ec changed the topic of #elliottcable to: #
cloudhead has quit [Ping timeout: 264 seconds]
alexgordon has quit [Quit: Computer has gone to sleep.]
sharkbot has quit [Remote host closed the connection]
sharkbot has joined #elliottcable
<joelteon> ok
<joelteon> i've concluded /fa/ is stupid and fashion trends make no sense
<devyn> so I just spent the last little bit getting my prompt to show a path relative to the project base dir if I'm working on a project
<devyn> i.e., normally, Firestar ~/blah>
<devyn> but if PBASE=~/Projects/devyn/maru then
<devyn> Firestar maru>
<devyn> and Firestar maru/subscriber>
<devyn> etc.
<devyn> think I might get rid of the host name
<devyn> yeah there we go
<joelteon> i've been spending the past hour trying to figure out WHAT the FUCK is happening inside pipes-safe
<joelteon> time to use TypeHoles
<joelteon> hory shit i fixed it
<joelteon> i have no idea how people write things that are so insanely complicated
<devyn> hah
<devyn> well, I finally made my prompt something nice
<devyn> wait
<devyn> tst
<devyn> ok
<devyn> [devyn] maru/subscriber - ls
<devyn> oh but the yellow isn't bold like that
<devyn> [devyn] maru/subscriber - ls
<devyn> there we go
<devyn> mirc color codes, eh?
<joelteon> neat
<joelteon> i have a colory prompt too
<devyn> you mean there are people who don't color their prompts?
<devyn> blasphemous!
<devyn> even elliott last time I checked was using something simple like just a colon that's green when $? == 0 and red otherwise
<devyn> he could be using something different now and I'm not sure if I even remember right but yeahy
<devyn> yeah*
<devyn> also my terminal title is set similar to my prompt
<devyn> [devyn] maru/subscriber
<devyn> when running a command:
<devyn> [devyn] maru/subscriber - vim public/scripts/main.js
eligrey has quit [Quit: Leaving]
<joelteon> well ec is dumb
<joelteon> you know I feel like
<joelteon> it would be helpful if there was a version of all the IO stuff in Haskell that works in MonadIO m, not just IO
<joelteon> oh hey this is a good place to use TemplateHaskell
<devyn> no
<devyn> TH is awful
<devyn> please don't
<devyn> :(
<micahjohnston> samurai champloo is awesome
<devyn> suddenly a micahjohnston appears
wudofyr has quit [Remote host closed the connection]
wudofyr has joined #elliottcable
yorick has joined #elliottcable
Sgeo_ has joined #elliottcable
Sgeo has quit [Ping timeout: 264 seconds]
alexgordon has joined #elliottcable
cloudhead has joined #elliottcable
cloudhead has quit [Ping timeout: 245 seconds]
cloudhead has joined #elliottcable
eligrey has joined #elliottcable
vil_ has joined #elliottcable
vil_ has quit [Client Quit]
vil has joined #elliottcable
eligrey_ has joined #elliottcable
eligrey has quit [Ping timeout: 245 seconds]
othiym23 has quit [Ping timeout: 264 seconds]
Sgeo_ has quit [Quit: No Ping reply in 180 seconds.]
Sgeo has joined #elliottcable
othiym23 has joined #elliottcable
eligrey_ is now known as eligrey
eligrey has quit [Quit: Leaving]
eligrey has joined #elliottcable
Sgeo has quit [Quit: No Ping reply in 180 seconds.]
Sgeo has joined #elliottcable
<alexgordon> ec: I forgot to ask
<alexgordon> what EXACTLY were you paying for for so many bitcoins O_O
<alexgordon> miner?
<devyn> drugs
<alexgordon> aahh
<alexgordon> I mean, 20 BTC is way more than a hooker
<devyn> by the way I've bought some cool things with bitcoins
<devyn> renewed my domain, bought the humble origin bundle,
<devyn> something else I can't remember
<devyn> lol
<purr> lol
<devyn> so all that mining did not go to waste
<alexgordon> devyn: someone just asked to buy chocolat with bitcoin
<devyn> and did you do it?
<alexgordon> yep!
<devyn> cool ^_^
<devyn> Difficulty 65,750,060.14908481
<devyn> holy fuck
<devyn> when I started mining it was 4 million
<devyn> lol
<devyn> and people thought Bitcoin would die, huh?
