00:00:08 <shachaf> "Distributed as source code in a pre-extracted tar file"
00:00:10 <kmc> KeithW is using it also for his awesome dynamic bitrate inferential flow control videoconferencing thingy
00:00:14 <kmc> https://github.com/keithw/alfalfa
00:00:26 <kmc> i think this may have started as a complete copy of the mosh repo...
00:00:32 <kmc> but now it's a videoconferencing program
00:00:49 <kmc> slash an academically relevant demonstration of something that could be turned into a videoconferencing program
00:01:34 <shachaf> I hope that's what SSP stands for.
00:01:38 <kmc> er, with one modification
00:01:50 <kmc> i forgot which though
00:02:22 <elliott> supplemental super proteins
00:02:42 <kmc> 'The major modification we made to SSP to support Alfalfa is that in Mosh, “diff” and “patch” are round-trip operations, so the protocol is lossless. In Alfalfa, the “diffs” are lossy and the available length for a diff varies according to Sprout’s forecast window size.'
00:03:07 -!- sivoais has quit (Ping timeout: 244 seconds).
00:03:24 <kmc> presumably you then re-apply the lossy diff locally, so that errors don't accumulate
00:03:52 <kmc> here "diff" = MPEG-2 P-frames
00:04:20 -!- nooga has quit (Ping timeout: 245 seconds).
00:04:27 <ion> Sanyo’s “SSP” is the most horrible protocol i’ve ever seen.
00:04:38 <kmc> it uses MPEG-2 not because that's actually a good choice for videoconferencing but because keith had to implement a custom variable bitrate encoder in like three days
00:05:08 <kmc> basically, it uses bayesian estimation to figure out the maximum bitrate it can send with only a 5% chance of a delay larger than 100ms (i think)
00:06:48 <ion> I wish all software did that.
00:07:56 <shachaf> kmc: How're your adventures in JavaScriptland, by the way?
00:07:59 <shachaf> Or wherever it is you are.
00:08:08 <kmc> pretty good
00:08:13 <kmc> i am enjoying my job
00:08:24 <kmc> we have been working on the thing for a month and a half and have made a ton of progress
00:08:34 <shachaf> Is the thing still secret?
00:19:15 <Arc_Koen> some days I just tell myself "hey let's go to bed now before it gets late" and then before I have successfully gone to bed I tell myself "hey I should take a look at that Emmental language it seemed really interesting" or anything of the like and it overwrites the previous resolution and IT'S 2:30AM ALREADY and today is one of those days
00:19:42 -!- Arc_Koen has quit (Quit: that's dr. turing to you, punk).
00:19:51 <shachaf> I woke up before 08:00 today!
00:22:14 <kmc> how'd you manage that?
00:23:30 <shachaf> Normally when I need to wake up early I set several alarm clocks and still have a lowish probability of it doing any good.
00:23:35 <kmc> how do you feel about this event?
00:23:55 <shachaf> When it happens on its own and I'm not tired.
00:29:22 <elliott> there is no good reason for hours less than 12 to exist
00:30:07 -!- monqy has joined.
00:30:11 <shachaf> elliott: 01:00 is a perfectly good time to be awake.
00:30:33 <shachaf> elliott probably believes in Random Standard Time.
00:30:51 <elliott> monqy: what is shachaf talking about
00:31:09 <shachaf> elliott: RST goes from 06:00 to 29:59
00:31:20 <shachaf> kmc probably knows more about it than I do.
00:31:43 <elliott> shachaf: shut up let me war with monqy in peace
00:32:59 <elliott> shachaf: 06:00 is wrong anyway
00:33:03 <elliott> since i'm often still up at that time
00:33:11 <elliott> 08:00 would be a reasonable roll-over probably
00:33:26 <lambdabot> how does this command even work again
00:36:15 <Sgeo> Oh, Phantom__Hoover elliott monqy et all, recent update if you didn't see it
00:37:36 <monqy> creative liberties, artistic license
00:37:47 <elliott> monqy: i didn't know you coded perl
00:38:13 <monqy> i used to know some perl but then i forgot it
00:38:22 <monqy> i think i';ve only written one perl thing in my entire life
00:38:34 <shachaf> iused to forgot all perl but thenr i learned it :"(
00:38:48 <elliott> monqy: http://en.wikipedia.org/wiki/Artistic_License
00:39:56 -!- fungot has quit (Ping timeout: 246 seconds).
