00:00:03 <AnMaster> btw is it even possible to write a non-trivial 2/3 polygot
00:00:16 <ehird> import sys; prunt = sys.stdout.write
00:00:20 <ehird> prunt('Hello, world!\n')
00:00:21 <AnMaster> ehird, maybe it wasn't, maybe it was just in #esoteric
00:00:25 <ehird> It's not hard if you keep within the boundaries.
00:00:48 <AnMaster> ehird, it is harder writing a C extension that works with both indeed
00:00:54 <ehird> AnMaster: so don't do that
00:00:55 <oklopol_> usually i want my randomization done in interesting and novel ways
00:00:59 <ehird> since nobody uses 3
00:01:09 <AnMaster> well nobody uses 3 because nothing supports it
00:01:18 <ehird> Untrue: work is underway.
00:01:20 <ehird> It's not too stable yet.
00:01:28 <AnMaster> the only way to make more people use it would be to support it
00:01:49 <AnMaster> ehird, well I found a crash bug in python 3 thanks to porting a C extension
00:02:39 -!- M0ny has quit ("Quit").
00:02:51 <AnMaster> because of a bug in one of the the C API functions that tried to convert a PyUnicode to PyByterray twice, in an unchecked way
00:04:58 <AnMaster> ehird, yes I tried to, but the bug reporting system never managed to send me the confirmation email
00:05:09 -!- Gracenotes has joined.
00:05:15 <AnMaster> probably the issue is at my isp
00:05:41 <ehird> Hey, when I get bogons I can be pretentious about my ISP.
00:05:43 <ehird> It's the little things!
00:06:11 <AnMaster> ehird, anyway I forgot what exact API call it was by now. I worked around it in some way iirc
00:07:09 <AnMaster> ehird, I did report it in #python with a note about that the bug tracker was broken for me.
00:08:08 <AnMaster> (really, I need to go up early tomorrow so...)
00:08:45 <ehird> Another installment in the "That classes as _music_?" series: http://www.youtube.com/watch?v=_po0RTKjsC8&fmt=18
00:19:12 <ehird> oklopol_: it goes on for 30 minutes
00:19:18 <ehird> apparently it reaches 120dB
00:19:48 <oklopol_> oh it actually sounds like that
00:21:30 <ehird> http://forum.slicehost.com/comments.php?DiscussionID=3234&page=1#Item_0
00:24:21 <ehird> As far as I can tell, bogons is just two people.
00:31:36 <bsmntbombdood> but grepping a tarball of the same directory only takes 15 seconds
00:33:23 <ehird> bsmntbombdood: directory iteration is slow
00:33:30 <ehird> and, well, files in general.
00:34:13 <bsmntbombdood> there's a stat,open,and close for every single file :/
00:34:56 <ehird> bsmntbombdood: don't grep a directory tree.
00:35:03 <ehird> consolidate multiple files.
00:40:42 -!- oerjan has joined.
00:40:48 <ehird> bsmntbombdood: What?
00:41:15 <ehird> bsmntbombdood: Well, there's no other solution.
00:41:19 <ehird> That's how the fs apis are.
00:42:11 -!- FireFly has quit ("Later").
00:42:27 <ehird> bsmntbombdood: Do you have a language impediment?
00:44:13 <ehird> bsmntbombdood: sorry to disappoint
00:44:42 <oerjan> <oklopol_> was just wondering could there be an infinite vector space for which there is no linearly independent subset which generates the space
00:45:09 <oklopol_> oerjan: so whether they always have a basis
00:45:24 <oerjan> 13:29:18 <Deewiant> Intuition says 'no'
00:45:24 <oerjan> 13:29:27 <Deewiant> But I guess with the AoC anything can happen ;-P
00:45:33 <oerjan> in fact AoC is used in the proof
00:46:24 <oklopol_> wait the axiom of death is used in it and it's on the same course as me
00:46:47 <ehird> the axiom of death: exists P. P&~P
00:47:34 <oerjan> or more directly, you use Zorn's lemma
00:48:10 <oklopol_> ah that's actually just a few pages down the material
00:48:19 <oklopol_> i think i'll leave it for next week
00:48:22 <oerjan> basically, if you have an independent set which does _not_ generate all, you can always extend it, and you can take the union of a chain of them
00:48:54 <oklopol_> and then you just take the smallest supremum
00:48:55 <oerjan> so Zorn's lemma applies to say a maximal chain generates all
00:49:21 <oerjan> there's no _the_ smallest supremum, because the extension is not unique
00:49:43 <ehird> does exists P. P&~P actually break anything
00:49:44 <oerjan> (also supremum is always smallest by definition
00:49:48 <ehird> I mean you still can't get to a P
00:50:17 <oklopol_> ehird: assume there exists P such that P&~P, then true=false => everything is true
00:50:42 <oerjan> oklopol_: that's not quite a first order proof though
00:50:43 <ehird> I was concentrating on the 'exists P.' part
00:51:16 <oklopol_> i want a "schaum's mathematical handbook with proofs"
00:51:44 <oerjan> oklopol_: also, Zorn's lemma is needed exactly to know that a maximal set exists
00:52:31 <oerjan> P => (A => P) is an axiom
00:53:44 <oerjan> so you get for that P, A => (P & ~P)
00:53:47 <ehird> oerjan: that's not true, because what if P is true by an axiom
00:54:18 <oerjan> ehird: em now i'm also using the P&~P hypothesis
00:55:02 -!- neldoreth has quit ("Lost terminal").
00:55:19 <oerjan> maybe this is overkill
00:56:07 <oerjan> and false => A is an axiom
00:56:51 <oklopol_> that's pretty much what i was going for
00:57:31 <ehird> yes but there's a reason I said 'exists'
00:57:34 <oerjan> so P & ~P => A (which could also be an axiom)
00:57:54 <oerjan> well then you need to use the quantifier rules
00:59:10 <oerjan> exists P. P & ~P => exists P. A
00:59:18 <oerjan> if that's not an axiom, it's close
00:59:34 <oerjan> finally, you can remove existentials if the variable is unused
01:02:12 <oerjan> of course since the set of axioms can vary by formulation, this is all a bit wishy-washy
01:02:44 <oerjan> *axioms/deduction rules
01:03:18 <ehird> P => ~P ← tip: Use "really untrue" things.
01:03:34 <oklopol_> yes kinda like a broken washing machine
01:03:53 <oerjan> P => ~P is equivalent to ~P, anyway
01:05:10 <oerjan> well in principle, but A => B = ~A \/ B (in classical logic), not &
01:05:36 <oklopol_> point is the redundancy was relevant to ehird's weird point
01:05:50 <oerjan> also P => ~P is equivalent to ~P with just intuitionist logic
01:05:55 <ehird> i was formulating new logic
01:06:00 <ehird> you have regular true things which are untrue
01:06:04 <ehird> so you have to use really untrue things
01:06:09 <ehird> really untrue = ~~~
01:06:39 <oklopol_> oerjan: what do you mean with just intuitionistic logic?
01:07:06 <oklopol_> actually isn't intuitionist logic a well-known logic, probably referred to that
01:07:08 <oerjan> intuitionist logic is a logic where the excluded middle doesn't hold, among other things
01:07:29 <oerjan> and where you cannot rewrite => & \/ in terms of each other
01:07:37 <oklopol_> i've just had one course in logic, and it was more about proving stuff than the general philosophical issues
01:07:39 <oerjan> although ~P = P => false still
01:07:50 <oklopol_> and by philosophy i mean other kinds of logics
01:08:37 <ehird> Under the Curry-Howard isomorphism, Peirce's law is the type of continuation operators, e.g. call/cc in Scheme[1].
01:08:41 <ehird> that's fucking hot
01:09:13 <ehird> oklopol_: ((p->q)->p)->q
01:09:19 <ehird> curry-howard isomorphism is like the best ting to happen to programming
01:09:44 <oerjan> intuitionistic logic corresponds to ordinary typed lambda calculus, while classical logic (with peirce's law etc.) corresponds to a lambda calculus with continuations
01:09:46 <oklopol_> hard to say, i don't understand it.
01:10:00 <ehird> modusPonens :: (p -> q, p) -> q
01:10:05 <ehird> modusPonens (f,x) = f x
01:10:30 <ehird> (where "data Void")
01:11:05 <oklopol_> yeah i've seen the definitions in many forms
01:11:13 <ehird> so how don't you understand it
01:11:19 <ehird> type = proposition, non-bottom implementation = proof
01:12:06 <oerjan> falseImpEverything :: Void -> a
01:12:12 <oklopol_> i mean just seeing a syntactical mapping between types and logic doesn't really tell me much.
01:12:17 <oerjan> falseImpEverything x = case x of {}
01:12:28 <ehird> that's a bottom implementation
01:12:46 <oerjan> because Void has no constructors
01:12:47 <ehird> (although you can't get any (Not a) with no-bottom rule, so maybe we need a better definition)
01:12:56 <ehird> (i mean maybe I am wrong)
01:13:25 <ehird> there's no value of type Void that isn't bottom
01:13:52 <ehird> oerjan: youc an't implement that
01:13:54 <oerjan> so you need to start with a contradiction
01:14:06 <oerjan> well you said _any_ Not a
01:14:17 <oerjan> there are specific a's you can get
01:14:41 <ehird> i guess you have to go into axiom land to get more
01:14:43 <oerjan> another would be ~(p & ~p)
01:14:47 <ehird> which involves bottom, I guess
01:14:58 <ehird> ((p, p -> Void) -> Void)
01:15:10 <ehird> prove with fix, I assume?
01:15:20 <ehird> how can you prove that
01:15:43 <ehird> oerjan: you don't have a (p,q)
01:15:47 <ehird> you have a ((p,q)->Void)
01:16:05 <ehird> ((p, p -> Void) -> Void)
01:16:12 <ehird> so how can you prove
01:16:15 <oerjan> f :: ((p, p -> Void) -> Void)
01:16:24 <ehird> you mean proving ~(p&~p)
01:16:33 <ehird> instead of proving Not ~(p&~p)
01:17:27 <oerjan> for anything of the form Not a, the intuitionist and classical tautologies are the same btw.
01:17:44 <ehird> oklopol_: yeah it's pretty hot isn't it, every piece of code you write is a proof of some proposition
01:18:37 <ehird> oklopol_: erm python doesn't really have types as such...
01:19:04 <oerjan> it works best with functional languages
01:19:23 <oklopol_> well i was mainly hoping for a fun pun answer
01:19:27 <ehird> functional strongly typed langs
01:19:54 <oerjan> oklopol_: the serpent calculus
01:20:15 <oerjan> i'm sure that is fun if you get the obscure logic reference
01:20:41 <oklopol_> but i'll just assume it is really something.
01:20:42 <oerjan> sequent calculus is a way of axiomatizing logics
01:21:54 <oklopol_> right was too busy correcting my keyboard's errors to see mine
01:21:56 <ehird> You agree that neither you nor any person using the Service with your permission shall use the Service: for any fraudulent, criminal or unlawful purpose; to send unsolicited advertising or promotional material (or to engage in any "spamming" activity); to transmit any virus, worm, trojan horse or other harmful material; for the purpose of intimidating, harassing or causing annoyance to any third party
01:22:02 <ehird> being annoying = illegal
01:25:12 <oklopol_> so err f (q, p) = p q proves f :: ((p, p -> Void) -> Void), but can't you prove the inverse, f :: (p, p -> Void), with simply having f be a pair of (some value, function that bottoms)
01:25:20 <oklopol_> no i don't think i still get it.
01:25:44 <ehird> oklopol_: you cant use bottom
01:27:24 <ehird> oklopol_: using bottom isn't allowed
01:27:39 <oklopol_> where did i use it incorrectly?
01:27:49 <oklopol_> i was thinking f = (value, mockingbird)
01:27:58 <ehird> (some value, function that bottoms)
01:28:11 <ehird> in any circumstance
01:28:31 <oerjan> only terminating functions allowed
01:28:45 <oerjan> otherwise your logic is inconsistent
01:30:08 <oklopol_> so how isn't f (q, p) = p q wrong then, p bottoms
01:30:13 <oerjan> also, the mockingbird doesn't type in ordinary haskell
01:30:20 <ehird> ... o rly oklopol_?
01:30:27 <oerjan> oklopol_: it's an argument
01:30:53 <oklopol_> ehird: wut? that was a question
01:30:56 <oerjan> yes, Void values bottom, so you cannot produce them, but you can still look at what happens if you get them in
01:31:43 <oklopol_> i mean (value, mockingbird) is a terminating implementation
01:32:01 <ehird> awesome, bogons.net cofounder posted about how he hates Phorm on a mailing list
01:32:02 <oerjan> it needs to be well-typed in the logic too
01:32:58 <oerjan> there _are_ ways to type the mockingbird, but they involve letting an expression have two types simultaneously
01:33:01 <oklopol_> oerjan: well okay mockingbird was a bad idea
01:33:50 <oerjan> (\a -> a a) :: (A /\ (A => B)) => B
01:34:02 <oerjan> but that's not a haskell or ML type
01:34:51 <oerjan> (/\ here has a different meaning than the (,) one, it means a should _actually_ be both types simultaneously)
01:39:11 <oklopol_> need to go give a friend a book at the railway station at exactly 8 am
01:39:19 <oerjan> "no sleep today, the milk has run awaaay"
01:39:28 <oerjan> or do i remember that wrong
01:40:05 <oerjan> i knew it, oklopol_ is a terrorist!
01:40:24 -!- oklopol_ has changed nick to oklopol.
01:40:55 <oerjan> who but a terrorist would show up at railway stations at exactly 8 am to give friends books
01:41:07 <oklopol> this guy has an exam on monday
01:41:15 <oerjan> i guess you _could_ be a secret agent
01:41:26 <oklopol> so it's somewhat crucial to get the book to him as early as possible
01:41:30 <oerjan> bomb making is no easy course
01:42:46 <oerjan> like, at least two weeks ago?
01:43:51 <oerjan> you don't need a railway station, you need a time machine
01:44:19 <oklopol> coolness is not only a top clicker, he's a dedicated reader too
01:45:03 <oerjan> #pokenet? is that some pokemon thing?
01:45:28 <oklopol> hmm err different coolness
01:45:39 <oklopol> i'm talking about the guy who played ehird's fun clickery game
01:46:44 <oklopol> he recently spent a week reading 16-20 hours a day, i have a long way to go.
01:50:22 <oklopol> and why are our university's pages constantly broken, i want to stare at my exam schedule :<
01:51:23 <oerjan> must be all the hackers doing exercises
01:52:30 <oklopol> sadly we don't have any explicit hacking courses
01:52:53 <oerjan> how can you be a successful terrorist university without hacking courses
01:53:38 <oklopol> i'm not sure i can parse that.
01:54:43 <oklopol> there are implicit hacking courses, you need to learn some non-hacking too
01:55:12 <oklopol> maybe it's so you can pretend you're not a terrorist when necessary
01:58:40 <oklopol> also maybe it's the university pretending that.
01:59:46 <oklopol> somehow deciding to stay awake because i'd only get a few hours of sleep reminds me of the story of the guy who threw all the eggs away because he dropped one.
02:00:57 * oklopol decides he probably shouldn't read because he didn't reach his quote anyway
02:03:06 <oklopol> wow i beat computer level 1 in reversi
02:22:22 -!- Sgeo has joined.
02:56:54 <kerlo> A campfire hanging by a knotted string.
02:58:58 <kerlo> oklopol: that's not a random association at all.
02:59:16 <kerlo> It's a precisely appropriate association.
02:59:41 <kerlo> Mot juste, in fact.
03:00:32 <kerlo> Or mō zhōos, as we reformers like to spell it. (The macrons are optional, but consider adding silent e's to compensate.)
03:02:25 <oerjan> what happened to the poor t
03:02:56 <kerlo> My English teacher failed to pronounce it.
03:03:02 <kerlo> I think the French don't or something.
03:03:41 <oerjan> i think your english teacher had a speech impediment
03:04:06 -!- Sgeo has quit (Read error: 104 (Connection reset by peer)).
03:04:13 <oerjan> http://en.wiktionary.org/wiki/mot_juste does not agree in either english or french
03:04:14 -!- Sgeo has joined.
03:05:14 <oerjan> nor does http://www.merriam-webster.com/dictionary/mot+juste
03:06:35 -!- olsner has joined.
03:09:41 <oklopol> kerlo: it was? i must be smarter than i thought.
03:10:14 <oklopol> in fact it is not the exact same thing, half of the reason not to sleep is being afraid of not waking up in time
03:10:22 * kerlo puts the Humble Hat on oklopol.
03:10:31 <oklopol> but the other half is that egg thing.
03:10:54 <oerjan> it is very bad to wake up with egg on your face
03:13:46 <oklopol> thought you meant the latter
03:14:15 <oerjan> in a self-referential way
03:14:58 <kerlo> What's an interesting use of o?
03:16:15 <oerjan> a use that is pronounced nowhere near lower back afaik
03:17:39 <oerjan> the french is of course an upper front rounded vowel
03:19:08 <kerlo> By "lower" and "upper", do you mean... well, no, I don't know what you mean.
03:19:28 <kerlo> Oh, low is open and high is close.
03:19:49 <kerlo> So the open-mid back rounded vowel is indeed near lower back.
03:19:56 <oerjan> terminology is so forble
03:20:32 <kerlo> And the "o" in "mot juste" is open-mid back rounded, if Wiktionary is trustworthy.
03:21:04 <oerjan> but the u is definitely not
03:21:11 <oerjan> that was the interesting part
03:21:11 -!- olsner has quit ("Leaving").
03:23:40 <kerlo> So the "u" isn't like the stereotypical "u" at all.
03:24:21 <oerjan> unless you mean stereotypically french, in which case it is
03:26:41 <kerlo> I guess the anglicisation preserved the roundedness rather than the frontness.
03:27:49 <kerlo> Using ōo to mean the vowel in "goose".
03:29:15 <oerjan> i don't read either wiktionary or merriam-webster that way.
03:29:39 <oerjan> although the latter has its own strange set of pronunciation symbols
03:30:17 * oerjan suddenly realizes his typing is nowhere near touch these days
03:30:27 <kerlo> Well, Wiktionary's representation of that vowel is more like ōō.
03:30:39 <kerlo> A long macron that goes right over both o's.
03:31:58 <oerjan> what i am saying is that neither actually says mot juste is pronounced with an oo sound in english
03:39:20 <kerlo> But hey, we're anglophones. We anglicise.
03:43:53 <oklopol> i think i'm going to save one egg by sleeping for an hour
03:44:24 <oklopol> wake me up at 7:30 if i'm not awake yet, current time being 5:54
03:45:15 <kerlo> oklopol, your clock is off by approximately 10 minutes.
03:45:19 <kerlo> Or maybe my clock sucks.
03:46:00 <oerjan> oklopol: and then you can have egg and bacon for breakfast! well, assuming you have bacon.
03:46:33 <oerjan> kerlo: it's a vacuum clock?
03:46:51 <oklopol> kerlo: three clocks here show the exact same time
03:47:03 <oklopol> so probably at least one of them is correct
03:47:43 <oerjan> actually i would assume the probability is either none, or exactly three are
03:47:44 <kerlo> No, they're all incorrect.
03:48:03 <oerjan> damn it's hard to touch type when you've slipped out of doing it properly
03:48:04 <kerlo> The correlation is probably due to a common cause that is not the correct time.
03:48:23 <kerlo> Like setting two of the clocks to the other.
03:48:32 <kerlo> Or setting all three clocks to an outside incorrect clock.
03:48:49 <oerjan> WHAT? WHO WOULD DO SUCH A SILLY THING?
03:48:51 <oklopol> the computer's clock was set a few years ago, and has been ticking silently in the corner of my room until i switched it on today
03:49:22 * oerjan wonders if the way he is usually typing is better for his wrists
03:49:23 <oklopol> my cell phone asks its time ...somewhere every time it boots
03:49:45 <oklopol> and the third clock was set by me, some time ago
03:50:04 <oklopol> kerlo: ideas for correlation?
03:50:14 <kerlo> Do you live in a very strange time zone?
03:50:20 <oklopol> well, me lying. but that's the easy way out
03:50:35 <oklopol> i don't think time zones get any stranger than 30 min off
03:50:39 <kerlo> UTC+3:50 or something?
03:50:42 <oerjan> YOUR HART IS FULL OF LYE!
03:50:58 <kerlo> What's your daylight saving time offset?