<alexgordon> devyn: have to admit, at the start I thought someone would find a crypto hole
<alexgordon> but now it seems pretty solid
<devyn> it's definitely gaining confidence
<alexgordon> hm which curve does bitcoin use
<devyn> so many people were like "durr going from $230 to $70 is going to make people not use it"
<alexgordon> they said that after the first crash too
<alexgordon> I remember the atmosphere on HN
<devyn> yeah
<alexgordon> it was like "ha, told you it couldn't last"
<devyn> it's a new market; of course it's volatile
<alexgordon> "$30 is just ridiculous, $10 is much more reasonable"
<devyn> it will get less volatile the more it's used
<devyn> haha
<alexgordon> the design of bitcoin is just amazing though
<alexgordon> whoever wrote it is a world class cryptographer
<devyn> apparently not the greatest programmer though; the code kinda sucks
<devyn> some pseudonym named Satoshi, though, yeah
<devyn> ...pseudonym named
<devyn> hm
<alexgordon> I'm sure the code's been cleaned up though :P
<alexgordon> all's well that ends well!
<devyn> oh I don't think it has, but I'm sure there are third-party clients that implement the protocol now
<devyn> that are better
<devyn> IIRC it's basically an RPC-style protocol
<devyn> that's fairly simple
<alexgordon> what does bitcoin use for crypto?
<alexgordon> openssl?
<alexgordon> (satoshi client)
* alexgordon finds the source
<devyn> possibly
<devyn> probably, even
<joelteon> it would be a lot better if String was just a type synonym for Text
<devyn> that would give them all of the SHA stuff too
<alexgordon> yeah it's strange that it uses sha-1
<prophile> alexgordon likes bitcoin?
<devyn> joelteon: but the purity of String = [Char]
<alexgordon> prophile: yes
<joelteon> :<
<prophile> never thought i'd see the day when we find something you actually like
<devyn> joelteon: anyway, you can use the polymorphic strings extension and just use Text as if it were a string as well
<devyn> joelteon: i.e. "blah" :: Text works
<alexgordon> prophile: hahaha
<alexgordon> prophile: I like myself!
<devyn> -loves alexgordon
<purr> devyn: alexgordon loves ragel, <3 <3 <3 <3 <3 ragel, devyn, everyone, Sorella, hoogle, micahjohnston, Chris Potter, your mom, raspberry pi, Kenny Garrett, Dewey Redman, mulgrew miller, kenny wheeler, Michael Brecker, pepper adams, Bush was an idiot, idiots, elly, tits, incomprehensibly, and people capable.
<prophile> insert non-triviality constraint here
<alexgordon> there we go
<prophile> trouble is
<alexgordon> that was my jazz saxophone phase
<prophile> i don't believe any of them
<prophile> well
<prophile> except michael brecker
<prophile> but everyone loves michael brecker
<devyn> well he had to enter them himself
<prophile> devyn: the brit lies
<devyn> this.love(context, "alexgordon", "profile", 1)
<purr> Let it be known that alexgordon loves profile.
<devyn> er
<alexgordon> -loves prophile
<purr> alexgordon: prophile loves applescript.
<devyn> this.love(context, "alexgordon", "prophile", 1)
<purr> Let it be known that alexgordon loves prophile.
<devyn> this.love(context, "alexgordon", "profile", -1)
<purr> Let it be known that alexgordon is indifferent to profile.
<devyn> lol
<purr> lol
<prophile> hah
<prophile> apparently i'm an applescript fiend now
<alexgordon> I feel extremely hurt by the implication that I don't like anything; coming from someone who ONLY LIKES APPLESCRIPT
<devyn> this.love(context, "prophile", "alexgordon", 1); channel.send("now kiss")
<purr> Let it be known that prophile loves alexgordon.
<purr> now kiss
* prophile rogers alexgordon politely
* alexgordon tries to think of a riposte to that that doesn't sound gay
<alexgordon> TITS
<devyn> haha
<alexgordon> anyway, the answer is yes
<alexgordon> it uses openssl
<alexgordon> at least, it does NOW
<prophile> ° ͜ʖ ͡°
<devyn> lol did it have a custom impl before?
<purr> lol
<alexgordon> devyn: no idea
<alexgordon> is there any way in git to go back to the earliest commit?