00:39:58 <monqy> never heard of it!
00:40:03 <monqy> plenty about the other artistic license though
00:40:05 <monqy> my favourite license
00:40:13 -!- sebbu has quit (Ping timeout: 246 seconds).
00:41:00 * Sgeo forgets whether there's a way for non-admins to privately message lambdabot to get it to say something in channel
00:41:22 <Sgeo> There's obviously someone in here who can do it, shachaf I think?
00:41:58 <shachaf> lambdabot is lying. It's me.
00:42:27 <Sgeo> shachaf vs shachaf, FIGHT!
00:43:09 <Sgeo> "Whether or not the original Artistic License is a free software license is largely unsettled. It was criticized by the Free Software Foundation as being "too vague; some passages are too clever for their own good, and their meaning is not clear.""
00:43:28 <shachaf> elliott: stop @admin-ing me :'(
00:43:31 <Sgeo> Too clever for their own good? Like some Perl code?
00:46:59 -!- sebbu has joined.
00:47:39 -!- sivoais has joined.
01:07:04 <FreeFull> shachaf: Not all haskell code is clever =P
01:07:17 <FreeFull> See: All Haskell code I've ever written
01:07:46 <shachaf> Not all things FreeFull says are clever.
01:08:07 <FreeFull> shachaf: I say none of them are
01:15:04 -!- copumpkin has joined.
01:26:41 -!- glogbackup has joined.
01:26:41 -!- slbkbs has joined.
01:26:41 -!- slbkbs has quit (Changing host).
01:26:41 -!- slbkbs has joined.
01:26:49 -!- lambdabot has quit (Ping timeout: 256 seconds).
01:26:52 -!- EgoBot has quit (Remote host closed the connection).
01:27:03 -!- EgoBot has joined.
01:28:05 -!- HackEgo has quit (Ping timeout: 255 seconds).
01:28:42 -!- HackEgo has joined.
01:30:22 -!- slbkbs has quit (Quit: adieu).
01:33:57 -!- TeruFSX_ has quit (Read error: Connection reset by peer).
01:34:46 -!- TeruFSX has joined.
01:36:53 <Sgeo> Common Lisp:Clojure::Tcl:???
01:36:59 * Sgeo wants the ??? language
01:39:10 <Sgeo> Clojure has CL-style macros, when I think I really like Tcl-style I-don't-know-what-to-call-it
01:39:40 <Sgeo> fexprs? Except, if Kernel is said to have fexprs, it's not quite it
01:40:27 <pikhq> Tcl: a Tcl-expr lisp.
01:41:09 <monqy> something about strings
01:42:42 <Sgeo> I didn't say I like the string aspect
01:43:06 <Sgeo> But the idea of being able to dynamically generate the code that a control structure will see
01:43:56 <Sgeo> Although admittedly it hurts compilability and any code-walking ability
01:44:24 <monqy> does it do anything good
01:44:40 <monqy> it'd be pretty sad if something so crazy was entirely useless and just a huge mess for losers!!!
01:46:16 <Sgeo> It comes to mind whenever I see someone complain about how some macro they want to use won't accept a variable containing a symbol or a vector in place of where the macro usually accepts a symbol or vector
01:51:41 -!- TeruFSX has quit (Remote host closed the connection).
01:53:04 -!- TeruFSX has joined.
01:57:33 <Sgeo> Wait, why exactly am I chatting on the W:A chat as though anyone actually chats?
01:58:01 <Sgeo> http://snooper.worms2d.info/
01:59:29 <elliott> why are you asking #esoteric, Sgeo
01:59:45 <shachaf> elliott: Twist: #esoteric is W:A caht.
02:09:23 <HackEgo> http://codu.org/projects/hackbot/fshg/index.cgi/raw-file/tip/paste/paste.17397
02:20:55 -!- Phantom__Hoover has quit (Read error: Connection reset by peer).
02:49:01 -!- Jafet has joined.
03:08:00 <Sgeo> I wonder if I should attempt to implement ALPACA in Clojure
03:12:45 <monqy> whatever it is, it's better than nothing, unless it isn't
03:26:38 <Sgeo> Why does catseye keep 404ing on me?
