00:02:07 -!- Jafet1 has joined.
00:02:16 -!- oerjan has quit (Quit: It's the nihilist!).
00:04:12 -!- Jafet has quit (Ping timeout: 276 seconds).
00:07:39 <kmc> shachaf: lenovo cancelled my X1 carbon order due to general failure
00:08:36 <kmc> maybe i'll get the i7 / 8GB model after all
00:08:53 <kmc> i could buy it now from the UK site and have it shipped to a friend in the UK, who might appear stateside early next year
00:19:37 -!- louishenri13 has joined.
00:27:53 -!- NihilistDandy has quit (Quit: ["Textual IRC Client: www.textualapp.com"]).
00:33:21 -!- Jafet1 has changed nick to Jafet.
00:45:21 -!- nooga has quit (Ping timeout: 252 seconds).
01:11:48 -!- intok has left ("PONG :asimov.freenode.net").
01:14:05 -!- louishenri13 has quit (Quit: ~ Trillian Astra - www.trillian.im ~).
01:14:11 <coppro> kmc: "general failure"
01:14:14 <coppro> FireFly: do you know what a homestuck is
01:25:40 <Arc_Koen> they create that new almighty invincible supervillain
01:25:47 <shachaf> kmc: Does general failure work with colonel panic?
01:25:59 <Arc_Koen> and manage to shoot him dead in the same episode
01:31:21 <Phantom_Hoover> this is still the good series of stargate you're watching right
01:32:03 <shachaf> @tell elliott non is broken?
01:32:26 <shachaf> @tell elliott Do you mean that it's not a proper isomorphism?
01:33:02 <shachaf> Phantom_Hoover: I don't know. elliott's not usually so nice.
01:33:24 <shachaf> Also you're obviously pretending.
01:33:27 <Phantom_Hoover> shachaf, your a fucking piece of shit, die and kill your family
01:33:35 <shachaf> Phantom_Hoover: You have the words, but you don't have the music.
01:34:00 <shachaf> `addquote <Phantom_Hoover> shachaf, your a fucking piece of shit, die and kill your family
01:34:05 <HackEgo> 867) <Phantom_Hoover> shachaf, your a fucking piece of shit, die and kill your family
01:40:35 -!- Phantom_Hoover has quit (Remote host closed the connection).
02:16:37 -!- Jafet1 has joined.
02:18:28 -!- Jafet has quit (Ping timeout: 245 seconds).
02:34:10 <kmc> "Luftwaffe fighters found it extremely hard to shoot down the Kukuruznik because of three main factors: the rudimentary aircraft could take an enormous amount of damage and stay in the air, the pilots used the defensive tactic of flying at treetop level, and the stall speed of both the Messerschmitt Bf 109 and the Focke-Wulf Fw 190 was similar to the Soviet aircraft's maximum cruise speed, making it difficult for the newer aircraft to
02:35:21 <Bike> it was too slow, nice
02:43:16 <kmc> basically the russians used extremely shitty aircraft to harass the german ground divisions at night, ensuring that they never got any sleep and were constantly on edge
02:43:35 <kmc> which worked so well that the germans copied the idea
02:44:23 <kmc> oh and the communists used it in the korean war too
02:44:41 <kmc> the allied fighters were unable to get a radar lock because the plane is made mostly of fabric and wood
02:45:06 <Arc_Koen> "we fire the whole bullet - that's 60 percent more bullet per bullet"
02:45:57 <shachaf> kmc: /script load splitlong.pl
02:48:51 <Arc_Koen> so this week's xkcd's what-if says the amount of thrust created by a rocket/machine gun is exactly mass * speed of the thing you're firing
02:49:06 <ion> kmc: mkdir -pv ~/.irssi/scripts && ln -s . ~/.irssi/scripts/autorun && ln -s /usr/share/irssi/scripts/splitlong.pl ~/.irssi/scripts/
02:49:39 <Arc_Koen> I remember in prepschool our physics teacher told us for a rocket the "loss in weight" created by burning up the fuel also added to the thrust
02:50:31 <Arc_Koen> she was never able to explain that properly so I have no idea what that means (basically she was saying "if you're holding a heavy bag and you drop it, you'll start to fly")
02:50:42 <Arc_Koen> but apparently Randall did not mention that
02:51:06 <Fiora> I think it's because if you lose mass, it's easier to accelerate?
02:51:16 <Fiora> so as you spend your fuel, the same amount of thrust pushes you farther
02:51:38 <shachaf> We're trying to solve a practical problem with lenses
02:51:49 <shachaf> And now edwardk is going on about Kan extensions and things.
02:52:07 <Arc_Koen> yes but in that case, you don't actually *get* any thrust by *having to* carry the fuel in order to burn it
02:52:56 <Arc_Koen> I mean, she was saying "losing mass" was actually producing something (that you wouldn't have if your mass was already low)
02:53:27 <Fiora> I'm guessing she probably just wasn't being clear or you misinterpreted?
02:53:33 <Arc_Koen> but I guess I'll satisfy myself with your explanation, as it is far better than any explanation she could come up with
02:53:48 <Fiora> I know there's a similar thing with how spaceships do gravitational assists around planets
02:54:14 <Fiora> where you burn the engine really at the closest point of your orbit
02:54:37 <Fiora> and your gain in velocity comes from the fact that you basically deposited mass inside the gravitational potential well (negative energy)
02:55:12 <Fiora> http://en.wikipedia.org/wiki/Oberth_effect aha
03:00:05 -!- Arc_Koen has quit (Quit: Gateship. It's a ship. It goes through the gate!).
03:40:04 -!- KALLISTI has quit (Ping timeout: 246 seconds).
04:01:08 -!- Jafet1 has changed nick to Jafet.
04:20:53 -!- DHeadshot has joined.
04:53:03 <monqy> is this about lens again
04:53:12 <shachaf> fromMaybe (error "oh no") blah
04:53:18 <shachaf> fromMaybe (error "oh dear") blah
04:53:29 <shachaf> (error "oh no", fmap coerce ff', j)
05:02:11 -!- variable has changed nick to constant.
05:07:14 -!- ion has quit (Ping timeout: 240 seconds).
05:14:33 <kmc> this oberth effect is crazy
05:15:37 -!- ion has joined.
05:17:20 <Fiora> rockets are really cool
05:18:28 <shachaf> monqy did you do this monqy
05:18:56 <Fiora> "So if a spacecraft is on a parabolic flyby of Jupiter with a periapsis velocity of 50 km/s, and it performs a 5 km/s burn, it turns out that the final velocity change at great distance is 22.9 km/s; giving a multiplication of the burn by 4.6 times."
05:19:13 <Fiora> I wonder what kind of modification you could get with a really heavy object. like a neutron star
05:19:17 <Fiora> er, multiplication
05:27:57 -!- epicmonkey has joined.
05:31:06 <kmc> just now i confused myself greatly about potential energy
05:31:26 <kmc> because if my potential energy is proportional to my distance from heavy things
05:31:37 <kmc> shouldn't it be really huge because i am so far away from all these black holes and neutron stars and shit
05:31:50 <kmc> but i think i understand now
05:32:19 <kmc> that the best way to measure potential energy for space travel is to say that being infinitely far from everything in the universe is zero potential energy
05:32:25 <kmc> and then it goes negative if you are near a heavy thing
05:32:27 <copumpkin> isn't that just on a short distance, and an approximation to the gravitational equation
05:32:29 <kmc> and that matches what you said above
05:33:11 <kmc> ah yeah it's not linear either
05:33:16 <Bike> well usually with potential energy you define it relative to the ground, which is arbitrary/whatever makes sense for what you're doing
05:33:16 <kmc> that's true
05:33:26 <copumpkin> I think the usual mgh is an approximation that works if you're close enough
05:33:33 <copumpkin> haven't done physics in years though so don't quote me :)
05:33:47 <kmc> yeah it is
05:34:15 <kmc> i guess you can also compute it from escape velocity
05:34:44 <Bike> no wait duh, it's roughly proportional to distance but it's inverse-square at some point.
