00:00:01 <ehird> so not attuned to makefile-and-.o ways
00:00:06 <AnMaster> ehird, do you trust Intel to get it right
00:00:06 <ehird> "linux compilation file for unix"; fix that
00:00:12 <AnMaster> Let me just state two examples:
00:00:26 <ehird> AnMaster: because intel are the only company that makes mistakes
00:00:29 * oerjan put links to his BCT interpreter in /// on the wiki
00:00:32 <ehird> want me to dig up a bunch of AMD gaffs?
00:00:39 <AnMaster> ehird, of course AMD made ones too
00:00:54 <AnMaster> I'm just saying I don't trust Intel to get it right. Nor AMD
00:01:05 <ehird> as long as it has a power supply I can yank out
00:01:13 <AnMaster> <ehird> clearly yudkowsky was not properly involved <-- and after googling I don't understand the reference
00:01:27 <ehird> Eliezer Yudkowsky, Research Fellow at the Singularity Institute for Artificial Intelligence.
00:01:34 <ehird> "Net income▼ $ -3.098 billion (2008) " — [[AMD]]
00:02:01 <AnMaster> ehird, Net income ▼ US$5.3 billion (2008) - [[Intel]]
00:02:16 <ehird> errrrrrrrrrrrrrrrrrrrrrrr
00:02:34 <ehird> AnMaster: Intel get $5.3 billion. AMD lose $3 billion.
00:02:41 <ehird> And down arrow = down from last year.
00:02:58 <ehird> So AMD's losing $3bn, Intel's gaining $5.3bn
00:03:03 <ehird> Not very ouch at all.
00:03:26 <ehird> AnMaster: yes, that just dropped it a billion or so
00:03:29 <ehird> still profiting quite a bit
00:04:17 <AnMaster> http://en.wikipedia.org/wiki/Intel#European_Union
00:04:32 <ehird> AnMaster: I am well aware of the fine.
00:04:43 <ehird> They make good products, I buy the products; I get my nice processor, they get money.
00:04:49 <AnMaster> ehird, support suspect marketing tactics?
00:05:05 <ehird> AnMaster: did you buy your microsoft keyboard/mouse?
00:05:12 <ehird> Why are you supporting their suspect marketing tactics?
00:05:16 <ehird> HOW FUN!11128126738213
00:05:22 <AnMaster> ehird, I don't have a Microsoft keyboard
00:05:36 <AnMaster> ehird, and no, I got it as a present
00:05:46 <ehird> so you'd refuse to buy it
00:06:37 <AnMaster> ehird, I'd probably test other ones first... But if it was the only one that worked ergonomically I'd probably use it
00:06:53 <ehird> So you'd buy a product from Microsoft, giving them money, "supporting suspect marketing tactics".
00:07:01 <ehird> And this does not make you a hypocrite how?
00:07:54 <AnMaster> ehird, I wouldn't buy an OS or a CPU though. Anyway I happen to know some Logitech mice are quite nice
00:08:12 <AnMaster> but if it did I would indeed be in a tricky dilemma
00:08:18 <ehird> CPUs are a magical thing. They're entirely different to mice in the context of a market, because *~SPARKLES!~*
00:08:28 <bsmntbombdood> i can't find documentation of sse functions for gcc
00:08:45 <AnMaster> ehird, yeah... And AMD ones work very well in my experience.
00:09:07 <ehird> AnMaster: So you seriously think that somehow, buying a CPU supports shady marketing, but buying a mouse doesn't, even though they both feed back into the company profits.
00:09:19 <ehird> You, sir, are on crack, and only attempting to justify how you would behave while criticizing me because you don't like the company.
00:09:52 <AnMaster> ehird, I certainly think both do
00:10:08 <ehird> AnMaster: But one is acceptable support of shady marketing practices?
00:10:20 <ehird> Shovel, meet AnMaster. Hole, meet shovel. Dig, dig, dig.
00:10:23 <AnMaster> ehird, but when there are no alternatives left that work...
00:10:34 <ehird> AnMaster: There is no AMD processor that can match the performance of an i7.
00:10:36 <AnMaster> you are sometimes forced to take the least bad action
00:10:44 <ehird> Therefore, there is no alternative left that even exists.
00:10:46 <AnMaster> ehird, then ask yourself: do you actually need that peformance
00:10:56 <ehird> AnMaster: I do not need anything but food, water, warmth, ...
00:11:10 <ehird> But what would I *like*? Yes, I'd like that performance.
00:11:11 <AnMaster> ehird, how much power does an i7 use btw
00:11:23 <ehird> I think that the EU fine is enough punishment for Intel.
00:11:42 <ehird> AnMaster: At maximum, around ~380 watts. At idle, about 100 watts.
00:11:56 <ehird> By the way, AnMaster? The Phenoms (AMD's high-end processors) suck up a ton of power.
00:12:02 <ehird> The i7s use way less juice.
00:12:13 <ehird> AnMaster: not just the processor
00:12:15 <ehird> that is the whole system
00:12:26 <ehird> the whole processor is 140W at absolutely maximum load, I believe.
00:12:53 -!- Dewi has joined.
00:12:54 <ehird> Compared to the Phenoms, it's so green your red and blue subpixels will stop functioning.
00:13:14 <AnMaster> ehird, same. I used one of those watt meter thingies on my system some time ago. When idle with ondemand CPU scheduler it was about 88-91W
00:13:37 <ehird> AnMaster: So... it's about as green as your low end CPU.
00:13:38 <AnMaster> was about 2 years ago I measured
00:13:40 <ehird> I think that's just fine.
00:14:16 <ehird> nooga: Why didn't you hardcode the standard raytracing example?
00:14:45 <nooga> ehird: ah, it's WIP
00:14:55 <ehird> AnMaster: Well, your system is still only a few dozens of watts less than mine at idle.
00:15:00 <ehird> And I'm using an 80+ efficient CPU, etc.
00:15:10 <ehird> (This is hypothetical; I cba to actually measure it when I get it.)
00:15:17 <nooga> i planned to make a raytracer with multisampling, reflections etc. and write a scene compiler in sadol
00:15:19 <ehird> (But I think I'm approximately on mark.)
00:15:45 <nooga> or maybe... monte-carlo path tracer ;D
00:16:15 <nooga> but after i finish my sadol compiler
00:16:58 -!- nooga has quit ("Leaving...").
00:17:17 <ehird> AnMaster: Do you think techniques applied to BF compilation could be applied to underload / unlambad?
00:18:01 <AnMaster> ehird, don't know enough about them to be sure
00:18:35 <AnMaster> stuff like constant folding and dead code elimination is pretty general
00:19:05 <AnMaster> stuff like balanced loops and such: well I don't know enough about ul and ul to be sure
00:19:46 <AnMaster> I'll resume the discussion once you worked out better acronyms!
00:20:15 <ehird> AnMaster: ul = unlambda
00:20:25 <ehird> oerjan: which way around?
00:22:05 <AnMaster> ehird, and I really don't know unl well enough to know what you could do in it
00:22:22 <ehird> it's quite similar to bf
00:22:24 <AnMaster> ul, well I don't know either exactly what would work
00:23:01 <ehird> let's say underload then
00:23:07 <ehird> it's quite similar to bf
00:23:18 <ehird> AnMaster: you can do a bit
00:23:27 <AnMaster> ehird, maybe. But a lot of the stuff in bf optimising is "figure out where the damn pointer is"
00:23:28 <ehird> but nobody would write code like that
00:23:44 <ehird> AnMaster: in this case, it'd be "figure out how large the damn stack i"
00:23:55 <AnMaster> well you could try to track that
00:24:09 <AnMaster> possibly relative some prior point
00:24:31 <AnMaster> like you lost track after something that you can't easily decide
00:24:39 <AnMaster> which for bf would be unbalanced loop
00:24:44 <AnMaster> not sure what it would be for ul
00:24:50 <ehird> AnMaster: pushing more than you pop
00:24:59 <ehird> ~ pops 1 (since it combines two elements)
00:25:30 <ehird> * pops 1, and combines
00:25:31 -!- Dewio has quit (Read error: 101 (Network is unreachable)).
00:25:37 <ehird> AnMaster: ~ swaps top two elements
00:25:40 <ehird> so no stack change
00:25:54 <AnMaster> ehird, is there anywhere you can't easily figure out the stack sise
00:26:02 <ehird> AnMaster: yes -- ^
00:26:04 <ehird> you can do, for instance:
00:26:10 <AnMaster> ehird, you can know it sometimes for ^
00:26:23 <ehird> but you can infer that too
00:26:32 <ehird> in fact, I'm not sure you can ever not infer it
00:26:44 <ehird> I might write an inferrer to see
00:26:49 <AnMaster> ehird, are there cases you can't infer
00:26:54 <ehird> AnMaster: i'm wondering
00:26:58 <AnMaster> when does this cross the line to compiler
00:27:11 <AnMaster> with compiling to a "output constant string"
00:27:18 <ehird> AnMaster: well, you'd never actually loop
00:27:22 <ehird> to infer the stack effect
00:27:34 <ehird> AnMaster: this wouldn't work if underload had user input, BTW
00:28:00 <AnMaster> you can't possibly infer there
00:28:51 -!- nooga has joined.
00:28:57 <ehird> AnMaster: no, you can:
00:29:15 <oerjan> ehird: it is easy to push an amount corresponding to a church numeral
00:29:22 <ehird> AnMaster: stack effect of :^ is x=[y x [y x → z] → z], where [foo] is a quotation with stack effect foo
00:29:29 <AnMaster> that would be a nice optimisation
00:29:32 <ehird> where x=foo means "foo can mention x"
00:29:43 <AnMaster> into some optimised representation
00:29:50 <ehird> ( x=[y x [y x → z] → z] )
00:29:55 <ehird> ( x=[y x [y x → z] → z] ) ( x=[y x [y x → z] → z] )
00:30:01 <ehird> and apply the types
00:30:07 <ehird> AnMaster: it just leads to an infinite type
00:30:13 <ehird> which could be handled, with enough care
00:30:15 <ehird> oerjan: do you think so?
00:30:40 <ehird> nooga: inferring underload stack effects
00:30:48 <ehird> oerjan: so we can always infer the stack effect of an underload program in finite time?
00:30:50 <ehird> that's really nice
00:31:01 <AnMaster> ehird, couldn't you optimise it into a bignum or whatever. I mean some representation that is fast (relatively to original) on a real computer
00:31:14 <EgoBot> Supported commands: addinterp bf_txtgen bfjoust daemon daemons delinterp fyb help info kill mush userinterps 1l 2l adjust asm axo bch bct befunge befunge98 bf bf16 bf32 bf8 bfbignum boolfuck c chiqrsx9p choo cintercal clcintercal cxx dimensifuck echo forth glass glypho google hello kipple lambda lazyk linguine malbolge ook pbrain perl qbf rail rhotor rot13 sadol sceql sh show slashes test trigger udage01 underload unlambda whirl yodawg
00:31:26 <ehird> oerjan: the opposite— why?
00:31:39 <ehird> but I want to go further
00:31:51 <ehird> oerjan: I don't see why you can't just handle recursive types like (:^), and thus infer any program
00:32:03 <oerjan> ehird: because you can calculate any recursive function it is impossible to calculate a church numeral for sure in finite time
00:32:07 <AnMaster> ehird, for UL the difference between interpreter and optimising compiler seems very very small
00:32:13 <ehird> oerjan: i'm not talking about church numerals‽‽
00:32:26 <AnMaster> ehird, just one generates a program that outputs the same thing
00:32:27 <ehird> oerjan: I'm talking about inferring the stack effect of underload programs
00:32:29 <oerjan> ehird: i was using them to prove you cannot do it
00:33:19 <oerjan> if you have a church numeral on the stack, then ((x))~^^ creates that number of (x) on the stack
00:34:20 <ehird> oerjan: you'd think once you infer (:^):^ properly everything else would be trivial :)
00:34:44 <ehird> oerjan: OTOH, isn't it possible to infer in a large number of cases?
00:35:33 <AnMaster> can't you infer from church numerals? I mean if you constant fold them. into some (church 4) representation.
00:35:54 <ehird> oerjan: explain to him properly :P
00:36:07 <oerjan> ehird: well sure it's probably quite analogous to bf balanced vs. unbalanced loops
00:36:23 <AnMaster> I meant, for many common cases
00:36:30 <oerjan> AnMaster: well sure if they're constants
00:36:46 <AnMaster> oerjan, yes of course you could build new ones
00:36:56 <AnMaster> and then you couldn't in general
00:37:37 <AnMaster> is optimising unbalanced loops
00:38:02 <AnMaster> like constant propagation to put a upper bound on the range of a [>]
00:38:21 <AnMaster> where I mean the constant propagation has the needed data sometimes
00:38:37 <ehird> AnMaster: does your compiler optimize >>>>>+<<<<<[>]+ into p[5] = 2;?
00:38:45 <AnMaster> ehird, not yet. But it is planned
00:38:58 <AnMaster> ehird, is 6th item in my private TODO
00:39:23 <AnMaster> ehird, currently in-between sometimes does better, and sometimes esotope does better
00:39:37 <AnMaster> often on different parts of the same program
00:40:05 <AnMaster> ehird, I would have finished polynoms today, but this headache got in the way
00:48:17 -!- oerjan has quit ("Good night").
00:50:25 <ehird> i love sse instructions
00:56:35 -!- inurinternet has quit (No route to host).
00:57:14 <nooga> porting MFC to iphone O_o
00:57:59 <coppro> that's the ABSOLUTE DUMBEST THING I have EVER heard
00:58:44 <ehird> coppro: he's porting some windows-shit application to the iphone
00:58:48 <ehird> and, um, apparently he's not porting it?
00:58:58 <ehird> nooga: i think you should go and choke on something before you do any more damage
00:59:04 <coppro> MFC is one of the worst libraries ever written
00:59:15 <coppro> it should not have been ported to Windows, much less the iPhone
00:59:50 * coppro goes and kills somebody
01:00:06 <ehird> coppro: don't you think you're overreacting a little
01:05:45 <ehird> http://en.wikipedia.org/wiki/Microsoft_Foundation_Class_Library
01:07:14 <pikhq> (and I mean no offense... To shit.)
01:07:33 <coppro> the wikipedia article failes to properly represent it
01:08:00 <ehird> *typeos, I suppose
01:08:05 <pikhq> I'm especially curious why one would use C++ on the iPhone.