<alexgordon> based on time
<devyn> yeah you can look at git log
<devyn> and then git checkout <commit>
<alexgordon> hm strange now they're using hmac_sha512
<alexgordon> (quite why they're using hmac_sha512 is a mystery)
<devyn> hmac_sha512 for what
<alexgordon> something in
<alexgordon> key.cpp
<alexgordon> void static BIP32Hash(const unsigned char chainCode[32], unsigned int nChild, unsigned char header, const unsigned char data[32], unsigned char output[64]) {
<alexgordon> unsigned char num[4];
<alexgordon> num[0] = (nChild >> 24) & 0xFF;
<alexgordon> num[1] = (nChild >> 16) & 0xFF;
<alexgordon> num[2] = (nChild >> 8) & 0xFF;
<alexgordon> num[3] = (nChild >> 0) & 0xFF;
<alexgordon> HMAC_SHA512_CTX ctx;
<alexgordon> HMAC_SHA512_Init(&ctx, chainCode, 32);
<devyn> srsly
<alexgordon> HMAC_SHA512_Update(&ctx, &header, 1);
<alexgordon> HMAC_SHA512_Update(&ctx, data, 32);
<alexgordon> HMAC_SHA512_Update(&ctx, num, 4);
<joelteon> u postin code in irc m8
<alexgordon> HMAC_SHA512_Final(output, &ctx);
<alexgordon> }
<joelteon> its happening
<alexgordon> oops that was longer than I expected :P
<alexgordon> -_-
<devyn> lol
<purr> lol
<joelteon> (._. )
<devyn> alexgordon: I think it might be for the new short-chaining thing where it doesn't have to have the entire history of the chain
<devyn> idk
<devyn> but for that they'd definitely want a very secure digest
<devyn> SHA512 works
<devyn> :p
<alexgordon> I guess
<alexgordon> it's a bit weird
<alexgordon> if you wanted it to be fast, why not use a modern HMAC like blake2
<alexgordon> otherwise might as well stick with HMAC-SHA-1, it's not terrible
<devyn> you don't necessarily want fast
<alexgordon> you almost always do
<alexgordon> :P
<alexgordon> for a MAC at least
<devyn> also, in cryptography, new = bad
<devyn> unless the old one is horribly broken
<alexgordon> well exactly, so if they were using HMAC-SHA-1 before no reason not to stick with it
<alexgordon> I think it might actually be faster, not sure
<devyn> well, SHA512 *is* more secure than SHA-1, provably
<alexgordon> so far, not in an HMAC
<alexgordon> as in, they're both more or less the same, security wise
<devyn> theoretically, with a secure algorithm like SHA, the more bits the more secure...
<devyn> so it doesn't matter if in the real world it doesn't matter
<devyn> :p
* devyn shrugs
<purr> ¯\(º_o)/¯
<alexgordon> blake2 is pretty much the state of the art right now, in software. It's likely more secure than SHA-512 and it's way quicker than doing HMAC-SHA-512
<devyn> *likely*
<devyn> cryptography is a time game
cloudhead has quit [Ping timeout: 256 seconds]
<devyn> same reason scrypt isn't used all that much yet
<alexgordon> well it was a SHA-3 finalist (or at least, the original BLAKE was)
<devyn> it hasn't been around long enough
<alexgordon> I don't think anybody seriously doubts it's secure.
<devyn> and I'm saying that that doesn't actually matter, because when it's absolutely critical that something be secure, it's best to use something that has withstood the test of time
<alexgordon> anyway this is way more than I intended to say on the matter :P
<devyn> :p
<alexgordon> just think it's a weird no-man's land
<devyn> even the most brilliant cryptographers miss things sometimes, and they only appear maybe a decade later
<devyn> anyway
<devyn> alexgordon: I think it's funny that what bitcoin might end up doing is actually making SHA unusable with this constant drive for better crypto hardware
<alexgordon> unusable how
<devyn> as in, if the crypto hardware gets good enough that bitcoin has to switch algorithms, that also means no one should really be using that algorithm,
<devyn> because the hardware is around to break it
<alexgordon> nobody's going to brute force sha-1 :P
<alexgordon> oh you mean for passwords?
<devyn> passwords would be one thing yeah
<alexgordon> yeah it's bad for all the people using PBKDF2/SHA-1 (or worse like SHA-1 on its own)
<alexgordon> because now you can buy off the shelf hardware that can do trillions of SHA-1's per second
<alexgordon> no doubt the NSA has this stuff
<alexgordon> but now everybody has it
<devyn> yeah. it's hilarious to me, because it's the same thing that happens with anything financial
<alexgordon> before the most most criminals could get was a FPGA
<devyn> resources get raped
<devyn> in this case, the resource is a crypto algorithm
<devyn> it's funny
<alexgordon> serves them right though :P
<alexgordon> scrypt for passwords!
<devyn> or bcrypt, but scrypt is almost certainly a better choice
<alexgordon> devyn: hm actually it is quite worrying because I would expect most encryption software uses PBKDF2/SHA-1 for turning a password into an encryption key
<devyn> I know
<devyn> oil vs. the environment all over again
<alexgordon> so any files encrypted with passwords are just going to edge closer to being broken!
<alexgordon> (short to medium passwords)
<alexgordon> i wonder if the NSA has partnerships with like intel for custom cracking hardware
<alexgordon> I doubt they have their own fabs!
<devyn> alexgordon: they probably subcontract all of the hardware, but idk
yorick has quit [Remote host closed the connection]