03:51:11 <oklopol> ah! ten minutes, that must be it
03:51:30 <oklopol> yeah so anyway take that into account when waking me up, sleep time
03:52:25 <oerjan> oklopol: actually there are some quarter zones
03:52:34 <kerlo> So wake you up at 00:20 local time.
03:53:12 <oerjan> http://en.wikipedia.org/wiki/List_of_time_zones
03:56:08 <oerjan> nepal, chatham islands, a village in australia (unofficial)
04:02:22 * kerlo separates sentences into chunks of unique letters
04:04:24 <kerlo> separ|ates |sent|enc|es into| chunks| of| uniq|ue l|et|ters
04:04:43 <kerlo> Touch typing challenge: type each of those chunks without moving the fingers between keys.
04:04:51 <kerlo> Only move fingers between keys between chunks.
04:08:11 <oerjan> next, keep one finger on backspace :D
04:08:50 <kerlo> I want a key that sends backspace about four times.
04:09:09 <oerjan> separates sentences damn
04:09:34 <oerjan> forgot space was not |
04:09:50 * oerjan gives up this exercise, which is stupid and STINKS
04:10:13 <kerlo> separsate ssentencesin tochukn of unique letters
04:10:18 <kerlo> I need more practice.
04:10:37 <oerjan> that sentence doesn't separse
04:12:25 <kerlo> I think I'll use more gentle chunking next time.
04:12:25 <oerjan> tochukn is probably a village in alaska
04:13:26 <kerlo> i[ th]i[nk ]il[]l [use] [more] ge[ntl]e chun[ki]n[g ]next[ ]time
04:13:33 <kerlo> Put a chunk break anywhere within the brackets.
04:17:32 <kerlo> i th|ink| il|l use| more| gentl|e chun|king| next| time
05:27:44 <oklopol> <oklopol> well, me lying. but that's the easy way out <<< in fact this turned out to be pretty close, i simply read the times wrong, the computer's clock is 10 min wrong :)
05:28:17 <oklopol> may still another one is wrong, i'll know only after more of it
05:28:58 -!- psygnisfive has joined.
05:30:48 <MizardX> psygnisfive: http://www.n-tv.de/img/1131671_src_path.E3IA.jpg
05:39:22 -!- bsmntbombdood has quit (Read error: 145 (Connection timed out)).
05:49:48 -!- bsmntbombdood has joined.
06:21:34 -!- jix_ has joined.
06:24:30 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
06:32:56 -!- GreaseMonkey has quit ("X-Chat -> http://xchat.org <- At least when I quit I don't look like a lamer").
06:36:04 -!- jix has quit (Read error: 110 (Connection timed out)).
06:45:53 <oklopol> animals are a weird concept
06:47:54 <oerjan> because i woke up 7 hours ago?
06:48:13 <oerjan> although i did consider taking another nap
06:48:36 <oklopol> i'm not really even tired anymore
06:48:48 <oklopol> but i have a happy hunch i'm going to be, soon.
06:49:07 <oerjan> IT'S LIKE AN ADDICTION I TELLYA
06:50:05 <oklopol> our drummer has holy days from his army duty, and suggested we play today; god i hope we don't, so i can just you know dozerate all day long
06:51:42 <oklopol> also the guy i gave the book to is probably going to do fine, we went over the contents, and it seems the course is a pretty thorough review of about 8 other courses
06:54:05 <oerjan> holy days, sacred nights
06:55:25 <oerjan> you want to drive around wrecking the neigborhood?
06:55:55 <oerjan> wait, you already gave him the book? darn timezones
06:56:19 <oerjan> clearly it should involve a bulldozer somehow, right?
06:56:38 <oklopol> ah i thought it made sense somehow
06:56:54 <oklopol> but couldn't get my fingers to touch it.
07:07:38 <fungot> SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM SPAM ...too much output!
07:09:24 <oklopol> thanks a lot, 10 min into the episode, an ad pops up, plays for 80 minutes and the rest of the episode won't play
07:16:18 <oklopol> what's the coolest round object in germany?
07:17:19 * oerjan is reminded of that hitler globe
07:19:47 * oerjan finds http://www.nytimes.com/2007/09/18/arts/design/18globe.html?_r=1
07:40:22 <oklopol> can you give me the cliff notes version
07:41:20 <oerjan> well the nazis made these huge globes, and hitler had several
07:41:43 <oerjan> and this guy is searching for the one that was at hitler's headquarters
07:41:59 <oerjan> and he doesn't think it's any of the ones at the known museums
07:42:58 <oerjan> well the one that claims to be it actually has a specially designed foot that matches one that was ordered for ribbentrof, not hitler
07:43:25 <oklopol> sounds like a pretty boring movie
07:43:31 <oerjan> i did not read it all very carefully
07:43:54 <oerjan> and there's a bit about chaplin in there
07:44:26 <oklopol> i read that one, recognized him in th epic
07:44:29 <oerjan> he apparently based a scene in The Great Dictator on parodying hitler and his globe
07:45:27 <oklopol> "uhh take that call me your führer"
07:45:44 <oklopol> well. likely or not, it's true
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:18:31 -!- kar8nga has joined.
08:30:04 -!- Mony has joined.
08:30:32 -!- Mony has changed nick to Guest71668.
08:30:48 -!- Guest71668 has changed nick to M0ny.
08:37:05 -!- neldoreth has joined.
09:00:01 -!- jix_ has quit (Read error: 113 (No route to host)).
09:03:04 -!- BeholdMyGlory has joined.
09:18:48 <AnMaster> <bsmntbombdood> there's got to be a good way to do it <-- what about opening the partition itself and implementing a file system driver in your app, and then search directly that way?
09:19:14 <AnMaster> bsmntbombdood, 1) needs root 2) complex and messy
09:19:31 <AnMaster> bsmntbombdood, or you could create an index of the directory then search in that index
09:20:14 <AnMaster> bsmntbombdood, if you are searching C source files apps like cscope and ctags might be useful, depending on what exactly you want to search for
09:20:50 <AnMaster> bsmntbombdood, but you would mind rolling your own file system driver in userspace?
09:21:16 <AnMaster> you wouldn't write using it anywya
09:23:27 <AnMaster> bsmntbombdood, didn't you say reading an uncompressed tar was fast?
09:24:03 <AnMaster> <bsmntbombdood> so i go into a huge directory and grep foo -R .
09:24:03 <AnMaster> <bsmntbombdood> this takes like 3 minutes
09:24:03 <AnMaster> <bsmntbombdood> but grepping a tarball of the same directory only takes 15 seconds
09:24:09 -!- kar8nga has quit (Read error: 60 (Operation timed out)).
09:24:10 <AnMaster> sounds quite a bit faster yeah
09:24:31 <AnMaster> bsmntbombdood, so use a file system that compresses on the fly?
09:24:57 <AnMaster> bsmntbombdood, use solid state drive or ram disk?
09:25:20 <AnMaster> bsmntbombdood, um, how large was that directory then...?
09:26:56 <AnMaster> $ dd if=/dev/zero of=/dev/null bs=1G count=2
09:26:56 <AnMaster> 2147483648 bytes (2,1 GB) copied, 2,7549 s, 780 MB/s
09:27:15 <AnMaster> bsmntbombdood, warning: bs must be smaller than free ram
09:27:57 <AnMaster> bsmntbombdood, it is my computer. From late 2005.
09:28:08 <bsmntbombdood> 2147483648 bytes (2.1 GB) copied, 0.0419093 s, 51.2 GB/s
09:28:13 <AnMaster> Linux tux.lan 2.6.27-gentoo-r8-1 #1 Sat Jan 31 04:55:36 CET 2009 x86_64 AMD Sempron(tm) Processor 3300+ AuthenticAMD GNU/Linux
09:28:26 <AnMaster> bsmntbombdood, that would be faster probably
09:28:39 <AnMaster> $ dd if=/dev/zero of=/dev/null bs=1M count=2K
09:28:39 <AnMaster> 2147483648 bytes (2,1 GB) copied, 2,34859 s, 914 MB/s
09:28:55 <AnMaster> bsmntbombdood, probably faster due to cache
09:29:09 <AnMaster> $ dd if=/dev/zero of=/dev/null bs=128K count=2K
09:29:10 <AnMaster> 268435456 bytes (268 MB) copied, 0,0579806 s, 4,6 GB/s
09:29:20 <bsmntbombdood> we are looking at a 4 order of magnitude difference here...
09:29:20 <AnMaster> the L2 cache on this CPU is 128 KB
09:30:16 <AnMaster> $ dd if=/dev/zero of=/dev/null bs=128K count=50K
09:30:16 <AnMaster> 6710886400 bytes (6,7 GB) copied, 1,34931 s, 5,0 GB/s
09:30:44 <AnMaster> bsmntbombdood, single core Sempron from 2005, what do you expect?
09:32:00 <AnMaster> bsmntbombdood, actually this should be one of the few things that a pentium is good at, it has larger cache than my CPU probably, there won't be a lot of branch prediction misses either
09:33:07 <AnMaster> bsmntbombdood, well memory bw could explain that. + X here is eating 10 % CPU atm... no clue why
09:33:41 <AnMaster> X kind of has issues when running for a few weeks without restart
09:35:59 <bsmntbombdood> also, comma instead of period for a decimal point is retarded
09:38:09 <AnMaster> a period instead of a comma is equally retarded
09:38:18 <AnMaster> bsmntbombdood, or do you have any justification?
09:39:03 <AnMaster> bsmntbombdood, how does something being European make that thing retarded?
09:39:32 <AnMaster> bsmntbombdood, why do you think Europe is retarded then?
09:40:25 <AnMaster> so you are just trolling then, right.
10:06:35 <AnMaster> I'm trying to fix your mycoedge so I can test if my exact bounds code work
10:07:00 <AnMaster> should it be (14,5) or (14,4) after removing (14,0) ?
10:07:17 <AnMaster> I think y should be 4, but mycoedge thinks it should be 5
10:08:25 <AnMaster> GOOD: y says initial minimal point is (14,0)
10:08:25 <AnMaster> BAD: after removing (14,0) y doesn't say minimal point is (14,5)
10:08:25 <AnMaster> BAD: after removing (14,5) y doesn't say minimal point is (15,5)
10:08:35 <AnMaster> but yes that is off by one for y I think
10:09:59 -!- sebbu has quit ("@+").
10:15:46 -!- oerjan has quit ("leaving").
10:28:32 -!- tombom has joined.
10:32:25 -!- MigoMipo has joined.
10:52:58 -!- sebbu has joined.
10:56:39 -!- M0ny has quit ("PEW PEW").
11:16:49 <AnMaster> Deewiant, the slowdown script is broken, It doesn't properly clear around 0,0
11:17:02 <AnMaster> there are still stuff left there
11:22:28 -!- KingOfKarlsruhe has joined.
11:23:28 -!- Judofyr has joined.
11:31:23 <AnMaster> Deewiant, I think you clear relative offset incorrectly or something
11:37:12 <AnMaster> Deewiant, also it causes errors in stack stack testing in mycology
11:37:41 <AnMaster> after manually clearing the static area and waiting for recalcuation of bounds (took a few minutes yeah)
11:52:35 -!- FireFly has joined.
12:05:35 -!- tombom has quit ("Peace and Protection 4.22.2").
12:14:53 <AnMaster> Deewiant, ah yes, mycology also tries to remove any stack stack when it detects it is not correct, thus causing huge bounds again
12:15:18 <AnMaster> at "WARNING: attempted recovery of stack stack, situation may be corrupt"
12:16:24 <AnMaster> so slowdown doesn't work 1) at all, because the clearing code is incorrect 2) for mycology, because mycology restores stack stack a bit before y.
12:19:05 <fizzie> It does work "at all", because it slows down things. It even works better than intended, as it slows down things even for interpreters that shrink bounds.
12:19:23 <AnMaster> fizzie, well, that is unfixable.
12:19:55 <AnMaster> btw I now have two stratergies, depending on how large the difference is
12:20:02 <AnMaster> if it is small I just try to shrink from the edges
12:20:38 <AnMaster> if it is large I instead iterate over all recorded values in the hash table and store max/min, then iterate over the static space and adjust those max/min if needed
12:21:12 <AnMaster> currently large is defined as: max-min > 0x10000
12:21:33 <AnMaster> will have to tune that probably
12:21:50 <fizzie> You could also select it based on the hash table size, I guess.
12:23:14 <fizzie> Since the hash table iteration time depends on how big it is, you could concievably tune your definition of a "large" difference based on how many entries there are.
12:23:28 <AnMaster> "large model" is way faster when difference is 10 000 000 or so (as in slowdown, after I manually cleared static space), but much worse when the bounds are smaller and I would only need to shrink with a few columns/row
12:24:05 <AnMaster> but first lets get rid of code duplication here.. brb
12:25:24 <Deewiant> 2) I suspected might happen but wasn't sure
12:26:14 <Deewiant> 1) is a bug I guess; which cells aren't cleared?
12:26:38 <AnMaster> Deewiant, those of the initial program at least.
12:27:08 <AnMaster> Deewiant, unknown, ccbi has a much better debugger, I would suggest trying with it?
12:27:11 <fizzie> Deewiant: Didn't you have a debugger, anyway? :p
12:27:22 <Deewiant> Hmm, wonder what I messed up there
12:27:53 <AnMaster> Deewiant, guess: you didn't compensate for the recently created storage offset when clearing?
12:28:08 <AnMaster> something like that sounds quite plausible to me
12:29:58 <Deewiant> No, it looks like I just forgot about an area
12:30:49 <Deewiant> Not sure how or why though, I'll take a look at it at some point, no time now
12:33:51 -!- olsner has joined.
12:39:11 <AnMaster> $13 = {32 <repeats 32832 times>, 118, 32, 32, 32, 32, 32, 32, 118, 32, 60, 46, 46, 47, 46, 46, 47, 116, 114, 117, 110, 107, 47, 109, 121, 99, 111,
12:39:11 <AnMaster> 108, 111, 103, 121, 47, 109, 121, 99, 111, 108, 111, 103, 121, 46, 98, 57, 56, 32 <repeats 469 times>, 380260801, -1325984155, 180, 795, 42, 107,
12:39:11 <AnMaster> 36, 62, 35, 94, 95, 48, 52, 52, 112, 62, 58, 35, 118, 95, 110, 52, 52, 103, 35, 118, 95, 118, 32 <repeats 484 times>, 380260981, -1325983360, 32,
12:39:12 <AnMaster> 32, 118, 32, 32, 32, 32, 32, 112, 48, 43, 97, 103, 32, 52, 52, 60, 32, 32, 32, 32, 32, 32, 32, 32, 62, 48, 97, 34, 46, 101, 109, 97, 110, 32, 101,
12:39:15 <AnMaster> 108, 105, 102, 32, 121, 116, 112, 109, 101, 110, 111, 110, 32, 100, 101, 101, 78, 34, 32, 32, 32, 118, 32 <repeats 452 times>, 118, 32, 71, 45, 62,
12:39:18 <AnMaster> 52, 52, 103, 49, 43, 52, 52, 112, 32, 32, 94, 32, 32, 32, 32, 32, 32, 32, 32, 32, 62, 32 <repeats 490 times>...}
12:39:23 <AnMaster> seems like you forgot quite a bit indeed
12:40:08 <Deewiant> AnMaster: Did a quick fix, try the one I just uploaded
12:47:22 -!- KingOfKarlsruhe has quit (Remote closed the connection).
13:00:32 <oklopol> AnMaster: comma as decimal point is retarded.
13:00:55 <oklopol> you use comma in them for a different purpose
13:01:28 <oklopol> in finland, the standard notation for tuples becomes (a; b) when either a or b contains a decimal point (comma)
13:01:35 <AnMaster> oklopol, and why should tuples use comma then? Why not make them use dots. Sure it would be a bit unusual, but why not?
13:01:40 <oklopol> a better solution would be to use ; all the time ofc
13:01:48 <AnMaster> after all lisp use . for pairs
13:02:29 <oklopol> AnMaster: clearly just swapping all notation is an automorphism.
13:02:49 <oklopol> but if tuples use comma, it shouldn't be the decimal point
13:03:47 <oklopol> also natural language uses commas in tuples, "x, y and z", but there commas as decimal points aren't really confusing, because numbers are less frequent; still pretty ugly tho
13:04:16 <oklopol> AnMaster: no use googling, i'm using my intuitive view of it
13:04:32 <AnMaster> oklopol, care to explain what you mean then?
13:04:38 <oklopol> if you knew what automorphisms were, you might understand what i meant, or, umm, not.
13:04:55 <oklopol> if you swap all notation, there will be no collisions.
13:05:06 <oklopol> unless you care about syntactical differences.
13:05:11 <AnMaster> oklopol, I just fail at knowing the English terminology
13:05:19 <fizzie> Also they use the same "do ; if the tuple item contains a ," rule: "foo, which means bar; baz, which means quux and zuul, which means zingobongo".
13:05:43 <AnMaster> I googled and I actually ran into it in Swedish, but I didn't know the English word.
13:06:05 <oklopol> lojban doesn't use notation for anything, it uses terms.
13:06:50 <oklopol> and automorphisms are a mathematical concept, a property of a mapping from a system to itself
13:07:05 <fizzie> Korpela lists another typical ;-use in lists of people-names in the "Surname, Given name" format.
13:07:41 <oklopol> you could consider it a handy way to do nested lists or depth 2
13:07:52 <oklopol> oklotalk does that, after all
13:08:10 <oklopol> (in some cases, in most cases not)
13:08:37 <oklopol> err, band training sessions time.
13:08:38 <fizzie> Yes; also you could extend it so that a ", with two dots over" would be the separator for lists of depth three, and so on.
13:09:22 <fizzie> ;̈ if you happen to get combining-characters rendered correctly.
13:09:23 <oklopol> except for some reason i like it when (especially) languages break the 0-1-infinity rule
13:40:59 -!- Asztal_ has quit (Read error: 104 (Connection reset by peer)).
13:47:00 -!- Mony has joined.
13:47:28 -!- MigoMipo has quit (Read error: 60 (Operation timed out)).
13:47:28 -!- Mony has changed nick to Guest99604.
13:49:28 -!- Guest99604 has changed nick to M0ny.
14:54:17 -!- Judofyr has quit (Remote closed the connection).
14:55:09 <Deewiant> AnMaster: Hmm, it worked for a small example
14:57:35 <Deewiant> Hmm, I wonder what's wrong, now it clears around 200 lines down for Mycology but not further
15:02:10 <AnMaster> Deewiant, since cfunge lacks a debugger beyond gdb I'll let you figure it out
15:02:47 -!- BeholdMyGlory has quit (Remote closed the connection).
15:03:15 <Deewiant> Alright, looks like it /should/ be clearing down to 249,836
15:03:22 <AnMaster> Deewiant, however, if I manually clear it from inside gdb then cfunge handles it fine (assuming EXACT_BOUNDS was turned on at compile time). Mycology's stack-stack check breaks however
15:03:28 <lifthrasiir> heh, i never knew that all lines in the unefunge source code should be concatenated into one line.
15:03:29 <Deewiant> Now let's see where it /is/ clearing down to
15:03:36 -!- MigoMipo has joined.
15:03:55 <AnMaster> lifthrasiir, oh yes they should. For Befunge I recommend handling form feed like that to
15:04:34 <AnMaster> that is, do not advance x, y or z when loading a file and you hit a form feed if you are in befunge mode and are not loading with i set to binary mode
15:04:39 <Deewiant> Argh, it looks like it was just a typo
15:05:29 <lifthrasiir> AnMaster: that means form feed should be ignored in unefunge and befunge, and newline should be ignored in unefunge, right? it seems the spec is almost clear about that.