03:26:43 <Sgeo> http://catseye.tc/projects/alpaca/doc/alpaca.html
03:26:56 <Sgeo> The Game of Life link and the WireWorld link both 404 on me
03:27:11 <Sgeo> Almost complained that ALPACA cheats, but that's REDGREEN not ALPACA cheating
03:27:50 <monqy> do you think you have a chance at winning the: catch of the year award???? I think you do
03:28:12 <monqy> you'll have to compete with the fishers but all they catch are fish
03:30:37 <elliott> monqy: sometimes they catch other things by accident
03:30:42 <Sgeo> I should do it in ClojureScript, not Clojure
03:30:59 <Sgeo> Although I kind of hate ClojureScript. A "lisp" with macros in a different language and no eval.
03:31:06 <monqy> do it in "clozure"
03:31:15 <elliott> monqy: is the joke clozure common lisp
03:32:13 <monqy> that was my final joke
03:32:26 <elliott> what's the point of living if you don't have jokes to tell.
03:32:37 <shachaf> i meant a pun but a pub would also do
03:33:34 <Sgeo> Is APALCA's only full documentation the Perl implementation?
03:35:59 <Sgeo> Oh, REDGREEN's source probably serves as documentation
03:36:30 <Sgeo> o.O cpressey writes useful code?
03:42:03 -!- TodPunk has quit (Read error: Connection reset by peer).
03:42:15 -!- TodPunk has joined.
03:42:23 <Sgeo> Can REDGREEN fish interact with the rest of the REDGREEN world?
03:42:46 <Sgeo> e.g. the GoL simulation interacting with WireWorld
03:44:19 <shachaf> kmc: newtype Rec a = InR { outR :: Rec a -> a } is just Curry's paradox.
03:48:09 <elliott> shachaf: curry's paradox doesn't work because you have to mix logic with common sense to get the best results
03:48:11 <elliott> a wise man once told me this
03:50:26 <shachaf> Those are the ingredients of a delicious curry.
03:55:31 -!- mig22 has joined.
03:57:03 -!- kinoSi0 has quit (Read error: Connection reset by peer).
03:57:31 -!- kinoSi has joined.
04:06:27 <Sgeo> DuctTape is flammable.
04:07:05 <monqy> i'll keep that in mind
04:07:16 -!- FireFly has quit (Ping timeout: 246 seconds).
04:12:40 <Sgeo> <ivan> let's train a new generation of programmers capable only of thinking functionally, leaving us with the high-paying procedural maintenance
04:19:56 -!- FireFly has joined.
04:26:02 <Sgeo> Oh, so Himera needs a server.
04:26:16 <Sgeo> (Himera is a web-based REPL for ClojureScript)
04:27:15 <Sgeo> Yes, this bothers me.
04:27:31 <elliott> thanks for letting us all know, Sgeo
04:38:19 -!- copumpkin has quit (Ping timeout: 260 seconds).
04:38:54 -!- copumpkin has joined.
04:45:22 -!- elliott has quit (Ping timeout: 246 seconds).
05:07:21 <Sgeo> Steam is Steamy. Good to know.
05:32:04 -!- copumpkin has quit (Ping timeout: 252 seconds).
05:32:14 -!- evincar has joined.
05:32:44 -!- copumpkin has joined.
05:36:34 <Sgeo> shachaf, this is surprising to you?
05:45:41 -!- ogrom has joined.
05:49:11 -!- evincar has quit (Read error: Connection reset by peer).
05:53:41 -!- evincar has joined.
06:03:27 <kmc> now running memtest86+ v4.20
06:03:34 <kmc> test memory everyday
06:06:02 -!- evincar has quit (Read error: Connection reset by peer).
06:08:25 -!- evincar has joined.
06:47:44 -!- epicmonkey has joined.
06:52:50 <kmc> achievement unlocked: mounted boot drive using only duct tape
06:54:24 * Sgeo keeps looking at redgreen.alp
06:54:48 -!- evincar has quit (Quit: leaving).