01:08:05 <nooga> but i'm porting massive library that strongly depends on MFC and other M$ shit
01:08:23 <pikhq> You've got Objective C as the normal language.
01:08:24 <nooga> and we decided to port some parts of MFC
01:08:25 <ehird> nooga: you're evilorrible
01:08:28 <pikhq> nooga: Rewrite it.
01:08:31 <ehird> please jump out of a window
01:08:37 <nooga> it was written for 3 years
01:08:57 <pikhq> nooga: 'Porting' MFC (reimplementing it) will take more time.
01:09:32 <nooga> just some core classes
01:09:46 <nooga> like CString CTime CSize CTimeSpan... etc etc
01:11:26 <ehird> Now think about what you've done!
01:11:38 <nooga> it's just my job lol
01:11:59 <nooga> i try to keep cool abt that
01:12:10 <nooga> besides, i never used mfc
01:12:46 <pikhq> I'd sooner rewrite the STL.
01:12:58 <pikhq> And the STL makes me want to gouge my eyes out.
01:25:51 <ehird> http://plasmasturm.org/log/542/
01:41:49 -!- Gracenotes has quit (Read error: 60 (Operation timed out)).
01:48:53 -!- olsner has joined.
01:50:02 -!- atrapado has joined.
01:54:44 -!- Gracenotes has joined.
02:00:35 -!- atrapado has quit ("WeeChat 0.2.6.2").
02:15:19 -!- olsner has quit ("Leaving").
02:18:24 -!- Gracenotes_ has joined.
02:21:08 -!- Gracenotes has quit (Nick collision from services.).
02:21:11 -!- Gracenotes_ has changed nick to Gracenotes.
02:23:23 <Gracenotes> would calling my ISP about my fucking slow connection do anything?
02:23:52 <Gracenotes> maybe I'll talk to my parents about pitching in for getting a better rate, if it's low atm, just for the summer
02:26:12 <ehird> "To pull something from L1 cache is like looking at something already on your desk. L2 cache is like asking the guy in the cube next to you. L3 cache is like driving to the other side of town and coming back with your answer. Main memory is like writing a letter to someone in Taiwan and waiting for the response."
02:44:18 <pikhq> bsmntbombdood: Accessing the disk is like ordering something from a mail-order catalog.
03:03:00 <bsmntbombdood> ok, i think i kind of know how to do a sort of odd/even sort with sse
03:25:18 -!- inurinternet has joined.
03:40:39 -!- pikhq has quit (Read error: 54 (Connection reset by peer)).
03:42:18 -!- pikhq has joined.
03:45:14 -!- inurinternet has quit (Success).
03:45:27 -!- bsmntbombdood has quit (Remote closed the connection).
03:52:07 -!- inurinternet has joined.
03:57:08 -!- pikhq has quit ("leaving").
04:11:29 -!- bsmntbombdood has joined.
04:12:55 -!- inurinternet has quit (Read error: 110 (Connection timed out)).
04:23:15 <bsmntbombdood> i am backing up data from my old drive...it seems that bzip2 is the bottleneck
04:28:12 -!- coppro has quit (Remote closed the connection).
05:07:52 <GregorR-L> And downloading something from the network is like what, reading smoke signals from Andromeda? :P
05:08:24 <bsmntbombdood> a local network can have a lot less latency than an hdd
05:13:19 <Gracenotes> okay. Here I was thinking that Firefly was okay so far, but... it suddenly got much much better
05:15:26 <bsmntbombdood> Gracenotes: pi = 3.14152625 after 1000000000 iterations
05:15:56 <Gracenotes> awesome! ... I still haven't gotten around to implementing it, darn it
05:16:46 <Gracenotes> forgot, lambdabot not in here >_> Haskell not esoteric enough
05:16:50 <bsmntbombdood> they are 64 bit ints from some mersenne twister implementation i pulled of the web
05:16:52 <Gracenotes> I think I meant to do logBase 10, too.
05:19:08 <Gracenotes> how much more do you think another billion would help?
05:19:53 <GregorR-L> ghci is GHC /interactive/, not GHC /interpreter/ :P
05:21:13 <Gracenotes> threads can't do much if there's one CPU context-switching all the time :)
05:22:57 <GregorR-L> You'd think you could do better with 2 billion iterations ....
05:24:07 <GregorR-L> My brain, after 1 iteration, pi = 3.14159265358979323
05:25:26 -!- Corun has quit ("Leaving...").
05:25:58 <Gracenotes> you might also have a limit with the floating point itself
05:26:19 <Gracenotes> interestingly, you could also try to predict how many out of the 10000000000 would be coprime
05:26:32 -!- GregorR-L has set topic: The interrational tub of esoteric programming language designers and soap. No soap, radio! Pass the http://tunes.org/~nef/logs/esoteric/.
05:27:20 <Gracenotes> which Haskell kindly tells me is 6079271019
05:29:05 <Gracenotes> I'd imagine a Haskell solution could be nicely optimized
05:29:43 <Gracenotes> it's not a difficult algorithm. n times, check GCD of two random large numbers; increment a counter if it's 1....
05:30:05 <Gracenotes> if we unbox strict field, or as I like to call it, funboxing. :/
05:30:59 <Gracenotes> in Haskell, there are primitive types like Int#
05:31:22 <Gracenotes> actually, no, it doesn't have to do with primitive types
05:31:42 <Gracenotes> or does it? Either way, it removes the constructor if it can
05:31:59 <Gracenotes> bsmntbombdood: you might be surprised :)
05:32:09 <bsmntbombdood> wait there's a better gcd algorithm than euclid or whatever's right?
05:33:03 <Gracenotes> bsmntbombdood: well... I'm sure there are more efficient ways to check if two numbers are coprime
05:33:19 <Gracenotes> Euler's algorithm is the most efficient GCD-taker I know of though
05:34:29 <bsmntbombdood> gonna go for another order of maginitude, i'll have a result in 90 minutes
05:34:36 -!- pikhq has joined.
05:45:04 <Gracenotes> bsmntbombdood: how many digits you listed is based on a value of pi you already have?
06:01:59 <bsmntbombdood> 23:01:43 up 1:53, 30 users, load average: 8.13, 8.00, 8.00
06:03:59 -!- inurinternet has joined.
06:10:13 <bsmntbombdood> gprof doesn't work will with multiple processes :(
06:20:12 <bsmntbombdood> ouch, looks like most of my time is spent calculating gcds
06:22:24 <Gracenotes> right. not much I can find on the internet about more efficient means of checking coprime, though
06:22:36 <Gracenotes> you're using the iterative version, of course?
06:22:49 <bsmntbombdood> i was thinking that generating the random numbers would be slower
06:23:08 -!- inurinternet has quit.
06:23:26 -!- inurinternet has joined.
06:23:51 <bsmntbombdood> the completely trivial one though--http://pastebin.ca/1429617
06:26:30 <Gracenotes> here is a possible optimization, a precheck mentioned: http://echochamber.me/viewtopic.php?f=12&t=22215
06:26:40 <bsmntbombdood> looks like there are algorithms that are asymptotically better, but n is relatively small here
06:27:57 <Gracenotes> well, we could use numbers that big. Would need special facilities though
06:28:21 <Gracenotes> if all you're doing is generating them to check if they're coprime, after all. Still, would be slower
06:28:56 <Gracenotes> anyway. gcd is efficient... enough. Some checks at the beginning (both even?) might help a bit, perhaps
06:29:05 <bsmntbombdood> http://www.springerlink.com/content/7042685247876171/
06:30:24 <Gracenotes> if any binary digits on the lower end are the same, they're not coprime
06:31:22 <Gracenotes> argh. But yeah, those come rarer for more and more 0s
06:34:32 <Gracenotes> hm... sort of a long-shot, but try making the function inline?
06:34:45 <Gracenotes> then again if you already have -O3 or suchlike it probably already is
06:43:39 <Gracenotes> well. anyway, the probabilistic algorithm would seem interesting
06:43:59 <Gracenotes> it might introduce a systemic bias in the pi calculation though
07:15:42 <Gracenotes> Perhaps you need higher numbers than 2^64-1 for better accuracy?
07:16:55 <Gracenotes> http://en.wikipedia.org/wiki/Coprime#Probabilities
07:17:14 <Gracenotes> as N -> infinity, the probability goes to 6/pi^2
07:18:25 <Gracenotes> there must be some way to calculate how many digits you can get from an N
07:18:50 <Gracenotes> related directly to the Riemann zeta function, particularly
07:19:28 <Gracenotes> bsmntbombdood: slower yet, but perhaps better results :D
07:20:33 <bsmntbombdood> i can't find the mpz function to convert from a bit array
07:27:45 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
07:46:37 -!- BeholdMyGlory has joined.
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:20:02 -!- Slereah has joined.
08:21:06 -!- upyr[emacs] has quit (Remote closed the connection).
08:29:35 -!- Slereah_ has quit (Read error: 110 (Connection timed out)).
08:41:34 -!- BeholdMyGlory_ has joined.
08:43:38 -!- M0ny has joined.
08:57:51 -!- BeholdMyGlory has quit (Read error: 110 (Connection timed out)).
09:02:06 -!- olsner has joined.
09:18:17 -!- Gracenotes has quit (Read error: 104 (Connection reset by peer)).
09:18:50 -!- Gracenotes has joined.
09:43:36 -!- BeholdMyGlory has joined.
09:59:09 -!- BeholdMyGlory_ has quit (Read error: 110 (Connection timed out)).
10:26:38 -!- bsmntbombdood has quit (Remote closed the connection).
10:46:04 -!- oerjan has joined.
10:47:59 -!- tombom has joined.
10:56:55 <EgoBot> Supported commands: addinterp bf_txtgen bfjoust daemon daemons delinterp fyb help info kill mush userinterps 1l 2l adjust asm axo bch bct befunge befunge98 bf bf16 bf32 bf8 bfbignum boolfuck c chiqrsx9p choo cintercal clcintercal cxx dimensifuck echo forth glass glypho google hello kipple lambda lazyk linguine malbolge ook pbrain perl qbf rail rhotor rot13 sadol sceql sh show slashes test trigger udage01 underload unlambda whirl yodawg
11:00:51 <oerjan> !c printf("%c", log(1000000000.0));
11:01:04 <oerjan> !c printf("%f", log(1000000000.0));
11:03:35 -!- olsner has quit ("Leaving").
11:04:52 <oerjan> <GregorR-L> ghci is GHC /interactive/, not GHC /interpreter/ :P
11:07:51 <oerjan> however, that takes an entire module iiuc
11:08:07 <oerjan> (not that adding main = is that much trouble)
11:08:54 <oerjan> GregorR-L: oh, also ghc -e
11:16:26 <oerjan> <bsmntbombdood> 10000000000 iterations, pi = 3.1415826392666
11:16:48 <oerjan> the precision may depend not just on the number of iterations, but also on the sizes of the numbers you test
11:18:22 <oerjan> the (probability of gcd = 1) being p^2/6 is a limit as number range size goes to infinity, after all
11:19:40 -!- Slereah_ has joined.
11:22:13 <oerjan> it comes from that product over an expression in primes
11:27:44 -!- Slereah has quit (Read error: 110 (Connection timed out)).
11:33:50 <oerjan> Main> product [1/(1-fromInteger p^^(-2)) | p <- take 100 primes]
11:33:50 <oerjan> 1.64451522172429 :: Double
11:33:50 <oerjan> 1.64493406684823 :: Double
11:36:34 -!- Sgeo has joined.
11:39:02 <oerjan> hm maybe that's relatively fast anyway
11:45:40 <oerjan> if the length of range of numbers selected is divisible by a prime, then the product term from that prime fits perfectly
11:46:42 <oerjan> and if the range is much huger than that, the probability of not fitting inside such a range becomes miniscule
11:47:00 <oerjan> *inside such a subrange
11:47:53 <oerjan> so for million digit numbers, primes < 10^100 say, should give practically no error
11:48:03 -!- BeholdMyGlory_ has joined.
11:48:17 <oerjan> and that's about 10^100/(100 ln 10) primes, by the prime number theorem
11:48:28 -!- BeholdMyGlory has quit (Nick collision from services.).
11:48:33 -!- BeholdMyGlory_ has changed nick to BeholdMyGlory.
11:48:46 <oerjan> someone probably calculated/tested this already.
11:50:05 <Sgeo> On some Facebook quiz: "Your best friend tells you (s)he's pregnant. What is your reaction?"
11:52:00 -!- KingOfKarlsruhe has joined.
11:55:11 <AnMaster> I just managed to transform some unbalanced loops into ifs
11:58:20 <AnMaster> ehird, you want to read Darth&Droids today... Pirate ninjas!
12:21:05 -!- MigoMipo has joined.
12:33:38 * oerjan swats AnMaster for spoiling -----###
12:33:52 <AnMaster> oerjan, thought you read it already
12:34:18 <AnMaster> oerjan, get a water proof computer :D
12:35:32 <AnMaster> I changed something completely unrelated
12:35:51 <AnMaster> related to loop -> if conversion
12:36:08 <AnMaster> for no sane reason as far as I can see
12:37:11 <oerjan> a harmless reordering?
12:37:48 <AnMaster> oerjan, sure, but it shouldn't happen
12:38:34 <AnMaster> anyway it is not very harmless. It is supposed to sort by offset when possible.
12:43:51 <AnMaster> it didn't properly recurse into all types of loops. Just the standard loop type.
12:44:28 <AnMaster> oerjan, anyway I think in-between is the first compiler that can optimise _some_ unbalanced BF loops into ifs.
12:45:03 <Slereah_> Would it be easy to program Limp on Mathematica?
12:45:19 <Slereah_> It's full of math shit that might be useful, so I wonder
12:46:54 <AnMaster> oerjan, btw... I have a question...
12:47:50 <AnMaster> oerjan, sorting BF instructions by offset has many restructions, for example it must be a stable sort ([-]+ isn't same as +[-]) and you can't swap two output or input instructions with each other.
12:48:14 <AnMaster> oerjan, currently I'm using something like a gnome sort. Which is a bit slow for lostking
12:48:34 <AnMaster> any idea of a better one that will still be able to handle the various restrictions
12:50:07 -!- BeholdMyGlory_ has joined.
12:50:27 -!- BeholdMyGlory has quit (Nick collision from services.).