15:05:36 <Deewiant> Easier to ask you than to print 1000x1000 areas into my terminal looking for chars ;-)
15:05:46 <AnMaster> $1 = {topLeftCorner = {x = 128, y = -910415652}, bottomRightCorner = {x = 2128881350, y = 967}, entries = 0x1958010, col_count = 0x1958050,
15:05:46 <AnMaster> row_count = 0x1958090, boundsexact = true, boundsvalid = true}
15:05:51 <AnMaster> $2 = {32 <repeats 32832 times>, 118, 32, 32, 32, 32, 32, 32, 118, 32, 60, 46, 46, 47, 46, 46, 47, 116, 114, 117, 110, 107, 47, 109, 121, 99, 111,
15:05:52 <AnMaster> 108, 111, 103, 121, 47, 109, 121, 99, 111, 108, 111, 103, 121, 46, 98, 57, 56, 32 <repeats 469 times>, 2128881171, -910415647, 180, 795, 42, 107,
15:05:52 <AnMaster> 36, 62, 35, 94, 95, 48, 52, 52, 112, 62, 58, 35, 118, 95, 110, 52, 52, 103, 35, 118, 95, 118, 32 <repeats 484 times>, 2128881351, -910414852, 32,
15:05:52 <AnMaster> 32, 118, 32, 32, 32, 32, 32, 112, 48, 43, 97, 103, 32, 52, 52, 60, 32, 32, 32, 32, 32, 32, 32, 32, 62, 48, 97, 34, 46, 101, 109, 97, 110, 32, 101,
15:05:55 <AnMaster> 108, 105, 102, 32, 121, 116, 112, 109, 101, 110, 111, 110, 32, 100, 101, 101, 78, 34, 32, 32, 32, 118, 32 <repeats 452 times>, 118, 32, 225, 251,
15:05:58 <AnMaster> 62, 52, 52, 103, 49, 43, 52, 52, 112, 32, 32, 94, 32, 32, 32, 32, 32, 32, 32, 32, 32, 62, 32 <repeats 490 times>...}
15:06:24 <Deewiant> Gah, this is getting really slow in CCBI
15:06:40 <AnMaster> Deewiant, this time it locks up before mycology output at all
15:06:48 <AnMaster> that is when it loads into that specific quadrant
15:06:51 <Deewiant> AnMaster: Doesn't lock up for me
15:06:57 <Deewiant> Gets up to the wraparound bit in CCBI
15:06:57 <AnMaster> Deewiant, well depends on where it is loaded
15:07:07 <AnMaster> Deewiant, needs to be in y < 0 x > 0
15:07:18 <Deewiant> It /does/ test whether it overlaps with itself
15:07:51 <Deewiant> It could do it incorrectly, of course :-P
15:08:10 <AnMaster> Deewiant, in fact only positive for both x and y seems to work correctly. No clue why
15:08:26 <Deewiant> AnMaster: That sounds like a bug in cfunge now
15:08:32 <AnMaster> Deewiant, with the previous version the first quadrant worked though
15:08:35 <Deewiant> Because it's definitely working for negative ones here
15:08:38 <AnMaster> and I haven't changed cfunge since then
15:09:25 <Deewiant> Positive x, negative y works in CCBI
15:09:28 <AnMaster> Deewiant, Hm is the offset in { incremental?
15:09:35 <AnMaster> and if yes. does mycology test that?
15:10:31 <AnMaster> what I said there made no sense
15:10:50 <AnMaster> Deewiant, does it properly clear the area around 0,0 for you or not?
15:11:02 <Deewiant> It clears some of it but not all
15:11:41 <Deewiant> Seems to be something else broken than just what I fixed
15:14:10 <AnMaster> Deewiant, well I set the bounds to a bit around the static area... and dumped
15:14:14 <AnMaster> http://rafb.net/p/yivjIS19.html
15:14:29 <AnMaster> that is the bounds were set to be possible to dump
15:14:42 <AnMaster> Deewiant, it does certainly look strange
15:15:15 <AnMaster> Deewiant, does it load mycology more than once?
15:15:37 <AnMaster> seems like one time, close to the core program isn't properly cleared
15:16:03 <Deewiant> Hmm, I asked my debugger for the stack
15:17:12 <AnMaster> Deewiant, anyway something is broken with delta there when it locks up too
15:17:36 <AnMaster> $1 = {topLeftCorner = {x = -652966086, y = 128}, bottomRightCorner = {x = 378, y = 230825290}, entries = 0x136a010, col_count = 0x136a050, row_count = 0x136a090,
15:17:36 <AnMaster> boundsexact = true, boundsvalid = true}
15:17:52 <AnMaster> it locked up before reaching mycology yes
15:18:15 <AnMaster> Deewiant, anyway original version does reach mycology, but not the last one
15:18:18 <Deewiant> It does clear down to 839 250 :-(
15:18:39 <Deewiant> AnMaster: The last one reaches it for me every single time.
15:18:49 <Deewiant> This might be broken, but so is cfunge. :-P
15:20:33 <AnMaster> Deewiant, the last one only reaches mycology when +x,+y, the previous one ("quick fix" some hours ago) broke for -x,+y, the first one worked for all quadrants
15:20:49 <Deewiant> AnMaster: The last one reaches it for me every single time, regardless of quadrant.
15:21:09 <AnMaster> Deewiant, well I tested with the first one and it works for all those, but the last one doesn't
15:21:30 <AnMaster> 1: fspace = {topLeftCorner = {x = 199, y = 173}, bottomRightCorner = {x = 349208971, y = 1373734736}, entries = 0x298a010, col_count = 0x298a050, row_count = 0x298a090,
15:21:30 <AnMaster> boundsexact = true, boundsvalid = true}
15:21:42 <AnMaster> but still doesn't clear original program
15:22:41 <Deewiant> Hmm, I'm going to pipe this debugger output to a file so I can see what original coordinates the p ones correspond to
15:23:42 <Deewiant> 10 seconds of CPU time to get to the start of Mycology :-P
15:24:09 <AnMaster> Deewiant, http://rafb.net/p/QtfJPW50.html
15:24:23 <AnMaster> note the dump does NOT start at 0,0 this time
15:24:35 <Deewiant> Yay let's print a 632545 element stack of 10-character integers
15:24:53 <Deewiant> AnMaster: Yes, I've found that area
15:25:16 <Deewiant> And I'd very much like to know why it's not getting p'd to oblivion
15:25:28 <Deewiant> Since it seems like the stack has all the coordinates it's supposed to...
15:26:29 <AnMaster> Deewiant, so would I, because I minimise the bounds even when wrapping if bounds are marked as fudgy AND difference is between max and min (for either x or y) is greater than 2^16 (that is power, not any bitwise operator)
15:29:28 <AnMaster> Deewiant, actually efunge should handle this well because it basically does a db search when finding next non-space cell iirc... Somewhat like this (but in erlang ets query syntax, instead of pseudo SQL) when going <: SELECT FROM fungespace WHERE x=pos.x AND y < pos.y AND value <> ' ' LIMIT 1
15:29:36 <AnMaster> (where it gets first non-matching yes
15:29:45 <AnMaster> or maybe that was only in some specific cases
15:30:16 <AnMaster> and I don't remember how I handled flying ips at all
15:32:11 <AnMaster> meh, seems I misremembered totally, since it doesn't implement t at all it just handles this same as z in the mainloop
15:33:06 <AnMaster> Deewiant, a bug fixed mycoedge would be nice too btw!
15:33:30 <AnMaster> Deewiant, hm now it works for -x,-y too
15:34:06 <Deewiant> AnMaster: Can you find me the coordinates of any char which isn't cleared
15:34:20 <AnMaster> Deewiant, in cfunge? not easily without a linear scan no
15:35:17 <AnMaster> Deewiant, I could find any columns or rows with non-space in though easily
15:35:27 <AnMaster> would just need to tweak a certain routine
15:35:43 <Deewiant> It'll be faster if I look for one in the debugger
15:38:08 <AnMaster> (gdb) print cfun_static_space[(64+71)+(64+254)*512]
15:38:38 <AnMaster> Deewiant, maybe it depends on which run?
15:39:21 <Deewiant> Cell ( 71 254 ): value 62 (0x3e), character '>'
15:39:24 <AnMaster> Deewiant, this is with 8 bit funge cells
15:39:33 <AnMaster> so maybe that makes a difference
15:40:59 <AnMaster> well in cfunge it is always cleared
15:51:38 <AnMaster> Deewiant, waiting for it to dump to disk...
15:52:56 <AnMaster> lets see if the pastebin accepts it
15:53:04 <AnMaster> Pasting > 800 kB often tend to fail with rafb. Use --verbose or --debug to see the
15:53:04 <AnMaster> error output from wget if it fails. Alternatively use another pastebin service.
15:53:04 <AnMaster> Pasting > 10000 lines often tend to fail with rafb. Use --verbose or --debug to see the
15:53:04 <AnMaster> error output from wget if it fails. Alternatively use another pastebin service.
15:53:40 <AnMaster> Deewiant, 114676 lines with the output
15:54:08 <AnMaster> since it is too big for pastebin
15:54:37 <AnMaster> Deewiant, http://omploader.org/vMWhhaA
15:54:44 <AnMaster> there is the dump of all set cells you asked for
15:55:47 <AnMaster> <Deewiant> AnMaster: Can you find me the coordinates of any char which isn't cleared
15:56:00 <AnMaster> I wrote the coded needed to do it now
15:56:05 <Deewiant> I didn't ask for a whole dump, just one char
15:56:07 <Deewiant> 2009-04-04 17:35:34 ( Deewiant) Ah, then forget it
15:56:07 <Deewiant> 2009-04-04 17:35:42 ( Deewiant) It'll be faster if I look for one in the debugger
15:56:16 <Deewiant> 2009-04-04 17:36:04 ( Deewiant) 71 254
15:56:18 <AnMaster> Deewiant, well it may be useful still
15:56:22 <AnMaster> since that one is cleared here
15:56:26 <Deewiant> Now it looks like I have x and y flipped
15:56:31 <AnMaster> Deewiant, 71 258 is set though
15:56:51 <Deewiant> I'm wondering what I've got wrong here
15:58:09 <AnMaster> Deewiant, I even dumped it in pseudo sexp and I don't get any "thanks" :(
15:59:08 <AnMaster> Deewiant, did you even look at it+
15:59:36 <AnMaster> it has separate sections for static array and hash
15:59:54 <AnMaster> Deewiant, and what do you think of the dump format? Nice I assume?
16:00:58 <AnMaster> well hash area is unsorted, but can't do anything about that
16:01:20 <Deewiant> Of course you can, you could sort it :-P
16:01:32 <Deewiant> But anyway, found it and fixed it
16:01:37 <Deewiant> 1000x1000 starting at (0,0) is clear
16:01:50 <AnMaster> Deewiant, not easily. I'm using an iterator struct thingy found in the hash library I use
16:02:00 <AnMaster> sorting it in C would be a pain
16:02:03 <Deewiant> I'd say making a temp array is easy
16:02:12 <Deewiant> Well alright, maybe not in C :-P
16:02:37 <AnMaster> Deewiant, + my funge space file is already over 1100 lines long
16:02:48 <AnMaster> I don't want to add more if I can avoid it
16:02:59 <Deewiant> This is only temporary, you know
16:03:05 <Deewiant> AnMaster: Uploaded one which seems to work
16:03:12 -!- oerjan has joined.
16:04:45 <AnMaster> (had to read it, since I saw oerjan join)
16:06:27 <AnMaster> Deewiant, right it still fails at -x,+y and +x,-y
16:06:36 <Deewiant> AnMaster: Right and that's cfunge's problem.
16:07:18 <AnMaster> Deewiant, -x,-y and +x,+y works correctly until mycology messes up storage offset
16:07:28 <oerjan> also, vampires and werewolves are clearly not so scary after all
16:07:37 <AnMaster> Deewiant, can you fix this mycology bug? ;P
16:07:38 <Deewiant> AnMaster: At what point is that, BTW?
16:07:48 <AnMaster> Deewiant, can't you try it in ccbi? :P
16:07:58 <AnMaster> GOOD: } reflects when stack stack has only one stack
16:07:58 <AnMaster> GOOD: u reflects when stack stack has only one stack
16:07:58 <AnMaster> GOOD: u with zero count does nothing
16:07:58 <AnMaster> BAD: u with a negative count reflects
16:07:58 <AnMaster> WARNING: attempted recovery of stack stack, situation may be corrupt
16:07:59 <AnMaster> y claims all of the following:
16:08:11 <Deewiant> Why does u with a negative count reflect
16:08:22 <AnMaster> Deewiant, it doesn't when located in the normal place
16:08:41 <AnMaster> Deewiant, doesn't ccbi has a better debugger? ;P
16:08:52 <Deewiant> AnMaster: But cfunge is the only one that's buggy
16:09:12 <Deewiant> u should only reflect if there's only one stack on the stack stack...
16:09:25 <AnMaster> Deewiant, so stack stack isn't "panic corrected" by mycology in ccbi?
16:09:38 <Deewiant> AnMaster: And that WARNING is /supposed/ to do exactly that, it tries to force the offset to (0,0) and to remove any extra stacks from the stack stack
16:10:22 <Deewiant> AnMaster: I don't know, I can try to let this run for an hour if you want to find out :-P
16:10:24 <AnMaster> u reflects if there is only one stack on the stack stack
16:10:32 <AnMaster> no idea why there is only one at that point
16:10:33 <Deewiant> So why is there only one stack on the stack stack
16:11:07 <AnMaster> Deewiant, can't you patch your bounds manually from inside gdb or whatever to make it work?
16:11:30 <AnMaster> as in set bounds=<calculated correct one>
16:11:41 <AnMaster> Deewiant, well what debugger do you use then?
16:12:03 <AnMaster> Deewiant, add a command in the ccbi debugger to hot patch bounds then?
16:12:43 <Deewiant> It's eating up around 950 megs of memory :-)
16:12:55 <Deewiant> Allocating a space cell for each one that it runs into
16:13:04 <AnMaster> Deewiant, is that on the stack or?
16:13:17 <AnMaster> Deewiant, you don't remove space cells?
16:13:26 <Deewiant> No, I add them for faster lookup
16:13:29 <AnMaster> Deewiant, I mean, in hash area setting to space is same as free
16:13:44 <AnMaster> because I can find if a cell is non-existing easily
16:13:59 <AnMaster> if ght_fspace_get returns a null pointer I know it is a sparse space
16:14:20 <AnMaster> Deewiant, sure it is faster by not doing sparse spaces?
16:14:24 <Deewiant> But I try to avoid having to do those checks
16:14:41 <Deewiant> AnMaster: Well, it allows me to do unsafeGets when I know I've just done a get
16:14:45 <AnMaster> Deewiant, but you loose cache locality when you have 950 MB data instead!
16:15:01 <AnMaster> the check overhead would be smaller
16:15:01 <Deewiant> I don't have locality anyway since it's a hash table
16:15:17 <AnMaster> Deewiant, well true, but parts of the hash table will be in hash
16:15:34 <AnMaster> but less so with all those spaces
16:15:56 <AnMaster> Deewiant, also did you profile the overhead of get?
16:16:18 <Deewiant> This is from before anybody really cared about speed :-P
16:16:19 <AnMaster> well then I did for my code, and found it was faster doing sparse
16:16:28 <AnMaster> because I had less items in the hash array
16:16:35 <Deewiant> I can't see the difference for Mycology time
16:16:39 <AnMaster> and unsafeget, eww, that breaks the funge space abstraction
16:16:50 <Deewiant> Funge-Space isn't abstracted here, sorry ;-)
16:17:00 <AnMaster> Deewiant, well in my case it is, much more than stack is
16:17:39 <Deewiant> unsafeGet is just part of the abstraction here ;-P
16:24:11 <AnMaster> Deewiant, I made cfunge print { u and } when the relevant instruction was run
16:24:19 <AnMaster> here is the result for normal run and a slowdown run:
16:24:23 <AnMaster> http://rafb.net/p/bmJPkK13.html
16:25:01 <AnMaster> Deewiant, "GOOD: u with a positive count transfers cells correctly" is missing for some reason
16:25:17 <AnMaster> maybe you copied mycology incorrectly? I don't know
16:25:36 <AnMaster> still I suggest testing it in CCBI to see if it works there
16:25:54 <Deewiant> Testing this in CCBI is somewhat hard as you know
16:25:56 <AnMaster> until then I have to write it off as a slowdown bug
16:26:26 <AnMaster> Deewiant, I recommended doing it by hot patching bounds as I said above
16:27:01 <AnMaster> anyway the load in quadrants where x and y are of different signs is a cfunge bug or a ccbi bug indeed
16:27:19 <AnMaster> hm test with pyfunge maybe... iirc it implements exact bounds
16:27:38 <AnMaster> lifthrasiir, there? what is the url to get the most recent pyfunge that works and have exact bounds?
16:27:57 <lifthrasiir> AnMaster: that's always available in http://hg.mearie.org/pyfunge/ .
16:28:13 <AnMaster> lifthrasiir, ok, I forgot how to use mercurial.
16:28:31 <lifthrasiir> well, hg clone http://hg.mearie.org/pyfunge/ and it will be done.
16:29:51 <AnMaster> lifthrasiir, does this do exact bounds when wrapping too or?
16:29:58 <AnMaster> because y isn't run at the point where it is needed
16:30:23 <AnMaster> lifthrasiir, I mean if that I'm trying to find if the bug is in slowdown.b98 or cfunge.
16:30:56 <AnMaster> lifthrasiir, iirc you said you did only calculate the exact bounds for wrapping when program used y
16:31:11 <AnMaster> lifthrasiir, oh ok... I need it to be kind of non-slow to be able to test...
16:31:29 <AnMaster> Traceback (most recent call last):
16:31:29 <AnMaster> File "./pyfunge.py", line 295, in <module>
16:31:29 <AnMaster> File "./pyfunge.py", line 213, in main
16:31:31 <AnMaster> TypeError: 'NoneType' object is not callable
16:32:37 <lifthrasiir> (and wait for moment, i'll enable hgweb's zip and tar.gz archive option)
16:33:10 <AnMaster> too slow to do average over 50 runs
16:33:14 <AnMaster> which is what I usually do for cfunge
16:33:40 <fizzie> Ten minutes (for 50 runs) is too slow for you? You're such a speed freak indeed.
16:34:01 <lifthrasiir> if you have psyco it will run faster (2-3x), but still slow
16:34:03 <AnMaster> because 50 runs for cfunge is something like 2-3 seconds
16:34:10 <AnMaster> lifthrasiir, does it work on x86_64?
16:35:16 <AnMaster> lifthrasiir, any option to trace program or such?
16:35:50 <lifthrasiir> hmm... it uses pyfunge_stopat environment variable for now
16:36:04 <AnMaster> lifthrasiir, somewhat like ./cfunge -t 5 (or higher, 3 5 and 9 are currently used levels for tracing in cfunge)
16:36:32 <AnMaster> lifthrasiir, I don't know where, it is random, I just want to know when it reaches somewhere far out in x and y
16:36:49 <AnMaster> Deewiant, how will slowdown.b98 calculate position for bignum funges?
16:37:15 <AnMaster> Deewiant, that is when mycology says " That the number of bytes per cell is -1"
16:37:45 <AnMaster> $ ./pyfunge.py ~/src/cfunge/trunk/tests/asdf.b98
16:37:45 <AnMaster> BAD: y says initial minimal point isn't (14,0)
16:37:45 <AnMaster> BAD: after removing (14,0) y doesn't say minimal point is (14,5)
16:37:45 <AnMaster> BAD: after removing (14,5) y doesn't say minimal point is (15,5)
16:37:50 <AnMaster> lifthrasiir, cfunge gets the first one right
16:38:05 <AnMaster> lifthrasiir, the other ones may buggy in the test
16:38:11 -!- Judofyr has joined.
16:38:48 <AnMaster> lifthrasiir, basically, do you set bounds to be 0,0 initially or such? if you do it is wrong for this program
16:39:14 <AnMaster> lifthrasiir, well ask Deewiant for a copy of mycoedge, the two latter tests are known to be broken, but the first one is correct.
16:39:23 <lifthrasiir> i think not, but i could have a problem with Space.putspace (eg. used for i or o) but not with Space.put
16:41:15 <Deewiant> It'll probably loop forever trying to generate a position in the range (0,0)-(0,0) which doesn't overlap with (0,0)-(70,44) :-P
16:42:00 <fizzie> Sometimes MATLAB's really too silly:
16:42:02 <fizzie> Error: The expression to the left of the equals sign is not a valid target for an assignment.
16:42:35 <Deewiant> Ooh, CCBI's been running for 18 minutes of CPU time and one wraparound is done
16:42:47 <Deewiant> It's at GOOD: 4k # jumps 4 times from k
16:42:50 <AnMaster> Deewiant, tell me what happens for u and such then
16:43:00 <Deewiant> AnMaster: Might not be done today
16:43:12 <AnMaster> Deewiant, a lot of wrap arounds is there?
