gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
<ottidmes> do you have to auth per se to play?
<infinisil> ottidmes: Not sure, but I know that you can just display your name as anonymous
<__monty__> infinisil: What's your time on the puzzle input?
<ottidmes> infinisil: I probably leave it for now, its more something I would like to do when learning a new programming language
<__monty__> ottidmes: I do think so. Haven't seen a way round it. It's probably so they have a constant seed to generate the puzzle inputs from.
<infinisil> __monty__: 0.00 :)
<__monty__> : o What cleverness are you using? Some sort of hash to determine the off-by-one IDs?
<infinisil> Okay it's ~0.01, hard to measure this correctly
<infinisil> __monty__: BK trees
<infinisil> (specifically the bktrees package)
<__monty__> infinisil: Hmm, you make a tree for each ID?
<infinisil> __monty__: Well I'm not concerned about the BK tree implementation. I just insert every id into a tree and check whether there's a similar one in it already while doing so
<__monty__> infinisil: Interesting. You knew about the data structure or you found it specifically for the puzzle?
<infinisil> I knew about it before already, using it for https://github.com/infinisil/soph
<__monty__> infinisil: Is the point deduplicating? Or is there another use of the imports?
<infinisil> __monty__: Really just deduplication
<__monty__> It's a cool application though. Does it use an advanced similarity metric?
<__monty__> Or is it simple convolution or something?
<ldlework> wouldn't any hash structure work
<infinisil> ldlework: "normal" hashes usually are very different for 2 similar inputs
<ldlework> ok I haven't gotten to the challenge so that's enough for me
<__monty__> ldlework: No, because most hash structures implement exact matching, not approximate.
<infinisil> > builtins.hashString "md5" "hello0"
<{^_^}> "4f6d054536d6613a91472139cc60f072"
<infinisil> > builtins.hashString "md5" "hello1"
<{^_^}> "203ad5ffa1d7c650ad681fdff3965cd2"
<infinisil> With blockhash, only a couple bits differ for similar inputs
<ldlework> jasongrossman: you mess with Perkeep any more?
__monty__ has quit [Quit: leaving]
<jasongrossman> ldlework: No. Low priority for me. Still looks neat, though.
pie__ has joined #nixos-chat
pie___ has quit [Remote host closed the connection]
<ldlework> jasongrossman: i found https://github.com/dbp/shoebox
<jasongrossman> ldlework: That gives me mixed feelings: happiness that people are working on interesting things, and unhappiness that there's no sign of standardisation of data formats.
<ldlework> well sandbox implements camli
<ldlework> err
<ldlework> shoebox
<infinisil> Awesome, my current solution for day2part2 can solve it for 100000 lines in about a minute :D
<ldlework> infinisil: that's the spirit
<jasongrossman> ldlework: Partly, but the sandbox page says they don't implement the same backends.
<ldlework> lol sandbox
<jasongrossman> ldlework: I'm not even saying they should, just that until there's more standardisation it's a research project rather than something to be evangelical about.
<jasongrossman> ldlework: Which is a pity because I enjoy being evangelical.
<ldlework> well i'm mostly particularly interested in having some that serves the purpose of a forever store much less about which one
<jasongrossman> ldlework: And I'm going to keep calling it sandbox. :-)
<ldlework> haha
lassulus has quit [Quit: WeeChat 2.2]
hedning has joined #nixos-chat
lassulus has joined #nixos-chat
lassulus has quit [Client Quit]
lassulus has joined #nixos-chat
lassulus has quit [Quit: WeeChat 2.2]
lassulus has joined #nixos-chat
ottidmes has quit [Quit: WeeChat 2.2]
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.1)]
jasongrossman has joined #nixos-chat
endformationage has quit [Quit: WeeChat 2.3]
jackdk has quit [Ping timeout: 244 seconds]
metastance has joined #nixos-chat
avn has quit [Ping timeout: 268 seconds]
avn has joined #nixos-chat
<colemickens> whoa shoebox is cool
<ldlework> is it?
<ldlework> why do you say so?