12:50:31 -!- BeholdMyGlory_ has changed nick to BeholdMyGlory.
12:53:42 <oerjan> it sounds more like a topological sort than an ordinary sort, since there are a lot of pair restrictions?
12:55:17 <AnMaster> oerjan, interesting. Yeah it seems so
12:55:30 <AnMaster> (after a quick read of http://en.wikipedia.org/wiki/Topological_sorting)
12:56:05 <AnMaster> maybe implementing it as a graph of cells using each other would be useful...
12:56:26 <AnMaster> wouldn't work at all with current code which uses linked lists to create a parse tree
12:56:46 <AnMaster> it would allow tracking dependencies nicely and so on
12:56:47 -!- BeholdMyGlory_ has joined.
12:56:57 -!- BeholdMyGlory has quit (Nick collision from services.).
12:57:03 -!- BeholdMyGlory_ has changed nick to BeholdMyGlory.
12:57:06 <AnMaster> not sure how to handle unbalanced loops in that though...
13:05:51 -!- BeholdMyGlory_ has joined.
13:06:08 -!- BeholdMyGlory has quit (Nick collision from services.).
13:06:11 -!- BeholdMyGlory_ has changed nick to BeholdMyGlory.
13:08:09 <oerjan> <bsmntbombdood> the completely trivial one though--http://pastebin.ca/1429617
13:08:24 <oerjan> wait, you are using _unsigned long_ for the selected numbers?
13:08:48 <oerjan> i'm not sure that is large enough for a good precision
13:08:59 <oerjan> (not sure it isn't, either)
13:13:42 <oerjan> as for shortcuts, you could select both numbers odd from the outset, and just multiply the probability by ... 3/4, i think
13:15:26 <AnMaster> oerjan, what is he using that gcd implementation for
13:15:43 <oerjan> calculating pi using random numbers
13:16:01 <AnMaster> ah that one mentioned yesterday
13:16:29 <oerjan> i am just responding to the logs
13:20:44 -!- Corun has joined.
13:33:02 -!- Judofyr has joined.
13:49:29 -!- oerjan has quit ("leaving").
14:25:45 -!- Corun has changed nick to Corun|away.
14:40:23 -!- Corun|away has changed nick to Corun.
14:44:05 -!- tombom has quit ("Peace and Protection 4.22.2").
15:08:27 -!- BeholdMyGlory_ has joined.
15:21:28 -!- BeholdMyGlory has quit (Nick collision from services.).
15:21:31 -!- BeholdMyGlory_ has changed nick to BeholdMyGlory.
15:51:12 -!- inurinternet has quit (Read error: 60 (Operation timed out)).
15:56:08 -!- MizardX has quit ("What are you sinking about?").
16:06:37 -!- inurinternet has joined.
16:07:01 -!- BeholdMyGlory_ has joined.
16:07:57 -!- BeholdMyGlory has quit (Nick collision from services.).
16:08:16 -!- BeholdMyGlory_ has changed nick to BeholdMyGlory.
16:11:33 -!- MizardX has joined.
17:23:23 -!- Corun has changed nick to Corun|away.
17:27:35 <ehird> 01:33 bsmntbombdood: ehird: what is accessing the disk then?
17:28:11 <ehird> waiting for time travel to be invented, going back to 1977, hopping on voyager 2, interpreting the geometry of the universe,
17:28:15 <ehird> waiting until you wrap around to earth,
17:28:24 <ehird> and then entering the bits with magnets
17:29:17 -!- oerjan has joined.
17:33:17 <ehird> oerjan: think that's accurate?
17:35:25 <ehird> 17:27 ehird: 01:33 bsmntbombdood: ehird: what is accessing the disk then?
17:35:25 <ehird> 17:28 ehird: waiting for time travel to be invented, going back to 1977, hopping on voyager 2, interpreting the geometry of the universe,
17:35:27 <ehird> 17:28 ehird: waiting until you wrap around to earth,
17:35:29 <ehird> 17:28 ehird: and then entering the bits with magnets
17:36:41 * oerjan had to look up bsmntbombdood's original too
17:37:16 <ehird> "USB FLOPPY DISK STRIPED RAID UNDER OS X" http://ohlssonvox.8k.com/fdd_raid.htm
17:37:42 <oerjan> i am not entirely sure that's accurate
17:38:09 <oerjan> what is the approximate ratio from main memory to main disk?
17:38:24 <ehird> oerjan: main memory is molasses-slow, disk is astronomically slow
17:39:01 <ehird> oerjan: i mean, to access the disk on a mechanical HD
17:39:06 <oerjan> that's equivalent to your _claim_, which i need numbers for in order to assess whether it's accurate
17:39:11 <ehird> you have to send the request across the quite slow connection
17:39:17 <ehird> it then has to seek
17:39:23 <ehird> which is so slow you can hear it!
17:39:30 <ehird> then it has to read it, and send the data back on the connection
17:39:33 <ehird> whereas, memory on a modern CPU:
17:39:53 <ehird> send the request across a fast transport, the RAM gets the data very quickly, and sends it back
17:40:18 <ehird> oerjan: memory takes something like 100 cycles. the fastest mechanical HD (VelociRaptor) can do a read from a random position of the harddrive in *7 milliseconds*
17:40:27 <ehird> so... no comparison
17:40:40 <oerjan> and i need numbers because (1) i have no intuition and little knowledge about hardware (2) my ballpark estimation ability is nearly nonexistent even for non-hardware subjects
17:40:59 <oerjan> i don't even know how much a cycle is nowadays
17:41:17 <oerjan> <- software guy at very abstract level at best
17:41:32 <ehird> oerjan: 7ms = 7,000,000 nanoseconds, ram goes at like 3nanoseconds/cycle
17:42:28 <ehird> oerjan: no, you've got two more zeros there
17:42:37 <ehird> disk is about a million times slower than RAM
17:42:52 <ehird> ok, so a letter to taiwan vs time travel, voyager 2, is more than a million
17:42:56 <ehird> but analogy-wise...
17:43:04 <oerjan> why couldn't you have said that when i asked for a fraction in the first place? :D
17:43:15 <ehird> cuz I had to google it
17:43:23 <ehird> oerjan: note that if you have an SSD, it's 0.1ms
17:43:32 <ehird> oerjan: = 100,000 nanoseconds
17:43:40 <ehird> still molasses slow, but, faster.
17:43:45 <ehird> I had to google the speeds
17:44:15 <oerjan> circle around earth ~= 1/7 light second iirc
17:44:25 -!- asiekierka has joined.
17:44:42 <oerjan> so, 1/3 million light seconds
17:44:54 <ehird> haha wow you're actually working it out
17:45:48 <oerjan> about 3.85 light days, barely out of the solar system
17:46:18 <oerjan> otoh there was that issue of snail mail vs. light speed
17:47:19 <oerjan> ok in time it's more impressive, say a week for mail to taiwan (see my lacking ability for ballparking)
17:47:58 <oerjan> that's about the age of the universe, right? ;D
17:49:01 <oerjan> dammit i forgot i'd cancelled the 7
17:53:59 * ehird burns an arch linux cd
17:54:10 <asiekierka> What's the hardest assembly language EVER
17:54:17 <asiekierka> that's the hardest to write in assembly
17:54:28 <asiekierka> Probably NOT x86, 6502, or any of deriatives of thes
17:54:29 -!- Corun|away has changed nick to Corun.
17:56:01 <oerjan> asiekierka: may i interest you in Malbolge Unshackled?
17:56:08 * oerjan cackles uncontrollably
17:56:17 <oerjan> oh right, that's machine code
17:56:25 <oerjan> a very *evil* machine's code
17:56:53 <AnMaster> ehird, so what did you think about Darth & Droids today
17:56:58 <asiekierka> Well, I still want to have a machine that does SUBLEQ
17:56:58 <ehird> I don't read that comic.
17:57:06 <AnMaster> ehird, you will love the one today
17:57:26 <AnMaster> ehird, http://www.darthsanddroids.net/
17:58:20 <ehird> you laugh now, but will you laugh when your home is destroyed by their power?
17:58:32 <AnMaster> ehird, you said before they couldn't be combined
17:58:45 <AnMaster> ehird, also I didn't laugh. It was you who did
17:58:49 <ehird> AnMaster: they couldn't be combined *safely*
17:58:57 <ehird> AnMaster: think strange quarks
18:00:12 <ehird> i'm going to install arch as a second os on this mac
18:00:22 <ehird> hope I don't fuck anything up
18:00:58 <AnMaster> ehird, as always: make backups
18:01:09 <ehird> i don't have any viable backup media
18:01:19 <ehird> and even if I did I'd wait until I get my new pc
18:01:23 <AnMaster> ehird, well, be careful with fdisk stuff
18:01:28 <ehird> and when I get my new pc, backing up will be unfeasible
18:01:42 <ehird> since enough media to back up 160GB+2TB for a long time incrementally is... expensive
18:01:49 <AnMaster> ehird, just buy almost as much external storage
18:01:56 <ehird> AnMaster: almost as much? you mean more
18:02:08 <AnMaster> ehird, depends on how you backup
18:02:21 <ehird> incrementally, but if I buy less storage then I can't use any more than that
18:02:23 <AnMaster> full every time: a bit less, no need to back up /tmp and /var/tmp
18:02:27 <ehird> which is redonkulous
18:02:43 <ehird> 18:02 AnMaster: full every time: a bit less, no need to back up /tmp and /var/tmp ← uhm that'd only handle one backup
18:02:55 <ehird> i want to keep backups forever, you see
18:03:11 <ehird> sure is if you buy a lot of storage
18:03:25 <AnMaster> you don't need to back up anything installed from package manager (exception: config files), since they can be restored by package manager easily
18:03:37 <AnMaster> nor most of the things in /var/cache and similiar
18:03:37 <ehird> AnMaster: i could just back up ~
18:03:40 <ehird> but I like snapshots
18:03:50 <AnMaster> ehird, I would probably do ~ and /etc
18:03:57 <ehird> keeping backups forever = ubiquitous revision control of all my computing
18:04:17 <AnMaster> ehird, hm, put your entire ~ in git?
18:04:37 <ehird> AnMaster: that requires manual effort to commit, annotate, etc, and storing .git locally would be infeasible
18:04:41 <AnMaster> possibly exclude ~/vcs-checkouts so you can check out other git projects and such
18:04:49 <ehird> i don't mind it being hard to flip through the backups later on
18:04:52 <ehird> as long as it's possible
18:05:06 <ehird> that's why, e.g. i prefer to archive in raw form things like irc logs, emails
18:05:13 <AnMaster> ehird, hm... I fail to see where you would find the needed storage
18:05:43 <ehird> AnMaster: 160GB + 12TB = 2208GB
18:05:53 <AnMaster> ehird, generally, raw for for irc is pretty verbose. more so than using a partly parsed for
18:05:53 <ehird> let's say that 10GB of the OS drive is used
18:06:10 <ehird> let's say that 500GB of the 2TB drive is used
18:06:20 <ehird> the os drive includes ~
18:06:25 <ehird> so let's say 60GB of the OS driv
18:06:28 <ehird> and 500GB of the 2TB drive
18:06:37 <ehird> so a non-incremental backup = 560GB
18:06:45 -!- ais523 has joined.
18:06:57 <ehird> 560GB for first backup
18:06:59 <AnMaster> ehird, but you forget that it will be compressed when backed up
18:07:05 <ehird> AnMaster: oh, of course
18:07:15 <ehird> the 500GB isn't very compressable
18:07:16 <AnMaster> ehird, but in worst case, yeah
18:07:18 <ehird> since it's movies/etc
18:07:27 <ehird> the 60GB could probably compress to, let's say, 45
18:07:32 <ehird> so 545GB for a non-incremental backup
18:07:41 <ehird> now, let's say that each monthly backup changes 1GB of data
18:08:21 <ehird> so ten years is 120GB
18:08:35 <ehird> now let's say that instead 10GB/mo changes, compressed
18:08:37 <ehird> that's excessive, but
18:08:45 <AnMaster> ehird, was this going to include /usr/bin and such?
18:08:54 <ehird> so 1200GB ten years
18:09:07 <AnMaster> ehird, but not /tmp. It wouldn't make sense to include /tmp
18:09:28 <ehird> i'd want to do non-incremental backups often
18:09:33 <ehird> and I won't keep this system for 10 bloody years
18:09:35 <ehird> but those two cancel out
18:09:46 -!- ais523_ has joined.
18:09:47 <ehird> now, a 2TB drive costs
18:10:15 <Deewiant> 2TB of storage costs around 200€
18:10:21 <Deewiant> In a single drive, more than thata
18:10:25 <AnMaster> ehird, yes, you could have two storage units, you don't want to overwrite your only full backup, what if it fails when you are doing the backup
18:10:28 <ehird> ok, a 1TB drive costs,
18:10:58 <ehird> so let's say we buy 4
18:11:03 <ehird> for a total of 4TB backup space
18:11:09 <ehird> plus the cost of an external drive enclosure
18:11:17 -!- ais523 has quit (Nick collision from services.).
18:11:23 <AnMaster> ehird, plus a fire proof wall mounted safe?
18:11:38 <ehird> AnMaster: unfortunately I'm not egotistical enough to convince myself my works are worth surviving that :)
18:11:44 <ehird> especially since I likely wouldn't
18:11:56 <AnMaster> ehird, depends on if you got out of the building in time
18:12:10 <ehird> i'd die trying to drag the safe with me :-P
18:12:31 <AnMaster> ah better to NOT have a safe, so you can easily take the backup with you
18:12:41 <ais523_> AnMaster: I don't get why the "wall mounted"
18:12:53 -!- ais523_ has changed nick to ais523.
18:13:01 <ehird> ais523: it looks coolr
18:13:02 <AnMaster> ais523_, maybe wrong word. I meant those that is mounted to the support structure, so it is hard to steal
18:13:11 <ais523> flash drives may be better for backups, if you aren't backing lots of stuff up
18:13:18 <ais523> because they easily fit in your pocket in an emergency
18:13:31 <ehird> ais523: er, it was in the context of buying 4TB of backup space
18:13:31 <AnMaster> ais523, he is talking about 500 GB+
18:13:53 <ehird> so unless e.g. SSDs suddenly turn really cheap, that's not feasible
18:13:59 <ehird> OTOH, mechanical disks are unreliable
18:14:07 <ehird> OTOOH, I've never had one fail me
18:14:23 <ehird> but that was my fault
18:15:03 <ehird> AnMaster: you are welcome to donate 4TB of tape to me
18:15:41 <AnMaster> hm... What languages would you say have sane macro systems btw?