16:43:14 <Deewiant> I mean, if it's, say, 15 minutes per wraparound this could take hours
16:43:28 <Deewiant> I'm not sure how many before the stack stack tests
16:43:32 <AnMaster> Deewiant, well, why not do as I said then, manually adjust bounds
16:43:37 <Deewiant> But I wouldn't be surprised if there were many
16:43:47 <Deewiant> AnMaster: That'd involve adding that to the debugger :-P
16:43:54 <Deewiant> lifthrasiir: No, since it isn't done
16:44:02 <Deewiant> If you want the broken one, ask AnMaster :-P
16:44:44 <Deewiant> AnMaster: Well, you're the one with the command-line paste tools
16:45:06 <AnMaster> Deewiant, it doesn't work well for spaces though, so I'll ompload it
16:45:14 <AnMaster> Deewiant, also doesn't arch have packages for those?
16:45:20 -!- FireyFly has joined.
16:45:31 <Deewiant> If it does, I wouldn't know how to use them
16:45:42 <AnMaster> lifthrasiir, http://omploader.org/vMWhhcw
16:45:56 <AnMaster> Deewiant, simple usually: $ <tool name> <file name>
16:46:10 <AnMaster> lifthrasiir, the two latter tests are off by one I think
16:46:59 <AnMaster> Deewiant, for example: $ wgetpaste mycoedge.b98 or $ ompload mycoedge.b98
16:47:07 <AnMaster> Deewiant, very simple to use yes :)
16:47:24 -!- FireFly has quit (Nick collision from services.).
16:47:26 -!- FireyFly has changed nick to FireFly.
16:48:01 <Deewiant> So maybe it's around 10 minutes per wraparound
16:48:48 * oerjan swats FireFly -----###
16:49:14 * oerjan swats FireFly's network -----###
16:50:01 <AnMaster> I just made cfunge print it out
16:50:07 <AnMaster> Deewiant, that is just before y
16:50:43 <lifthrasiir> http://pastie.org/436879 hell, what's going on then?
16:50:44 <AnMaster> Deewiant, there are wrapping in the {} test code though
16:51:10 <Deewiant> AnMaster: You can bother to find this out but not what's going on with cfunge there? :-P
16:51:29 <AnMaster> Deewiant, because that would include trying to follow mycology code flow
16:51:48 <lifthrasiir> AnMaster: as i said pyfunge does update rectmin and rectmax (i.e. bounds) only when sysinfo is used, and it should call Space.updaterect but it works correctly!
16:51:51 <AnMaster> INTERCAL is worse yes than that yes, but most other things are not
16:52:14 <AnMaster> lifthrasiir, you set initial bound to 0,0?
16:52:36 <lifthrasiir> AnMaster: yes, before sysinfo is executed.
16:52:37 <Deewiant> AnMaster: You could paste a trace of what happens after cell 145,83 (relative to the new offset, of course)
16:52:52 <lifthrasiir> when sysinfo is used it updates (or should update) to correct bounds
16:52:53 <AnMaster> I mark bounds as "not valid" until first non-space have been loaded from initial file
16:54:10 <AnMaster> Deewiant, hm... ok. it is a bit hard, because once cfunge passed into new mycology it is so fast. How would I find out where the new offset it
16:54:21 <lifthrasiir> omg, there was a bug. how did i manage such a bug... :S
16:54:33 <AnMaster> Deewiant, give me a fixed place to break on and info where new x,y is stored!
16:54:58 <Deewiant> The third or fourth one is at (-2,0) relative to new offset
16:55:47 <AnMaster> Deewiant, so the third k I hit?
16:56:05 <Deewiant> The first one that is far away from (0,0) :-P
16:56:50 <AnMaster> Deewiant, which I don't hit for the "different sign for x,y" case
16:56:54 <Deewiant> CCBI is up to GOOD: } transfers cells correctly
16:57:44 <AnMaster> Deewiant, 54571405+2+145,1419676713+83 right?
16:57:50 <AnMaster> if those large numbers is the k
17:00:58 <Deewiant> GOOD: u with zero count does nothing
17:02:14 <AnMaster> Deewiant, http://omploader.org/vMWhhdw
17:02:30 <AnMaster> that is from the position you said to y
17:02:49 <Deewiant> Cheers, I'll look at that after CCBI fails/succeeds
17:03:24 <lifthrasiir> Deewiant: are you going to add new test (i suppose that's related to boundary of Funge space)?
17:03:37 <AnMaster> Deewiant, well the position in y isn't relative storage offset afaik? right?
17:03:40 <Deewiant> lifthrasiir: Mycoedge is supposed to be that at some point
17:04:30 <AnMaster> Deewiant, is the least point as reported by y relative storage offset?
17:04:59 <Deewiant> Depends on whether o's input is :-P
17:05:30 <AnMaster> Deewiant, if mycology would reach that far when loaded offset it would probably barf on that
17:05:39 <Deewiant> BAD: u with a negative count reflects
17:06:54 <AnMaster> Deewiant, So is storage offset relative or not in ccbi?
17:07:11 <Deewiant> Seems to me like it shouldn't be, but I don't remember
17:07:13 <AnMaster> Deewiant, So is storage least point relative offset or not in ccbi?*
17:07:17 <Deewiant> I.e. it'd make more sense if it wasn't
17:07:44 <AnMaster> y is not relative here, but o and i are
17:07:48 <Deewiant> Since what y pushes is supposed to work directly as an input to o.
17:08:06 <AnMaster> as in "go make everything relative offset"
17:08:38 <AnMaster> Deewiant, because mycology never complained about it I probably didn't even remember about that
17:10:19 <Deewiant> Check the spec if it says anything
17:15:05 <AnMaster> "1 vector containing the least point which contains a non-space cell, relative to the origin (env)"
17:15:13 <AnMaster> Deewiant, depends on what origin means
17:16:35 <Deewiant> So hmm, would it make sense for i and o to be absolute then
17:17:28 <lifthrasiir> that's a spec bug, which became a feature now
17:17:29 <Deewiant> Since you get the storage offsets right next to that when you do y anyway
17:17:35 <Deewiant> lifthrasiir: Yeah, probably :-)
17:17:45 <Deewiant> AnMaster: Taking notes for -109? ;-P
17:18:49 <AnMaster> but since no one else but Mike Riley *is* interested...
17:19:29 -!- kar8nga has joined.
17:21:05 <fizzie> There's the "These two vectors are useful to give to the o instruction to output the entire program source as a text file." note there too, which sort-of breaks if i/o are relative but y bounds are not; OTOH you can't really have relative y bounds since it explicitly says "relative to the origin".
17:27:41 <AnMaster> fizzie, iirc Deewiant said i and o should be relative some months ago
17:28:23 <AnMaster> hm seems gcc 4.3 has inter-procedural optimising thingy
17:28:27 <fizzie> Yes, it does feel logical to me too that they would be relative; this is just one point that suggests otherwise.
17:32:21 <AnMaster> anyway tracking bounds add around 0.010 wall clock to a normal mycology run (optimised build, no concurrent funge, empty environment, redirecting to /dev/null). With more env and/or concurrency it adds more of course. The overhead is not constant rather it is relative number of writes to funge space (and to a lesser degree also ip wrap check)
17:35:30 -!- kar8nga has quit (Remote closed the connection).
17:35:34 -!- kar8nga has joined.
17:36:51 <fizzie> Hmpfh, Octave doesn't seem to do much auto-vectorization, based on the fact that it's using almost exactly 50 % of CPU on this two-core system.
17:37:48 <fizzie> Or maybe it does in some operations, since it has now climbed up to 60 %.
17:38:27 <fizzie> Auto-parallelization, I mean.
17:40:14 <AnMaster> fizzie, for gcc there is an option to do that in 4.3 iirc
17:40:23 <AnMaster> maybe it was added in 4.2 or so, not sure
17:41:15 <fizzie> Well, it is doing *something* sensible, since the CPU usage has climbed up to 69 %.
17:47:02 <AnMaster> Deewiant, found why it locks up, because it still fails to clear the static area when funge space is loaded in -x,+y or +x,-y quadrants. Or it locks up before it cleared it.
17:47:16 <AnMaster> anyway it is after the program was loaded
17:47:28 <AnMaster> I think the main program is wrapping
17:47:33 <fizzie> I think I've been misreading those CPU usage numbers, actually; I guess the ps-reported one should climb to >100 for parallelistics, but it's some sort of longer-term average, and that's why it's changing so slowly. Since "top" shows exactly 100 in the "%CPU" column, but something like "48.6%us" in the total CPU usage line.
17:47:52 <AnMaster> Deewiant, ah yes that it... slowdown itself is wrapping there, before it cleared itself
17:49:45 <AnMaster> Deewiant, hm what the hell happened then...
17:50:09 <Deewiant> Slowdown doesn't wrap at all after loading the file for the second time
17:50:16 <Deewiant> Since otherwise it could run into the other file's code
17:51:40 <AnMaster> Deewiant, so if I want to make slowdown print something just before that, where would be the safe place to insert it
17:51:54 <AnMaster> and I can't break there easily tnen
17:52:03 <Deewiant> AnMaster: That code is in the loop which pushes a space 600000 times
17:52:47 <AnMaster> Deewiant, would putting '1, or on the side of "Begin magic" be safe
17:53:25 <Deewiant> That code isn't very self-modifying, you can mess with it quite freely
17:53:36 <Deewiant> Just make sure everything remains aligned :-P
17:54:24 <AnMaster> meh why does the code before that take so long...
17:55:12 <AnMaster> Deewiant, I think there is some issue with buffered output...
17:55:17 -!- neldoret1 has joined.
17:55:56 <AnMaster> when I insert a newline there I get the output I wanted
17:56:57 <AnMaster> Deewiant, that x is it from the right side?
17:57:04 -!- ehird has left (?).
17:57:06 <AnMaster> I mean x ;JUMP; is where it actually jumps right?
17:57:13 <AnMaster> and that is entered from the right side
17:57:30 * AnMaster adds some debugging code in the blank space there
17:59:34 -!- ehird has joined.
18:00:34 -!- neldoreth has quit (No route to host).
18:00:43 <AnMaster> lets see how long a working one takes
18:00:50 <AnMaster> so I can know when to give up with valgrind...
18:01:01 <AnMaster> (this is super-slow under valgrind btw)
18:02:02 <AnMaster> and here comes a broken one...
18:02:14 * AnMaster made i print new bounds at end
18:04:43 <ehird> 08:40 AnMaster: logical flaw there...
18:04:43 <ehird> 08:40 AnMaster: so you are just trolling then, right.
18:04:50 <ehird> AnMaster's deduction engine runs in verbose mode
18:05:54 <Deewiant> ehird: BTW, about me and 22 dB, I assume your allegation was another instance of shit
18:06:10 <ehird> Deewiant: Everything I say is an instance of shit.
18:06:19 <Deewiant> ehird: Alright, I won't respond properly then
18:06:29 <ehird> Deewiant: You're going to put me out of business
18:06:33 <AnMaster> <ehird> Deewiant: Everything I say is an instance of shit. <-- that explains it...
18:06:39 <Deewiant> AnMaster: Mycology assumes (0,0) storage offset there, yes.
18:06:46 <Deewiant> AnMaster: The error message is wrong though.
18:07:09 <AnMaster> Deewiant, ah ok. Is this a mycology bug?
18:07:24 <AnMaster> also now I have a case of broken offsets
18:07:37 <Deewiant> AnMaster: The error message is: I've apparently forgotten to move a ; after inserting some code
18:07:54 <Deewiant> AnMaster: The offset thing isn't IMO, since I can't assume that y works there so I can't know the correct offsets.
18:08:15 <Deewiant> AnMaster: The thing that u complains about.
18:08:39 <Deewiant> 2009-04-04 20:06:45 ( Deewiant) AnMaster: The error message is wrong though.
18:08:43 <AnMaster> <Deewiant> AnMaster: The error message is: I've apparently forgotten to move a ; after inserting some code <-- this must mean "BAD: I've apparently forgotten to move a ; after inserting some code"
18:09:08 <Deewiant> AnMaster: If you want, I could probably make it just print the storage offsets with UNDEF
18:09:09 <AnMaster> Deewiant, it was a lame joke yes
18:09:26 <AnMaster> Deewiant, well, it is usually bad
18:09:45 <Deewiant> It'd make Mycology a step closer to being slowdown-clean :-)
18:09:47 <AnMaster> Deewiant, slowdown is non-conforming anyway
18:10:27 <AnMaster> Deewiant, because mycology isn't position independent. And I think having relative i and o is completely valid
18:10:37 <AnMaster> and y is defined as relative 0,0
18:10:46 <ehird> http://www.arctic-cooling.com/catalog/product_info.php?cPath=2_&mID=244&language=en ← Yum.
18:10:49 <Deewiant> AnMaster: How does any of that make slowdown non-conforming
18:11:17 <AnMaster> Deewiant, because it modifies the interpreter so the program that will run in it won't run in a conforming environment
18:11:29 <Deewiant> Well of course, but the program itself isn't
18:11:39 <Deewiant> It isn't guaranteed to work for all programs: that's obvious
18:11:39 <AnMaster> Deewiant, make slowdown mycology clean instead by interpreting instead!
18:12:22 <AnMaster> slowdown should be a befunge98 in befunge98 interpreter, it would store it's interpreter code near the program to avoid the slow wrap thing or something
18:12:29 <ehird> 18:09 AnMaster: Deewiant, slowdown is non-conforming anyway
18:12:45 <AnMaster> Deewiant, still it would work better
18:13:12 <AnMaster> Deewiant, also this is odd, the quadrant bug is *not* due to memory corruption
18:13:13 <Deewiant> But it'd be a crapload of work
18:13:32 <AnMaster> at least not detectable by either valgrind or mudflap
18:14:14 <AnMaster> Deewiant, can you provide a trace of ccbi running it in +x,-y or -y,+x that lists each instruction executed in some format?
18:14:22 <AnMaster> I don't have a working D compiler as you know
18:14:49 <AnMaster> Deewiant, can I just download ldc, compile it against llvm 2.5 and it will work out of box on x86_64?
18:14:59 <AnMaster> because my llvm is pure 64-bit host backend only
18:15:38 -!- BeholdMyGlory has joined.
18:16:28 <AnMaster> Deewiant, last release or some svn or whatever version
18:16:41 <Deewiant> AnMaster: You'll need latest hg, hang on
18:16:56 <Deewiant> AnMaster: Didn't you get pyfunge? :-P
18:17:21 <AnMaster> Deewiant, well I don't know url for ldc
18:17:50 <Deewiant> LDC's URL is on its home page, of course
18:18:08 <Deewiant> hg clone http://tar.us.to:8000
18:18:41 <Deewiant> LDC's URL is on its home page, of course
18:20:17 <AnMaster> bzr is faster these days, it used to be a bit slow for remote operations on large repos, but since 1.13 it is blazing fast
18:20:36 <Deewiant> The server is slow, not the program
18:20:53 <AnMaster> Deewiant, should I get D2 or tango?
18:21:05 <Deewiant> AnMaster: Read the LDC build instructions, also on the home page.
18:21:14 <AnMaster> Fetch the tango runtime for D1
18:21:14 <AnMaster> svn co http://svn.dsource.org/projects/tango/trunk ldc/tango
18:21:14 <AnMaster> Fetch the druntime for D2 (this is experimental)
18:21:14 <AnMaster> svn co http://svn.dsource.org/projects/druntime/trunk ldc/druntime
18:21:37 <Deewiant> One is experimental, one isn't
18:21:58 <Deewiant> I don't hook into the compiler :-P
18:21:59 <AnMaster> so I shouldn't fetch either then?
18:22:09 <AnMaster> Deewiant, you needed tango didn't you say?
18:22:14 <Deewiant> Do you know what a language runtime is?
18:22:28 <AnMaster> but iirc you said D2 wasn't compatible
18:22:56 <Deewiant> But as far those runtimes are concerned it really shouldn't matter
18:23:09 <Deewiant> Since it isn't experimental and will work
18:23:15 <Deewiant> The other one may or may not, since it is experimental.
18:23:17 <AnMaster> Deewiant, but tango vs. phobos *did* matter? "But as far those runtimes are concerned it really shouldn't matter" meh
18:23:28 <Deewiant> AnMaster: I reiterate: Do you know what a language runtime is?
18:23:53 <Deewiant> So you should also realize that if they are both conforming to the language spec, it makes absolutely no difference which one is used
18:23:53 <AnMaster> Deewiant, it would be like libc, or libstdc++ or such
18:24:20 <ehird> 18:22 Deewiant: Are you intentionally dense?
18:24:25 <ehird> my thoughts exactly.
18:24:29 <AnMaster> Deewiant, but then I should be able to use phobos too, yet you needed tango. Or is phobos non-conforming?
18:24:40 <ehird> Deewiant: he doesn't know what a language runtime is.
18:24:42 <Deewiant> AnMaster: Those are libraries, not part of the language spec, and not language runtimes.
18:24:55 <Deewiant> They both provide a language runtime.
18:25:06 <Deewiant> druntime is only a language runtime.
18:25:12 <Deewiant> Tango is a language runtime + a library.
18:25:17 <Deewiant> Phobos is a language runtime + a library.
18:25:28 <Deewiant> You can use the Phobos library with the Tango runtime.
18:25:46 <Deewiant> You can not use the Tango library with the Phobos runtime because the Phobos runtime is crap.
18:25:58 <Deewiant> You can use both libraries with druntime, at least in theory.
18:26:27 <ehird> One plus two is two.
18:26:29 <ehird> Two plus two is four.
18:26:34 <AnMaster> Deewiant, hm does out of tree builds work for ldc?
18:26:39 <ehird> "One plus two is two."
18:26:54 <ehird> AnMaster: There's no "make install".
18:26:59 <ehird> You have to put it in your PATH.
18:27:06 <ehird> Deewiant: It doesn't work properly
18:27:09 <ehird> I know this because I tried
18:27:15 <Deewiant> Maybe not on OS X, but it worked for me. :-P
18:27:15 <ehird> also, it isn't documented
18:27:22 <ehird> the documented method is to put it in your PATH
18:27:26 <Deewiant> That's probably because it doesn't work properly. ;-)
18:27:38 <AnMaster> I plan to install to ~/local/ldc
18:27:45 <ehird> AnMaster: Well, don't "install".
18:27:48 <ehird> Just move the build tree there.
18:27:53 <ehird> If you're doing it that way anyway.
18:30:03 <AnMaster> Deewiant, do I want any of BUILD_BC_LIBS, BUILD_SHARED_LIBS, BUILD_SINGLE_LIB ?
18:30:24 <ehird> are you bsmntbombdood
18:30:28 <AnMaster> BUILD_SINGLE_LIB is on by default
18:33:12 <Deewiant> AnMaster: You don't have to touch any of the settings.
18:33:34 <AnMaster> Deewiant, I do need to touch linker flags because otherwise llvm doesn't work, need to add rpath and such
18:33:38 <ehird> Deewiant: Why did you do that? We could have spent half an hour answering 100 questions.
18:34:21 <AnMaster> svn: '/home/arvid/src/llvm/llvm-2.5' is not a working copy
18:34:41 <AnMaster> that's right. The directory doesn't even exist
18:34:43 <ehird> AnMaster: Did the rest of the build fail?
18:34:50 <AnMaster> I don't have the llvm source checkout around any more
18:34:53 <ehird> Also, that's because you specified a path wrong.
18:34:56 -!- neldoret1 has quit (Client Quit).
18:35:01 <ehird> that error is expected
18:35:04 <ehird> but it has to exist
18:35:09 <ehird> so you did it wrong.
18:35:21 <AnMaster> ehird, err, I only keep installed llvm around, not the source
18:35:26 <AnMaster> and the source was an out of tree build
18:35:33 <ehird> But the path has to exist.
18:35:39 <ehird> It's checking whether it's an svn version.
18:35:42 <ehird> But the path has to exits.
18:35:46 <ehird> Otherwise it'll break later on.
18:35:50 <ehird> When it tries to use that.
18:35:52 <AnMaster> ehird, that is a bug in the build system clearly
18:36:09 <ehird> "If something went wrong it's obviously *their* fault."
18:36:23 <AnMaster> ehird, I hope it won't write anything in it at least?
18:36:50 <ehird> AnMaster: redo cmake and specify the paths right.
18:37:05 <AnMaster> ehird, yes llvm install path exists. it is ~/local/llvm/
18:37:12 <ehird> Deewiant: kill me.
18:37:15 <AnMaster> but the SOURCE PATH FROM WHICH LLVM WAS INSTALLED IS NO LONGER AROUND
18:37:23 <AnMaster> how hard is that to understand
18:37:32 <Deewiant> AnMaster: Did something go wrong, exactly?