06:55:58 <kmc> shachaf: that is roughly all i know about random standard time
06:56:26 <kmc> what you said
06:56:45 <kmc> it goes from 06h to 29h, with the numbers under 24 lining up with normal time
06:57:04 <kmc> actually I think in some real world contexts both 00h and 24h are used
06:57:09 <kmc> to denote beginning or end of day resp.
07:30:27 -!- carado has joined.
07:46:18 -!- carado has quit (Quit: Leaving).
07:59:31 <Sgeo> Why is my college so full of idiots
07:59:58 <Sgeo> Apparently, a bunch of students ignored a fire drill, then ignored a seminar that they were required to attend because they didn't participate in the fire drill.
08:00:03 <Sgeo> They are now expelled from the dorms.
08:00:56 -!- monqy has quit (Quit: hello).
08:20:44 <Sgeo> Oh hey, I think I found a mistake in cpressey's REDGREEN implementation
08:21:44 <Sgeo> Fish in water are supposed to follow GoL rules. But one fish surrounded by 8 fish will turn into air (not a GoL cell state) rather than water (GoL dead)
08:22:12 <Sgeo> Unless the precedence of the state transitions is opposite of what I think it is
08:30:55 -!- Arc_Koen has joined.
08:33:55 <Sgeo> No, I think I'm right and cpressey is wrong.
08:36:57 <fizzie> I do think it's first-matching, too.
08:37:23 <fizzie> Though you could just try it out.
08:38:21 <fizzie> Arc_Koen: REDGREEN's ALPACA spe
08:38:41 <fizzie> I was going to make that "Arc_Koen: Fish.", to be more mysterious.
08:38:46 <Sgeo> fizzie, looking at the generated Perl also suggests it's first-matching
08:38:46 <fizzie> Now I spoiled the whole thing.
08:39:38 <fizzie> I just looked at the Perl generating the Perl, but it did, too.
08:39:54 <Arc_Koen> fizzie: according to the logs the discussion about alpaca and redgreen started four hours ago
08:40:33 <Arc_Koen> wait that's actually what you were talking about ?
08:41:26 <Sgeo> The Perl generating the Perl gives me a headache of confusion
08:49:37 -!- fungot has joined.
08:57:46 <Sgeo> Does fungot speak REDGREEN?
08:57:46 <fungot> Sgeo: and the same buttons to page up and page down in aterm. oh, you are just making a bad optimization for fun
08:58:00 <Sgeo> bad optimizations for fun!
09:04:45 -!- Phantom_Hoover has joined.
09:11:32 -!- lambdabot has joined.
09:12:12 -!- Jafet has quit (Quit: Leaving.).
09:24:57 -!- Jafet has joined.
09:25:57 -!- nooga has joined.
09:31:58 <Arc_Koen> I think I accidentally erased the brainfuck interpreter I had written in thue
09:36:43 <Arc_Koen> FreeFull: fortunately I pasted it to sprunge.us and copied the link here and it's in the logs
09:39:58 <Arc_Koen> oh, it turns out § is not ascii
09:41:12 <Arc_Koen> I didn't think that was the problem
09:41:59 <Arc_Koen> replaced § with ¬ß (which is not ascii either, that's weird)
09:56:33 -!- FireFly has quit (Changing host).
09:56:33 -!- FireFly has joined.
10:22:09 <ogrom> let's not screw files, shall we?
10:28:47 <fizzie> Seven bits is not enough for miracles. At least ASCII has both upper- and lowercase.
10:35:51 -!- AnotherTest has joined.
10:43:01 <Arc_Koen> you need only 6 bits for A-Za-z0-9, who needs the rest
10:57:42 -!- ogrom has quit (Quit: Left).
10:59:47 <fizzie> Sometimes a bit of punctuation is good too.
11:08:25 <oklofok> get access to GAME BREAKING BUGS
11:12:18 -!- Jafet has quit (Quit: Leaving.).
11:20:05 -!- sivoais has quit (Read error: Connection reset by peer).
11:20:24 -!- Jafet has joined.
11:20:48 -!- sivoais has joined.
11:32:10 -!- sivoais has quit (Read error: Connection reset by peer).
11:32:53 -!- sivoais has joined.
12:06:48 -!- itidus21 has joined.