05:34:46 <kmc> the amount of kinetic energy you need in order to escape
05:34:57 <kmc> i.e. to have your position not converge as t → ∞
05:37:01 <kmc> apparently the energy from a small nuclear bomb is enough to propel me to escape the milky way galaxy
05:37:15 <kmc> if converted perfectly to kinetic energy
05:37:22 <Bike> good luck with that
05:37:29 <kmc> i'll wear my bike helmet
05:37:40 * copumpkin imagines kmc in a cannon with a nuke behind him
05:37:57 <Bike> did you hear the story of when they originally came up with that?
05:38:07 <Bike> they like, left a plate next to a nuke and then set it off, with a high speed camera
05:38:13 <shachaf> kmc: does mosh work from outside the milky way
05:38:17 <Bike> only had it in one frame, but by gum was it moving fast probably!
05:38:29 <kmc> yeah, that gives a pretty good lower bound on the speed :)
05:46:34 -!- epicmonkey has quit (Ping timeout: 240 seconds).
05:49:29 <shachaf> kmc: Guess what everyone in that other channel is doing right now. :-(
05:50:58 <kmc> which other channel
05:51:02 <kmc> also what are they doing
05:54:42 -!- Frooxius_ has quit (Ping timeout: 264 seconds).
05:58:45 -!- nortti_ has changed nick to nortti.
06:15:59 -!- Jafet has quit (Quit: Leaving.).
07:02:17 -!- nooga has joined.
07:51:50 <shachaf> monqy: http://imgur.com/a/pCdd7
07:59:04 -!- ais523 has quit.
08:14:48 <Fiora> kmc: isn't gravitational potential energy always negative?
08:15:30 <Fiora> I'm reminded of a theory stating that the total mass/energy of the universe is the same as the gravitational potential energy (except positive), I think that's what it was
08:15:39 <Fiora> meaning the total energy of the universe is zero
08:18:05 -!- epicmonkey has joined.
08:25:59 <Fiora> http://en.wikipedia.org/wiki/Zero-energy_universe ah, this thing I think
08:44:54 -!- epicmonkey has quit (Ping timeout: 240 seconds).
08:52:57 -!- carado has joined.
08:59:28 -!- carado has quit (Ping timeout: 246 seconds).
09:06:27 -!- oklo has joined.
09:09:46 -!- oklopol has quit (Ping timeout: 268 seconds).
09:12:06 -!- oerjan has joined.
09:12:51 -!- Bike has quit (Quit: leaving).
09:14:53 -!- augur_ has joined.
09:16:14 -!- augur has quit (Ping timeout: 252 seconds).
09:18:39 <ion> A Jolla live event or something. http://jolla.com/
09:22:36 -!- oerjan has quit (Quit: leaving).
09:28:53 -!- FreeFull has quit (Ping timeout: 245 seconds).
09:32:53 <Sgeo> Fiora, monqy. List list list list listlist list list list list list list list list list list EVERYBODY
09:35:04 -!- ais523 has joined.
09:36:08 <Fiora> "You can't keep down the clown."
09:37:20 <Fiora> this is a staggering number of panels dedicated to repeatedly killing gamzee
09:41:37 -!- monqy has left.
09:43:39 -!- epicmonkey has joined.
10:12:03 -!- Phantom_Hoover has joined.
10:34:06 -!- Jafet has joined.
10:57:20 -!- Phantom_Hoover has quit (Ping timeout: 255 seconds).
12:12:53 -!- oerjan has joined.
12:13:20 -!- sirdancealot has quit (Ping timeout: 252 seconds).
12:31:40 -!- sirdancealot has joined.
12:58:27 -!- ais523 has quit.
13:34:29 -!- sirdancealot has quit (*.net *.split).
13:43:50 -!- EgoBot has quit (Ping timeout: 246 seconds).
13:44:25 -!- EgoBot has joined.
13:50:14 -!- DHeadshot has quit (Ping timeout: 240 seconds).
13:50:45 -!- ogrom has joined.
13:51:27 -!- DHeadshot has joined.
13:54:03 -!- elliott has joined.
14:00:20 -!- oerjan has quit (Quit: leaving).
14:10:18 -!- boily has joined.
14:13:30 <elliott> kmc: did you know the apple retina stuff does software magic so that bitmap images get scaled up to be a reasonable size but text and stuff doesn't?
14:22:35 <fizzie> Did you know that using the Retina stuff can cause your retina to fall off?
14:30:55 -!- Phantom_Hoover has joined.
14:33:10 -!- atriq has joined.
14:41:01 -!- FreeFull has joined.
14:50:00 -!- Sanqui has quit (Ping timeout: 257 seconds).
14:50:05 -!- Sanky has joined.
14:51:57 <atriq> I'm not enjoying Legostar Galactica
14:52:04 <atriq> Should I stop reading it?
14:54:38 <atriq> I could persevere in the hope it improves
14:55:03 <atriq> I'm only up to November 2005
14:55:25 -!- ogrom has quit (Quit: Left).
15:03:00 <atriq> I enjoy El Goonish Shive now a lot more than I did a couple of years ago
15:06:09 -!- Frooxius has joined.
15:14:16 -!- Arc_Koen has joined.
15:16:14 -!- atriq has quit (Quit: Page closed).
15:20:15 -!- sirdancealot has joined.
15:53:43 -!- kashifpak has joined.
15:54:23 <HackEgo> kashifpak: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page. (For the other kind of esoterica, try #esoteric on irc.dal.net.)
15:54:51 -!- kashifpak has left.
15:58:35 <kmc> elliott: yeah
15:58:40 -!- Frooxius has quit (Read error: Connection reset by peer).
15:59:38 <kmc> it's like they render to a virtual screen with a lower resolution
16:00:03 <kmc> but that rendering is a combination of raster and vector elements
16:00:06 <shachaf> elliott: Did you know that GHC's Derive{Functor,Foldable,Traversable} has a bug that makes them quadratic?
16:00:17 <shachaf> http://hackage.haskell.org/trac/ghc/ticket/7436
16:00:32 <elliott> now I just have to wait for a company that I feel happy (well, happier) giving money to to copy it
16:00:48 <elliott> I guess you can do the same with Linux already sort of
16:01:24 <elliott> set your dpi in X high enough (and hacks to fix stuff that ignore that) that fonts are big, get your browser to scale images up, and use a gtk theme without too many bitmap elements
16:01:42 -!- Frooxius has joined.
16:01:56 <elliott> shachaf: "An eta-expanded definition like foldr becomes asymptotically worse for some reason. Maybe this is expected behavior for this function, since f gets eta-expanded at each iteration?"
16:02:01 <elliott> this is pretty much semantically required
16:02:12 <shachaf> elliott: THEN WHY DOES SPJ FIND IT SO SURPRISING
16:02:19 <elliott> since eta-retraction is invalid in Haskell
16:02:35 <shachaf> elliott: Doesn't stop GHC from doing it.
16:02:59 <elliott> yes GHC has invalid optimisations and this is bad
16:03:00 -!- copumpkin has quit (Quit: Computer has gone to sleep.).
16:03:18 <elliott> probably GHC should compile this example efficiently because it does such optimisations
16:03:37 <shachaf> Oddly enough, GHC's invalid optimizations tend to eta-expand, not eta-contract
16:03:50 <shachaf> Well, the only time I've ever bothered to catch them
16:04:24 <shachaf> 08:01 <edwardk> i think its because ghc won't inline a function unless it has been given all of its arguments
16:04:27 <shachaf> 08:01 <edwardk> so they tend to eta expand like mad in hope they'll give it all its args
16:05:24 <elliott> if only we had edwardk's perfect proprietary Haskell compiler
16:06:43 <shachaf> elliott: Anyway if you've ever used DeriveFoo, I hope you're glad that your code is slow. :-(
16:07:05 <kmc> shachaf: did I ever tell you about my shell / python coroutine idea?