18:37:52 <AnMaster> Deewiant, no but ehird seems to say that you need the llvm source around, I just asked why "svn: '/home/arvid/src/llvm/llvm-2.5' is not a working copy" happened
18:38:04 <ehird> When I did it it gave that error about my actual llvm directory.
18:38:06 <AnMaster> I don't keep source around for installed programs
18:38:12 <ehird> I infer that it should at least exist.
18:38:17 <Deewiant> It checks whether you have a recent enough version, which includes checking for SVN revision.
18:38:25 <Deewiant> It assumes it's recent enough if it's not SVN.
18:38:41 <AnMaster> well the path in question is not listed in ccmake
18:38:51 <AnMaster> it only mentions /home/arvid/local/llvm there
18:39:24 <Deewiant> Where is that mentioned? LLVM_INSTDIR?
18:39:33 <Deewiant> In my case it's /usr which obviously doesn't contain llvm-2.5
18:39:56 <AnMaster> LLVM_CONFIG:FILEPATH=/home/arvid/local/llvm/bin/llvm-config
18:39:56 <AnMaster> LLVM_INSTDIR:PATH=/home/arvid/local/llvm
18:40:34 <AnMaster> how hard is this to understand
18:41:04 <ehird> Deewiant: Whatever you're doing, do it more so he headdesks more.
18:41:10 <Deewiant> And what are you complaining about, the build is working, right?
18:41:20 <AnMaster> Scanning dependencies of target gen_revs_h
18:41:20 <AnMaster> [ 1%] Generating revisions.h and llvm-version.h
18:41:20 <AnMaster> svn: '/home/arvid/src/llvm/llvm-2.5' is not a working copy
18:41:24 <AnMaster> I was wondering about that yes
18:41:38 <Deewiant> And if it continued on to 2% there isn't a problem?
18:42:08 <AnMaster> [ 3%] Generating dmd/impcnvtab.c
18:42:23 <AnMaster> I hope that means it won't take too long...
18:42:55 <ehird> FUCK! WE HAVE A SKIPPING 2% SITUATION IN THE HOUSE!
18:43:01 <ehird> Oh god get a SWAT team
18:43:13 <ehird> AnMaster: National emergency
18:43:17 <ehird> I declare martial law
18:43:22 <AnMaster> [ 47%] Building CXX object CMakeFiles/ldc.dir/ir/irlandingpad.cpp.o
18:43:22 <AnMaster> Linking CXX executable bin/ldc
18:43:22 <AnMaster> /usr/lib/gcc/x86_64-pc-linux-gnu/4.3.2/../../../../x86_64-pc-linux-gnu/bin/ld: BFD (GNU Binutils) 2.18 assertion fail /var/tmp/portage/sys-devel/binutils-2.18-r3/work/binutils-2.18/bfd/elf64-x86-64.c:2548
18:43:22 <AnMaster> /usr/lib/gcc/x86_64-pc-linux-gnu/4.3.2/../../../../x86_64-pc-linux-gnu/bin/ld: BFD (GNU Binutils) 2.18 assertion fail /var/tmp/portage/sys-devel/binutils-2.18-r3/work/binutils-2.18/bfd/elf64-x86-64.c:2548
18:43:45 <ehird> Tools just get more and more informal
18:44:06 <ehird> "im in ur supposed program, confused about this missin {"
18:44:08 <AnMaster> Deewiant, is this a known issue with binutils 2.18?
18:44:20 <ehird> "Yo dawg, I herd u liek recursion but you just blew the stack."
18:44:21 <Deewiant> AnMaster: I don't know, I'm not an LDC developer
18:44:33 <ehird> "I herd u liek segfaults"
18:44:44 <AnMaster> Deewiant, well, then I'd rather track stable in gentoo binutils
18:45:44 <AnMaster> at least there is a binary there
18:46:10 <ehird> AnMaster: do you know what '[100%] Built target ldc' means?
18:46:16 <ehird> you do know that make gives up on errors, right?
18:46:49 <AnMaster> but the word assertion mislead me
18:47:06 <AnMaster> since that usually means it will exit, but I guess this isn't assert()
18:47:22 <AnMaster> Deewiant, btw any small test program in D to check with?
18:47:31 <ehird> AnMaster: I've had worse ld fuckups -- it seems to only sort symbols to so many characters, so if you have long mangled names that are same for a long prefix -- it spits out 'atom sorting error'
18:47:32 <Deewiant> AnMaster: Did you make runtime?
18:47:35 <ehird> Got a few hundred? Have fun EVERY LINK TIME <3
18:47:49 <AnMaster> object.d: Error: module object cannot read file 'object.d'
18:47:53 <AnMaster> make[3]: *** [runtime/dcrt/bitmanip.o] Error 1
18:47:55 <Deewiant> import tango.io.Stdout; void main() { Stdout("Foo").newline; }
18:48:09 <Deewiant> AnMaster: Did you download a runtime?
18:48:22 <ehird> "What's a runtime?"
18:48:22 <AnMaster> Deewiant, I did "svn co http://svn.dsource.org/projects/tango/trunk ldc/tango" yes
18:48:51 <Deewiant> One option at this point is to try and follow the instructions and not do an out-of-tree build
18:49:12 <Deewiant> Yes, following instructions does.
18:49:21 <AnMaster> Deewiant, you said it worked though
18:49:25 <ehird> Nobody's gonna tell me what to do!
18:49:28 <Deewiant> AnMaster: It used to; maybe it doesn't any more.
18:49:28 <ehird> Deewiant: you LIED to him
18:49:32 -!- Judofyr has quit (Remote closed the connection).
18:49:35 <ehird> you said it worked. you LIAR! You malicious... lying thing.
18:52:02 <AnMaster> ehird, you seem edgy today ehird.
18:52:15 <ehird> AnMaster: Actually, this is lighthearted.
18:52:25 <ehird> The more lighthearted, the more fun I have poking fun at you.
18:53:12 <Deewiant> "Edgy" would be /ignore and /part and arguing with you; yelling at others is lighthearted
18:53:36 <ehird> Deewiant: I appoint you my public relations officer.
18:53:56 <AnMaster> Deewiant, no that would be "raging" I think
18:54:04 <ehird> Deewiant: No choice.
18:54:17 <Deewiant> AnMaster: Yeah, I guess the former two at least would be.
18:59:21 <AnMaster> (as in out of tree builds are better)
18:59:47 <oerjan> <ehird> I declare martial law
19:00:16 <oerjan> you must all grow antennas, and drink only 1 liter of water per year
19:00:24 <AnMaster> Deewiant, how do I build ccbi then?
19:00:25 <ehird> I used to never say k
19:04:25 <AnMaster> Deewiant, well? ldc *.d */*.d or something?
19:04:48 <Deewiant> AnMaster: ldc won't rm -rf $HOME if you do something wrong
19:05:02 <ehird> Deewiant: Actually it will, I added that a few days ago
19:05:05 <Deewiant> You can try that, I don't know if that works
19:05:06 <AnMaster> Deewiant, well but it was better asking you
19:05:12 <ehird> It's in pebkac_fix.d
19:05:14 <AnMaster> Deewiant, how do you build ccbi?
19:05:26 <Deewiant> AnMaster: The 'easy' way is to get rebuild; but it doesn't know about LDC and requires a bit of hacking to get to compile.
19:05:43 <Deewiant> AnMaster: Can you run 32-bit executables?
19:06:04 <Deewiant> I can't, so I had to hack it a bit and build a 64-bit version.
19:07:03 <AnMaster> http://rafb.net/p/tMWfK914.html
19:07:48 <Deewiant> Let's see if I can reproduce that
19:08:05 <Deewiant> AnMaster: What Rebuild does is build each file to a .o and then links them
19:08:27 <Deewiant> AnMaster: Also, Mycology updated, you can see if cfunge gets any further under slowdown.
19:09:34 <Deewiant> AnMaster: Doesn't happen for me :-/
19:09:42 <Deewiant> Upgrading binutils might help, perhaps :-P
19:09:52 <Deewiant> ehird: iki.fi/deewiant/files/befunge/programs/slowdown.b98 IIRC
19:11:43 <AnMaster> Deewiant, hm llvm is 64 bit only, but otherwise I can run 32 bit binaries yes
19:12:01 <ehird> http://m1.2mdn.net/2233165/300x250-proggit.gif ← Best ad ever.
19:12:01 <Deewiant> AnMaster: So you should just get Rebuild (part of DSSS), like I said.
19:12:31 <Deewiant> ehird: Doesn't look quite 100% to me. :-P
19:12:32 <AnMaster> Deewiant, so rebuild doesn't need 32-bit ldc?
19:12:45 <Deewiant> AnMaster: Like I said it isn't aware of LDC.
19:12:47 <ehird> Deewiant: The best part is that it blinks.
19:12:58 <Deewiant> AnMaster: Like I said to get it to build implies hacking.
19:13:17 <ehird> Deewiant: Don't say dirty words like "binary".
19:13:26 <AnMaster> ehird, that is some javascript code dumped into the page in that image...
19:13:27 <ehird> There are children here.
19:13:31 <ehird> AnMaster: No shit sherlock!
19:13:43 <ehird> Mosaic renders the contents of <script>.
19:13:45 <AnMaster> ehird, so 100% compatible I don't agree with
19:13:59 <ehird> AnMaster: "What is failure to understand sarcasm?", Alex.
19:14:28 <lifthrasiir> i recently thought about two esolangs, one is designed to be harder than malbolge to write code in, one is designed to be graphical and no textual input possible.
19:14:38 <ehird> AnMaster: GregorR made DSSS. It isn't LDC-savvy, as far as I know.
19:14:46 <Deewiant> AnMaster: Yes, your IRC client doesn't corrupt data.
19:15:00 <ehird> Sure it doise7*!^*~&%^
19:15:06 <AnMaster> Deewiant, for me all of the first page on google searching for DSSS is related to "direct-sequence spread spectrum"
19:15:07 <ehird> Well I mean mayE*(!^& it dos%@
19:15:08 <lifthrasiir> ...maybe one day i will come with them, but i'm still just planning
19:15:16 <Deewiant> AnMaster: Consider searching for DSSS rebuild.
19:15:30 <Deewiant> Or you know, anything at all related.
19:15:57 <Deewiant> Better than plain DSSS, though not necessarily by much :-P
19:16:02 <ehird> DSSS child pornography
19:16:10 <ehird> Gives the best results.
19:16:27 <Slereah_> Is your experience that of a child pornographer
19:16:38 <ehird> Slereah_: I prefer "mini pornographist"
19:16:55 <ehird> HEY CLOG DELETE THE PAST FEW LINES OK THANKS
19:17:39 <Slereah_> Silently judging you behind my internet.
19:19:04 -!- kar8nga has quit (Read error: 110 (Connection timed out)).
19:25:23 <AnMaster> dsss-0.78-x86-gnuWlinux.tar.bz2
19:26:45 <ehird> it's hacka' slang.
19:26:48 <ehird> get wit the prorgam.
19:27:45 <Deewiant> AnMaster: Read the LDC instructions for setting it up properly.
19:27:54 <oerjan> so, without the itty bitty
19:27:54 <ehird> Deewiant: I thought we'd been over not using instructions.
19:28:04 -!- Judofyr has joined.
19:28:19 <ehird> "instructions? harumph. meh. fiiine."
19:29:08 <oerjan> instructions are so archaic. modern programming languages should use suggestions instead.
19:29:36 <oklopol> yes, you're not the boss of your computer, just a guide
19:29:45 <AnMaster> "ln -s `pwd`/ldc-posix-tango $HOME/.rebuild" <-- in the ccbi build dir, in the ldc bin directory?
19:29:58 <AnMaster> also you said I didn't need shared tango
19:30:03 <AnMaster> but it seems I do with rebuild
19:30:19 <oerjan> STOP THE CPU OPPRESSION
19:30:24 <ehird> AnMaster: you didn't say you were going to use dsss
19:30:26 <AnMaster> <AnMaster> Deewiant, do I want any of BUILD_BC_LIBS, BUILD_SHARED_LIBS, BUILD_SINGLE_LIB ? <ehird> no <Deewiant> AnMaster: You don't have to touch any of the settings.
19:30:38 <Deewiant> What are you talking about, shared tango?
19:30:40 <ehird> AnMaster: not my fucking fault then, so shut up
19:30:41 <AnMaster> "Rebuild has some advantages to using a static Tango library (as described above). It compiles and links only to the modules you actually need, and changing the compiler options, for all the modules your application needs, is easy."
19:30:50 <AnMaster> http://www.dsource.org/projects/ldc
19:30:53 <Deewiant> AnMaster: You do realize that DMD wasn't even capable of building shared libraries.
19:31:04 <AnMaster> Deewiant, then what the hell does that paragraph mean
19:31:05 <Deewiant> Or is, but they don't work, or something. Not sure.
19:31:21 <Deewiant> AnMaster: It means that instead of linking to the .lib it builds the individual .d you need and links to those .o.
19:31:31 <Deewiant> AnMaster: Sorry, .a for the first on linux.
19:34:36 <AnMaster> Deewiant, trace.d(13): module container cannot read file 'ccbi/container.d'
19:34:48 <AnMaster> something is up with path in ccbi
19:35:36 -!- M0ny has quit ("PEW PEW").
19:35:38 <Deewiant> Reading instructions, perhaps?
19:35:48 <Deewiant> I mean, either you haven't read them or something is wrong.
19:35:53 <AnMaster> ~/funges/ccbi-trunk $ ~/local/dsss-0.78-x86-gnuWlinux/bin/rebuild -dc=ldc-posix-tango *.d
19:36:15 <AnMaster> only changelog.txt and license.txt
19:36:18 <Deewiant> You have clearly read neither the CCBI page nor Rebuild's manual
19:36:22 <Deewiant> AnMaster: There is my web site.
19:36:57 <AnMaster> Deewiant, all other packages I have seen so far on *nix have either an INSTALL README or something like that with building instructions
19:37:03 <ehird> Deewiant: I feel for you.
19:37:05 <AnMaster> that is what linux users expect
19:37:06 <Deewiant> And CCBI, in fact, is not a package!
19:37:24 <ehird> Deewiant: I know a good therapist.
19:38:06 <Deewiant> AnMaster: CCBI is distributed from exactly one place, which has building instructions.
19:38:26 <AnMaster> Deewiant, how is that relevant?
19:38:39 <Deewiant> It means that, if you have CCBI, you have seen the building instructions.
19:38:51 <Deewiant> And hence, there is no need for a README.
19:39:06 <AnMaster> Deewiant, what do you have against README or INSTALL?
19:39:09 <Deewiant> When I'll start hosting the repositories publicly I'll add READMEs.
19:39:13 <ehird> hahahaha /me facepalm
19:39:17 <Deewiant> AnMaster: Nothing, I just never felt the need for one.
19:39:27 <ehird> Deewiant: you have an entitlement to AnMaster, don't you know
19:39:59 <oklopol> Deewiant: that's just not cool man
19:40:03 <AnMaster> Deewiant, so after a quick look on that page and the ldc page it seems the correct way should be: ~/local/dsss-0.78-x86-gnuWlinux/bin/rebuild -rfccbi.rf -dc=ldc-posix-tango
19:40:09 <oklopol> what use is a program without something to read
19:40:23 <ehird> oklopol: read the program
19:40:31 <Deewiant> AnMaster: You can also link $HOME/.rebuild/default to ldc-posix-tango so you don't need that -dc= argument
19:40:40 <AnMaster> Deewiant, ah ok thanks for that tip
19:40:42 <oklopol> okay sorry Deewiant i must be high.
19:40:43 <ehird> oklopol: i just revolutionized reading and computing in one.
19:41:07 <oklopol> ehird: yes the world is of very different color now
19:41:08 <AnMaster> $ ~/local/dsss-0.78-x86-gnuWlinux/bin/rebuild -rfccbi.rf
19:41:08 <AnMaster> ccbi.d(20): module instructions cannot read file 'ccbi/instructions.d'
19:41:16 <oklopol> also god i suck at board games
19:41:22 <Deewiant> AnMaster: Are you in a directory called ccbi?
19:41:32 <Deewiant> That's something I should fix but haven't bothered
19:41:38 <AnMaster> Deewiant, no ccbi-trunk is the checkout directory, it is my naming scheme here
19:41:47 <Deewiant> AnMaster: Then make a ccbi directory under that
19:42:00 <AnMaster> Deewiant, seems a bit odd depending on directory name?
19:42:21 <Deewiant> AnMaster: So that the .d files would be found under ../ccbi/whatever
19:42:39 <Deewiant> Instead of having to have ccbi/ccbi/*.d I can have ccbi/*.d
19:42:44 <AnMaster> Deewiant, ln -s . ccbi *seems* to work so far
19:42:52 <ehird> infinity fuck yeah
19:43:06 <AnMaster> Deewiant, why would you need any part of the directory name to be "ccbi" anyway? Just wondering...
19:43:12 <AnMaster> there is probably some logical reason
19:43:14 <Deewiant> AnMaster: Because the module names start with ccbi.
19:43:43 <AnMaster> Deewiant, link time errors, about undefined reference to ncurses.
19:44:11 <Deewiant> 2009-04-04 21:07:02 ( AnMaster) http://rafb.net/p/tMWfK914.html
19:44:15 <Deewiant> arvid@tux ~/funges/interpreters/ccbi-trunk $ /home/arvid/local/ldc/bin/ldc -output-bc -of=ccbi.bc *.d */*.d */*/*.d
19:44:19 <Deewiant> Remember to link with an ncurses library.
19:44:28 <AnMaster> Deewiant, well I thought rebuild handled that..
19:44:48 <AnMaster> Deewiant, after reading on your website didn't mention this
19:45:11 <Deewiant> Is "Remember to link with an ncurses library." somehow ambiguous?
19:45:14 <AnMaster> Deewiant, well since users who download ccbi source would read there to find out how to build
19:45:22 <ehird> Deewiant: well why did you call it Deewiant's Mega Source of Rebuild Info For Dummies
19:45:27 <AnMaster> so that means everyone will get the build error
19:45:41 <ehird> "Remember to link with an ncurses library."
19:45:43 <Deewiant> AnMaster: 2009-04-04 21:44:11 ( Deewiant) 2009-04-04 21:07:02 ( AnMaster) http://rafb.net/p/tMWfK914.html
19:45:44 <ehird> "Remember to link with an ncurses library."
19:45:45 <ehird> "Remember to link with an ncurses library."
19:45:45 <ehird> "Remember to link with an ncurses library."
19:45:47 <ehird> "Remember to link with an ncurses library."
19:45:49 <ehird> "Remember to link with an ncurses library."
19:45:52 <ehird> "Remember to link with an ncurses library."
19:45:54 <ehird> "Remember to link with an ncurses library."
19:45:54 <Deewiant> AnMaster: That's /your/ paste.
19:45:54 <AnMaster> Deewiant, rebuild *didn't* output it this time?
19:45:55 <ehird> The usual 3 do not quite seem enough.
19:46:09 <Deewiant> AnMaster: It cached the ncrs.o probably
19:46:17 <Deewiant> And didn't rebuild it, because it's smart that way.
19:46:57 <AnMaster> Deewiant, find . -name '*.o' -delete
19:46:59 <AnMaster> Deewiant, find . -name '*.bc' -delete
19:47:07 <AnMaster> yeah I guess my find is broken
19:47:18 <Deewiant> Since, if ncrs.d is built, that message is output.
19:47:22 <AnMaster> Deewiant, I have it all in scrollback, not enough link time errors
19:47:47 <Deewiant> And if ncrs.o is linked without a curses library, those link errors will happen.
19:48:05 <Deewiant> Since ncrs.o being linked implies ncrs.d having been built, the message must have been output.
19:50:12 <AnMaster> Deewiant, it wasn't. Maybe there was a bug somewhere? http://rafb.net/p/IzTmPs26.html
19:50:40 <AnMaster> anyway now I have a ccbi binary finally
19:51:03 <AnMaster> Deewiant, "Assuming 32-bit chtype... correct ccbi.fingerprints.jvh.ncrs.chtype to ushort if link errors ensue." seems to indicate it was built?
19:51:04 <Deewiant> But as you say, it doesn't matter.
19:51:22 <AnMaster> or maybe that wasn't in ncrs... *shrug*
19:51:30 <AnMaster> and I won't waste more time on this indeed
19:52:11 <Deewiant> Hmm, maybe you're right, those are in the same file.
19:53:21 <AnMaster> -s, --script Begin execution on the second line if the first line
19:53:36 <ehird> Not cool because in a shebang you can only specify one command line argument
19:53:38 <AnMaster> Deewiant, do you actually use that feature?