18:16:27 <AnMaster> Deewiant, that is open for discussion.
18:16:31 <ehird> I've never seen another sane one.
18:16:37 <ehird> Deewiant: I'm using "good" as the definition.
18:16:51 <pikhq> There's not many that are all that good.
18:17:07 <pikhq> ehird: If you can call Forth's words 'macros', then Forth, too.
18:17:08 <Deewiant> GHC's Template Haskell is semi-good
18:17:19 <ehird> Deewiant: Yeah but it's not really elegant.
18:17:23 <Deewiant> I.e. modulo implementation restrictions
18:17:39 <AnMaster> ehird, I think the erlang one is pretty sane, for being text replace based... All uses of macros must be prefixed with ?. And they are fairly limited.
18:17:47 <Deewiant> Like the fact that the macros have to be defined in a different module than where they're used, coupled with the fact that only one module per file is allowed
18:17:49 <ehird> "Text replace based"
18:17:52 <ehird> You are disqualified.
18:18:09 <ehird> AnMaster: can you install pkgs from arch livecd
18:18:10 <Deewiant> D has handy string macros, but yeah... string macros.
18:18:11 <AnMaster> since blah blah makes no sense
18:18:28 <ehird> AnMaster: to install irssi to chat while I install.
18:18:31 <AnMaster> ehird, I only ever done net install
18:18:32 <ehird> s/chat/bug you all/
18:18:44 <ehird> containing a minimal linux setup
18:18:55 <AnMaster> ehird, and no idea. Why not create a cheap VM and try it ;P
18:19:04 <ehird> cuz it's cheaper to reboot!
18:19:18 <ehird> see you later green suckers of amazing
18:19:59 <AnMaster> "<ehird> ←←←←←←←←←→" <-- Hm... Assuming that they are matching... He is more "here" than ever before
18:21:50 <ais523> AnMaster: I believe presence in a channel is saturating
18:22:12 -!- ehird_ has joined.
18:22:22 <ehird_> it's better than kubuntu.
18:22:35 <ehird_> it recognized my wireless kb and mouse!
18:22:39 <ehird_> without doing anything!
18:23:09 <ehird_> wait mouse isn't wireless, w/e, kb is wireless w/ bluetooth and it goddamn recognized it
18:23:25 <ehird_> i can almost forgive it for being text-based
18:23:46 <ehird_> AnMaster- can you hear me?
18:24:10 <ais523> also, what OS is that?
18:24:30 <ehird_> Arch Linux install CD.
18:24:51 <ehird_> ofc, it's the bios emulation recognizing my bluetooth KB, but dammit, kubuntu didn't do it!
18:25:24 <ehird_> there's a possibility it won't work in X, but I doubt it
18:25:46 <ehird_> OK, time to break everything by partitioning
18:25:56 <ehird_> god nc is a terrible irc client
18:26:11 <pikhq> You should install a copy of RawIRC.
18:26:23 <ehird_> heh, heh, heh, cfdisk refuses to run 'cuz of GPT. Tells me to use parted.
18:26:32 <ehird_> time to see if pacman works on a ramdisk
18:27:14 <AnMaster> AnMaster, not with a - after no
18:27:48 <ehird_> anmaster it says protocol error when it tries to download the package list when i do pacman ;;;;(
18:28:05 <AnMaster> ehird_, um... sounds like network is broken?
18:28:10 <AnMaster> ehird, maybe mirror list is wrong
18:28:17 <AnMaster> it is in /etc/pacman.d/mirrorlist
18:28:20 <ehird_> how can i change what mirror I'm using? /etc/- ah
18:28:34 <ehird_> /etc/pacman.conf right
18:28:42 <AnMaster> ehird_, err that is not for the mirror list
18:28:46 <AnMaster> that is for the over all config
18:28:57 <AnMaster> mirror list is in /etc/pacman.d/mirrorlist or /etc/pacman.d/mirror_list
18:29:06 <ehird_> I want to force one particular mirror.
18:29:06 <AnMaster> not sure, and arch computer is turned off due to cleaning
18:29:19 <ehird_> "add your preferred mirrors here" --pacman.conf
18:29:21 <AnMaster> ehird, then comment the other ones out and just uncomment the one you want
18:29:35 <AnMaster> ehird_, never done it that way
18:29:51 <ehird_> ah works now, luck of the draw
18:30:38 <ehird_> *fuck*, it already has parted, haha
18:30:59 <ehird_> [1~[5~argh, I can't scroll up.
18:31:38 <AnMaster> nc doesn't use pageup to to scroll
18:31:46 <ehird_> AnMaster: it's a console, you silly.
18:31:57 <ehird_> it works in none of the
18:32:10 <AnMaster> ehird, you may want to use the key combo for linux vt scroll then. which only seems to work until you switch vt
18:32:15 <ehird_> AnMaster: can you provide start/end to parted as anything other than sectors or whatever?
18:32:37 <AnMaster> ehird_, no idea. I tend to use fdisk or gparted
18:33:27 <ehird_> Apple's partition table thing != BIOS's.
18:33:29 <AnMaster> and if you already have free space at the end, which I assume, since you probably used bootcamp to free up some space?
18:33:42 <ehird_> nah, I have the drives from my old Kubuntu
18:34:00 <ehird_> AnMaster, could you "parted\nhelp" and tell me how to lit partitions?
18:34:01 <AnMaster> ehird, well ok, can't you just do that and mkfs on them
18:34:29 <AnMaster> ehird_, how short is your scrollback but ok
18:34:42 <ehird_> what is the vt scroll combo?
18:35:00 <AnMaster> print [devices|free|list,all|NUMBER] display the partition table, available devices, free space, all found partitions, or a particular partition
18:35:35 <AnMaster> it probably only works for frame buffer consoles
18:35:57 <ehird_> so vga. AnMaster, what is the cmd to add a partition? add odnes't work
18:37:59 <AnMaster> ehird_, sorry was afk, had to clean my glasses
18:38:06 <AnMaster> mkfs NUMBER FS-TYPE make a FS-TYPE file system on partititon NUMBER
18:38:06 <AnMaster> mkpart PART-TYPE [FS-TYPE] START END make a partition
18:38:06 <AnMaster> mkpartfs PART-TYPE FS-TYPE START END make a partition with a file system
18:38:27 <AnMaster> ehird_, it can probably only handle basic ones
18:38:32 <ehird_> lolwtf, it wants a partition name :-)
18:38:58 <AnMaster> ehird, also that help is LESS than 25 lines.
18:39:04 <AnMaster> how can you not be able to read it
18:39:33 <AnMaster> wait ok, some lines are longer than 80
18:40:35 <ehird_> Support for creating ext3 filesystems is not sdfksdf yet. lol ok
18:40:53 <AnMaster> ehird, just create partition then manually create file system
18:41:13 <ehird_> yeah. swap partitions have no fs right?
18:41:40 -!- tombom has joined.
18:42:01 <AnMaster> ehird_, just adds some header think. NOTE: mkswap doesn't ask for confirmation!
18:42:10 <ehird_> The installer will do that. AnMaster: wtf, parted is daemanding a partition type even for @"mkpart" and i dunno what to give it, it won't accept swap
18:42:23 <AnMaster> ehird, err that is the number think I assume
18:43:05 <ehird_> AnMaster: do i need to run any command to save? or does just quit work
18:43:06 <AnMaster> PART-TYPE is one of: primary, logical, extended
18:43:10 -!- asiekierka has quit.
18:43:21 <ehird_> i'm using the prompt version, and that doesn't apply to GPT iirc
18:43:32 <ehird_> if you omit params it prompts you
18:43:37 <AnMaster> ehird, I'm using the prompt too: help mkpart
18:43:38 <ehird_> but yeah is there a save/write command?
18:44:11 <AnMaster> ehird_, btw parted has tab completion
18:44:54 <AnMaster> ehird_, fun, help doesn't list all commands
18:45:36 <AnMaster> like "mktable" (don't do it, help mktable says it is for changing between GPT, MBR and so on)
18:45:38 <ehird_> it's installing time!!
18:46:19 <ehird_> i have an itch vto install w3m. bloody internet addiction
18:48:09 <AnMaster> ehird_, might be useful in case you need have issues with X and need to google to find out how to solve it
18:48:46 <ehird_> i mean on this ramdisk :-)
18:49:32 <ehird_> but I have every faith in Xorg's auto-detection. probably. i mean, evdev is the only thing that actually works with one of the mice i have :-)
18:51:57 <ehird_> THIS IS THE X86 CD NOT THE X86_64 ONE
18:54:36 <ehird> raaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaage
18:54:36 -!- ehird_ has quit (Read error: 104 (Connection reset by peer)).
18:55:03 * ehird downloads the Real Fucking CD
18:56:22 <ehird> AnMaster: excuse me, be annoyed
18:56:57 <ehird> well it'd work just fine
18:57:02 <ehird> but not very 64bitlovefesty
18:57:18 <ehird> http://www.crooksdesign.co.uk/stock-photography-misuse.html ← ah, the BNP
18:57:31 <ehird> such irritating xenophobes
18:57:42 <ehird> (british national"ist" party)
18:57:45 <AnMaster> hm, is it possible to make a tail recursive version of the _Extended_ GCD.
18:57:53 <ehird> AnMaster: paste the code
18:58:28 <ehird> AnMaster: not truly tail recursive, no, it seems
18:58:29 <AnMaster> just based it on pseudo code found on wikipedia
18:58:33 <ehird> ofc, you could continuation-passing-style
18:58:38 <AnMaster> ehird, there is an iterative style
18:58:39 <ehird> but that just sucks up another part of memory
18:58:45 <ehird> AnMaster: show the iterative style
18:58:58 <AnMaster> ehird, would be flooding on irc. sec for link
18:59:05 <AnMaster> http://en.wikipedia.org/wiki/Extended_Euclidean_algorithm#Iterative_method
18:59:17 <ehird> AnMaster: yes, you could do that
18:59:29 <ehird> AnMaster: change every variable in the pseudocode to a parameter
18:59:34 <AnMaster> ehird, would be rather inelegant though
18:59:34 <ehird> to change a variable, recurse with its new value
18:59:51 <ehird> AnMaster: so do it :)
19:00:46 <ehird> http://www.crooksdesign.co.uk/images/bnp_poster.jpg ← this is a polish plane
19:01:12 <AnMaster> ehird, looks like an UK one to me.
19:01:22 <ehird> http://www.theregister.co.uk/2009/03/04/polish_spitfire/
19:01:40 <AnMaster> ehird, Spitfire is originally from UK though
19:02:02 <ehird> but for a party that goes around talking about how polish immigrants are ruining the UK with their devil fire and Muslins...
19:05:07 <ehird> AnMaster: Muslin is a type of fibre.
19:05:18 <ehird> It is destroying the world.
19:05:21 <AnMaster> what has that got to do with Polish
19:05:28 <ais523> ehird: if muslin is involved, then would polish immigrants refer to making things shinier?
19:05:42 <ehird> AnMaster: right wingers often misspell muslim as muslin
19:05:46 <ehird> because they're ignorant, see.
19:06:21 <ehird> "with their devil fire and Muslims"
19:06:27 <ehird> i.e., they're collaborating with muslims to end the world
19:06:39 <ehird> it's a joke, see. about the BNP's "send all immigrants back!!12182374" policy.
19:07:03 <ehird> fire personally provided by the devil?
19:07:54 <ehird> AnMaster: to basically sum up the BNP, their founder once said "Mein Kampf is my bible"
19:08:01 <ehird> and correct, but I made up "devil fire"
19:10:24 <ehird> "Text adventure blockbusters Zork I II III (PC/MAC) now free from Infocom!"
19:10:30 <oerjan> AnMaster: the table method listed on wp can be made to work tail recursive
19:10:35 <ehird> that's the old pirate site, I think
19:10:42 <ehird> what a disappointm—
19:10:46 <AnMaster> oerjan, which is the most efficient one though
19:11:03 <oerjan> AnMaster: don't ask me such questions :D
19:12:35 <oerjan> because accurate speed and efficiency is incompatible with how my brain works. or something.
19:17:55 <oerjan> AnMaster: oh, the table method and the iterative method may be equivalent. i was confused by how the description of the latter seemed to calculate the quotients first, but the pseudocode seems to interleave them
19:18:13 <pikhq> I don't want an efficient method.
19:18:16 <pikhq> I want a fast method.
19:18:40 <ehird> pikhq: i may have to slap you for that practical remark
19:18:41 <pikhq> (10 trillion CPUs for a single calculation is absurdly fast. It sure as hell isn't efficient. :D)
19:18:55 <pikhq> ehird: Not practical.
19:19:10 <pikhq> It's a desire for a room of blinkenlights.
19:19:12 <AnMaster> huh, the iterative one return one more result than the recursive one
19:19:15 <ehird> THEORY TRUMPS EVER— oh okay
19:19:16 <Deewiant> I don't want a practical method.
19:19:32 <ehird> is that good or bad
19:20:56 <oerjan> AnMaster: perhaps that a is just the gcd
19:21:15 <oerjan> the recursive one only returns the coefficients
19:21:51 <pikhq> Deewiant: Get 100 of the uber-build that ehird and I built.
19:21:54 <AnMaster> gcd_test:extended_gcd_rec(3, 9).
19:22:01 <ehird> pikhq: it wasn't all that fast
19:22:09 <ehird> isn't gonna really do anything for speed
19:22:20 <oerjan> AnMaster: right, the latter doesn't return the actual gcd. easy to modify though.
19:22:30 <pikhq> ehird: It had 32 cores.
19:22:37 <AnMaster> oerjan, yeah. But I'm going for iterative one. Now just to clean it up
19:22:39 <pikhq> That's ridiculously fast.