16:07:50 <elliott> IMO kmc should write a Haskell compiler that doesn't suck
16:08:10 <kmc> i started writing a haskell compiler but "doesn't suck" was not one of the design goals
16:08:10 <shachaf> elliott: imo "its called thc havnet u heard"
16:08:26 <elliott> what were the design goals
16:08:36 <kmc> shachaf: so python would suck as an interactive command shell, but it's superior to bash for certain data munging tasks
16:09:00 <kmc> so i want each interactive bash to have an associated python interpreter in the background
16:09:04 <kmc> so that i can do like
16:09:08 <kmc> $ cmd1 | cmd2 | py -i foo
16:09:35 <kmc> >>> bar = ' - '.join(x.strip().split(':')[3] for x in foo)
16:09:44 <kmc> $ py -o bar | cmd3 | ...
16:10:10 <kmc> i think this is not too hard to implement so, has it been done, and is it crazy?
16:10:17 <shachaf> Is there really such an advantage over just running a command?
16:10:24 <shachaf> Well, other than "Python doesn't support one-liners"
16:10:39 <Jafet> I wonder when people will realize that using bash as a standard language has been a terrible mistake
16:10:53 <kmc> shachaf: yeah there's that
16:11:01 <kmc> although the same property makes interactive python kind of limited as well
16:11:16 <elliott> if you are goin gto do that might as well use a nicer interface than bash for the shell part too :P
16:11:31 <kmc> maybe it's almost as good to do python -c' <newline> ... python script
16:11:32 <elliott> also have you seen ipython
16:11:34 <elliott> iirc it does similar things to this
16:11:38 <kmc> i use ipython all the time
16:11:41 <elliott> as in i think ls | cat ... works in ipython
16:11:48 <elliott> ignore the fact that that pipeline makes no sense
16:11:54 <shachaf> "imo this is reason enough to switch to ruby"
16:11:54 <elliott> ls | cat - /dev/random | ... makes sense
16:12:10 <kmc> i don't think it implements full shell syntax though
16:12:22 <kmc> that would be kind of crazy
16:12:31 <kmc> <elliott> if you are goin gto do that might as well use a nicer interface than bash for the shell part too
16:12:38 <kmc> this is that #haskell fallacy that i hate
16:12:53 <kmc> "if you're going to fix a bug in any program you might as well rewrite that program in haskell and also your operating system"
16:13:31 <kmc> if you're going to try to convert me to zsh or whatever then fine
16:13:31 <elliott> well my point is that this is a relatively huge engineering effort, so basing it on bash would be a shame when there are things like fish (fish isn't perfect but I think fixing the flaws would probably be less effort in total esp. since I bet bash's codebase is awful)
16:13:50 <kmc> no i don't think this would require extensive changes to bash
16:13:52 <elliott> well zsh is like bash but with marginally nicer interactive features :P
16:14:02 <elliott> oh you are doing "py -o bar"
16:14:06 <elliott> rather than just "bar | ..."
16:14:19 <kmc> in fact i would hope to make it not depend on the shell you use
16:14:19 <kmc> or only minimally
16:14:24 <shachaf> What about python 'blah' < named pipe
16:14:42 <kmc> i would have a bashrc line that starts up the python process and puts in an env var a path to some socket or something where it can be talked to
16:14:46 <elliott> i am not sure that the py -i / py -o model is bset here
16:14:59 <elliott> it seems better if you instead have somethin glike py -a <function>
16:14:59 <kmc> and then 'py' would use that
16:15:01 <elliott> which runs function with the input
16:15:06 <kmc> yeah i was thinking about that too
16:15:08 <elliott> so you would do cmd1 | cmd2 | py -a foo | cmd3 | ... instead
16:15:20 <elliott> by defining bar = lambda foo: ' - '.join(x.strip().split(':')[3] for x in foo)
16:15:21 <kmc> but there are def. cases where you want something more complicated than that
16:15:24 <shachaf> perl has -e. Also oneliners.
16:15:27 <elliott> but one-line python is painful a lot
16:15:32 <elliott> yes i think perl is better for this
16:15:39 <kmc> with perl i think it's not really needed
16:15:47 <shachaf> "this is why i use ruby more than python"
16:16:02 <shachaf> kmc: Clearly the answer is to use node.js with asynchronous callback functions instead of coroutines.
16:16:12 <shachaf> "coroutines? more like slowroutines"
16:16:21 <Jafet> "My init scripts are webscale"
16:16:24 -!- ogrom has joined.
16:17:37 <shachaf> > set (every 5) 'Q' "this is just some text to say hi monqy with to elliott"
16:17:39 <lambdabot> "Qhis Qs juQt soQe teQt toQsay Qi moQqy wQth tQ ellQott"
16:18:56 <lambdabot> (Integral a1, Applicative f, TraversableWithIndex a1 t, Indexed a1 k) => a1 -> k (a -> f a) (t a -> f (t a))
16:19:16 <elliott> > every 5 ~. 'Q' $ "is this right"
16:19:22 <elliott> > every 5 .~ 'Q' $ "is this right"
16:19:30 <elliott> > every 0 .~ 'Q' $ "is this right"
16:19:46 <shachaf> the zero police is coming to take you away
16:19:53 <kmc> in particular this python shell would be persistent
16:20:02 <elliott> > every (-1) .~ 'Q' $ "is this right"
16:20:06 <kmc> you can jump back and forth and keep its state
16:20:14 <elliott> > every (-2) .~ 'Q' $ "is this right"
16:20:18 <elliott> does it use abs or something
16:20:20 <kmc> which i think makes it a lot more powerful than python -c, perl -e, etc.
16:20:37 <shachaf> every k = iwhere (\i -> i `mod` k == 0)
16:21:08 <elliott> > iwhere (== 'q') .~ 'Q' $ "abcdqe"
16:21:10 <lambdabot> Couldn't match type `GHC.Types.Int' with `GHC.Types.Char'
16:21:16 <elliott> > where (== 'q') .~ 'Q' $ "abcdqe"
16:21:19 <lambdabot> <hint>:1:1: parse error on input `where'
16:21:24 <lambdabot> (Applicative f, TraversableWithIndex i t, Indexed i k) => (i -> Bool) -> k (a -> f a) (t a -> f (t a))
16:25:32 <elliott> what is the equivalent of iwhere for values
16:26:06 <shachaf> Breaks the traversal laws unless you use it carefully.
16:27:07 <elliott> what's wrong with filtered (== 'q') .~ 'Q'
16:27:21 <shachaf> over f . over g = over (f.g)
16:28:11 <shachaf> > over (filtered even) (+1) . over (filtered even) (+1) $ 2
16:28:12 <shachaf> > over (filtered even) ((+1) . (+1)) $ 2
16:28:30 <lambdabot> Setting s t a b -> (a -> b) -> s -> t
16:28:49 <ion> stabOf eye
16:31:03 <elliott> shachaf: ugh, is cheater talking a lot in #haskell right now
16:31:18 <shachaf> elliott: "how would i know"
16:31:44 <elliott> well you just said something in #haskell
16:31:48 <elliott> so presumably you are paying some attention to it
16:32:01 <ion> The alot is talking cheater in #haskell
16:32:25 <shachaf> Well, I've had him on /ignore for a while now.
16:32:31 <shachaf> I think he's the only person I've ever /ignored
16:32:54 <shachaf> I wonder whether he just hates me personally or what.
16:33:35 -!- sirdancealot has quit (Read error: Operation timed out).