19:53:41 <ehird> so you can't use --script and /usr/bin/env
19:53:46 <Deewiant> AnMaster: No, but someone requested it.
19:54:10 <AnMaster> Deewiant, did you do it without any justification or?
19:54:23 -!- Sgeo has joined.
19:54:34 <Deewiant> I can't remember the justification off the top of my head.
19:55:02 <AnMaster> remember who? Because I wonder who would want to run scripts written in befunge...
19:55:41 -!- tombom has joined.
19:56:31 <AnMaster> Deewiant, hm you changed to use a switch like cfunge? last I looked you used an array of function pointers...
19:57:06 <AnMaster> for main dispatch of instructions
19:57:20 <AnMaster> oh well I guess not only I copied then ;P
19:57:40 <Deewiant> I didn't copy anything, that was just when cfunge started being faster than CCBI :-P
19:57:55 <Deewiant> I haven't actually looked at cfunge's source at all, I don't think, apart from snippets you've pastebinned
19:58:04 <AnMaster> Deewiant, so you deny you are violating the GPL3 license used for cfunge?~
19:58:15 <AnMaster> (I wouldn't bother for that small anyway)
19:58:36 <Deewiant> It's not like switch statements are exactly cfunge's property anyway
19:58:39 <AnMaster> Deewiant, and I remember how we discussed this long ago
19:59:32 <AnMaster> Deewiant, is "ip" some global variable pointing to the current ip or such?
20:00:45 <AnMaster> reallyGoEast <-- lovely function name (yes I see it is some special thing for MODE or such indeed)
20:03:16 <AnMaster> Deewiant, your string handling is odd, I can't figure out how it handles the double space thing
20:03:47 <AnMaster> where does it handle b98 spaces?
20:16:12 <AnMaster> Deewiant, this http://rafb.net/p/e2Spd512.html
20:16:31 <AnMaster> Deewiant, that was all I needed
20:18:06 <AnMaster> Deewiant, anything wrong with it
20:18:46 <AnMaster> Deewiant, wrong format for me :)
20:19:29 <AnMaster> Deewiant, this is easier though
20:19:31 <ehird> no learn awk or perl
20:19:42 <AnMaster> yeah I would use awk if I had to
20:20:01 <Deewiant> I won't accept a patch for functionality that already exists :-P
20:20:29 <AnMaster> Deewiant, right, better format though IMO
20:22:03 <AnMaster> Deewiant, ... stack pushing is slow in ccbi?
20:22:19 <Deewiant> 15 million iterations of space[x, y].
20:22:29 <Deewiant> Well, actually a bit more since there's g and p there.
20:23:00 <AnMaster> ehird, not found here, what does it do?
20:23:06 <Deewiant> And, that's obviously slower with a hashtable than with a static array.
20:23:17 <Deewiant> I haven't profiled it, but I suspect that's by far the main reason.
20:23:21 <Deewiant> AnMaster: Did you try the new Mycology?
20:23:23 <AnMaster> ah found it in package manager
20:23:32 <ehird> AnMaster: It's grep(1), but it highlights matches, does a recursive search of . if you just do 'ack term', lets you filter by prorgamming language, automatically omits VCS directories, ... http://betterthangrep.com/
20:23:38 <Deewiant> ehird: Doesn't support Haskell.
20:23:49 <ehird> (also, 'ack term' greps stdin if it has an stdin. I guess it checks for terminalosity)
20:23:54 <ehird> Deewiant: you mean no --haskell?
20:24:06 <ehird> it's not like it's some innate mega clever "support"
20:24:11 <ehird> it's just file extension matching, iirc :-P
20:24:24 -!- kar8nga has joined.
20:24:38 <AnMaster> Deewiant, I'm waiting for trace dump from ccbi atm...
20:24:56 <AnMaster> Deewiant, will try mycology later
20:25:34 <AnMaster> Deewiant, also I think mycology is correct there about offset 0,0 before. It is what specs says in fact
20:25:58 <Deewiant> That doesn't mean I can't make it more amenable to slowdown.
20:26:09 <AnMaster> oh? you mean by not resetting?
20:26:43 <AnMaster> Deewiant, btw your yes s idea was even slower than my -D when it came to dumping
20:27:13 <Deewiant> 15 million iterations, like I said.
20:27:30 <Deewiant> 30 bytes per line - about right, no?
20:28:17 <AnMaster> Deewiant, hm my ASCII code dump was wrong there...
20:28:28 -!- oerjan has quit ("leaving").
20:29:39 <fizzie> Who cares about speed: (incoming 8-line paste)
20:29:40 <fizzie> fis@eris:~/irclogs/freenode/#esoteric$ egrep -i 'slow|fast|speed|optimi|profil' 200[678]* | grep '] <' | cut -c 37- | sed -e 's/>.*//' | sort | uniq -c | sort -nr | head
20:29:41 <AnMaster> Deewiant, any idea why the number there is NOT ASCII code?
20:29:48 <Deewiant> AnMaster: Or it will, but it won't do what you think.
20:30:00 <ehird> fizzie: Note: my lines are mostly ridiculing AnMaster.
20:30:03 <AnMaster> Deewiant, I copied from the unknown instruction dump thing
20:30:08 <Deewiant> AnMaster: The 1 there means 'argument number 1'
20:30:11 <AnMaster> Deewiant, what should I use instead?
20:30:17 <Deewiant> AnMaster: Obviously you did :-P
20:30:26 <Deewiant> AnMaster: So it's outputting the x-coordinate there.
20:30:34 <ehird> fizzie: Do one of them to see who ends sentences with ? the most
20:30:40 <ehird> I bet AnMaster or maybe sgeo or smth
20:30:47 <ehird> s/sentences/messages/
20:31:05 <AnMaster> fizzie, do one for repeated lines
20:31:23 <ehird> AnMaster: Me, Deewiant, end list.
20:31:27 <AnMaster> from same person after the other
20:31:46 <ehird> AnMaster: Interestingly, the two people who have to deal with your questions most.
20:32:07 <fizzie> Well, for 2006-2008 absolute number of questions top-3 is:
20:32:13 <AnMaster> ehird, ais also have to deal with it
20:32:26 <ehird> s/have/has/; and he has the patience of... of a mammoth.
20:32:35 <AnMaster> fizzie, yeah, if you never ask you never learn
20:32:47 <fizzie> Although tusho+ehird == 3864 too. Relative number of ?s would be more interesting, but I don't have a script ready for that.
20:33:02 <ehird> Deewiant: 'how many % of msgs are qs'
20:33:08 <AnMaster> and you just made your number go up
20:33:29 <Deewiant> AnMaster: You'll be careful to not use ? from now on then, I take it?
20:33:30 <AnMaster> fizzie, what exact line did you use there to find the question marks<question mark goes here>
20:33:54 <AnMaster> Deewiant, or use it in the middle of lines or something, depending on what command fizzie used
20:33:56 <fizzie> "looks like a privmsg" + matches /\? *$/
20:34:11 <fizzie> The " *" there is simply because I self always add a trailing space. :p
20:34:13 <ehird> fizzie: what if you let the spacers go free
20:34:14 <AnMaster> ah right... so in the middle of line is fine
20:34:40 <fizzie> There was a bot configured to "answer" all questions.
20:35:32 <fizzie> Numbers are pretty same even without the trailing-space thing, although one of oklopol's questions disappears. I'm not in the top-ten anyway.
20:35:57 <AnMaster> Deewiant, hm you break the spec, you don't reflect on stdout failing it seems
20:36:40 <AnMaster> Deewiant, "Although the standard input and output are generally displayed in some sort of interactive user terminal, they needn't be; many operating systems support redirection. In the case of an end-of-file or other file error condition, the & and ~ both act like r. "
20:36:49 <fizzie> I don't always remember the space. But I have 481 questions with /\? *$/ and only 206 with /\?$/.
20:37:32 <AnMaster> ok you don't break the spec, but you are inconsistent
20:38:45 <AnMaster> Deewiant, rebuild ignores ^C ....
20:39:12 <ehird> --[no]haskell .hs .lhs
20:39:16 <ehird> — http://betterthangrep.com/
20:40:57 <ehird> Deewiant: --type-set d=.d
20:41:09 <ehird> --type-add befunge=.bf,.b98
20:41:13 <ehird> --type-set befunge=.bf,.b98
20:41:17 <ehird> type-add adds to an existing type
20:41:34 <ehird> also, for one-offs, -G '\.b98$'
20:41:43 <ehird> and yes --type-(set|add) are persistent
20:44:29 <ehird> ack allows you to define your own types in addition to the predefined
20:44:29 <ehird> types. This is done with command line options that are best put into
20:44:31 <ehird> an F<.ackrc> file - then you do not have to define your types over and
20:44:33 <ehird> over again. In the following examples the options will always be shown
20:44:34 <ehird> on one command line so that they can be easily copy & pasted.
20:44:39 <ehird> from the "defining your own types" manual section
20:44:56 <ehird> When defining your own types in the F<.ackrc> file you have to use
20:44:57 <ehird> --type-set=eiffel=.e,.eiffel
20:45:00 <ehird> or writing on separate lines
20:45:06 <ehird> The following does B<NOT> work in the F<.ackrc> file:
20:45:07 <ehird> --type-set eiffel=.e,.eiffel
20:49:16 <ehird> Deewiant: Did you last use it a year ago or sth?
20:49:24 <ehird> I'm getting that vibe
20:49:24 <Deewiant> I haven't looked at it since the release announcement
20:50:04 <ehird> Yeah, everything is massively new. :D
20:50:14 <Deewiant> Or maybe it was a 1.x release which was somehow big
20:50:22 <AnMaster> Deewiant, is this really correct:
20:50:24 <AnMaster> "Unimplemented instruction '{}' ({1:d}) (0x{1:x}) encountered at ({}, {}).",
20:50:39 <Deewiant> AnMaster: Yes, it is, you can try it and it'll work.
20:50:53 <AnMaster> Deewiant, where does i end up?
20:51:08 <Sgeo> I'm not in the top 3 for questions?
20:51:19 <Deewiant> AnMaster: Also, & and ~ do reflect on stdin failing
20:51:21 <ehird> {} (we're at 0, now at 1)
20:51:27 <ehird> {1:d) (we're at 1, now at 2)
20:51:32 <ehird> {1:x} (we're at 1, now at 2)
20:51:39 <ehird> {} (3 -> we're done)
20:52:29 <fizzie> Sgeo: You have just 726 questions.
20:52:43 <fizzie> Well, more now, of course. But I was counting just 2006-2008 anyway.
20:56:12 <ehird> [ehird:~] % ack -i \bsyn\b
21:15:12 <ehird> `w3m -dump` that is
21:28:00 <AnMaster> fizzie, you said ehird+tusho was number two right<q>
21:28:12 -!- ehird has set topic: topic appoppic http://tunes.org/~nef/logs/esoteric/?C=N;O=D aproppotic.
21:28:45 <AnMaster> what about adding in ehird_ ehird` and so on fizzie~question~
21:28:48 <fizzie> Those unterminated q tags are upsetting my sense of balance.
21:28:57 <ehird> ???????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????
21:28:59 <ehird> ?????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????
21:29:04 <ehird> ???????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????
21:29:06 <AnMaster> fizzie, I wasn't using HTML anyway.
21:29:09 <ehird> ??????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????
21:29:13 <ehird> ??????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????
21:29:17 <ehird> Oh wait, it's per-line.
21:29:37 <fizzie> It still looks somehow unbalanced. Anyway, I don't think the nick-variants affect things that much.
21:31:00 <AnMaster> I think that is the correct syntax.(insert a question mark here, also I'm trying to make too many variants to make it easy to special case that when grepping)
21:31:40 <Deewiant> AnMaster: You'll lead the stats for years to come anyway; why bother?
21:32:13 <Deewiant> AnMaster: No, really, you will. :-P
21:32:32 <AnMaster> Deewiant, better start early IMO.
21:32:53 <oklopol> fizzie: my questions don't always have a question mark.
21:33:42 <fizzie> oklopol: I'm not writing a regexp to recognize things that semantically speaking are questions, if that's what you mean.
21:33:47 <oklopol> yes but i do it mucho mucho.
21:34:29 <oklopol> maybe everyone else does it mucho mucho too
21:37:45 <AnMaster> and if I miss it, I correct it by adding on the next line
21:37:53 <AnMaster> so I guess that means I'm over-represented
21:38:04 * ehird downloads wikipedia, lynx -dumps it all, and cats it all into one file called tome.txt
21:38:27 <AnMaster> ehird, doing it with xml dump or?
21:38:33 <ehird> AnMaster: static html dump
21:38:42 <ehird> w/ a perl script to filter out the common glump
21:38:57 <ehird> http://static.wikipedia.org/new/wikipedia/en/articles/m/a/i/Talk%7EMain_Page_649b.html scroll to the bottom and gawp
21:39:20 <ehird> wikipedia-en-html.ta..>21-Jun-2008 16:44 14G
21:39:31 <ehird> can I plz just have the text only :||||||||
21:39:37 <AnMaster> they provide that too<ascii code 63>
21:39:50 <ehird> but not pre-rendered
21:39:53 <Deewiant> ehird: I wouldn't be surprised if that took more space :-P
21:40:07 <ehird> Yar, I just want HTML without all the imags
21:40:22 <ehird> The current dump is the June 2008 edition. This dump has no image snapshot, it's just HTML. Due to performance problems when compressing millions of files with 7-zip, the archives are now packaged as a 7-zipped tar file.
21:40:29 <ehird> I thought it was just lik
21:40:50 <AnMaster> I never seen that "based on work" thing before
21:40:56 <ehird> AnMaster: it's the static version
21:41:02 <ehird> gfdl legally requires it
21:41:09 <ehird> AnMaster: it's the static version
21:41:14 -!- sebbu2 has joined.
21:41:19 <ehird> and note the lack of history/edit/etc
21:41:45 <fizzie> The XML-based "current revisions, all pages" dump I used for fungot (since they didn't provide talk pages separately) was something like nine gigabytes. I think. I threw it out already.
21:41:46 <fungot> fizzie: do you have to write a text-based braincopter/ brainloller first, then
21:41:57 <ehird> Oh well, I have gbs
21:42:25 <ehird> This makes me !happy.
21:43:02 <fizzie> You people are all so impatient; AnMaster wants all programs to execute in less than a second, and you're complaining about a rather reasonable 13-hour download.
21:43:15 <ehird> 2009-03-13 01:27:21 done Articles, templates, image descriptions, and primary meta-pages.
21:43:15 <ehird> 2009-03-13 01:27:20: enwiki 8251357 pages (61.202/sec), 8251357 revs (61.202/sec), 59.5% prefetched, ETA 2009-03-15 15:29:29 [max 21919559]
21:43:17 <AnMaster> ehird, that would be a RISC OS binary called happy I think...<question>
21:43:17 <ehird> This contains current versions of article content, and is the archive most mirror sites will probably want.
21:43:20 <ehird> pages-articles.xml.bz2 4.6 GB
21:43:27 <ehird> 4.6gb seems quite acceptable.
21:43:34 <ehird> I can render it locally.
21:43:45 <fizzie> I would've gotten pages-articles, too, but it didn't have talk pages.
21:44:45 <AnMaster> Deewiant, 431 MB output log...
21:45:04 <AnMaster> funny when less says: Calculating line numbers... (interrupt to abort)
21:45:46 <AnMaster> Deewiant, was that a question $QUESTIONMARK
21:46:13 <fizzie> "pages-meta-current.xml.bz2 8.6 GB". There seems to be approximately even split between article-content and user-content/discussion.
21:46:23 <AnMaster> Deewiant, didn't you forget something at the end (ascii code 63)
21:46:40 <Deewiant> AnMaster: No, the question mark was earlier, the question came later.
21:47:04 <AnMaster> Deewiant, then you should use the relevant unicode symbol
21:47:25 <Deewiant> You're just travelling in the wrong direction in time.
21:47:48 <ehird> Deewiant: reminds me of mostly harmless's Guide
21:48:21 <AnMaster> Deewiant, so you read what the next line *directed at you* after this one will be then<question> If so what will it be#ascii code 63#
21:48:22 <ehird> "How many of me are there?"
21:49:14 -!- psygnisf_ has joined.
21:49:15 <AnMaster> ¿I don't talk spanish<64-1 in ASCII>
21:49:16 <ehird> Watch me travel the other way in time.
21:49:38 <ehird> AnMaster, you now must ask a question to avoid breaking temporal reality.
21:49:47 <ehird> 21:49 Tuscane: ehird, there's a perl script called wikiprep, my mistake, not xslt.... 46 hours for a full dump to run
21:49:50 <ehird> 46 hours to render wp.
21:50:21 <ehird> AnMaster: Wrong question.
21:50:30 <ehird> The question I answer is much further in the future.
21:50:43 <AnMaster> Deewiant, yes but it was worth it this time
21:50:56 <Deewiant> Oh, only important questions get a question mark from now on? :-P
21:51:10 <AnMaster> Deewiant, possibly. I'm not sure yet
21:51:14 <fizzie> Deewiant: That smiley-appending trick is also quite clever.
21:51:30 <AnMaster> fizzie, you could filter around that too.
21:51:32 <Deewiant> fizzie: Why not just grep for plain \?
21:51:51 <fizzie> Someone said "lines ending in ?" originally. No real reason, though.
21:52:07 <AnMaster> fizzie, and combine ehird and tusho please
21:52:19 <fizzie> fungot: How often do you ask questions here?
21:52:19 <fungot> fizzie: ran as fast as, say, staying here
21:52:31 <fungot> Available: agora alice darwin discworld europarl ff7 fisher ic irc* lovecraft nethack pa speeches ss wp
21:52:35 <fungot> Selected style: alice (Books by Lewis Carroll)
21:52:35 <AnMaster> fizzie, never seen a question mark from it
21:52:43 <ehird> fungot: You should be questiony now, despite not having ? in your grammar.
21:52:44 <fungot> ehird: taking fnord fnord fnord
21:52:45 <fungot> AnMaster: " oh, thank you!"
21:52:51 <fungot> Deewiant: 14. all unwarlike nations are uncivilised. and as to being in a fright. " oh," said the duck.
21:52:58 <fungot> AnMaster: there comes a happy pause, for human passion madly yearning! o weary air of dumb despair, from marble won, to marble turning! " leave us not thus!" we can but stand aside, and let us suppose all the new ones to be put before men who do know what fnord is, if i could only just see the forms of the children, " come, there's no place for ignorance in these times! and all fnord agree that the time to learn is when you're
21:52:59 <Deewiant> fungot: Not being cooperative today, I see
21:53:00 <fungot> Deewiant: " it is just as i feared!" and we all went out into the snow! and you'd have fnord it, as if it really wasn't worth noticing. but, when he dipped it they fnord a little farther off, and the slithy toves did gyre and gimble in the wabe; all mimsy were the borogoves, and the fall was over.
21:53:07 <ehird> "oh," said the duck.
21:53:16 <Deewiant> fungot: Tell me about the duck
21:53:16 <fungot> Deewiant: " lions don't fnord," she added, as an actual fact"--which does not at all imply that no x can be y") the two terms, which contain these contradictory attributes, may be called ' the premisses', the square would be ' occupied'.
21:53:22 <fizzie> It should have a ? token in there, though.
21:53:22 <fungot> Deewiant: the following is a little harder, perhaps: he heard once more the baron looked fnord to the smaller, that there are no gentiles in the garden.
21:53:27 <fungot> AnMaster: the professor referred to a card he held in his hand, fnord silence said, " that that's his line. military genius, indeed! i'll vex her finely, though!"
21:53:31 <fungot> Deewiant: the gloomy look returned to arthur's face: and i could not at the moment, that was miles away from the door.
21:53:36 <fungot> Deewiant: " hurted mine self welly much!" i said, " but i know that, because i've held up one of the fnord clearly depends on his fnord face, as he fnord over the fnord, the fnord, full of tricks and fun.
21:53:47 <fungot> ehird: we all granted that. well, let's see how it would look: this led to a fnord, sometimes." make up your mind. now, only fnord you believe fnord put some of it in!"
21:54:38 <fizzie> I don't remember how to parse that language model format, so I can't really check if the question mark token is used in any of them.
21:54:47 -!- AnMaster has changed nick to AnMaster_ipv6.
21:54:54 <ehird> fungot: don't green
21:54:55 <fungot> ehird: " yea, each to each was worse than foe: thou, a scared fnord, fnord away almost like snow, as they got it in a newspaper." " all new cakes are nice.