19:22:48 <ehird> that doesn't cost $80k
19:23:00 <AnMaster> http://rafb.net/p/0LVnQZ12.html is the "raw" translation of it
19:23:05 <AnMaster> there are some pointless variables there
19:24:44 <AnMaster> http://rafb.net/p/Oy8qYb13.html that is better
19:24:54 <AnMaster> of course it could be merged into the call if we wanted
19:25:27 <ehird> AnMaster: extended_gcd(A, B, X, Y, LastX, LastY) -> Quot = A div B, extended_gcd(B, A rem B, LastX-Quot*X, LastY-Quot*Y, X, Y)
19:26:13 <oerjan> AnMaster: note that calculating div and rem simultaneously can be faster iirc if you have a function for it
19:26:22 <AnMaster> and since erlang doesn't actually need newlines there you could do it all in one line, apart from the initial -module(gcd_test). which needs to be on a separate one
19:26:29 <ehird> oerjan: dude he's using erlang
19:26:52 <AnMaster> oerjan, don't think I have... but it is possible erlang optimises that
19:26:57 <oerjan> ehird: i don't know if erlang does, i know haskell has it
19:27:15 <ehird> oerjan: i meant, microöptimization hardly helps
19:27:35 <ais523> ehird: haha, I love your technically-speaking-correct Englih
19:27:54 <oerjan> ehird: but if the bignum routines take most of the time, still...
19:27:58 <ehird> ais523: yes, the diaeresis is a true friend of mine
19:28:13 <ehird> what's its use called in English?
19:28:19 * oerjan recalls that his brain is not supposed to care about speed and efficiency
19:28:21 <ehird> AnMaster: preëmptive
19:28:32 <ehird> micro-optimization used to be microöptimization
19:28:35 <ehird> it's the only accent english as
19:28:36 <ais523> AnMaster: in theory, you can use a diaresis (looks the same as an umlaut) to separate two syllables in English
19:28:45 <ais523> but nobody's used it for years
19:28:51 <ais523> so it's more or less died out from the language
19:28:56 <ais523> you can see it in old books, sometimes
19:29:02 <Deewiant> I use it in stuff I don't intend others to read
19:29:10 <ais523> and in a few names, like Zoë
19:29:20 <ehird> is Zoë an english name?
19:29:46 <ehird> GregorR-L: I meant with.
19:29:51 <ehird> Also, not an umlaut.
19:30:15 <ehird> GregorR-L: I meant _with_.
19:30:28 <GregorR-L> And I'm saying it is, but only without :P
19:30:39 <ehird> ais523: defend yourself!
19:30:43 <pikhq> You mean most people don't think of the diaeresis?
19:30:56 <ehird> pikhq: most people don't even know it's valid in English...
19:30:57 <GregorR-L> I've seen the name, but only without.
19:31:04 <oerjan> poor Zoë and her diarrhea
19:31:09 <ehird> ais523: that Zoë is a valid english name, not just Zoe. GregorR-L seems to think not.
19:31:14 <ais523> GregorR-L: that's because most English people never bother to learn how to type it
19:31:18 <pikhq> I suppose next you'll tell me that people don't think of "print'ed" making the e pronounced.
19:31:24 <GregorR-L> There's no such thing as an "invalid" name.
19:31:38 <ais523> GregorR-L: some countries have lists of legal names to give children
19:31:46 <ais523> although the UK doesn't, nor does the US as far as I know
19:31:53 * pikhq uses psuedo-dead keys instead of proper compose keys
19:31:56 <ehird> "He has proposed that a one-way trip to Mars could be a viable option."
19:31:58 <GregorR-L> Of course the US doesn't, otherwise nobody would be named Shaniqua
19:32:09 <pikhq> ais523: This is the land of people named "Gonnorhea".
19:32:15 <oerjan> norway used to have such a list, but i think it was abolished
19:32:30 <pikhq> There is neither a list of valid names *nor* any sense in choosing names.
19:32:38 <GregorR-L> However, I'd bet that far, far more people are named Zoe, as the parents don't know how to write a diacritical mark any more than anyone else :P
19:32:50 <pikhq> And I'm pretty sure that there is nothing requiring your name to even use English phonemes.
19:33:05 <ais523> also, nowadays there's a big advantage to pure-ASCII names
19:33:10 <GregorR-L> pikhq: Tell that to the artist formerly known as the artist formerly known as Prince.
19:33:16 <oerjan> it's probably still illegal to give a child a name that "can be a burden to them", though
19:33:22 <AnMaster> Sweden used to have such a list at least. Not sure if it still does.
19:33:25 <ais523> which is that many websites you might have to use don't know how to handle encodings properly
19:33:30 <ehird> GregorR-L: Unicode should have the Prince symbol
19:33:38 <pikhq> GregorR-L: I suspect it has to be representable in Latin.
19:33:43 <ehird> after all, it DOES have a valid use: entering albums under that name into a music player
19:33:51 <pikhq> Erm. The Latin alphabet...
19:33:57 <pikhq> Including diaretics.
19:34:01 <ais523> actually, even on paper forms, sticking to ASCII is helpful because the values are likely to be written into computers eventually, by people who don't know what they're doing
19:34:21 <AnMaster> ais523, if Zoë is valid... is that same as Zo-e ??
19:34:28 <ais523> AnMaster: pronunciation-wise, yes
19:34:53 <ais523> GregorR-L: Zoee would probably be the nearest representation without the accent
19:35:02 <ais523> but people just memorise how words are meant to be pronounced
19:35:07 <oerjan> avoid dianetics in your latin alphabet
19:35:08 <ais523> AnMaster: invalid, C isn't a vowel
19:35:10 <pikhq> GregorR-L: ASCII representation of the Artist's symbol: O(+>
19:35:16 <GregorR-L> Except that Zoe is pronounced like Zo :P
19:35:16 <oerjan> (says someone anonymous)
19:35:39 <ais523> AnMaster: although hyphens are sometimes used to replace diaereses, not all hyphens are used for that purpose
19:35:41 <ehird> more words need to have a place to use a ¨ in, in English
19:35:58 <ais523> GregorR-L: yes, people memorise the correct pronounciations of words, nowadays, rather than memorising the correct spelling
19:36:01 <pikhq> Why the *hell* do the umlauts show up when GregorR-L types it, but not for someone else typing it?
19:36:07 <GregorR-L> ais523: That's because this is English :P
19:36:10 <pikhq> Someone else familiar with urxvt care to enlighten?
19:36:10 <ehird> I AM GOING TO REBOOT AND REINSTALL ARCH ! !!!!
19:36:27 <ais523> pikhq: I wonder if GregorR-L is using Latin-1?
19:36:39 <AnMaster> and my client does auto detection
19:36:39 <pikhq> AnMaster: Nominally.
19:36:42 <GregorR-L> Mine may be Latin-1, I haven't configured it on this laptop.
19:36:48 <ais523> This line is written in Latin-1: premptive
19:36:56 <AnMaster> ais523, how did you pull that off
19:37:03 <ais523> This line is written in UTF-8: preëmptive
19:37:09 <ais523> AnMaster: I have an encoding setting in my IRC client
19:37:10 -!- sebbu has quit (Read error: 60 (Operation timed out)).
19:37:11 <pikhq> So why the crap does rxvt-unicode only handle Latin-1?
19:37:12 <AnMaster> ais523, my client converts input too
19:37:19 <AnMaster> ais523, ah, switching it, right
19:37:31 <ais523> I /typed/ both in Unicode, but one of them I told my client to send in Latin-1 in
19:37:33 -!- sebbu has joined.
19:38:00 <pikhq> ... Why the crap is LC_ALL=C?
19:38:25 <GregorR-L> Because if English was good enough for Jesus, it's good enough for pikhq?
19:38:29 <pikhq> I've made a point of making this en_US.UTF-8.
19:39:13 <GregorR-L> Apparently there's a genus of amoeba that can be as large as 5mm long.
19:39:36 <GregorR-L> I wonder if an amoeba farmer could use artificial selection to get amoeba that are 1cm long :P
19:39:49 <pikhq> ... And inexplicably, my system doesn't have the en_US.UTF-8 locale.
19:41:49 <pikhq> Now, any idea where to go for a system-wide locale setting in Gentoo?
19:42:49 <ehird> the iso downloaded md5 != the right one
19:42:52 <ehird> kekkekekkekekekekekeke
19:43:05 <ehird> how big are amoebas anyway
19:43:55 <ehird> i wanta pet amoeba
19:44:01 <pikhq> Deewiant: Lemme guess: not obvious.
19:44:09 <Deewiant> I think the biggest is 1 mm or so
19:44:18 <ehird> 18:39 GregorR-L: Apparently there's a genus of amoeba that can be as large as 5mm long.
19:44:30 <Deewiant> pikhq: The 'something' bit? Not necessarily, I can't remember
19:44:52 <ehird> pikhq: /etc/profile?
19:44:53 <AnMaster> lifthrasiir, there? What are the members "value" and "target" for the Cond class in Esotope-BFC?
19:45:07 <Deewiant> pikhq: http://www.gentoo.org/doc/en/guide-localization.xml sez /etc/env.d/02locale
19:45:16 <AnMaster> <pikhq> Now, any idea where to go for a system-wide locale setting in Gentoo?
19:45:27 <AnMaster> pikhq, well.. what about the /etc/locale.gen thing
19:45:37 <GregorR-L> Most amoeba are microscopic, in the 10-20um size
19:45:43 <AnMaster> pikhq, to make sure the locales are generated at all
19:45:48 <GregorR-L> But Chaos (an awesome name for a genus) are as big as 5mm.
19:46:04 <ehird> why are gentoo docs .xml? they're html 4.01 transitional
19:46:14 <pikhq> AnMaster: I just checked that to make sure it's set right, and have gone ahead and edited 02locale.
19:46:14 <AnMaster> ehird, they are auto translated on server iirc
19:46:22 <GregorR-L> (And, technically, are amoebazoa, not amoeba, but anyway :P )
19:46:25 <ais523> ehird: is it valid XML as well?
19:46:26 <ehird> AnMaster: yes but the file you get just isn't XML in any form :P
19:46:35 <ais523> it would be very funny if it was valid HTML, and valid XML, but not valid XHTML
19:46:36 <pikhq> ehird: Crappy autogeneration.
19:46:36 <ehird> ais523: it has <br>
19:46:56 <ais523> well, I hope at least that it's valid SGML
19:47:05 <pikhq> Seems to be well-formed XML.
19:47:23 <AnMaster> ehird, you can add ?style=printable
19:47:50 <ehird> plus, the doctype doesn't point to a sgml doctype
19:47:57 <pikhq> I bet the file is stored on the server as XML, then.
19:48:14 <ehird> it points to an sgml doctype
19:48:22 <pikhq> ehird: http://www.w3.org/TR/html4/loose.dtd is an SGML doctype.
19:48:27 <ehird> 19:48 ehird: it points to an sgml doctype
19:48:28 <ehird> 19:48 ehird: not an xml one
19:49:55 <GregorR-L> But it also has </br>, making it invalid HTML and XHTML, and valid XML :P
19:50:08 <ehird> 91228e6b71d74e7a52269f1aaf225a6d archlinux-2009.02-ftp-x86_64.iso—mirrorservice.org
19:50:09 <ehird> % md5sum archlinux-2009.02-ftp-x86_64.iso
19:50:10 <ehird> ec8d390ce27fa5bc3125004ba3e7c2f0 archlinux-2009.02-ftp-x86_64.iso
19:50:18 <ehird> AAAAAAAAAAAAAAAAAAAAAAAAA
19:50:19 <ehird> AAAAAAAAAAAAAAAAAA
19:50:20 <pikhq> GregorR-L: </br> is valid XHTML in context.
19:50:24 <ehird> AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
19:50:29 <pikhq> <br></br> is perfectly valid.
19:50:30 <ehird> pikhq: it has content after <br>
19:50:42 <pikhq> Well, it's well-formed XML.
19:50:52 <AnMaster> http://www.gentoo.org/doc/en/xml-guide.xml
19:50:57 <AnMaster> that seems to be what it is about
19:51:45 <pikhq> Actually, it has some <br> tags that it doesn't close.
19:51:46 <Deewiant> ehird: Just use it, who cares about the checksum
19:51:57 <pikhq> So, it's only valid SGML.
19:52:02 <ehird> Deewiant: Are you being sarcastic? But I did. The crc failed on decompression while booting
19:52:42 <AnMaster> ehird, if that is the minimal one, doesn't it take like 2-3 minutes to download it iirc
19:52:45 <Deewiant> I wasn't; some errors might not matter, and the expected sum could be incorrect :-P
19:52:50 <AnMaster> or maybe that is the gentoo minimal that does
19:53:09 <GregorR-L> ehird's modem is so slow it's described in baud.
19:53:18 <ehird> AnMaster: its 153MB
19:53:21 <ehird> so about 4 minutes
19:53:34 <ehird> AnMaster: yes but I'm on to my third cd
19:53:36 <AnMaster> I remember when I downloaded a 500 MB ISO on 512 kbps!
19:53:53 <ehird> AnMaster: I remember when I downloaded a 30MB file on 56kbps.
19:54:03 <Deewiant> I've downloaded several 600+MB files on 256kbps
19:54:04 <ehird> It took all day. And since it was 56kbps, couldn't load anything else while I waited.
19:54:09 <AnMaster> ehird, I never did. Since it was pay per time.
19:54:22 <ehird> In those days, if a file was >1MB I canceled it immediately.
19:54:23 <AnMaster> ehird, you could load other stuff, it would just be very slow
19:54:27 <GregorR-L> I remember when I downloaded a 3TB file on a 2880 baud modem!
19:54:27 <ehird> I just couldn't download it fast enough.
19:54:36 <ehird> AnMaster: Too slow for usabiilty.
19:54:53 <pikhq> GregorR-L: Shaddup.
19:54:56 <ehird> GregorR-L: I remember downloading every internet of every infinite planet via TCP/IP over carrier pigeon over space rocket.
19:55:06 <pikhq> You know as well as I do that that was 3TBs of 0s, and it was RLE'd.
19:55:06 <ehird> IT TOOK INFINITY TO THE POWER OF INFINITY YEARS
19:55:29 <ehird> another mirror works
19:55:40 <AnMaster> ehird, report the broken mirror?
19:55:42 <pikhq> GregorR-L: Also... Have you though of making a Javascript LLVM backend?
19:55:51 <ehird> AnMaster: It's mirrorservice.org, quite a big thing
19:55:54 <ehird> Maybe they'll notice themselves.
19:56:03 <ehird> Not as fun as emulating CPUs.