16:36:01 -!- Vorpal has joined.
16:36:28 -!- ogrom has quit (Ping timeout: 255 seconds).
16:47:17 <elliott> shachaf: what's the difference
16:47:26 <elliott> wait why does Getting have another parameter there
16:52:17 -!- sirdancealot has joined.
16:59:22 <ion> type Getting r s t a b = (a -> Accessor r b) -> s -> Accessor r t
16:59:26 <ion> Accessor ~ Const
17:00:34 -!- augur_ has quit (Remote host closed the connection).
17:00:51 <ion> > ("foo", "bar") ^. both
17:00:55 <ion> > ("foo", "bar") ^.. both
17:01:02 -!- augur has joined.
17:01:21 <ion> > (Sum 4, Sum 5) ^. both
17:01:27 <ion> > (4, 5) ^.. both
17:05:31 -!- augur has quit (Ping timeout: 260 seconds).
17:18:40 -!- Bike has joined.
17:22:55 <lambdabot> Source not found. This mission is too important for me to allow you to jeopardize it.
17:24:26 <elliott> hah, expecting @src to be reasonable
17:24:37 <elliott> (it's both f (a,b) = (,) <$> f a <*> f b)
17:25:55 <ion> No, both f (a,b)
17:26:14 <ion> like traverse f [a,b]
17:26:31 <lambdabot> Applicative f => (a -> f b) -> (a, a) -> f (b, b)
17:34:46 -!- augur has joined.
17:48:42 <Vorpal> hm... how do I make git restore the working tree to a pristine state. git reset --hard seems to keep ignored files, but I want those gone too
17:49:08 <Vorpal> Deewiant, .... which is not even listed in "git help"
17:49:22 <Fiora> git clean -x I think
17:49:33 <Deewiant> git help lists only common commands
17:49:35 <Deewiant> "The most commonly used git commands are:"
17:50:03 <Vorpal> so I did git clean -x -f aaand wait for it:
17:50:04 <Vorpal> Not removing arch/arm/include/generated/
17:50:25 <Vorpal> I wanted a *pristine* checkout
17:50:42 <Deewiant> -d says "remove untracked directories in addition to untracked files"
17:51:11 <ion> I wonder if any IRC client follows the RTL override codepoint?
17:51:41 -!- Nisstyre has quit (Remote host closed the connection).
17:52:01 <elliott> and other things that use external stuff to render text
17:53:59 <kmc> yeah i think i have seen it in xchat
17:54:08 <kmc> not just override but following the bidi algorithm
17:54:15 <kmc> so if you write arabic it will show up RTL
17:55:14 <Vorpal> kmc, xchat behaves strangely if you try to select RTL text though
17:55:28 -!- Phantom_Hoover has quit (Read error: Operation timed out).
17:55:37 <Vorpal> as in, the text reverts to LTR in the selection iirc
17:59:04 <Vorpal> well it did last I checked at least, which was probably over a year ago by now
17:59:19 <elliott> xchat behaves really weirdly with proportional fonts also
18:00:34 -!- epicmonkey has quit (Ping timeout: 246 seconds).
18:01:26 -!- Nisstyre has joined.
18:03:02 <Vorpal> elliott, I never even tried that
18:03:14 <Vorpal> but out of interest, what does it do when you use them?
18:04:02 -!- carado has joined.
18:04:46 <elliott> but if you select text it redraws often
18:04:48 <elliott> and that sometimes looks weird
18:09:14 -!- DHeadshot has quit (Ping timeout: 240 seconds).
18:11:21 -!- DHeadshot has joined.
18:16:14 <lambdabot> Setting s t a b -> (a -> b) -> s -> t
18:37:41 <lambdabot> Ambiguous type variable `b0' in the constraints:
18:37:49 <lambdabot> Ambiguous type variable `b0' in the constraints:
18:37:59 <elliott> adding .0 never helps ambiguity
18:38:13 <ion> > (-1)**0.5 :: Complex CReal
18:39:16 <FreeFull> A graph of the real part of (-1)**x is a cosine wave periodic in 2, and of the imaginary part is a sine wave periodic in 2
18:40:33 <fizzie> It goes around and around and around and around.
18:40:41 <ion> > cycle "and around "
18:40:43 <lambdabot> "and around and around and around and around and around and around and arou...
18:41:20 -!- augur has quit (Remote host closed the connection).
18:43:01 -!- augur has joined.
18:50:26 <fizzie> Not for *that* long, I'm sure.
18:52:53 <elliott> > cycle "and around " & every 2 .~ "q"
18:53:07 <elliott> > cycle "and around " & every 2 .~ "q"
18:53:08 <lambdabot> Couldn't match expected type `[GHC.Types.Char]'
18:53:33 <elliott> > cycle "and around " % every 2 .~ "q"
18:53:49 <elliott> shachaf: oh, it's not standard?
18:55:54 -!- DHeadshot has quit (Ping timeout: 240 seconds).
18:58:40 -!- carado has quit (Ping timeout: 246 seconds).
19:00:02 -!- DHeadshot has joined.
19:00:43 -!- carado has joined.
19:05:36 -!- epicmonkey has joined.
19:08:40 -!- monqy has joined.
19:18:47 <ion> http://literallyunbelievable.org/post/36209732360/can-you-beat-this-a-facebook-user-has-uploaded
19:20:58 <olsner> hmm, 12 million photos over 6 days, that's 2 million a day, and a day has less than 100000 seconds, so that's about 20-30 fps video over the entire trip
19:21:32 <olsner> I think that should take a lot less space than 15 TB
19:22:16 <pikhq_> Motion JPEG is not very efficient.
19:22:20 <Fiora> http://www.theonion.com/articles/area-woman-finally-uploads-all-12-million-pictures,30443/ ^^;;
19:22:26 <Fiora> it's an onion piece
19:23:13 <pikhq_> 15TB is faintly plausible for high quality motion JPEG though.
19:24:22 <olsner> but if you videotaped your entire vacation, would you really use motion jpeg?
19:24:37 <FreeFull> literallyunbelievable.org catalogues people's responses to theonion.com pieces
19:24:50 <Gregor> So does #esoteric , apparently.
19:24:55 <FreeFull> Surprising how many people think they're real
19:25:18 <olsner> aren't we responding to the literallyunbelievable piece though?
19:26:13 <olsner> for one, I think we've concluded that it's definitely not literally unbelievable to take 12 million pictures of your vacation
19:26:20 <Sgeo> "Stories from The Onion as interpreted by Facebook"
19:26:31 <Sgeo> I did not know that India Today was a part of Facebook.
19:26:44 <olsner> ... as interpreted by Literally Unbelievable as interpreted by #esoteric
19:28:07 <Sgeo> http://literallyunbelievable.org/post/35986654788/wow-osama-os-still-alive
19:30:26 <Sgeo> http://literallyunbelievable.org/post/35773802529/r-we-for-real ... I think ... The Onion actually taught people something?
19:31:18 <ion> “India Today is a part of Facebook!”
19:31:23 <fizzie> olsner: Individually captioning those 12 million pictures in 8 months does mean you couldn't spend more than about 1.7 seconds per caption, assuming you spent 24 hours per day captioning. (Or 1.15 seconds, assuming a more reasonable 16 hours/day schedule.)
19:32:23 <ion> sgeo: The war one :-D
19:32:40 <olsner> oh, they were individually captioned? that's literally unbelievable!