21:55:12 <AnMaster_ipv6> ipv4 to freenode, because that is much more stable
21:55:13 <ehird> AnMaster_ipv6: in a day or two I will have fucking NATIVE STATIC IPV6 fuck yeah
21:55:25 <ehird> AND I WILL SUBVERT YOUR SHITTY REGIME OF... FAKITUDE!
21:55:29 <ehird> Of course I have to lose my bouncer
21:55:38 <ehird> But 'tis a small price to pay for ipv6 ego-masturbation
21:55:38 <Deewiant> fizzie: Do you know if any Finnish ISP offers IPv6?
21:55:44 <ehird> Deewiant: Nova or sth
21:55:46 <fizzie> Deewiant: Nebula. As I've said many times.
21:56:03 <Deewiant> fizzie: Well sorry, I haven't asked you. :-P
21:56:15 <AnMaster_ipv6> anyway this should spread out the question marks over some nicks
21:56:19 <ehird> AnMaster_ipv6: Parent. And I'm not sure, sometimes these things just come together. :P
21:56:20 <fizzie> Deewiant: No, I mean, it's been the topic on the channel several times. You should be logreading.
21:56:31 <ehird> everyone should logread
21:56:52 <Deewiant> Nebula is too expensive for my blood
21:56:58 <fizzie> Deewiant: Anyway, they give you this static /64 block, and are also polite enough to point the reverse-DNS zone NS delegation to any name server(s) I happened to want.
21:57:32 <fizzie> Deewiant: Yes, it's a bit more expensive than the mean, and the v6 and static-IP stuff is also only in the "advanced" product, not the cheaper "basic" one.
21:57:33 -!- sebbu has quit (Connection timed out).
21:57:33 -!- sebbu2 has changed nick to sebbu.
21:57:33 <Deewiant> I don't need a static address. I'd just like IPv6 support.
21:57:50 <ehird> Move to the UK and get bogons.
21:57:56 <ehird> They give you static IPV6 for free. :|
21:58:13 <ehird> How expensive is nebula, fizzie?
21:58:19 <Deewiant> ehird: http://www.nebula.fi/nebulazone.php
21:58:28 <ehird> I don't read moonlanguage
21:58:43 <Deewiant> ehird: You can read '24M/1M ADSL 59.90€' right?
21:58:45 <fizzie> You probably can interpret the table on the right side.
21:59:15 <ehird> well, £45/mo if you want 800kbps upload (as opposed to 400 or so)
21:59:40 <fizzie> This is 1Mbps up by default.
21:59:42 <Deewiant> Guess I'll just wait for the major ISPs to jump on the IPv6 bandwagon.
21:59:47 <fizzie> More expensive for the 3M variant.
21:59:54 <ehird> fizzie: Well, yeah, also 24M down.
22:00:09 <ehird> This area doesn't have ADSL2+ so we don't get the fancy 24M speeds.
22:00:15 <ehird> And the only adsl2+ isp is Be, iirc
22:00:28 <fizzie> The 8M/1M is 50 eur/month, and 4M/1M is 40 eur. I'm not sure what that is in your fancy money.
22:00:51 <ehird> 40 dirty european money = 36 proper british money
22:00:57 <ehird> They're, uh, pretty similar.
22:01:18 <ehird> "1 Euro = 0.909931183 British pounds"
22:01:26 <ehird> The 1 penny makes all the difference
22:01:36 <fizzie> Deewiant: It was some sort of multiple-connections-load-balancing thing, I think.
22:02:04 <fizzie> Deewiant: Yes, it's just ADSL2+ port bonding to get higher speeds.
22:02:13 <ehird> Bogons don't even give you a router or anything.
22:02:41 <Deewiant> What speed is their VDSL2, I wonder?
22:02:45 <fizzie> "ADSL2+ port bonding is also known as G.998.x or G.Bond" ← http://en.wikipedia.org/wiki/ADSL2+
22:02:59 <ehird> SHUT UP ABOUT ADSL2+ YOU'RE MAKING ME WANT TO MOVE TO .Fi
22:03:51 <Deewiant> ehird: EuroDOCSIS 3.0 for the win.
22:03:58 <fizzie> Obviously I don't actually get 24Mbps here.
22:04:01 <fizzie> styx> wan adsl chandata
22:04:01 <fizzie> DSL standard: ADSL2+ Mode
22:04:01 <fizzie> near-end bit rate: 19235 kbps
22:04:01 <fizzie> far-end bit rate: 1020 kbps
22:04:02 <Deewiant> 110M/5M for around 45-50 euros.
22:04:10 <ehird> Deewiant: Oh my god.
22:04:12 <Deewiant> Crippled upload but it's still sweet.
22:04:27 <ehird> What I mean to say is, I want that.
22:04:49 <fizzie> Deewiant: Isn't it still a bit shared between people, though? At least the earlier cable-modem stuff used to be.
22:05:03 <ehird> 13mb/sec... damn that's hot.
22:05:04 <Deewiant> It's not marketed that way, anyway.
22:05:08 <ehird> That's like *hot hot hot*
22:05:37 <fizzie> I think Sonera and Elisa are also offering 100M/10M stuff in some very randomly selected locations.
22:05:44 <fizzie> Basically places they've bothered to dig fiber to.
22:05:57 <AnMaster_ipv6> fizzie, you said those ISK added up, well, not really I'm at 211 ISK...
22:06:20 <ehird> fizzie: do many .fi isps do traffic shaping bullshit?
22:06:24 -!- psygnisf_ has changed nick to psygnisfive_.
22:06:28 <Deewiant> ehird: Not any that I know of.
22:06:59 <fizzie> Deewiant: Quite many do some sort of port-filtering, though. At least on SMTP ports, anyway. Not that it usually matters much.
22:07:03 <Deewiant> Some ISPs have opted in to the police's stupid blocklist but that's about all they do, I think.
22:07:12 <ehird> Deewiant: Heard of Phorm?
22:07:18 <ehird> .uk isps have no respect for... well, anyone.
22:07:32 <ehird> The company drew attention when it announced it was in talks with several United Kingdom ISPs to deliver targeted advertising based on user browsing habits by using deep packet inspection.[3] It is one of several companies developing behavioral targeting advertising systems, seeking deals with ISPs to enable them to analyse customers' websurfing habits in order to deliver targeted advertising to them. Others include NebuAd and Front Porch.[4]
22:07:41 <ehird> They convinced virgin media and other major isps to deploy it
22:07:51 <ehird> Phorm is working with major US[5] and British ISPs including British Telecom, Virgin Media, and TalkTalk on a targeted advertisement service to monitor browsing habits and serve relevant advertisements to the end user. Phorm say these deals will give them access to the surfing habits of 70% of British households with broadband.[1]
22:08:03 <ehird> Hope they choke on a dick
22:08:15 <ehird> one of the owners of bogons posted on a mailing list about how they suck :-D
22:08:41 <fizzie> A *very* large percentage of Finnish ISPs state in their service agreement things something like "you are not allowed to have any sort of servers connected to your pipe", but I haven't heard of anyone actually enforcing that rule.
22:09:13 <Deewiant> And if they did, I think people would start complaining quite loudly.
22:09:20 <Deewiant> Not that they'd probably care.
22:09:48 <ehird> The bogons AUP is, uh, "if you do illegal things we can shut your service off and we can shut your service off regardless too"
22:09:50 <AnMaster_ipv6> fizzie, so do my ISP, and I recorded some port scans on 25, 80 and 8080
22:09:51 <ehird> http://www.bogons.net/aup.shtml
22:10:19 <AnMaster_ipv6> from an IP in the same range as the DNS servers of my ISP
22:10:36 <fizzie> Deewiant: Oh, and Elisa's formulation of that prohibition was really really bizarre.
22:10:45 <fizzie> Deewiant: "Ellei Elisan kanssa ole nimenomaan muuta sovittu, asiakas ei saa pitää laajakaistaliittymässä palvelimia tai käyttää verkkopalvelua palvelujen tarjoamiseen laajakaista- tai internet-palvelujen käyttäjille sijoittamalla palvelinlaitteita tai -sovelluksia tietoliikenneyhteyteen."
22:10:52 <ehird> tietoliikenneyhteyteen
22:10:59 <ehird> laajakaistaliittymässä
22:11:03 <fizzie> AnMaster_ipv6: Finnish ISP. Well, a phone company. Something like that.
22:11:10 <Deewiant> ehird: Those aren't very long words. :-P
22:11:18 <ehird> Deewiant: show me a long word i love long words
22:11:22 <Deewiant> Longer than average but not very long IMHO.
22:11:40 <Deewiant> ehird: I think they're bad for your health.
22:12:08 <AnMaster_ipv6> ehird, What about Swedish then... technically industriarbeterarfackårstämmepenna would be a valid word
22:12:31 <ehird> 'industry are better are fuckers stamina pen'?
22:12:32 <Deewiant> fizzie: If things like that were parsed literally that'd mean nothing in terms of the law. Too bad they're not.
22:12:45 <ehird> Deewiant: translate?
22:12:47 <Deewiant> I mean, 'sijoittamalla .. tietoliikenneyhteyteen'? What?
22:12:49 <fizzie> Anyway, the text doesn't partially even make sense. "käyttää verkkopalvelua palvelujen tarjoamiseen laajakaista- tai internet-palvelujen käyttäjille sijoittamalla palvelinlaitteita tai -sovelluksia tietoliikenneyhteyteen" equals approximately "to use the network service to provide services for broadband- or internet-service users by placing server equipment or software into the data communication link".
22:13:22 <AnMaster_ipv6> ehird, "(((industry worker (workers union)) yearly meeting) pen)
22:14:11 <fizzie> Wikipedia has for Swedish: "nordöstersjökustartilleriflygspaningssimulatoranläggningsmateriel-underhållsuppfölningssystemdiskussionsinläggsförberedelse-arbeten"
22:14:13 <Deewiant> ehird: You can probably google for 'finnish long words' and come up with some.
22:14:27 <ehird> TRANSLATE FUCKER!!!!!!!!!!!!!!!!!!!!!!!
22:14:52 <fizzie> And for Finnish, "kumarreksituteskenteleentuvaisehkollaismaisekkuudellisenneskenteluttelemattomammuuksissansakkaankopahan" which is not a compound word, just suitable suffixes. It's not really parseable.
22:15:13 <AnMaster_ipv6> fizzie, translate all the Finnish in the last few screens.
22:15:14 <Deewiant> fizzie: What page does Wikipedia have that on?
22:15:18 <ehird> Finnish: lojban for finns
22:15:23 <Deewiant> AnMaster_ipv6: Everything has been translated.
22:15:33 <AnMaster_ipv6> Deewiant, "<Deewiant> I mean, 'sijoittamalla .. tietoliikenneyhteyteen'? What?" <-- that too<q>
22:15:41 <Deewiant> AnMaster_ipv6: It's a quote from fizzie's translation.
22:15:42 <fizzie> Deewiant: I don't know, since it was on a Finnish forum under a "found in Wikipedia" label. Might not even be there any more.
22:16:03 <AnMaster_ipv6> <fizzie> And for Finnish, "kumarreksituteskenteleentuvaisehkollaismaisekkuudellisenneskenteluttelemattomammuuksissansakkaankopahan" which is not a compound word, just suitable suffixes. It's not really parseable.
22:16:16 <Deewiant> AnMaster_ipv6: "Not really parseable."
22:16:20 <Deewiant> AnMaster_ipv6: Means nothing, it's crap.
22:16:24 <AnMaster_ipv6> <fizzie> Wikipedia has for Swedish: "nordöstersjökustartilleriflygspaningssimulatoranläggningsmateriel-underhållsuppfölningssystemdiskussionsinläggsförberedelse-arbeten"
22:16:39 <Deewiant> Yes, I can parse that as well, to a reasonable degree of accuracy.
22:17:03 <AnMaster_ipv6> what is nordöstersjön in English though... What is it in Swedish<q>
22:17:08 <ehird> antidisestablishmentarianism
22:17:11 <fizzie> Compound words are easier. There are "reasonable" extremely long Finnish compound words, too. But the suffix-adding only goes so far.
22:17:30 <Deewiant> AnMaster_ipv6: You should know, right? :-P
22:17:33 -!- psygnisfive has quit (Read error: 110 (Connection timed out)).
22:17:37 <ehird> AnMaster_ipv6: North Baltic Sea
22:18:01 <Deewiant> AnMaster_ipv6: Googling suggests it's only used in that word.
22:18:33 <Deewiant> 'Men "Nordöstersjö" kan ju inte vara ett använt ord - norra östersjön i all ära.'
22:18:38 <AnMaster_ipv6> ehird, so do you want "nordöstersjökustartilleriflygspaningssimulatoranläggningsmateriel-underhållsuppfölningssystemdiskussionsinläggsförberedelse-arbeten" translated or not <ascii 62+1>
22:18:53 <ehird> AnMaster_ipv6: Yes, and I hope it ends with a question mark.
22:19:21 <Deewiant> ehird: Words don't usually end in punctuation.
22:19:24 <Deewiant> AnMaster_ipv6: coast artillery
22:21:13 <AnMaster_ipv6> north eastern sea coast artillery aircraft surveillance simulator <anläggning> material - maintenance follow-up[not sure about that] system discussion <inlägg> preparation - work
22:21:27 <AnMaster_ipv6> and I'm not sure what the best word for those two non-translated words are
22:21:43 <Deewiant> 'anläggning' is something like 'structure', right?
22:21:55 <fizzie> The Wikipedia "Compound (lingustics)" page has a Finnish example, "lentokonesuihkuturbiinimoottoriapumekaanikkoaliupseerioppilas". I would translate, but I'm not quite sure of the army-related terminology.
22:21:57 <AnMaster_ipv6> BeholdMyGlory, that would fit in this case probably yeah
22:21:59 <Deewiant> Facility might be better, not sure.
22:22:29 <Deewiant> fizzie: Aeroplane jet turbine engine assistant mechanic non-commanding officer student
22:23:12 <AnMaster_ipv6> BeholdMyGlory, the Swedish or the Finnish one<ascii 3*3*7>
22:24:40 <AnMaster_ipv6> if you can't think in RPN then you should, uh... learn it or something...
22:25:06 <Deewiant> Of course I can, I wrote Mycology.
22:25:31 <AnMaster_ipv6> fizzie, in fact wouldn't you say it would be a good idea for everyone to know both RPN and prefix notation(/ (* 3 3 7 2) 2)
22:25:53 <fizzie> I can believe that the Finnish one might be used once or twice, if there actually happens to be such a post in the army. "lentokonesuihkuturbiinimoottori" sounds reasonable, and "apumekaanikkoaliupseerioppilas" too (for a did-not-go-to-army-person), but I'm not sure they have that sort of specialization.
22:26:08 <Deewiant> AnMaster_ipv6: No(+ 23 33 10 -)
22:26:20 <fizzie> Another wp quote. "The Finnish word saippuakivikauppias (soap-stone vendor) is claimed to be the world's longest palindromic word in everyday use. A meaningful derivative from it is saippuakalasalakauppias (soapfish bootlegger). An even longer effort is saippuakuppinippukauppias (soapdish batch seller)."
22:26:32 <ehird> soapfish bootlegger
22:26:45 <fizzie> Yes, I speak of soap-stone vendors every day.
22:27:09 <Deewiant> AnMaster_ipv6: It's a mixture, and ambiguous.
22:27:26 <AnMaster_ipv6> Deewiant, I didn't suggest you would mix them duh!
22:27:50 <Deewiant> I wasn't insinuating you were.
22:28:10 <AnMaster_ipv6> also I'm running out of original ways to write question mark in...
22:28:50 <Deewiant> Consider using the traditional instead.
22:29:15 <AnMaster_ipv6> ah, interobang should relieve some of the question mark pressure I think‽
22:29:17 <fizzie> Consider using the interrobang all the time, it would make you sound suitably high-strung.
22:29:23 <fizzie> Gah. You idea-stealer.
22:29:33 <ehird> ‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽
22:29:44 <Deewiant> It implies you're yelling all the time, which sucks.
22:30:00 <AnMaster_ipv6> FireFly, what does high-strung mean. It doesn't sound nice, or does it‽
22:30:16 <fizzie> 1. edgy, high-strung, highly strung, jittery, jumpy, nervy, overstrung, restive, uptight -- (being in a tense state
22:30:19 <Deewiant> Is there something fundamentally difficult about tab completion?
22:30:26 <ehird> ‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽
22:30:29 <ehird> ‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽‽
22:30:34 <AnMaster_ipv6> fizzie, I certainly am now that you done those stats
22:30:50 <fizzie> General punctuation page also has the question-mark-combinations: ⁇, ⁈ and ⁉. They have similar shoutery implications mostly.
22:31:05 <ehird> is there an upside down ‽
22:31:12 <AnMaster_ipv6> fizzie, still waiting for results about question mark anywhere in line where ehird and tusho were combined
22:31:20 <ehird> fizzie: make a list of the top unicoders
22:31:25 <ehird> (= everything outside of printable ascii)
22:32:07 <ehird> ←→↑↓↖↗↙↘⌃⌥⌘⇧¹²³⁴⁵⁶⁷⁸⁹⁰☺☹‽
22:32:09 <Deewiant> AnMaster_ipv6: Still waiting for results of cfunge on the latest Mycology, tell me when you've done that
22:32:39 <oklopol> saippuakuppinippukauppias xD
22:32:57 <AnMaster_ipv6> please, can we get an audio recording of saippuakuppinippukauppias
22:33:38 <ehird> just record half then reverse
22:33:42 <Deewiant> AnMaster_ipv6: Alternatively, still waiting for an updated cfunge in the public repo so I can run it myself.
22:33:45 <fizzie> Deewiant: ITYM ¿pɹ!ɥə -- There *is* a real inverted i somewhere, but I can't find it right now.
22:34:05 <Deewiant> fizzie: Nah, I didn't actually mean that, although I considered it.
22:34:33 <AnMaster_ipv6> Deewiant, will be in the feature branch on launchpad
22:34:55 <Deewiant> AnMaster_ipv6: Not rage.kuonet.org?
22:35:23 <Deewiant> AnMaster_ipv6: So how do I get to it, I'm bzr-ignorant
22:35:25 <ehird> http://www.justlinux.com/forum/showthread.php?t=147959
22:35:53 <Deewiant> ehird: 'more Linux since added plus one Solaris and 2 BSD'
22:36:32 <fizzie> Question marks anywhere:
22:36:46 <fizzie> The ehird there has had tusho added into it.
22:37:03 <fizzie> I don't have a microphone handy.
22:37:11 <oklopol> i have a broken microphone handy
22:37:12 <Deewiant> I don't have a microphone at all.
22:37:19 <AnMaster_ipv6> Deewiant, seems like lp have connection issues atm
22:38:02 <fizzie> I have a headphones/microphone headset thing at work, but people might look if I start repeating "saippuakuppinippukauppias" there to myself.
22:38:03 <Deewiant> oklopol: The times I might use it are mostly like these; and with my parents sleeping in the next room, I wouldn't use it anyway.
22:38:31 <fizzie> "I'm training our recognizer to recognize palindromes better."
22:38:37 <Deewiant> Hmm, actually my phone can probably record audio, so I could use that.
22:38:56 <Deewiant> fizzie: Or "I'm taking a 3-minute break off work"? :-P
22:39:00 <oklopol> Deewiant: well i just wanted to use the word a third time, but yes understandable.
22:40:11 <oklopol> saippuakuristajat, sirukauppias
22:40:45 <oklopol> yes, but it was a pretty palindrome, and i didn't want to erase it.
22:40:46 <fizzie> Saippuakiinnostunut sonni-iKauppias.
22:40:46 <Deewiant> If you're coming up with these in real time, props though
22:41:01 <oklopol> Deewiant: well yes naturally
22:41:11 <Deewiant> fizzie: First word fails at wordness.
22:41:13 <oklopol> well, umm, saippuakullipillukauppias
22:41:32 <oklopol> my randomizer is a bit childish.
22:41:34 <Deewiant> saippuakupanapukauppias in the same vein.
22:41:50 <AnMaster_ipv6> Deewiant, bzr branch http://rage.kuonet.org/~anmaster/bzr/cfunge/cfunge-speedup
22:41:51 <Deewiant> Although I guess that fails a bit at wordness too, with the genitive and all.
22:41:52 <fizzie> Deewiant: You mean "saippuakiinnostunut"? It's obviously anyone who's interested in soap.