19:56:12 <ehird> AnMaster: So everyone will have to use another mirror
19:56:17 <ehird> It'll fix next release, I assume
19:56:17 <Deewiant> I think the biggest file I downloaded on 56kbps was around 60-70 MB
19:56:18 <pikhq> Technically, he could just make LLVM target MIPS and do that.
19:56:44 <GregorR-L> It would presumably be more efficient to compile directly.
19:56:55 <GregorR-L> And my MIPS-JS AOT compiler is stalled :P
19:57:14 <AnMaster> whatever was the size of netscape 3 was the largest file I downloaded on 28.8k
19:58:05 <ehird> http://www.theweek.com/article/index/96719/Dick_Cheney_2012 ← No. NO. NO, NO AND MORE FUCKING NOS FOREVER UNTIL YOU RUN OUT OF THE INFINITE SUPPLY OF NOS. AND THEN AGAIN! AND THEN, FINALLY, ONE MORE NO!
19:59:14 <GregorR-L> Don't worry, Cheney'll be dead by then :P
19:59:52 <ehird> Wow, my drive is whirring.
19:59:55 <ehird> WHIRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRR
20:00:31 <ehird> MOTOR-FUCKING-CYCLE TIME, WHIRRRRRRRRRRRRRR I'M GOING TO ARCH UP THE POODLES
20:00:44 -!- bsmntbombdood has joined.
20:00:48 <ais523> !bfjoust ais523_defend0 +*9999
20:00:49 <EgoBot> Score for ais523_defend0: -10 (maximum 10)
20:02:18 <ais523> wow, according to report.txt, defend0 drew with every single program in there at the moment
20:02:22 <ais523> which is slightly surprising
20:02:27 <pikhq> GregorR-L: Something else nice would be being able to compile Java to LLVM and then run Java in a browser via Javascript.
20:02:29 <AnMaster> <Deewiant> Probably less than 10 MB <-- took all day
20:02:34 <ais523> !bfjoust ais523_defend0_1 .+*9999
20:02:35 <EgoBot> Score for ais523_defend0_1: -11 (maximum 11)
20:03:01 <Deewiant> AnMaster: Googling suggests it's between 3 and 6 MB depending on the precise version
20:03:30 <ais523> another drew-with-everything
20:03:39 -!- ehird_ has joined.
20:03:44 <ais523> every program's ending with an error
20:03:47 <ehird_> good MORNING america I'm feeling FABULOUS
20:03:50 <ais523> !bfjoust ais523_defend0 (+)*9999
20:03:59 <ais523> !bfjoust ais523_defend1 .(+)*9999
20:04:07 <ehird_> because AMERICA, AMERICA!3487 only the thing
20:04:10 <EgoBot> Score for ais523_defend0: -5 (maximum 11)
20:04:20 <ehird_> you, see, you see, you not see? okay.
20:04:22 <EgoBot> Score for ais523_defend1: -6 (maximum 12)
20:04:45 <ehird_> And so, the sage said: ,,Perl!''
20:05:01 <ais523> ehird_: you're not making a whole lot of sense
20:05:01 <AnMaster> Deewiant, might have been 68k running in some sort of backward compat mode, not sure if those old macs had that
20:05:16 <AnMaster> Deewiant, hm I think I downloaded Netscape 4.something too on that... Again for Mac, PPC.
20:05:42 <AnMaster> Deewiant, switched to 56k soon after that
20:05:56 <ehird_> AnMaster: what were you using before 56k
20:06:33 <Deewiant> AnMaster: Netscape 4 is already 10 MB and up.
20:06:44 <ais523> !bfjoust ais523_defend0_1 .(+)*9999
20:06:46 <pikhq> I believe that a good Morse keyer would have similar speed.
20:06:56 <EgoBot> Score for ais523_defend0_1: -6 (maximum 12)
20:06:57 <ehird_> yeah pikhq it was powered by a morse keyerer
20:07:13 <AnMaster> Deewiant, wait. this is going in the wrong direction...
20:07:25 <ais523> !bfjoust ais523_defend1 [>+[]<(.)*258(+)*127]
20:07:27 <ehird_> Deewiant: 9 is based on mozilla isn't it? 6MB for a mozilla browser? how novel
20:07:39 <Deewiant> AnMaster: Netscape 4 was quite crap, you know :-P
20:07:53 <bsmntbombdood> ehird: i put the side on this case and the gpu temp went up to like 75
20:07:56 <pikhq> ehird_: Netscape since 4.0 was rebranded Mozilla.
20:08:03 <ehird_> bsmntbombdood: that's fine
20:08:09 <EgoBot> Score for ais523_defend1: -6 (maximum 12)
20:08:11 <ehird_> GPUs run up to like 85C
20:08:17 <AnMaster> Deewiant, all netscape after 3 was
20:08:23 <ehird_> bsmntbombdood: or was this at idle?
20:08:29 <Deewiant> AnMaster: 4.5 for PPC is 13.6 MB
20:08:29 <pikhq> 9 was based on Firefox, 6, 7, and 8 were based on the suite.
20:08:53 <pikhq> AnMaster: 4, you mean. Mozilla got the Netscape 4 source and rewrote the whole thing.
20:09:02 <pikhq> Netscape 6 was Mozilla 1.0.
20:09:05 <ehird_> someone gimme a hostname for this arch installation
20:09:22 <pikhq> ehird_: Roll a dice.
20:09:30 <ehird_> bsmntbombdood - just put a fan on that thing
20:09:39 <pikhq> AnMaster: And you're wrong.
20:09:45 <ehird_> Deewiant - brilliant. crcerror, though, no -
20:09:51 <pikhq> AnMaster: After *4*.
20:10:00 <Deewiant> pikhq: Right you are, that'd explain why it was good.
20:10:04 <AnMaster> Netscape Communicator 4 was bad
20:10:17 <Deewiant> ehird_: Some phrase from the CRC error you got, then
20:10:20 <AnMaster> I'm forced to assume you didn't
20:10:20 <pikhq> The Mozilla project was created when Netscape dumped the Netscape 4.5 source.
20:10:30 <ais523> wohoo, I wrote a program that attack5 doesn't beat
20:10:33 <ehird_> Deewiant - there was none. pikhq - IE5 was betetr than NS4
20:10:34 <AnMaster> pikhq, yes, and they rewrote it
20:10:34 <Deewiant> ehird_: "not-91228e6b71d74e7a52269f1aaf225a6d' also
20:10:37 <pikhq> AnMaster: Yes, but that doesn't change that Netscape 4 was not based on Mozilla.
20:10:57 <pikhq> AnMaster: Mozilla didn't exist until Netscape 4 was well and done.
20:11:03 <Deewiant> Although I guess that kind of goes against the whole point of hostnames by making it a long & hard to remember number
20:11:08 <AnMaster> pikhq, I can't figure out if you are agreeing or disagreeing with me...
20:11:12 <pikhq> Mozilla was created instead of Netscape 5.
20:11:15 <ehird_> Mozilla was the name of Netscape's open-source-NS4 project, AnMaster.
20:11:19 <pikhq> AnMaster: I'm saying you're being dumb.
20:11:23 <ehird_> No "basing on" - direct descendant.
20:11:28 <AnMaster> ehird_, I know what Mozilla is
20:11:47 <pikhq> AnMaster: And you still said NS4 was based on Mozilla, when it's the other way around.
20:12:02 <AnMaster> pikhq, it was you who talked about Mozilla
20:12:31 <pikhq> I completely misunderstood you when you said "all netscape after 3 was".
20:12:35 <pikhq> Yes, NS4 was crap.
20:12:38 -!- ehird_ has quit (Read error: 104 (Connection reset by peer)).
20:12:55 <AnMaster> pikhq, tried NS6 too, was a memory whore, and buggy
20:12:58 <pikhq> NS6 through 9 were taking Mozilla's latest and making it into crap.
20:13:23 <pikhq> That's because NS6 was based on a Mozilla RC or a beta...
20:13:28 <pikhq> 1.0 RC, I think it was?
20:13:43 <AnMaster> iirc I tried NS9 too, never think I used NS8
20:13:45 <pikhq> And nobody in their right minds would use Mozilla 1.0.
20:13:56 <pikhq> NS9 was based on a pre-1.0 Firefox.
20:14:22 <Deewiant> I used Mozilla 1.0 for a brief period.
20:14:51 <Deewiant> I think I was using IE6 when I switched to Phoenix 0.4.
20:15:05 <pikhq> Ah, NS8 was based on pre-1.0 Firefox. NS9 was pre-2.0 Firefox.
20:15:41 <pikhq> Netscape had this odd tendency to use browser versions before the stable release.
20:18:43 <Deewiant> Pre-2.0 was stable, unless you mean it was a 2.0 RC?
20:19:28 <ehird> Deewiant: btw by no - i meant crcerror not crc-error
20:19:31 <ehird> not "not crc-error"
20:19:44 <pikhq> And NS9 didn't last very long.
20:19:56 <ehird> but efi doesn't realise its bootable
20:20:12 <ehird> i installed grub onto the partition i installed arch on to
20:23:10 <AnMaster> ooh I got esotope to traceback
20:23:16 <AnMaster> on one of my test programs for in-betwee
20:23:47 <AnMaster> +>+>+<<[>]>> ,[ >[-]+ [++>--<]< ,] ,[ >[-]++ [++>--<]< ,] ,[ >[-]++ [+++>--<]< ,]
20:24:11 <AnMaster> Traceback (most recent call last):
20:24:11 <AnMaster> File "./esotope-bfc", line 74, in <module>
20:24:13 <Deewiant> "segfault" in a Python context
20:24:22 <AnMaster> TypeError: object of type 'CGenerator' has no len()
20:24:39 <Deewiant> Dynamic typing ruins the day yet again
20:24:39 <AnMaster> too messy and undercommented code to have any idea what caused that
20:25:09 <AnMaster> Deewiant, erlang has dynamic typing. But it has a tool to verify static contracts. Optionally.
20:25:23 <AnMaster> and it can also find some of those bugs without static contracts
20:25:27 * ehird installs rEFIt, puts grub on the global mbr, runs rEFIt's gptsync
20:25:33 <AnMaster> dialyzer is the name of the tool
20:25:42 <ehird> ais523: syncs gpt/mbr
20:25:50 <ehird> gpt = GUID partition table
20:26:07 <AnMaster> does something called "successive typing", I have no idea how it works.
20:26:42 <AnMaster> bsmntbombdood, are you on linux? there may be help if so
20:26:51 <ehird> bsmntbombdood: which disks
20:27:01 -!- MizardX has quit ("What are you sinking about?").
20:27:03 <ehird> bsmntbombdood: use the new one?
20:27:15 <AnMaster> search for "Automatic Acoustic Management"
20:27:39 <AnMaster> wonder why man sometimes add two spaces
20:30:03 <AnMaster> in python, if I do print foo, where foo is a variable. And I get {2} back. What type is that...
20:30:25 <AnMaster> bsmntbombdood, isn't that { key : value } ?
20:31:24 <AnMaster> that overwides repr stuff yeah
20:43:04 -!- MizardX has joined.
20:47:39 -!- M0ny has quit ("reboot").
20:52:08 -!- M0ny has joined.
20:56:35 -!- nooga has quit (Client Quit).
21:09:52 <AnMaster> meh, too many abstractions, too little documentation, I can't make any sense out of the code
21:11:22 <pikhq> Too many abstractions? No such thing.
21:11:41 * pikhq abstracts away the computer into a butler.
21:11:58 <AnMaster> pikhq, there is nothing anywhere indicating what a block of BF code devided by an integer does.
21:12:11 <AnMaster> well one: "# This is a part of Esotope Brainfuck-to-C Compiler."
21:12:37 * pikhq abstracts away AnMaster
21:14:11 <AnMaster> Gracenotes, optimising compiler duh
21:14:40 <AnMaster> Gracenotes, it constant folds the hello world program on the wiki into PUTS("Hello, World!")
21:15:41 <AnMaster> Gracenotes, that program uses loops
21:15:46 <pikhq> Gracenotes, it does similar crazy shit to LostKingdom.
21:16:29 <AnMaster> pikhq, lostkingdom is easy to optimise. Eliminating dead code is the most important bit for LK
21:16:43 <AnMaster> apart from basic sorting and offset and such
21:16:57 <pikhq> AnMaster: It still does some crazy stuff to it.
21:18:38 <bsmntbombdood> heh, AnMaster can't even compile the output of his compiler
21:18:54 <AnMaster> bsmntbombdood, well nor can I compile the stuff esotope outputs for lostking
21:19:39 <AnMaster> also I already do better than esotope for some stuff
21:21:22 <AnMaster> from esotope: p[8] += 3; PUTC(p[8]); p[8] -= 7; PUTC(p[8]); I output that like p[8]-=4; oc(p[8]+7) oc(p[8]);
21:21:38 <AnMaster> for lostking this allows quite a bit of folding adds
21:21:50 <AnMaster> even sometimes leading to being able to figure out the constant value for it
21:25:35 -!- KingOfKarlsruhe has quit (Remote closed the connection).
21:26:49 <Deewiant> AnMaster: Can either of these compilers optimize multiplication algorithms to direct multiplications
21:27:10 <Deewiant> Or similar; other primitives than +-
21:27:11 <AnMaster> Deewiant, mine can't currently, but I'm working on doing that right atm!
21:27:19 <AnMaster> if you mean like that in bf text gen
21:27:31 <EgoBot> 102 ++++++++[>++++>+++++++++++++>+><<<<-]>>.---.+++++++..+++.<.>++++++++.--------.+++.------.--------.>++. [287]
21:27:39 <AnMaster> [>++++>+++++++++++++>+><<<<-] <-- if you mean like that
21:27:46 <Deewiant> I can't read brainfuck at all :-P
21:27:56 <Deewiant> But if that does a multiplication, then yes ;-)
21:27:57 <AnMaster> Deewiant, then what algos did you mean
21:28:23 <Deewiant> Whatever is used to do multiplication in brainfuck; if that's it, then that
21:28:56 <pikhq> Deewiant: That's just a few constants.
21:29:17 <AnMaster> pikhq, ,[>++<-] multiplies by two
21:29:31 <pikhq> AnMaster: Well, yes.
21:29:56 <AnMaster> pikhq, the fact that it multiplies by a constant is, overall, irrelevant!