19:33:25 <Bike> well the captions were like "cute!!!" according to the article
19:33:29 <Bike> gotta use all the information here
19:37:27 <kmc> elliott: btw ls | cat can make sense
19:37:48 <kmc> it disables colors and gives you one file per line
19:37:54 <kmc> which is useful for copy-pasting
19:38:02 <kmc> of course there's some ls flag for that
19:38:24 <kmc> i could remember that, or i could use the tools i already know
19:38:58 <kmc> of course forcing a command's output to be not a tty is useful for things other than ls
19:39:02 <FireFly> ls | xclip -i is even more useful for copy-pasting
19:47:47 <fizzie> cat | something | cat # double action
19:49:21 <kmc> all the way
19:49:43 <kmc> i also write "cat foo | bar" rather than "bar < foo" to preserve the left to right flow of data
19:49:55 <kmc> of course i've been told this is proof that my unix penis is very small
19:50:10 <kmc> think of all the resources wasted by that extra cat process
19:50:16 <kmc> that could have gone towards feeding the hungry
19:50:28 <kmc> you can also write "< foo bar" but that's just fucked up
19:51:25 <Nisstyre> kmc: you could be running folding@home, those extra cpu cycles could go towards curing cancer
19:52:11 <Vorpal> hm is cpio just a storage format like tar or is compression built in?
19:54:38 -!- atriq has joined.
19:55:23 <Vorpal> wait a second... this makes no sense. Flashing this zip depends on the order the files are put in to the zip file. I.e. zip foo.zip a b or zip foo.zip b a
19:56:16 <olsner> yeah, how could anyone make software depend on irrelevant details like the order of input files
19:56:48 <Vorpal> olsner, it depends on the order of which the files are stored in the sodding zip file
19:56:56 <Vorpal> which doesn't make much sense to me
19:57:04 <Vorpal> the update script has to go first.
19:57:54 <olsner> although it may look like a zip file to you, you're probably looking at a very specific format for delivery of software updates :P
19:58:37 <olsner> kind of like how DOS' bootup files have to be the first files on the file system for the MBR to find them
19:59:16 <Vorpal> olsner, well, I would expect a custom open source recovery on an android phone to not care about the order though
19:59:26 <olsner> fizzie: which iirc only works if you don't already have files on C:
20:01:41 <kmc> Nisstyre: hmm true
20:01:52 <fizzie> olsner: It's more complicated than that. http://support.microsoft.com/kb/66530 has the detailed requirements.
20:01:56 <kmc> would actually make me feel really guilty
20:02:48 <fizzie> MS-DOS 4.0 and later seems to have a SYS.COM capable of rearranging things.
20:02:51 <Vorpal> <kmc> you can also write "< foo bar" but that's just fucked up <-- you know, I have actually seen that in serious code
20:04:02 <fizzie> And the at-the-start requirements also seem slightly relaxed.
20:04:53 <olsner> hmm, if 4.0 and up can fix that automagically, I wonder why I'm familiar with the problem at all
20:07:35 <fizzie> I remember it breaking my Windows installation.
20:07:53 -!- sirdancealot has quit (Read error: Operation timed out).
20:08:02 <Vorpal> I grew up on mac btw, so I remember ResEdit and such instead
20:08:05 <fizzie> There's some kind of an issue w.r.t. wfw3.11 and a dblspace'd c: drive.
20:08:40 <fizzie> It's a possible setup, but there was a way to make it not work, and I hit on that.
20:09:12 <fizzie> Still, free (modulo speed) disk space was nothing to scoff at.
20:10:45 <fizzie> HLFSPACE stores all data on disk twice, for redundancy. (Not really.)
20:13:03 <Vorpal> okay, now I got it building a kernel at amm
20:13:57 <Vorpal> now to check if I can actually make it do what I want though
20:14:02 <fizzie> What are you flashing there?
20:15:47 <Vorpal> fizzie, I'm trying to build a custom kernel for my phone with ftrace support. Since some tools in the SDK need that (systrace specifically) but the stock samsung kernel lacks it. Also the Siyah kernel (which is rather nice, offers dual booting for example) lacks it. Had some issues building a modified version of that kernel though
20:16:01 <Vorpal> seems to be working now, but now to try with ftrace enabled as well
20:16:33 <Vorpal> first issue was getting the initramfs building correctly
20:17:07 <Vorpal> you would think that the automated build script would actually work. Given the errors in it I'm surprised it worked for anybody
20:17:28 <fizzie> Hey, now that there's perhaps a different subset of people awake... why doesn't "dropbox puburl" work for me? Is it supposed to work at all? (I mean, it "works" in the sense that it prints URLs for ~/Dropbox/Public/ stuff, but those URLs just give a 404 when accessed.)
20:17:48 <Vorpal> yeah it is impossible it would have worked with bash
20:18:00 <Bike> fizzie: it works fine for me.
20:18:12 <fizzie> Bike: I don't know what I'm doing worng, then.
20:18:36 <Bike> what's "dropbox status"?
20:18:58 <Vorpal> fizzie, I just get this:
20:18:59 <Vorpal> Another instance of Dropbox (25520) is running!
20:19:27 <Vorpal> $ file ~/.dropbox-dist/dropbox
20:19:27 <Vorpal> /home/arvid/.dropbox-dist/dropbox: ELF 64-bit LSB executable, x86-64, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.8, stripped
20:19:30 <fizzie> Bike: http://sprunge.us/hYEW
20:19:37 <Vorpal> but if I hit ctrl-c I get a python traceback
20:19:40 <Vorpal> how the hell does that work
20:20:07 -!- oerjan has joined.
20:20:14 <fizzie> Vorpal: file $(which dropbox) instead?
20:20:21 <Vorpal> fizzie, it is not in my path?
20:20:33 <Bike> oh, I just use the python script
20:20:35 <kmc> maybe it execs a python script or embeds a python interpreter
20:20:45 <Vorpal> fizzie, I don't know, I just followed the install instructions way back when, when I installed it
20:21:00 <Vorpal> I don't think my distro has it anyway
20:21:04 <Vorpal> so it is just from dropbox.com
20:21:18 <Vorpal> elliott, yeah sure, why not? I got 50 GB free for 2 years with my phone
20:21:27 <Vorpal> it is excellent for *encrypted* backups
20:22:06 <fizzie> I suppose it runs some Python; I do get Python stuff for ^C if I try to run that binary manually.
20:22:21 <Vorpal> elliott, why are you surprised I use dropbox? The backup program I use on my phone has built in support for syncing with it, so it is convenient.
20:22:43 <Vorpal> fizzie, where is the dropbox in your path from then?
20:23:17 <elliott> i'd expect some rant about relying on an external corporation for your data and closed source and how using scp to your own server is just as easy. :p
20:23:20 <fizzie> Vorpal: I went the package manager way.
20:23:57 <Vorpal> speaking of which I need to upgrade this laptop from 10.04 at some point, kind of soon
20:24:04 <Vorpal> iirc support ends at the end of the year
20:24:30 <fizzie> I only have dropbox because of that Space Race thing, but I was thinking I could use it for stupid-images-to-share-on-IRC use; it's just that the puburl just isn't working. Doing a "share link" thing (via the webface) did work, but it makes a different sort of thing.
20:24:31 <lambdabot> `traverseBits' (imported from Data.Bits.Lens),
20:24:38 <Vorpal> elliott, nah, I wouldn't put unencrypted backups there though. Unless I wanted to make them public
20:24:56 <elliott> :t (1,2) % fromWithin traverse
20:24:58 <lambdabot> Couldn't match expected type `(t0, t1)' with actual type `a0 :> b0'
20:24:58 <lambdabot> Actual type: (a0 :> b0) -> (a0 :> b0) :> c0
20:24:59 <Vorpal> also I really need more RAM in this laptop, building a kernel makes the input lag unless I close firefox
20:25:04 <elliott> Vorpal: maybe you have changed :p
20:25:05 <elliott> :t (1,2) % fromWithin both
20:25:07 <lambdabot> Couldn't match expected type `(t0, t1)' with actual type `a0 :> b0'
20:25:08 <lambdabot> Actual type: (a0 :> b0) -> (a0 :> b0) :> c0
20:25:19 <lambdabot> SimpleLensLike (Control.Lens.Internal.Bazaar c c) b c -> (a :> b) -> (a :> b) :> c
20:25:22 <fizzie> I tried a lazy google for dropbox puburl 404 but didn't see too many complaints.