<infinisil> Man, advent of code day 3 took me a while to finish
<ldlework> I'm still trying to learn enough haskell to do day 1
<ldlework> :)
<infinisil> ldlework: feel free to ask me anything about Haskell if you want
<ldlework> thanks homie
<etu> infinisil: 3.1 was quite easy to me, but 3.2 ate too much memory when basing that solution on my first solution :-D
<etu> I did it fast before work, so it was a very fast-to-write but not efficient way to solve 3.1 :)
<infinisil> etu: I pretty much used the same approach for both, how long did 3.1 take for you to finish running??
<ldlework> infinisil: I started writing a guide on lambda calculus from my first investigations into Haskell, hehe http://ldlework.com/projects/lojban/la-lamda/
<etu> infinisil: about .215s
<etu> I'm also doing php because lazy
<etu> But was at work when I started it and really have work to do :-p
<infinisil> Well my inefficient solution took more than 1 minute to run (after that I Ctrl-C'd it)
<etu> woha
<etu> My 3.2 doesn't even build up the grid for me :-p
<etu> With the actual input
<etu> I'll rework it some other day
<etu> Or just increase memory-limit ;D
<infinisil> Hehe
<infinisil> My solution is online as well, not sure if I should link it because I don't want to spoil anybody :))
<etu> Ok, got the answer now
<etu> :-)
<etu> Well, the top person on the global leaderboard generally post his solutions before I even read the task :p
__monty__ has joined #nixos-chat
ninjin has quit [Ping timeout: 256 seconds]
ninjin has joined #nixos-chat
qyliss has quit [Quit: bye]
qyliss has joined #nixos-chat
pie___ has joined #nixos-chat
pie__ has quit [Read error: Connection reset by peer]
lnikkila has joined #nixos-chat
<gchristensen> HaikuOS adds "Packages" -- immutable directories just for individual packages which get merged together at startup -- neat! https://osvoyager.wordpress.com/2018/11/30/what-makes-beos-and-haiku-unique/ / https://lobste.rs/s/24gnnu/what_makes_beos_haiku_unique
<jasongrossman> gchristensen: Yes, interesting.
ottidmes has joined #nixos-chat
<gchristensen> people say "you should just use iso8601" like that is easy, and it makes me grumpy
<gchristensen> yes, today is 2018-337, why can't your iso8601 parser handle that??
<samueldr> how about --12-03?
<gchristensen> hell yah that is a good one too
<__monty__> gchristensen: That's a valid iso8601 date?
<gchristensen> you bet it is
<gchristensen> so is 2018-W49-1
<__monty__> Oh noooooo, unexpected complexity.
<gchristensen> and 20181203
<manveru> ruby handles those fine at least :)
<gchristensen> ooh! even --12-03?
<manveru> yeah
<gchristensen> +1 ruby -- PHP does all but that one
<manveru> go on the other hand...
<manveru> i dunno, RFC 3339 format is my go-to, and it's complex enough already
<gchristensen> ah! iso 8601 3rd ed. dropped the truncated representationo
pie__ has joined #nixos-chat
pie___ has quit [Remote host closed the connection]
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.1)]
ottidmes has quit [Ping timeout: 246 seconds]
ottidmes has joined #nixos-chat
disasm has quit [Ping timeout: 250 seconds]
disasm has joined #nixos-chat
<__monty__> infinisil: Any clever tricks for today's puzzles? Takes my solution 7.4s. More BKTrees?
sir_guy_carleton has joined #nixos-chat
<infinisil> __monty__: No BKTree's this time :P
<infinisil> __monty__: I do have a much faster solution though (0.16s), want a hint?
<__monty__> infinisil: Sure, probably won't implement it today but I always want to learn more.
<infinisil> I'll PM you
dsx has quit [Ping timeout: 252 seconds]
pie___ has joined #nixos-chat
dsx has joined #nixos-chat
pie__ has quit [Ping timeout: 260 seconds]
endformationage has joined #nixos-chat
<joepie91> gchristensen: yeah, this seems like a poor decision
<gchristensen> https://www.youtube.com/watch?v=Wq8kS6UoOrQ 2min30s to spacex launch
<joepie91> gchristensen: I also can't help but feel that it's reactionary towards the app removal by Apple, and essentially a way of saying "we don't really want to invest in real content moderation"
<gchristensen> I used to work for Tumblr and my experience was invested pretty heavily in content moderation
<gchristensen> but yes
<joepie91> gchristensen: quite a few people were reporting long-running problems with things like child pornography though, and complaining about their reports not being acted upon
<joepie91> so I don't know where things went wrong exactly, but there were certainly serious moderation issues
<gchristensen> I can tell you the two major points at which things went wrong
<gchristensen> yahoo, oath ;)
<joepie91> lol
<gchristensen> oath would hold all-hands meetings where a major topic was getting kids hooked on their apps to sell more ads
<gchristensen> part of why I left
<joepie91> welp, that sounds about right
<simpson> Happens.
