ideas for this one in my mind for a while now
22:50:59 <bsmntbombdood> like?
22:51:12 <ehird`> like some various optimization techniques that i have not seen before
22:52:33 <bsmntbombdood> like?
22:52:40 <ehird`> like x[x] -> dowhile loop
22:52:43 <ehird`> (i haven't seen that before)
22:53:10 <bsmntbombdood> that's not any faster
22:56:16 <ehird`> it's still a reasonable optimization
22:56:20 <ehird`> and helps with code size
22:58:02 <bsmntbombdood> the obvious one is to pre execute all the code up to the first input
22:58:16 <ehird`> which is actually a really sucky optimization
22:58:25 <ehird`> what about infinite loops?
22:58:29 <ehird`> if you put a timelimit,
22:58:37 <ehird`> what if my code takes 5 minutes to execute but isn't an infinite loop?
22:58:47 <ehird`> do i get different, unoptimized code jsut because of the system i'm on is slow?
22:59:00 <bsmntbombdood> not a timelimit, an instruction limit
22:59:09 <ehird`> still ridiculous
22:59:21 <ehird`> just because my code is a certain way it gets compiled in a different, perhaps slower way
22:59:27 <bsmntbombdood> duh...
22:59:35 <bsmntbombdood> any optimization is like that
22:59:46 <bsmntbombdood> what about x<>[x]
22:59:53 <ehird`> also, i don't want my factorial program that calculates the factorial of 10 to be executed at compiletime
22:59:57 <bsmntbombdood> oops, your optimization fails
23:00:01 <ehird`> and no
23:00:12 <ehird`> the x[x] optimization will be run after the useless-instruction removal
23:00:20 <ehird`> "<>" = "", so it'd be x[x] again
23:02:27 <bsmntbombdood> just an example
23:02:41 <bsmntbombdood> there's other ways to make it the same but unrecognizable
23:02:41 <ehird`> still
23:02:59 <ehird`> btw, the look-ahead and code-pattern-matching codes would be pretty advanced
23:03:08 <ehird`> it'd be damn hard to make e.g. the x[x] optimization fail
23:04:25 -!- pikhq has quit (Remote closed the connection).
23:04:41 -!- pikhq has joined.
23:15:37 -!- ehird` has quit.
23:52:22 -!- RedDak has quit (Remote closed the connection).
←2007-09-17 2007-09-18 2007-09-19→ ↑2007 ↑all