20:25:33 <elliott> oerjan: have you learned about lens yet
20:25:35 <elliott> oerjan: it is super fancy!
20:25:45 <fizzie> "Please note: New Dropbox accounts created after October 4, 2012 no longer have a Public folder. Don't worry! Every account created prior to this date will still have a Public folder. If you would like to enable a Public folder on a new account, see the instructions below (Creating a Public folder).
20:25:48 <oerjan> not any concrete implementation, no
20:25:51 <fizzie> However, all the extra functionality provided by the Public folder is now accessible anywhere in your Dropbox. Now all you need to do to share and preview files and folders in your Dropbox is select Share link via your computers, phones, and tablets."
20:25:56 <elliott> oerjan: no i mean the "lens" package
20:25:58 <Vorpal> elliott, I doubt I would use google drive btw, I try to not put everything in the hands of one company
20:25:59 <elliott> which is what i am using here
20:26:02 <fizzie> I suppose my ~/Dropbox/Public isn't actually a real public folder.
20:26:05 <Vorpal> google already knows way too much
20:26:08 <Bike> fizzie: welp, glad I got in before that
20:26:24 <fizzie> Bike: It can still be enabled, there's instructions.
20:26:26 <oerjan> i've heard about that van lairwoofer thing
20:26:26 <elliott> oerjan: it has lenses (incl. ones which can change the type i.e. polymorphic fields) and traversals (hence partial lenses) and effectful versions of all of this and indexed stuff and generic zippers and
20:26:38 <elliott> it is edwardk's new megapackage
20:26:43 <fizzie> "Dropbox links give you everything you need to share and preview files and folders. However, if you’re just a diehard fan of Public Folders, click here to enable one on your account."
20:26:47 <Bike> fizzie: yeah but that sounds pretty hard for something I indeed mostly use for stupid pictures for irc
20:26:56 <fizzie> It's hard to click on a link?
20:27:05 <elliott> has some nice examples like (zipper ("hello","world") % down _1 % fromWithin traverse % focus .~ 'J' % rightmost % focus .~ 'y' % rezip)
20:27:05 <Vorpal> <fizzie> However, all the extra functionality provided by the Public folder is now accessible anywhere in your Dropbox. Now all you need to do to share and preview files and folders in your Dropbox is select Share link via your computers, phones, and tablets." <-- wait, can you see which folders are public and so on then?
20:27:16 <Vorpal> or do stuff just automatically become public?
20:27:31 <fizzie> I'm not sure I'm a diehard fan of public folders, but I am a fan of getting links from the command-line client as opposed to navigating the webs.
20:27:45 <fizzie> (I don't think the command-line thing can create "shared links".)
20:28:00 <fizzie> You need to "Share link" specific files for them to be public.
20:28:08 <fizzie> But those files don't need to be in any specific folders.
20:28:13 <Vorpal> fizzie, can you revoke the publicness of them?
20:28:24 <Vorpal> also can you make an entire folder, with a directory listing, public?
20:28:28 <fizzie> I suppose you can delete the links.
20:28:33 <fizzie> They appear separately in the webface.
20:28:47 <fizzie> And there was something about sharing an entire folder that I hit on the last Google.
20:29:01 <fizzie> Though it might've been just a note that it's not possible, plus third-party things to make it happen.
20:29:26 <fizzie> Okay, I clicked on the "enable public folder" link, and got "Error: Something went wrong. Don't worry, your files are still safe and the Dropboxers have been notified. Check out our Help Center and forums for help, or head back to home."
20:29:36 <fizzie> Perhaps just because I wasn't signed in.
20:29:42 <Vorpal> so hm, the ftrace option breaks the kernel, oh well
20:30:19 <Vorpal> wow, /proc/last_kmsg is binary garbage
20:30:26 <fizzie> "Your Public folder is now enabled. Your existing Public folder was renamed to "Public (old)"."
20:30:53 <oerjan> <elliott> it is edwardk's new megapackage <-- did you see his recent reddit comment about how he found haskell? with him 2006 seems so short ago...
20:31:19 <atriq> That was about the same time I found the joys of the internet
20:31:39 <elliott> oerjan: iirc yes, I think I knew that date already
20:31:44 <elliott> oerjan: I think Hackage is actually post-2006 though!
20:31:56 <atriq> Vorpal, I knew it existed beforehand
20:32:06 <fizzie> Yay, now the puburl works. (Except that of course the .jpg isn't EXIF-autorotated, unlike it was when I "Share link"ed it. (It has a previewy kind of a thing.))
20:32:10 <atriq> But that would have been when I joined the UMMF
20:32:15 <elliott> oerjan: I think Hackage is ~2007
20:32:16 <atriq> I got linked IWC from there
20:32:22 <elliott> oh, looks like late 2006 maybe
20:32:24 <atriq> And thence to everywhere else
20:32:32 <elliott> IIRC edwardk has said Hackage came around right as he was about to give up on Haskell
20:32:45 <atriq> Unofficial Murderous Maths Forum
20:32:56 <Vorpal> atriq, so you used internet before 2006?
20:33:32 <atriq> Take in mind, I was 11, 12?
20:33:33 <Vorpal> did you used it back in 96 though?
20:33:56 <atriq> The earliest I remember using a computer would have been 98, I think
20:34:11 <Vorpal> I remember the 28 kbit modem my dad got in either late 95 or early 96
20:34:19 <fizzie> I don't think our university is going to be getting that 25 GB. :/ :/ :\ (Unless the Space Race is going to continue for a long long time.)
20:34:47 <Vorpal> <fizzie> Yay, now the puburl works. (Except that of course the .jpg isn't EXIF-autorotated, unlike it was when I "Share link"ed it. (It has a previewy kind of a thing.)) <-- that is because your browser fails
20:35:40 <Vorpal> that would be pretty cool
20:35:41 <fizzie> Vorpal: 25 GB of dropbox space.
20:35:56 <fizzie> https://www.dropbox.com/spacerace -- that thing.
20:36:03 <fizzie> It's a kind of a University-specific thing.
20:36:12 <fizzie> At least we're at the top of the Finland Leaderboard.
20:36:25 <fizzie> And we were on the global top 10 for like several hours!
20:36:33 <Vorpal> fizzie, what is the cost for 50 GB btw, as a private customer? It doesn't show it to me for some reason, just the 100 GB level
20:36:56 <fizzie> Now it's all MIT and National University of Singapore and so on, they have a slight advantage in the number of students.
20:37:00 <Vorpal> maybe it is because I already get 50 GB from that promo that came with my phone
20:37:07 <Vorpal> or did they remove the 50 GB level?
20:37:26 <fizzie> I'm trying to figure out where you can see those anyway.
20:37:48 <fizzie> Doesn't show anything else than 100/200/500 for me either.
20:38:03 <elliott> Open University is at #10 in the UK. That's a bit silly.
20:39:19 <fizzie> It's a bit disingenuous that the "Get free space!" link leads to a page where you can pay for space.
20:39:28 <fizzie> Though I suppose it's still more "free space" as in empty.
20:40:08 <elliott> oerjan: surely types like "type IndexedTraversal i s t a b = forall f k. (Indexed i k, Applicative f) => k (a -> f b) (s -> f t)" must tempt you
20:40:16 <fizzie> I could get a total of 3*125M for "connecting" Facebook and Twitter accounts and following Dropbox on Twitter.
20:40:20 <fizzie> Doesn't quite seem worth it.