21:30:11 <AnMaster> it just means it can constant fold the multiplication later
21:30:41 <Deewiant> ,[>++<-]. should become putc(getchar() << 1)
21:30:43 -!- sebbu has quit (Connection timed out).
21:30:56 <AnMaster> or earlier, since previous passes will already have transformed the loop in ++++++++[>++++>+++++++++++++>+><<<<-] to a "repeat 8 times"
21:31:10 -!- Gracenotes_ has joined.
21:34:44 -!- Gracenotes has quit (Nick collision from services.).
21:34:58 -!- Gracenotes_ has changed nick to Gracenotes.
21:35:29 -!- comex has joined.
21:36:32 <comex> I want to "partially emulate" a binary
21:36:53 <comex> run some parts of the code in emulation, but others on the actual cpu
21:37:04 <ais523> IIRC qemu does something like that
21:37:14 <comex> I tried using softx86 but it appears to suck
21:37:36 <pikhq> Deewiant: DOSemu only emulates peripherals.
21:38:00 <pikhq> On x86_32, it executes the DOS code using virtual x86 mode.
21:38:15 <comex> well, I just want it to use real (user mode) memory and stuff
21:38:19 -!- M0ny has quit ("Read error: 182 (Connection reset by beer)").
21:39:05 <comex> pikhq: how do I use qemu to run parts of the code on the actual cpu
21:39:31 <ais523> don't you need kqemu installed for that?
21:39:42 <ais523> but if you have it installed, and run qemu as root, it does automatically
21:39:44 <pikhq> ais523: kqemu lets you run kernel mode code on the actual CPU.
21:39:44 <comex> does that work for user-mod eemulation?
21:40:07 <ais523> I don't know, I'd imagine that would either work even without kqemu, or fail in both cases
21:40:17 <pikhq> Without kqemu, it only can run user-space code.
21:40:20 <AnMaster> ** exception error: no function clause matching io_lib:string_char(latin1,356,34,"!\"")
21:40:28 * AnMaster tries to figure out what on earth that means
21:41:08 <pikhq> That's, and dynamic recompiling, is why qemu is faster than Bochs. ;)
21:41:19 <comex> too bad I get errors like
21:41:21 <comex> qemu: Unsupported syscall: 240
21:41:23 <comex> and skype doesn't run
21:41:33 <comex> (probably because it's doing something weird on purpose, but nevertheless)
21:42:04 <AnMaster> Gracenotes, it is a class of exception in erlang
21:42:38 <comex> but I don't even want it to emulate system calls
21:42:43 <comex> just pass through for 99% of code
21:42:56 <pikhq> It's supposed to pass through the system calls.
21:43:07 <comex> the plan was to mprotect the 1% of code, then use a SIGSEGV handler to emulate it
21:44:54 <AnMaster> Gracenotes, there are three classes: 'error' (run-time errors, like division by zero, bad match and similar), 'exit' (when a process calls exit(Term)), 'throw' from when you just do throw(Term)
21:45:09 <AnMaster> I'm not sure why the difference between the last two is useful
21:46:31 <comex> but yeah, softx86 doesn't seem to do protected mode
21:59:09 <AnMaster> ah found it, there are reasons for the difference, related to linked processes in erlang, and supervising processes.
21:59:42 <AnMaster> too related to the details of Erlang and OTP to explain to a non-erlang programmer
22:08:31 <ehird> 12:27:33 <AnMaster> space messed up by man there
22:08:31 <ehird> 12:27:39 <AnMaster> wonder why man sometimes add two spaces
22:08:58 <AnMaster> a bit silly to do justified margins with a mono-space terminal
22:09:04 <AnMaster> but meh, can't have everything
22:09:22 <AnMaster> hm... idea: mixed formatting terminal
22:09:28 <AnMaster> like a normal terminal most of the time
22:09:49 <AnMaster> but with control codes that can change it to use different fonts in a section and such
22:10:27 <ehird> you could emulate a fb in x
22:10:46 <AnMaster> ehird, and usually if you exit the app that uses the framebuffer all the stuff from it is gone from scrollback
22:10:59 <AnMaster> ehird, which wouldn't happen with my idea
22:11:09 <AnMaster> rather you could scroll back to that
22:11:26 <ehird> AnMaster: plan9 graphics system
22:12:14 <ehird> [[‘We did not know that child abuse was a crime,’ says retired Catholic archbishop. ]]
22:12:44 <ehird> AnMaster: [[We all considered sexual abuse of minors as a moral evil, but had no understanding of its criminal nature. ]]
22:12:49 <ehird> is the actual quote
22:13:02 <ehird> http://freethinker.co.uk/2009/05/21/‘we-did-not-know-that-child-abuse-was-a-crime’-says-retired-catholic-archbishop/
22:13:35 <ehird> Bit detached from reality ther
22:14:15 <ehird> Deewiant: Oh you mean somewhat
22:14:19 <ehird> IT'S NOT TWO WORDS <__<
22:14:32 <Deewiant> ehird: "What" isn't "wut" either
22:14:45 <Deewiant> I was using the space as emphasis, actually
22:14:56 <ehird> As the flipside of that, here is the title of the post right below it:
22:14:57 <ehird> [[My school blocks www.atheist.net and categorizes it as adult language, is this a suppression of alternate beliefs?]]
22:14:58 <Deewiant> Like a pause between "some" and "what"
22:15:03 <ehird> THEY ARE TRYING TO SUPPRESS YOU
22:15:08 <ehird> IT IS NOT AN ERROR IN THE CONTENT FILTER AT ALL
22:15:20 <ehird> REBEL! REBEL! FLYING SPAGHETTI MONSTER LOLOLOLOLOL
22:22:58 <ehird> now remove the tape accesses
22:23:17 -!- nooga has joined.
22:23:18 <AnMaster> ehird, need to fix "crash due to outputting a char with the value of 823" first.
22:23:45 <AnMaster> which is [ X rem 256 || X <- Str ]
22:24:32 <AnMaster> ehird, and yes DCE from end is planned. Part of the code needed for it is already in place.
22:24:55 <AnMaster> ehird, are you on Arch or OS X now?
22:25:11 <ehird> AnMaster: os x, arch only booted via the grub console on the livecd and x didn't work first time
22:25:16 <nooga> GG (polish IM) servers are dead
22:25:19 <ehird> then I spilled some coke and decided to give up
22:25:34 <AnMaster> ehird, you aren't persistent enough IMP
22:25:35 <ehird> nooga: Gadu-Gadu's dead?
22:25:42 <ehird> AnMaster: i spent a good while on it fyi
22:25:53 <comex> anmaster: what language is that
22:25:54 <AnMaster> ehird, only because of wrong cd
22:25:57 <nooga> everything, all servers are down
22:26:07 <ehird> nooga: shut down or just glitch
22:26:57 <nooga> nope, since 20:00 all of 40 servers are dead
22:27:04 <ehird> shut down or just glitch
22:27:15 <AnMaster> thing with erlang and bf compiling: Erlang only have bignums. Thus you need to manually correct values to be in range
22:27:22 <nooga> http://gg.thinkspire.org/
22:27:29 <ehird> AnMaster: *waves Haskell and its Int8 type in your face*
22:27:54 <AnMaster> ehird, yeah sure, I'm not saying other languages doesn't have it
22:28:15 <AnMaster> ehird, and if I wanted to be sadistic I could do all my calculations inside 8 bit binaries
22:28:58 <ehird> masochistic you mean
22:29:36 <AnMaster> it would look like: <<RealValue:8/integer-native-unsigned>> = <<(FirstValue + Second Value):8/integer-native>>
22:30:10 <AnMaster> ehird, and why do you think I treat - as "add 255" not "substract 1"
22:30:25 <ehird> AnMaster: add -1, foo
22:30:44 <AnMaster> ehird, when I see a - in the input I make a node {add, 255}
22:31:03 <AnMaster> ehird, simplifies calculations
22:31:38 <AnMaster> <ehird> AnMaster: *waves Haskell and its Int8 type in your face* <-- shouldn't that be UInt8?
22:31:50 <ehird> AnMaster: that's not mathematical!
22:32:05 <AnMaster> ehird, but how do I decide if I want signed or unsigned
22:32:22 <oerjan> AnMaster: you mean Word8
22:32:35 <AnMaster> well, BF uses unsigned 8 bit cells
22:33:14 <oerjan> also, signed or unsigned is completely equivalent for wrapping BF
22:33:52 <oerjan> AnMaster: BF cannot distinguish whether it is using signed or unsigned
22:34:09 <oerjan> all it cares about is how many incs/decs until you return to 0
22:34:10 <ehird> AnMaster: -128 to 127
22:34:18 <ais523> ooh, interesting: IRC is officially assigned TCP port 194
22:34:24 <ais523> even though everyone runs it on 6667 anyway
22:34:34 <oerjan> AnMaster: that's just a matter of conversion to characters
22:34:39 <ais523> as in, it isn't news at all
22:34:44 <ais523> but I only just discovered it
22:34:47 <ehird> ais523: i doubt anyone else knew
22:34:54 <ehird> i didn't, for instance
22:34:56 <ehird> so it's news to me
22:35:16 <ais523> it's not like now is a particularly appropriate moment to announce that IRC is on port 194
22:35:21 <ehird> AnMaster: that's nice.
22:35:27 <ais523> which is why I'm not treating it as news
22:35:36 <AnMaster> I know various ircd developers who know
22:36:02 <AnMaster> I also know some ircd developers who _didn't_ know, which worries me more
22:36:21 <AnMaster> ehird, you should know the protocol you are working with
22:36:22 <ehird> why is knowing that the useless tcp assignment for irc is 194 helpful to implementing the IRC protocol
22:36:26 <ehird> apart from elitism?
22:36:28 <ehird> AnMaster: it's not the protocol
22:36:35 <ehird> the protocol is described in its entirety by a few RFCs
22:36:39 <ehird> its port assignment is irrelevant
22:36:52 <ehird> the only argument for "ircd developers must know it" is elitism
22:36:55 <AnMaster> ehird, no it isn't. Nobody fully follows the RFCs
22:37:01 -!- BeholdMyGlory has quit (Remote closed the connection).
22:37:06 <ehird> AnMaster: guess what
22:37:08 <AnMaster> a lot of stuff aren't documented in the RFCs, but is used by everyone
22:37:10 <ehird> AnMaster: NOBODY USES THAT PORT
22:37:18 <AnMaster> ehird, actually, I seen it used
22:37:21 <ehird> *snap! logical snare set by yourself*
22:37:26 <ehird> AnMaster: and I'm sure there are pure-IRC servers too
22:37:38 <ehird> but port 194 is irrelevant as implementing purely what the RFC says.
22:37:43 -!- MigoMipo has quit (Read error: 60 (Operation timed out)).
22:37:45 <nooga> oerjan: explain please -> http://www.insults.net/html/swear/norwegian.html
22:37:45 <AnMaster> and even them aren't 100% pure
22:38:40 <AnMaster> nooga, there are translations...
22:39:14 <nooga> but why "penisfuteral" :C
22:39:15 * oerjan didn't know about "fleskepanne"
22:39:45 -!- MigoMipo has joined.
22:40:11 <nooga> what's the difference between: snurebass and snabel?
22:40:11 <oerjan> nooga: i'm not sure all of those are quite often used
22:40:30 <AnMaster> the search is broken on that site
22:40:31 <oerjan> nooga: they're both just amusing words
22:41:40 <oerjan> the first really means "spinning top" (as in toy), the second means "trunk" (as in elephant)
22:42:32 <AnMaster> the latter word is the same in Swedish. But I never heard it as a alias for that thing
22:42:44 <AnMaster> as in snabel == Elephant trunk
22:43:19 <ehird> dra meg hardt i rompehara /
22:43:19 <ehird> knull meg hardt og fort - fuck me hard and fast
22:43:20 <ehird> ↑ this is also the norweigan mating ritual
22:43:33 <ehird> runknisse = wanking gnome
22:43:47 <oerjan> "dra meg baklengs inn i fuglekassa" is afaik by a bird character from a norwegian puppet movie
22:43:51 <ehird> mus = mouse (i.e. vagina)
22:43:51 <ehird> mouse, vagina, what's the difference
22:44:15 <AnMaster> ehird, the difference between a rat and a cat
22:44:40 <oerjan> "dra meg hardt i rompehåra" actually means "pull my ass hairs hard"
22:44:50 <ehird> [[Jeg spretta søstra di mens den tilbakestående
22:44:50 <ehird> faren din sto bak og fumlet med ballene mine.
22:44:51 <ehird> = I stretched your sisters ass
22:44:53 <ehird> while your retarded dad stood
22:44:55 <ehird> behind me and fumbled my balls]]
22:49:13 <AnMaster> os("\nBrainfuck Edition v0.11\n\nTo read the back-story enter '!'.\nFor a list of commands enter '?'.\n");
22:49:26 <ehird> AnMaster: not exactly hard on generated code is it
22:50:00 <AnMaster> ehird, It constant folded quite a few strings in lostking, not all yet
22:50:07 <AnMaster> since I can't handle all polynoms
22:50:13 <AnMaster> ehird, and that one was lostking
22:50:19 <pikhq> ehird: Except that there's only so many ways for *most* people to do stuff in Brainfuck. ;)
22:50:29 <ehird> but lostkng is not realistic
22:50:57 <AnMaster> ehird, doing partial loop unrolling halved the output size btw
22:51:00 <pikhq> AnMaster: Would you like the PEBBLE and PFUCK source to test with, as well?
22:51:24 -!- MigoMipo has quit ("QuitIRCServerException: MigoMipo disconnected from IRC Server").
22:52:04 <AnMaster> and I can't yet unroll if the constant is unknown, since I haven't yet coded a generic polynom node. This is for the simpler ones.
22:52:14 <AnMaster> the more complex ones I'll work on tomorrow
22:52:36 -!- tombom has quit ("Peace and Protection 4.22.2").
22:55:18 <nooga> anybody tried to do genetic programming in bf? ;o
22:55:45 <EgoBot> 49 ++++++++++[>++++++++++>+>><<<<-]>++.+++++++++..>. [60]
22:56:13 -!- Judofyr has quit (Remote closed the connection).