22:42:02 <Deewiant> fizzie: "saippuasta" or no cigar?
22:42:09 <oklopol> saippuakaverirevakauppias :D
22:42:09 <fizzie> Deewiant: That's just a variant. :p
22:42:11 <Deewiant> AnMaster_ipv6: Can I do that in the existing repo?
22:42:24 <Deewiant> fizzie: Yes, well, I'm thinking that it's the only correct one. :-P
22:42:33 <oklopol> i guess palindromes are a bit trivial in languages you know natively
22:42:38 <Deewiant> AnMaster_ipv6: I mean, I have that rage.kuonet.org branch.
22:42:44 <oklopol> so maybe i'll just stop, i could probably spam them all night
22:43:34 <oklopol> (saippuakinolemumelonikauppias)
22:43:35 <AnMaster_ipv6> you create a shared repo instead and create branches under that
22:43:36 <Deewiant> AnMaster_ipv6: Okay, so it's completely separate.
22:44:02 <Deewiant> AnMaster_ipv6: So how do I go about that locally?
22:44:05 <AnMaster_ipv6> but bzr don't do the confusing thing that git does where you switch between branches
22:44:05 <fizzie> Deewiant: Saippuakisaan aasikauppias. I just *had* to generate at least one you accept as wordy enough. I hope you don't have anything against the soap competition.
22:44:29 <Deewiant> fizzie: Nah, that's fine, if a bit strange. :-P
22:44:34 <AnMaster_ipv6> Deewiant, create a directory cfunge, go into it, bzr init-repo .
22:44:42 <oklopol> saippuakisa-aasikauppias would be a compound
22:44:45 <AnMaster_ipv6> bzr branch http://rage.kuonet.org/~anmaster/bzr/cfunge/trunk
22:44:49 <AnMaster_ipv6> bzr branch http://rage.kuonet.org/~anmaster/bzr/cfunge/cfunge-speedup
22:45:09 <AnMaster_ipv6> so you would have to switch over your trunk branch
22:45:28 <fizzie> oklopol: I'm not sure there's really a market for donkeys competing in the soap competition.
22:45:48 <AnMaster_ipv6> anyway now I'm trying to find out why loggerhead crashed
22:46:50 <oklopol> fizzie: horses can be made soap out of, according to some sources, maybe soap competition is an euphemism for that
22:46:57 <oklopol> and donkeys are close enough to horses
22:47:15 -!- tombom has quit ("Peace and Protection 4.22.2").
22:47:22 <fizzie> oklopol: Isn't it glue they make out of horses? Well, maybe soap too, then.
22:47:54 <oklopol> those are a pair in my head
22:48:54 <AnMaster_ipv6> Deewiant, haven't fixed the broken quadrant bit yet
22:49:04 <Deewiant> AnMaster_ipv6: I noticed, it gets stuck a lot. :-P
22:50:35 <Deewiant> AnMaster_ipv6: Yes, I saw that as well.
22:50:55 <oklopol> so who here have passed mycology?
22:51:06 <oklopol> a reasonable version of it
22:51:31 <oklopol> i haven't, but it's my dream to do it one day
22:51:39 <fizzie> It was too long, I ran out of patience.
22:51:54 <fizzie> But I hope one day to get an interpreter that does.
22:52:02 <AnMaster_ipv6> oklopol, it is the bignum one coded in erlang yeah
22:52:22 <Deewiant> AnMaster_ipv6: What's line count for efunge vs. cfunge btw? (Ignoring fingerprints of course)
22:52:46 <Deewiant> An unfinished part of Mycology.
22:53:07 <AnMaster_ipv6> Deewiant, depends on which version of efunge, the one I was rewriting as a set of OTP style processes was considerably larger than the trunk one
22:53:12 <Deewiant> It evidently took on that name since the temporary file I originally made it in had/has that name.
22:53:21 <Deewiant> AnMaster_ipv6: Well, all, I don't care.
22:53:24 -!- BeholdMyGlory has quit (Remote closed the connection).
22:53:27 <fizzie> "Ah, it must be myco*edge* because it's bleeding-*edge* software."
22:53:56 <Deewiant> AnMaster_ipv6: Looks like 3559 for CCBI; still a bunch of TRDS handling and such in there, of course.
22:54:24 <Deewiant> Oh, and 519 of that is the debugger.
22:54:53 <Deewiant> So the core of CCBI is actually shorter than that version of efunge :-P
22:55:28 <Deewiant> Ah, you've got to include .h, too, unless they're autogenerated.
22:55:38 <Deewiant> Essentially, all the stuff you've written.
22:55:42 <AnMaster_ipv6> Deewiant, some of my *.h are auto generated, some are *.c
22:55:56 <Deewiant> AnMaster_ipv6: In the core, too?
22:56:11 <AnMaster_ipv6> one auto generated C and one auto generated h in core
22:56:32 <AnMaster_ipv6> Deewiant, well src/instructions/safe_env.c for example
22:58:41 <fizzie> One of my matrices is a lot smaller than the others. :/ They should all be 64x5119, but that one is 64x1025. And it's number 7 out of a set of 39; there should be nothing special there. How strange.
22:58:45 <AnMaster_ipv6> Deewiant, 5062 lines for the speedup branch, 2987 lines code:
22:58:50 <AnMaster_ipv6> ---------------- ----- --------- --------- --------- --------- ---------
22:58:50 <AnMaster_ipv6> ---------------- ----- --------- --------- --------- --------- ---------
22:59:00 <Deewiant> Yay for pasting 80% pointless lines
22:59:12 <Deewiant> dmd 10 1964 401 17.0% 675 3040
22:59:42 <Deewiant> AnMaster_ipv6: --------- ------- isn't particularly important, nor is the total of a single number.
23:00:14 <Deewiant> dmd 14 1859 361 16.3% 544 2764
23:00:28 <AnMaster_ipv6> Deewiant, efunge trunk passes mycology and is still smaller
23:00:48 <Deewiant> dmd 17 2522 554 18.0% 784 3860
23:00:55 <Deewiant> I was wondering why it was so small :-P
23:01:27 <fizzie> 4527 total lines in jitfunge (.cc + .hh); I don't have any fancy line-counting tools.
23:01:30 <AnMaster_ipv6> >;Deewiant, so how much of ccbi2 works now;337**.@
23:01:37 <Deewiant> AnMaster_ipv6: It doesn't compile currently.
23:01:46 <Deewiant> Before that, it worked except for fingerprints.
23:02:21 <Deewiant> Or actually, I think I had a few fingerprints working as well.
23:03:08 <Deewiant> AnMaster_ipv6: But yeah, it's left in a fugue state where I was trying to clean something up and ran into the compiler bug.
23:03:27 <Deewiant> The annoying thing is that even if the bug was fixed it'd take me probably quite a long while to figure out what exactly I was trying to do.
23:03:49 <Deewiant> AnMaster_ipv6: No, it'd loop forever.
23:04:29 <Deewiant> CCBI 2 could do it rather easily.
23:04:41 <Deewiant> CCBI 1 is too full of globals to even consider it. :-P
23:04:46 <AnMaster_ipv6> Deewiant, the REPL or that way to write the string<q>
23:04:54 <Deewiant> I did get minifunge to work though. Never again.
23:05:19 <Deewiant> AnMaster_ipv6: Yes, but not as hacky.
23:05:31 <Deewiant> Or actually, I'm not sure if it's currently in a working state.
23:05:59 <AnMaster_ipv6> Deewiant, it does have quite a few fingerprints...
23:06:08 <fizzie> Deewiant: jitfunge is more like debug.com. :p
23:06:16 <Deewiant> AnMaster_ipv6: GNU ed, then. :-P
23:06:43 <AnMaster_ipv6> efunge is like heirloom ed run under an interpreting emulator on a i286
23:06:52 <Deewiant> By the way, objdump doesn't like my DOBELA interpreter :-(
23:06:59 <Deewiant> It seems to think that ELF files should have sections
23:07:10 <Deewiant> AnMaster_ipv6: No, they're optional.
23:07:42 <Deewiant> AnMaster_ipv6: http://www.muppetlabs.com/~breadbox/software/tiny/teensy.html -- "the section header table is optional for executables -- but is almost always present! "
23:08:11 <Deewiant> Linux does run the executable, after all. :-P
23:08:29 <fizzie> They could be defining optional rather loosely.
23:08:58 <Deewiant> Anyways, I need to hand-write the ELF header for my program.
23:08:59 <fizzie> Deewiant: Have you tried "readelf" on it?
23:09:22 <Deewiant> "There are no sections in this file.
23:09:38 <AnMaster_ipv6> Deewiant, don't do the "store program in unused fields in ELF header" please
23:10:03 <fizzie> Deewiant: Well, I mean, what do you need objdump for, if readelf can read it?
23:10:34 <Deewiant> fizzie: Disassembly? Also, I think making objdump work would make gdb a lot happier with it as well.
23:10:57 <fizzie> Actually it's rather reasonable that *obj*dump wants sections, since it's "display information from object files" and all.
23:11:10 <Deewiant> Scratch the first two words actually; the reason I tried to get disassembly to work was to get gdb to work.
23:11:54 <AnMaster_ipv6> <Deewiant> Also, I think making objdump work would make gdb a lot happier with it as well. <-- <Deewiant> Scratch the first two words actually
23:12:07 <Deewiant> AnMaster_ipv6: "error while reading ELF: Zero count for section headers"
23:12:18 <Deewiant> AnMaster_ipv6: Yes, it's what I meant, you just skipped the first word.
23:12:36 <fizzie> Deewiant: It's rather retro.
23:12:47 <Deewiant> AnMaster_ipv6: It's a hex editor?
23:13:00 <fizzie> It's also a disassembler-and-stuff.
23:13:04 <AnMaster_ipv6> Deewiant, there is F6 or something to put it in ELF mode
23:13:07 <Deewiant> Maybe it would be, if it liked my executable.
23:13:29 <Deewiant> AnMaster_ipv6: Doesn't work, since it starts at 0 instead of where the code starts.
23:14:09 <Deewiant> Of course it does, since cfunge is probably made by GNU ld.
23:14:20 <fizzie> Since it's trying to be so tool-tool, one would think you could press some magic keys to tell it some offsets.
23:14:35 <AnMaster_ipv6> Deewiant, actually llvm-ld in this case, but it probably uses gnu ld below
23:14:41 <fizzie> Disclaimer: I don't even have it installed.
23:15:29 <Deewiant> fizzie: I can actually tell it to look at a certain offset; now if only I knew which one. :-P
23:15:35 -!- pikhq_ has joined.
23:15:38 <fizzie> Deewiant: Didn't you write it yourself. :p
23:15:39 <Deewiant> I could ask my assembler, I guess.
23:15:52 <fizzie> Deewiant: Ah, not yet, that is.
23:16:20 <Deewiant> When I do, I'll add a section table, even though it'll bump this over 4 Kio.
23:16:30 <fizzie> You can ask readelf for the entry point and compute from that, if your assembler's not talking to you.
23:16:42 <Deewiant> AnMaster_ipv6: That's just the magic number at the beginning.
23:16:48 <Deewiant> I'm talking about the whole 300 bytes or whatever.
23:17:41 <fizzie> Well, is your entry point at the start of the code?
23:18:03 <Deewiant> AnMaster_ipv6: Do tell me how.
23:18:16 <fizzie> The readelf-printed program header table probably should help you.
23:18:30 <Deewiant> fizzie: Yes, 0x78 or 0xe58 are my best guesses.
23:18:30 <AnMaster_ipv6> Deewiant, then search for that in your program to find the ELF header
23:18:51 <Deewiant> AnMaster_ipv6: I'm not looking for the ELF header; it must be at 0.
23:18:54 <AnMaster_ipv6> Deewiant, and in ht I just press F6 and select elf/image
23:19:11 <Deewiant> AnMaster_ipv6: Of course you do, since ht thinks your file is valid ELF.
23:19:26 <Deewiant> Even though it actually probably is, just a bit unusual.
23:19:30 <AnMaster_ipv6> you said "<Deewiant> AnMaster_ipv6: Do tell me how."
23:19:56 <Deewiant> Yes, you did, and it didn't help me at all, since it assumes that ELF is autodetected correctly.
23:19:58 <AnMaster_ipv6> Deewiant, tell it where the section headers is or something, that would probably help
23:20:15 <Deewiant> AnMaster_ipv6: There are no section headers, as I said.
23:20:56 <Deewiant> AnMaster_ipv6: Yes, that's writing my own ELF headers, which I said I plan/need to do.
23:21:00 <fizzie> Deewiant: Incidentally, what sort of tool made you that ELF file without section headers?
23:21:23 <Deewiant> Its 'ELF executable' mode is evidently supposed to be more of a debugging tool than anything you're meant to use.
23:21:53 <Deewiant> Maybe I could just turn my source files into *.o and use a linker.
23:22:13 <AnMaster_ipv6> <Deewiant> Its 'ELF executable' mode is evidently supposed to be more of a debugging tool than anything you're meant to use. <--- what
23:22:23 <Deewiant> AnMaster_ipv6: BTW, I can't force ht into x86-64 mode so this isn't helpful even if I know the offset, so so much for that.
23:22:37 <Deewiant> AnMaster_ipv6: Something unclear?
23:22:59 <Deewiant> AnMaster_ipv6: Note: "for you" makes no difference at all, since you have a different executable.
23:23:10 <Deewiant> AnMaster_ipv6: I was speaking about fasm.
23:23:28 -!- MigoMipo has quit (Read error: 110 (Connection timed out)).
23:25:40 <Deewiant> I'll take this break in the discussion as an opportunity to go to bed ->
23:26:40 <fizzie> AnMaster_ipv6: Some sort of french page suggests those wrappers are generated for dynamically linked functions. Since the editor won't even try to do dynamic linking or figure out the function names.
23:28:09 <oklopol> i'd pass mycology in one coding streak
23:28:44 <oklopol> but noooo i need to have an exam tomorrow
23:28:54 <fizzie> *Why* is that stubborn matrix so small? Don't get it.
23:28:57 <oklopol> that was a bit of a finnishm
23:29:13 -!- pikhq has quit (Connection timed out).
23:30:07 <AnMaster_ipv6> also I have decided to be proud of my question marks
23:30:10 <fizzie> The one I was complaining about earlier. I have deliberately been vague about it.
23:31:13 <AnMaster_ipv6> fizzie, where is this matrix used, what software or such
23:32:02 <fizzie> It's in Octave. I'm analyzing the [colloquial word for excrement] out of that matrix.
23:32:45 <fizzie> Yes, I'm analyzing poo out of the matrix.
23:32:54 <oklopol> hmm hope i didn't point to the wrong scientist
23:33:20 <fizzie> This is not work-related research, though.
23:33:28 <oklopol> AnMaster_ipv6: are you a scientist?
23:33:49 <fizzie> (Except for very tortured definitions of "related".)
23:34:01 <oklopol> AnMaster_ipv6: in what subject
23:34:13 <oklopol> what are your interests?????????????????
23:34:35 <AnMaster_ipv6> oklopol, CS hopefully, Starting at uni this autumn.
23:34:40 -!- kar8nga has quit (Read error: 145 (Connection timed out)).
23:35:19 <oklopol> the good parts of computer science don't really have much to do with science
23:35:42 <oklopol> terminology clash there prolly, i consider only inexact science science
23:36:04 <fizzie> oklopol: Also you consider only oko matters "good", I guess.
23:36:27 <oklopol> although i guess there is no term for exact science then, because math is reserved for a random subset of it.
23:36:58 <oklopol> fizzie: well only the things that are timeless
23:37:38 <oklopol> yes it sounds a bit silly as an english term
23:37:54 <fizzie> Anyway, we had a guy from TUT (Tampere University of Technology, where Tampere is a Finnish city) doing some random babbling, and there was this thing about non-negative matrix factorization for sound source separation, I wanted to play around with it a bit.
23:39:18 <oklopol> what's matrix factorization
23:39:39 <oklopol> i wish i hadn't skipped that linear algebra course, matrices are everywhere
23:40:04 <fizzie> It's when you have a matrix X, and then you figure out matrices Y and Z so that Y*Z is something a bit like X, maybe, sort-of.
23:40:32 <oklopol> err well okay, i guess that was kinda guessable
23:40:50 <fizzie> Non-negative matrix factorization is when the elements of X, Y and Z are all non-negative. :p
23:41:38 <oklopol> AnMaster_ipv6: err wait you would actually be interested in a career in science? i always thought of you as the silent nerdy talented code monkey type
23:42:35 -!- neldoreth has joined.
23:42:35 <AnMaster_ipv6> like a citation needed statement replaced with a reference to another wikipedia page
23:43:40 <oklopol> doesn't sound very wikipediaish
23:43:47 <AnMaster_ipv6> oklopol, what sort of job would I be able to get otherwise than in science, if I was "the silent nerdy talented code monkey type".
23:44:12 <AnMaster_ipv6> I don't fit FSF, because I lack a great big beard and the charisma.
23:44:13 <oklopol> AnMaster_ipv6: well i was thinking you'd sit in a cubicle and make code for a big company
23:44:19 <oklopol> then again i guess you don't like companies
23:44:33 <GregorR> The beard isn't strictly necessary, but the charisma is.
23:45:06 <AnMaster_ipv6> oklopol, rather I'd sit in a great big room and write a thesis on O(1) wait free algorithms that actually worked well in practise
23:45:26 <fizzie> Anyway, in this sound source separation application X is a collection of magnitude spectrums for short time windows of sound, and if you limit the number of columns in B (which == rows in G) to be suitably small, doing NMF estimation for X=B*G gives you a collection of "basis vectors" that are things like single notes of instruments, and a G matrix which represents the whole original sound as a linear combination of those.
23:46:12 <fizzie> Well, assuming you also have a sensible cost function when optimizing the X=B*G estimation. There's something about temporal continuity here.
23:46:36 <AnMaster_ipv6> GregorR, BUT: RMS, A. Cox and so on all have beards.
23:46:43 <GregorR> It's not necessary, but it helps.
23:46:51 <oklopol> i've never seen him do anything but bitch :D
23:47:12 <GregorR> So, if I connect to #esoteric with my BeagleBoard, is that cooler than AnMaster_ipv6 connecting via IPv6?
23:47:54 <oklopol> AnMaster_ipv6: true, was just going to say i've probably only seen the bad stuff.
23:47:55 <AnMaster_ipv6> (because I have a runway process eating ram so can't run firefox too)
23:47:56 <fizzie> The clever part here is that if you then cluster the vectors in B completely blindly, you end up with clusters that actually represent different sound sources. Then, if you want, you can just pick up the basis vectors of one source, the corresponding rows from G, and reconstruct the signal from the magnitude spectrums (there's a trick for that too) you get the single source out. Clever!
23:48:44 <GregorR> AnMaster_ipv6: It's a small embedded ARM-based SBC.
23:48:44 <AnMaster_ipv6> <AnMaster_ipv6> (because I have a runway process eating ram so can't run firefox too)
23:48:48 <GregorR> AnMaster_ipv6: That runs Linux.
23:49:03 <oklopol> fizzie: do i just accept that, or ask for details?
23:49:04 -!- pikhq_ has quit (Read error: 110 (Connection timed out)).
23:49:12 <fizzie> oklopol: You get to choose.
23:49:29 <oklopol> fizzie: ah the beauty of conversing
23:49:42 <GregorR> AnMaster_ipv6: It can run Debian :P
23:49:59 <GregorR> Unless it's megashit code, it should work.
23:50:14 -!- fizzie6 has joined.
23:50:31 <fizzie6> hoy, there's a ipv6-enabled freenode server in sweden now, too.
23:50:45 <fizzie6> I don't think there was, last I looked.
23:50:59 <oklopol> fizzie: get a single source out?
23:51:16 <fizzie6> the only one in Finland (orwell) is a v4 thing.
23:52:03 <fizzie> oklopol: Yes. If you happen to have something like a musical piece with two instruments; if you do the trickery I described, you get them separated. Well, in theory, anyway.
23:53:09 <fizzie6> It's sad that this v6 connection doesn't really look any different from the v4 one to other people. There should be some sort of fancy star in the nickname or something.
23:53:27 <oklopol> fizzie: sounds like something worth experimenting with
23:53:48 <fizzie6> deliberate breakage does not sensible sound.
23:53:54 -!- fizzie6 has quit (Client Quit).
23:56:03 <fizzie> I think I have some sort of whitelisting in my firewalling rules. Maybe I'll let ehird to do that sort of stuff when he gets that native-v6.