12:10:51 <itidus21> * Topic for #esoteric is: BEWARE THE O/RJANIST MO/O/SE |
12:11:02 <itidus21> I, for one, welcome our new hash function overlords |
12:11:11 <itidus21> E5081A06F9E364E179B336A2C6D6831D4B50CD7739C7E156
12:11:19 <itidus21> 5E03EBF2 | ZARDOZ created ZARDOZ; all else is the work of
12:11:30 <itidus21> ZARDOZ | New channel logs: http://5z8.info/hookers_j0l4yf_nak
12:11:42 <itidus21> edgrandmas.jpg | Old-style channel logs: http://codu.org/logs/_es
12:12:01 <itidus21> * Topic for #esoteric set by zzo38!~zzo38@24-207-49-17.eastlink.c
12:12:06 <olsner> itidus21: it is looking more odd the further you go
12:12:38 <itidus21> the arbitrary line breaks are new information
12:13:32 <olsner> oh, zzo probably put weird ascii in there
12:13:58 <itidus21> uh.. thats just the linebreaks based on the width of my xchat window
12:14:03 <Arc_Koen> I was implementing the . instruction in my brainfuck interpreter in thue
12:14:08 <itidus21> Look at Dave. He's like "Okay. You can do this. DON'T hit the cymbals as hard as you fucking can!"
12:14:12 <olsner> itidus21: so what is the odd thing?
12:14:44 <itidus21> the odd thing was me re-posting the topic with the linebreaks based on the width of my xchat window
12:14:50 -!- itidus21 has left ("Leaving").
12:15:00 <Arc_Koen> so I though "I'd do characters 32-126 manually, and characters above 128 can be detected easily since they have a 1 as their most significant digit
12:15:12 <Arc_Koen> so I do all characters 32-126 manually
12:15:36 <Arc_Koen> I realize I did all that with the . marker right to the number
12:15:46 <Arc_Koen> and the most significant digit is left to the number
12:16:39 <olsner> hmm, so the oddest part about that whole thing was that nothing was odd in the first place
12:16:40 <Arc_Koen> does that mean I have to do it all over again?
12:16:41 <olsner> ... except itidus, I guess he's a bit odd
12:17:57 <Arc_Koen> @tell itidus21 about that null stack you were so impressed about: I thought maybe a copy stack that would do exactly the opposite: instead of destroying everything that's pushed onto it (and being always empty) it keeps copy of everything that's popped from it (and is never empty)
12:19:19 <Arc_Koen> ok one cheap way to cover my mistake without redoing everything manually is to consider ascii characters mod 128
12:19:45 <Arc_Koen> that is, characters 128-255 would be equivalent to characters 0-127
12:20:17 <Arc_Koen> or I could just say I don't care and trying to output a character outside the range 0-127 causes the program to halt
12:20:25 <Arc_Koen> yeah that should be good enough
12:20:51 -!- spamdribble has joined.
12:21:44 -!- spamdribble has left.
12:26:03 -!- Frooxius has joined.
12:32:36 <Arc_Koen> wouuuuuuuuh works with the Hello World program from the esolangs.org brainfuck page!
12:34:14 <olsner> Arc_Koen: congratulations
12:34:37 <Arc_Koen> I think I'm gonna send an email to the guy who wrote the thue interpreter I'm using
12:35:27 <Arc_Koen> it allows to run the thue program step by step or at any speed and to see the current state of the program and that's a very valuable debugging tool
12:45:13 -!- AnotherTest has quit (Read error: Connection reset by peer).
12:45:27 -!- AnotherTest has joined.
13:01:47 -!- AnotherTest has quit (Quit: Leaving.).
13:03:27 -!- AnotherTest has joined.
13:08:24 -!- Jafet has quit (Quit: Leaving.).
13:28:33 <ion> Absolutely No Words http://youtu.be/pquYAEfbBFM
13:37:27 <olsner> that's slightly disturbing
13:40:24 -!- iconmaster has joined.
13:46:01 -!- Vorpal has joined.
14:21:06 -!- Jafet has joined.
14:24:54 -!- iconmaster has quit (Ping timeout: 260 seconds).
14:32:35 -!- nooga has quit (Ping timeout: 256 seconds).
14:42:02 -!- Frooxius has quit (Quit: ChatZilla 0.9.89-rdmsoft [XULRunner 1.9.0.17/2009122204]).