20:40:33 <oerjan> <Phantom_Hoover> why does everyone leave the second we welcome them <-- with all the people who have left that way you'd think there should be enough for a big other kind of esoteric channel somewhere...
20:41:00 <oerjan> (last i asked the dalnet one was tiny)
20:41:39 <oerjan> elliott: tempt me to run away screaming, you mean?
20:45:07 <kmc> did you all see http://blog.burtonthird.com/?p=81
20:45:13 <elliott> oerjan: but it's cool!! You can compose lenses with just function (.)
20:45:21 <kmc> i don't think this is really interesting enough to merit such a detailed writeup, but it's kinda amusing
20:45:37 <elliott> oerjan: and do things like both f (a,b) = (,) <$> f a <*> f b
20:45:48 <elliott> > ("abc", "def") % both .~ "q"
20:47:03 <elliott> "First things first, we needed a way to automatically generate lots of email addresses. They didn’t actually need to be MIT email addresses for us to earn points, but what fun would it be if they weren’t? We’re fortunate that MIT keeps a relatively open network, so students are able to create their own @mit.edu mailing lists without any approval."
20:47:06 <elliott> what could possibly go wrong
20:47:15 <fizzie> Vorpal: Speaking of 25 Gbps network connections, I did hear a rumour that there was some kind of a project to start expanding the 10Gbps backbone further inside the computer science building. The 1Gbps switches are sometimes a bit overloaded when e.g. someone starts a 150-task Condor job and all those tasks attempt to fetch the same file over NFS.
20:47:37 <elliott> > ("abc", "def") % both %~ reverse
20:47:39 <oerjan> ...i guess you don't really need that Data.Ratio operator
20:47:49 <elliott> oerjan: it's being renamed to (&) because of complaints about the overlap
20:47:53 -!- augur has quit (Remote host closed the connection).
20:47:59 <fizzie> Man, these sysadm group meeting minutes are always so depressing. "Triton's MATLAB is broken. Triton's disk system needs more memory."
20:48:01 <elliott> (but (/) is a thing so I don't really like Data.Ratio's (%) anyway)
20:48:24 <lambdabot> Prelude (&&) :: Bool -> Bool -> Bool
20:48:24 <lambdabot> Data.Bool (&&) :: Bool -> Bool -> Bool
20:48:24 <lambdabot> Control.Arrow (&&&) :: Arrow a => a b c -> a b c' -> a b (c, c')
20:48:28 <atriq> There was a bunch advocating (|>)
20:48:36 <atriq> But that overlaps with Data.Sequence
20:49:02 <Vorpal> fizzie, also couldn't you use multicast to deal with that
20:49:19 <olsner> nfs probably doesn't do multicast
20:49:44 <fizzie> Vorpal: Condor is this "use idle time of regular workstations as a computation grid" queue/etc. system.
20:49:46 <Vorpal> olsner, well sure, but couldn't you replace it?
20:49:56 <Vorpal> fizzie, ah nice, is it open source?
20:49:56 <elliott> "The next issue was that we asked MIT’s network administrators how we could programmatically create lists, and they told us we couldn’t. Making 30,000 mailing lists the next day probably wouldn’t look too good, so we came up with an alternative."
20:50:22 -!- augur has joined.
20:50:33 <fizzie> Vorpal: I believe it is, yes. We've got I think around 140 workstations, most of them quad-core Xeons, and there's usually quite a lot of idle capacity there.
20:50:40 <Vorpal> elliott, wait, I want to read the context to that, where is it from?
20:51:00 <fizzie> "HTCondor was formerly known as Condor; the name was changed in October 2012 to resolve a trademark lawsuit."
20:51:25 <elliott> Vorpal: http://blog.burtonthird.com/?p=81
20:51:37 <elliott> which kmc linked above, many times the size of your scrollback ago
20:52:09 <Vorpal> elliott, yeah probably, I'm using a small laptop atm
20:52:50 -!- augur has quit (Remote host closed the connection).
20:52:57 <fizzie> "Dropbox’s founder and MIT Alum, Drew Houston, tried to lessen our emotional damage by creating a “United States Leaderboard” where we still held the #1 position, but the damage was already done."
20:53:05 <fizzie> I like to think they also thought of us while doing that.
20:53:09 <fizzie> (They probably didn't.)
20:53:12 <elliott> kmc: I was hoping for some kind of story after the tech details. :/
20:57:13 -!- Phantom__Hoover has joined.
20:59:25 <Vorpal> elliott, some interesting scams there
20:59:45 <Vorpal> hopefully the hack was reverted
21:02:09 <elliott> sounds more or less impossible to revert
21:02:15 <elliott> they deserve it for initiative anyway
21:12:36 <elliott> Phantom__Hoover: rocks fell & everybody died
21:13:06 <atriq> Scam MIT pulled off, something to do with Dropbox
21:13:51 <elliott> atriq: that was not terribly helpful :P
21:14:06 <atriq> Although technically true
21:14:07 -!- epicmonkey has quit (Ping timeout: 246 seconds).
21:14:22 <elliott> no i mean the #haskell thing
21:14:45 <atriq> One problem at a time
21:15:09 <atriq> It fixed his problem, but didn't answer why he had a problem
21:15:11 <olsner> oh, are people still in #haskell?
21:15:23 <elliott> well his "problem" is that his GHC is broken
21:15:32 <atriq> That is a big problem
21:15:37 <atriq> olsner, at least me and elliott
21:16:47 * olsner goes check out this "haskell" thing
21:19:36 <oerjan> olsner: no one goes there any more, it's too crowded
21:20:13 <atriq> It's people who don't know what they're doing (eg, me) helping people who don't think they do
21:23:21 <kmc> imo that dropbox thing is not really a "hack"
21:23:26 <kmc> in either MIT or general usage
21:23:55 <kmc> "check out this awesome hack where i got people to buy herbal viagra from me by automatically posting the link to a ton of blogs and wikis!"
21:33:02 -!- augur has joined.
21:41:52 -!- sirdancealot has joined.
21:52:40 <Vorpal> <elliott> well his "problem" is that his GHC is broken <-- how is it broken?
21:52:58 <Vorpal> never mind, misread that
21:54:07 <Vorpal> kmc, well okay, whatever the terminology it is a scam
21:56:09 <oerjan> i'm sure elliott would be more than capable of arguing that ghc is broken, but would find it boring. am i close?
21:57:25 <elliott> depends what you mean by broken
21:57:29 <elliott> it has optimisations that break Haskell semantics
21:58:28 <Vorpal> oh, that seems a bit silly
21:58:39 <ion> International Bakery http://www.itslenny.com/recording.php?file=de45d6e156191bbe92c872cab04c8824
21:58:48 <Vorpal> elliott, anything you would run into normally?
21:59:03 <oerjan> has anyone ever been bitten in practice by build/foldr fusion theoretically not being compatible with seq?
21:59:16 <elliott> oerjan: that wasn't the example I was thinking of
21:59:20 <elliott> Vorpal: well, lens ran into it :)
21:59:28 <elliott> admittedly, in an edge-case example
21:59:34 <elliott> basically the problem is that GHC does invalid eta-expansions
21:59:45 <elliott> like its optimiser can equate f and (\x -> f x)
21:59:49 <elliott> but these are not the same in the presence of _|_
21:59:54 <oerjan> (well, parametricity not being entirely compatible)
22:00:01 <elliott> because _|_ = _|_ but (\x -> _|_ x) =/= _|_
22:00:10 <Vorpal> I was just about to say that seemed sensible, but right, it is not in that case
22:00:37 <Vorpal> why does ghc do that? Does it offer such a great speed advantage?
22:01:32 <elliott> spj doesn't know of a way to get decent optimisation without it
22:01:39 <elliott> it is about inlining and stuff I think
22:02:51 <Vorpal> I can't personally think of a non-contrived way of running into that issue though... but I guess at least one of them exists.