<joepie91> yeah, understandable
<joepie91> boingboing's phrasing
<joepie91> "Even the cold dead embrace of a Yahoo! acquision could not end Tumblr, such was the power of fandom gathered there. But Yahoo never knew what it owned in Tumblr and was indifferent to its continued existence. The management of new Yahoo owner Verizon, however, has a pulse. It knows what Tumblr is and it hates it. It will hack it down until a perfectly clean advertising- and appstore-friendly traffic center remains."
<gchristensen> perfect
<gchristensen> tumblr makes lots of money
<gchristensen> it was profitable, but execubots wanted more money
<gchristensen> and tumblr can't really monetize porn
<gchristensen> but also porn was a big expense
<joepie91> gchristensen: due to video cost, I'm assuming?
<gchristensen> yeah
<__monty__> Why are advertisers so squeamish when it comes to porn? Beer and cigarettes for example?
<simpson> The law, at least in the USA. Same reason that beer ads never actually feature *drinking* of beer.
<__monty__> It's illegal to show ads next to porn? I'm not talking about putting porn in the ads.
<gchristensen> Brand Safety is just a hell of a thing
<gchristensen> Budweiser probably doesn't want to be The Beer Of Weird Porn On Tumblr
<__monty__> Suggesting sex is fine but innocent ads next to porn (porn the viewer probably looked for) are not?
<gchristensen> yeah, no
<joepie91> puritan nonsense...
dsx has quit [Quit: dsx]
<gchristensen> nice, ldlework
<ldlework> yeah what a song
<ldlework> thile isn't the greatest vocalist and i got crazy shivers a few times
<ldlework> and then that partita at the end
<steveeJ> SSH'ing into my laptop and trying to use sudo ask's me to swipe my finger across the fingerprint reader. nice one :D
<ldlework> heh
<ldlework> i found that direnv has nix integration. never used direnv. wonder if it is useful.
<gchristensen> it is! it basically solves all the problems of "how do I make my editor aware of nix?"
<ldlework> well that's nice
<manveru> direnv++
<{^_^}> direnv's karma got increased to 1
<ldlework> lmao
<ldlework> trump++
<{^_^}> trump's karma got increased to 1
<manveru> there's an emacs direnv package too :)
<manveru> it's not 100% reliable, since some modes insist on searching for executable paths on startup... but you can fix those usually
<ldlework> One thing I thought about is somehow integrating nix with org-babel, so that each org-babel language had an associated nix-shell configuration so that babel could execute those snippets
<manveru> that'd be neat
<gchristensen> hah
<joepie91> seems puritanbot is a little buggy still
<gchristensen> it is bound to be buggy
<joepie91> you know that, I know that
<joepie91> whoever made this call probably doesn't
<joepie91> lol
<joepie91> there's not even really a business case to be made for this
<gchristensen> ad dollas
<joepie91> yeah but that doesn't work if your non-NSFW users get pissed at misfires
<gchristensen> yeah but in my trivial model where "porn-based users don't even see ads" we'll get rid of a bunch of freeloaders and make tons of money
lnikkila has quit [Ping timeout: 244 seconds]
<manveru> more people for mastodon :)
jackdk has joined #nixos-chat
__monty__ has quit [Quit: leaving]
<colemickens> discussion of stickers the other day, are these any good? https://www.redbubble.com/shop/nixos+stickers cc: gchristensen
<gchristensen> I can mail you some if you want
<samueldr> colemickens: not the worst, but the print quality leaves to be desired; the angles lines a jagged
<samueldr> not the cut, but the printed lines
<samueldr> (though you have to be close to see it)
dmc has quit [Quit: WeeChat 2.3]
hedning has quit [Quit: hedning]
dmc has joined #nixos-chat