22:57:45 <AnMaster> also I have plans to in the future read everything into a graph of some kind. And then prune unreferenced nodes. It seems like an interesting way to handle constant propagation
22:57:53 <EgoBot> 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/
22:58:00 <AnMaster> the source for that one is in there somewhere
22:58:18 <AnMaster> nooga, however it is coded in java, to generate the bf
22:58:27 <AnMaster> not sure if that is what you meant
22:59:42 <nooga> but still, some optimizer should be applied and clean up things like >><<<< to <<
23:00:25 <oerjan> nooga: oh that point is part of a fixed template
23:00:39 <oerjan> it always uses 4 cells in that loop
23:00:54 <nooga> bud doing >><< is for nothing
23:01:20 <oerjan> nooga: it will be used for any complicated text
23:01:35 <oerjan> it's just because foo is so short
23:02:06 <EgoBot> 126 +++++++++[>++++++++>+++++++++++>+++++>+<<<<-]>.>++.+++++++..+++.>-.------------.<<+++++++++++++++.>.+++.------.--------.>+.>+. [350]
23:02:09 <oerjan> also, EgoBot doesn't run the genetic algorithm very long, so it's not optimal even for that
23:02:25 <EgoBot> 126 +++++++++[>++++++++>+++++++++++>+++++>+<<<<-]>.>++.+++++++..+++.>-.------------.<<+++++++++++++++.>.+++.------.--------.>+.>+. [317]
23:02:39 <EgoBot> 126 +++++++++[>++++++++>+++++++++++>+++++>+<<<<-]>.>++.+++++++..+++.>-.------------.<<+++++++++++++++.>.+++.------.--------.>+.>+. [706]
23:02:58 <oerjan> it seems to be deterministic...
23:03:08 <nooga> !bf_textgen i eat slugs or hedgehogs
23:03:20 <nooga> !bf_txtgen i eat slugs or hedgehogs
23:03:23 <EgoBot> 183 +++++++++++++++[>+++++++>++++++++>+++++++>++<<<<-]>>>.>++.<<<----.----.>----.>>.<<-.>+++.<++.<++++++.>--.>>.<+++.<-.>>.<<----------.---.-.<.>+.<+.>>.<++.>++++.>----------------------. [515]
23:03:25 <oerjan> i think it has that cached as a special case :D
23:03:38 <oerjan> or maybe it's because of length
23:04:05 <oerjan> AnMaster: i didn't notice nooga misspelled the command
23:04:05 <nooga> yea, i though that it's running so long
23:04:29 <oerjan> the old EgoBot used to run it much slower
23:04:57 <oerjan> !bf_txtgen i eat slugs or hedgehogs
23:05:00 <EgoBot> 185 +++++++++++++++[>+++++++>++++++++>++>+++++++<<<<-]>.>>++.<<----.----.>----.>.<-.>>+++.<<++.<++++++.>--.>.>+++.+++.<.<-----------.---.-.+++.--.<+.+++++++.>++.>>+.<----------------------. [579]
23:05:11 <oerjan> ah not identical this time
23:05:16 <oerjan> !bf_txtgen Hello, World!
23:05:19 <EgoBot> 126 +++++++++[>++++++++>+++++++++++>+++++>+<<<<-]>.>++.+++++++..+++.>-.------------.<<+++++++++++++++.>.+++.------.--------.>+.>+. [560]
23:05:46 <oerjan> but always the same on that one
23:06:26 <oerjan> i think the number after is number of generations or something
23:07:25 <oerjan> so strangely it reached the same code for Hello, World! 4 times, but after different numbers of generations
23:07:48 <oerjan> perhaps that really is optimal for that method
23:08:08 <nooga> !bf_txtgen iiiiiiii
23:08:11 <EgoBot> 47 ++++++++[>+++++++++++++>+>><<<<-]>+........>++. [513]
23:08:34 <oerjan> i think it includes the newline
23:08:42 <oerjan> new EgoBot tends to do that
23:10:27 <fizzie> That particular 126-character version seems to be the one it ends up with a very high probability.
23:10:36 <fizzie> Though sometimes rather slowly.
23:10:36 <ais523> maybe it should use annealing, or something?
23:10:38 <fizzie> 126 +++++++++[>++++++++>+++++++++++>+++++>+<<<<-]>.>++.+++++++..+++.>-.------------.<<+++++++++++++++.>.+++.------.--------.>+.>+. [5700]
23:10:49 <nooga> +++[>+<-----]>++........
23:13:00 <oerjan> !bf +++[>+<-----]>++........
23:14:01 <nooga> and who's awesome?
23:15:27 <nooga> we need better string generator
23:15:57 <ais523> nooga: are you messing around with wraparound there?
23:16:36 <oerjan> !bf ++[>-<+++++++]>+.----.
23:17:28 <nooga> aww my shrips have rotten
23:18:01 <fizzie> "shrip is an application for ripping and encoding DVDs into AVI, OGM, MKV, or MP4 files."
23:18:18 <oerjan> nooga: you rotten pirate you
23:18:19 <fizzie> It might've been about ripped DVDs and bitrot.
23:18:35 <AnMaster> ib_block_info.erl:335: The pattern 'false' can never match the type 'unknown' | {'false',pos_integer()} <-- I think dialyzer went insane. As far as I can see the type is wrong.... And isn't listed anywhere
23:18:36 <nooga> now ÿ is the smiley
23:18:57 <ehird> combining diaeresis v
23:19:47 <ehird> AnMaster: [combining diaeresis] [v]
23:20:05 <AnMaster> ehird, all I saw was a plain v
23:20:14 <ehird> AnMaster: that's all i typed
23:20:24 <ehird> i can't type a combining diaeresis
23:20:27 <ehird> so i provided a recipe
23:20:27 <AnMaster> thought you mean it was there on the line
23:20:34 <nooga> ehird: combining diaeresis v
23:20:49 <AnMaster> ehird, hm I tried that here, didn't work
23:20:59 <AnMaster> my such key only works for a few...
23:21:11 <ehird> AnMaster: its not combining
23:21:18 <ehird> if you type ë it gives the one character
23:21:29 <ehird> those are single chars
23:21:32 <ehird> instead of [combiner][char]
23:21:41 <fizzie> You put the Unicode combining characters after the to-be-combined thing, though.
23:22:05 <fizzie> v̈ should work if you renderize it right.
23:22:24 <nooga> how to type the thing
23:22:38 <ehird> v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈v̈
23:22:51 <ehird> fizzie: is there a ) on side character?
23:23:32 <nooga> http://img.moronail.net/img/0/3/903.jpg
23:25:10 <fizzie> You mean ") rotated 90 degrees", not "combining ) that's on one side of the target"?
23:25:36 <fizzie> Something like in a͡a?
23:25:52 <ehird> fizzie: that's a (
23:26:06 <fizzie> Depends on which way you rotate it. But there's the other too.
23:26:17 <ehird> fizzie: that, but lower, with a ¨ on top
23:26:49 <fizzie> Oh, right, you don't need a combining thing. Well, that's easier.
23:28:27 <fizzie> I'm not sure which block it would be in... there's a very low thing, but it's also a bit wide, I'm not sure where the diaeresis would render. ∼̈
23:29:41 <ehird> fizzie: the umlaut gives a ridge to the ) which is up at the top
23:29:45 <nooga> how do you type it?
23:30:26 <ehird> "SeekFind.org is a unique “Christian-content-only” search engine. The major search engines such as Google, Yahoo, MSN, and Ask often produce quality results for searches related to Christian terms. However, mixed in with these search results will be results from pages attacking the Christian faith and/or presenting unbiblical views. For example, a search for “Jesus Christ” at Google will result in page 1 listings from the Mormon church, a genealo
23:30:28 <ehird> gical service, and a secular history of views about Jesus. "
23:30:30 <fizzie> Well, there's a boring old bottom parenthesis, ⏝. But that's pretty low.
23:30:37 <ehird> fizzie: no, that's perfect
23:30:45 <ais523> it isn't in my font...
23:31:11 <fizzie> Hmm, I don't remember where I got the bottom parenthesis from. ⍨
23:31:23 <fizzie> (That last one is "APL FUNCTIONAL SYMBOL TILDE DIAERESIS = smirk"
23:32:12 <fizzie> It's a bit sideways in my font, but...
23:32:22 <ehird> fizzie: http://imgur.com/f0gbm.png
23:32:52 <nooga> http://img.moronail.net/img/3/2/1232.jpg
23:32:57 <fizzie> Right, it's not very centered. Well, maybe it's a style issue.
23:33:13 <ehird> fizzie: it looks nicer like that
23:33:17 -!- bsmntbombdood has quit (Remote closed the connection).
23:33:52 <fizzie> You can do the same with the "lower half circle" too, though not sure how that'd look like. ◡̈
23:34:00 <ais523> !bfjoust ais523_attackslow [>[-]-[+]+]
23:34:01 <EgoBot> Score for ais523_attackslow: -1 (maximum 13)
23:34:54 <ehird> fizzie: http://imgur.com/4f04t.png
23:35:01 <fizzie> The small ⌣ is "U+2323 SMILE" from the "miscellaneous technical" set, one would think that's a very smileable character.
23:35:55 -!- coppro has joined.
23:36:50 <fizzie> I wonder who's been thinking up those also-known-as aliases for the APL symbol set. ⍤ is known as "hoot", ⍥ as "holler".
23:36:54 <ais523> hmm... EgoBot seems to be counting draws as losses, then halving the result
23:37:12 <ehird> coppro: 23:35 fizzie: ⌤ looks very unhappy.
23:37:40 <ais523> ehird: that character isn't in this font
23:37:49 <ehird> ais523: your loss!
23:37:56 <ehird> one with more smilies
23:38:03 -!- bsmntbombdood has joined.
23:38:14 <fizzie> It's actually not a smiley; it's the "up arrowhead between two horizontal bars", used to represent the enter key.
23:38:17 <nooga> http://img.moronail.net/img/8/4/1184.jpg AnMaster: is that correct?
23:38:34 <fizzie> It's right next to the option key symbol ⌥, which isn't very facey.
23:38:47 <ehird> nooga: I'm sure AnMaster has no idea what you're talking about.
23:39:01 <ehird> fizzie: right, I had that on my old apple keyboard
23:39:04 <ehird> more legible though
23:39:12 <AnMaster> nooga, it isn't what you usually see. But I guess it happened at least once.
23:39:38 <nooga> i've been to sweden and i can tell there are MANY of nice females there
23:39:41 <AnMaster> one of them has an url on the back of the clothes it seems
23:39:52 <ehird> 23:39 AnMaster: one of them has an url on the back of the clothes it seems ← this is so AnMaster
23:39:53 <AnMaster> but if you could it might help trace it
23:39:54 <ehird> so, so very AnMaster
23:40:11 <ehird> actually it's so AnMaster it wrapped around, stopped being AnMaster, but then went a whole lot further and became very very AnMaster again
23:40:12 <nooga> http://img.moronail.net/img/7/5/1175.jpg
23:40:14 <AnMaster> ehird, the one with the dark skin
23:40:23 <ehird> nooga: i see you have discovered a new domain name.
23:40:27 <ehird> unfortunately, nobody else cares.
23:40:46 <nooga> i'm not that naive :P
23:40:58 <ehird> AnMaster: Did you also know: the flying spaghetti monster doesn't really exist?
23:41:02 <nooga> ehird: i don't care if anyone cares
23:41:09 <ehird> Tune in for more facts, tomorrow.
23:41:09 <ais523> ehird: why should I believe you?
23:41:12 <ehird> Same bat time, same bat channel!
23:41:20 <ehird> ais523: Because, as above, I am batman.
23:41:33 <ehird> AnMaster: you don't have enough slack.
23:41:34 <oerjan> ehird: of course it does, the orbital teapot told me so!
23:41:47 <ehird> AnMaster: http://en.wikipedia.org/wiki/Church_of_the_SubGenius
23:41:55 -!- ais523 has quit (Remote closed the connection).
23:45:00 <AnMaster> would using a DAC for representing a stretch of BF code and the dependencies in it be sane?
23:45:27 <ehird> graphs are always right, just ask oklopol
23:45:38 <AnMaster> ehird, well the question is: What sort of grap
23:45:54 <ehird> the most pure and eso one you can find
23:46:19 <nooga> now i'm SubGenius follower
23:46:39 <oerjan> ehird: oklopol is gravely mistaken. nearly half of all graphs are in fact, left.
23:46:39 <ehird> someone introduce nooga to a less high-brow joke religion to save them
23:46:48 <AnMaster> ehird, nah, I want something practical, useful and easy to work with
23:47:00 <ehird> AnMaster: then gtfo #esoteric and go to #BORING or something
23:47:27 <AnMaster> ehird, if it means I can compile LostKing into a readable C program? ;P
23:48:01 <nooga> what's the LostKink?
23:48:19 <oerjan> i think there must some typo energy here today
23:48:21 <ehird> AnMaster: i think,
23:48:23 <ehird> that he is trying to be funny
23:48:29 <ehird> that is my in-depth analysis
23:48:38 <ehird> nooga: also correct
23:48:42 <ehird> we're on a roll of correctness toda
23:49:11 <AnMaster> oerjan, best graph to represent BF instructions with dependencies and so on
23:49:26 <AnMaster> IO needs to be ordered in this properly somehow
23:50:02 <oerjan> AnMaster: erm you put every instruction as a vertex, and put an edge between if they cannot be swapped
23:50:33 <nooga> yay, gg servers started
23:50:51 <AnMaster> oerjan, for unbalanced ones I think the best you can do is to start building a new such network after the loop
23:51:05 <AnMaster> for balanced ones you SHOULD be able to do something better
23:51:05 <oerjan> might have to treat loops as substructures which are graphs themselves, as well as vertices in the large one
23:52:45 <oerjan> for balanced ones you should at least be able to detect which cells they can touch, and swap the whole loop with things that don't touch that
23:53:09 <AnMaster> oerjan, I can do that already by tracking a dict of how cells are touched for each loop
23:53:21 <AnMaster> that is how I can figure out if a loop can turn into if and such
23:54:03 -!- bsmntbombdood has quit (Remote closed the connection).
23:54:22 <oerjan> AnMaster: also, that topological sort page mentioned scheduling as an application, so i am sure there is already research on how to apply this to programming languages
23:54:38 <AnMaster> oerjan, yes it is what got me thinking...
23:57:04 <oerjan> nooga: so i take it gg wasn't a victim of the economic crisis yet?
23:59:26 <nooga> how come that you know what gg is
23:59:40 <nooga> it's polish only invention ;p