22:13:04 <kmc> perhaps forcing something of function type should be a n-oop
22:13:12 <kmc> that would probably break other shit though
22:15:24 -!- Vorpal has quit (Ping timeout: 276 seconds).
22:19:21 -!- copumpkin has joined.
22:23:03 <ion> Wrong Number http://www.itslenny.com/recording.php?file=7c89a6e66edca1256a35c7cb3bc6f794
22:23:04 <oerjan> it means the code to evaluate an expression needs to _know_ whether it is a function type, which i don't think is currently the case
22:23:58 <kmc> probably not in all cases
22:24:03 <kmc> PAP objects are known to have function type
22:25:08 <kmc> so is a FUN_*
22:25:13 <oerjan> ...and those are considered already whnf, are they not?
22:27:20 <oerjan> i vaguely suspect that keeping track of that information will be equivalent to reinstating the old Seq type class
22:27:45 <elliott> kmc: the solution is the one older Haskell versions had
22:27:56 <elliott> which was dropped for bad reasons
22:27:58 <elliott> (just like fail not being in Monad)
22:28:56 <kmc> oh is to put seq in a class
22:29:06 <oerjan> moreover however, you might very well _want_ to pre-compute a long evaluating resulting in a function.
22:29:55 <elliott> Haskell is a tale of musts over wants though :P
22:30:06 <elliott> "it would be convenient to put side-effects here..."
22:30:18 <elliott> oerjan: anyway it is safe to provide a *variant* of seq on functions, I think
22:30:41 <oerjan> i assume the old Seq class worked for functions too.
22:33:41 -!- boily has quit (Quit: Poulet!).
22:33:44 <oerjan> nope, http://www.mat.uc.pt/~pedro/cientificos/funcional/haskell-report-1.4-html/basic.html#sect6.2.7
22:33:46 -!- atriq has quit (Ping timeout: 245 seconds).
22:35:06 <ion> Insurance Company http://www.itslenny.com/recording.php?file=c969235df6839ae39eb119d303bfa727
22:35:10 <elliott> anyway the solution is to not have an impure language
22:35:49 <oerjan> remember the heap of haskell data types these days that are newtype wrappers over functions, it would be a total mess if seq couldn't be applied to any of them
22:36:37 <elliott> that's what causes the much-promoted "violation" of monad laws for Reader/IO/etc.
22:36:46 <elliott> standing by my real solution though
22:37:25 <kmc> you could use a non-newtype newtype-like data to do that
22:37:36 <kmc> but it's a hack
22:37:51 -!- monqy has quit (Quit: hello).
22:37:54 <elliott> more importantly that's slow :p
22:38:31 -!- Jafet has quit (Quit: Leaving.).
22:52:16 -!- Jafet has joined.
23:01:08 <shachaf> And the difference is that one gives you a list.
23:05:45 <shachaf> > sumOf _1 ("hello","there")
23:06:12 <lambdabot> Ambiguous type variables `s0', `t0', `a0', `b0' in the constraint:
23:06:21 <lambdabot> (Functor f, Field1 s t a b) => (a -> f b) -> s -> f t
23:06:39 <FireFly> That's a lot of type variables
23:06:49 <lambdabot> Functor f => LensLike (Control.Lens.Internal.Context a b) s t a b -> LensLike (Control.Lens.Internal.Context a' b') s' t' a' b' -> ((a, a') -> f (b, b')) -> (s, s') -> f (t, t')
23:09:35 <shachaf> elliott: "what would it even mean"
23:09:41 <shachaf> (To have a Traversal with Monad.)
23:09:55 <shachaf> (I've wondered about that and haven't come up with anything interesting.)
23:12:24 -!- dedis3 has joined.
23:14:34 <elliott> is there a good way of saying "I don't want to change the subject because I'm confused"
23:17:44 <shachaf> elliott: How do you feel about DeriveFunctor?
23:18:09 <elliott> apparently the point was that he didn't know not every functor was applicative
23:22:13 <shachaf> ion: http://hackage.haskell.org/trac/ghc/ticket/7436
23:23:13 <shachaf> "very surprised" "Truly bizarre" -- Simon Peyton Jones, Haskell implementor
23:26:14 <ion> from Microsoft™
23:27:53 <Jafet> That's not what he said about the exponential time instances
23:28:05 <Jafet> http://hackage.haskell.org/trac/ghc/ticket/1544
23:28:57 <shachaf> Jafet: I like how all the progress on this bug has been changing the milestone for every release.
23:29:09 <lambdabot> http://hackage.haskell.org/package/gread
23:29:18 <lambdabot> http://hackage.haskell.org/package/greads
23:29:30 <ion> @hackage greadses
23:29:30 <lambdabot> http://hackage.haskell.org/package/greadses
23:31:51 <lambdabot> Data.Generics.Text gread :: Data a => ReadS a
23:31:52 <lambdabot> System.Posix.IO NonBlockingRead :: FdOption
23:32:04 <Jafet> @hackage ChristmasTree
23:32:04 <lambdabot> http://hackage.haskell.org/package/ChristmasTree
23:32:14 <Jafet> That is the worst package name ever
23:32:26 <elliott> Changing Haskell's Read Implementation Such That by Mainpulating ASTs it Reads Expressions Efficiently
23:35:17 <kmc> no that is the best package name in the history of civilization
23:35:31 -!- carado has quit (Ping timeout: 246 seconds).
23:35:37 <kmc> that's a great backronym
23:35:44 <kmc> it is grammatical and does not cherry pick letters very much
23:36:02 <oerjan> inventing cool acronyms requires unusual skill
23:36:57 <Bike> is that a reference? I swear I've seen that joke before
23:38:09 <elliott> oerjan early rises: just another noodle
23:40:55 <ion> Very amusing, gleefully inventing neat acronyms.
23:41:34 <elliott> noodles: good for acronyms
23:41:58 <oerjan> Bike: sorry, i got stuck finding something giving "deja vu"
23:42:14 <shachaf> https://www.humblebundle.com/double-fine
23:42:43 <elliott> Phantom__Hoover: please, hoover, a noodley tell-tale oratory making hinders only ocular veracity, evaporating reason
23:43:42 <shachaf> Phantom__Hoover: attention or money
23:45:02 <elliott> kmc: künstlerroman, makes child
23:45:35 <elliott> not much to work with but i tried
23:45:57 <oerjan> at least you avoided any strained words
23:47:14 <Phantom__Hoover> significant hassling 'alarming' correspondents; harming a friend?
23:48:23 <elliott> freaky indecency: zany Zulu impales elephant
23:48:30 -!- copumpkin has quit (Ping timeout: 250 seconds).
23:49:02 -!- copumpkin has joined.
23:49:14 <Jafet> > map (`lookup`zip['a'..'z'](['n'..'z']++['a'..]) "shachaf"
23:49:15 <lambdabot> <hint>:1:58: parse error (possibly incorrect indentation)
23:49:20 <Jafet> > map (`lookup`zip['a'..'z'](['n'..'z']++['a'..])) "shachaf"
23:49:22 <lambdabot> [Just 'f',Just 'u',Just 'n',Just 'p',Just 'u',Just 'n',Just 's']
23:49:55 <lambdabot> Prelude lookup :: Eq a => a -> [(a, b)] -> Maybe b
23:49:55 <lambdabot> Data.List lookup :: Eq a => a -> [(a, b)] -> Maybe b
23:49:55 <lambdabot> Data.HashTable lookup :: HashTable key val -> key -> IO (Maybe val)
23:50:29 <shachaf> elliott: non is making it into lens, though.
23:50:58 <elliott> i told edwardk i don't like it
23:51:17 <elliott> right so its type signature is wrong
23:51:33 <elliott> I also told him I hated _head! but then I thought of a solution to that and he told me to open an issue