14:42:14 -!- Frooxius has joined.
14:46:31 -!- iconmaster has joined.
15:05:16 -!- hagb4rd has joined.
15:06:34 -!- iconmaster has quit (Ping timeout: 256 seconds).
15:07:57 -!- iconmaster has joined.
15:12:05 -!- FireFly has quit (Excess Flood).
15:19:05 -!- FireFly has joined.
15:20:26 -!- FireFly has quit (Excess Flood).
15:23:44 -!- FireFly has joined.
15:26:49 -!- mig22 has quit (Quit: mig22).
15:29:29 -!- nooga has joined.
15:42:45 -!- elliott has joined.
16:12:06 -!- iconmaster has quit (Ping timeout: 265 seconds).
16:13:20 -!- sirdancealot7 has quit (Read error: Connection reset by peer).
16:15:03 -!- iconmaster has joined.
16:38:40 -!- AnotherTest has quit (Ping timeout: 246 seconds).
16:38:44 -!- iconmaster has quit (Ping timeout: 246 seconds).
16:43:52 -!- iconmaster has joined.
16:47:31 -!- nooga has quit (Ping timeout: 244 seconds).
16:56:23 -!- iconmaster has quit (Ping timeout: 245 seconds).
17:02:15 <Gregor> Facebook just told me that somebody I don't know accepted /my/ friend request.
17:05:39 <ion> :-D Someone had made a friend request to me ages ago, but i didn’t recognize the name immediately and didn’t get around to looking up who he was until today. It turns out it was you and i accepted.
17:07:04 <Gregor> Dafaq, why would I friend you, I hate you. I MEAN, NOTHING.
17:07:26 <Gregor> More tot he point, I don't remember ever doing this, or for that matter, ever friending anyone from #esoteric other than pikhq >_>
17:08:46 <shachaf> Hmm, now I know what ion looks like. Just like I imagined, really.
17:08:55 <pikhq> Gregor: Only me? Hah.
17:09:23 <Gregor> I don't have “friends”, I only have “acquaintances I tolerate”.
17:09:42 <ion> Date: Wed, 6 Jun 2012 17:36:14 -0700 Subject: Gregor Richards haluaa olla kaverisi Facebookissa.
17:10:06 <shachaf> haluaa olla kaferisi Facebookissa
17:10:21 <Gregor> June? May as well be 1960.
17:10:47 <pikhq> haruā orra kaferishi fēsubūkkissa
17:11:29 -!- Phantom__Hoover has joined.
17:11:35 -!- iconmaster has joined.
17:12:39 -!- ogrom has joined.
17:14:04 <elliott> pikhq: Hey, if I break my system, fix it.
17:14:15 <elliott> I have no idea what I'm doing, so I probabiy will break it.
17:14:34 -!- Phantom_Hoover has quit (Ping timeout: 240 seconds).
17:14:38 <pikhq> Computers love coffee.
17:14:45 <pikhq> You should pour a whole pot on it.
17:15:29 <elliott> pikhq: Weeeell, I'm removing my separate ext2 /boot partition and merging it into my JFS / at the same time as switching from MBR to GPT.
17:15:43 <elliott> To make room for a 2 MiB partition GRUB2 apparently needs on GPT systems.
17:15:54 <elliott> So I'm fairly certain my system is going to stop booting in a few minutes.
17:16:14 <olsner> that does sound like something almost guaranteed to break your system
17:16:22 <pikhq> You *might* want to make a boot disk of some sort.
17:16:41 * FreeFull wonders what's up with people trying out GPT
17:16:46 <elliott> pikhq: I think I have a USB drive with an Arch thing on it.
17:16:59 <elliott> So probably I can just boot from that if it goes wrong.
17:17:03 <elliott> No idea where it *is*, mind you.
17:18:05 <Gregor> FreeFull: I switched to GPT when my 3TB HD forced me to ;)
17:18:19 <elliott> OK; writing new GUID partition table (GPT) to /dev/sda.
17:18:19 <elliott> Warning: The kernel is still using the old partition table.
17:18:19 <elliott> The new table will be used at the next reboot.
17:18:19 <elliott> The operation has completed successfully.
17:18:25 <elliott> Hey does anyone know how I can force the kernel to reload the partition table?
17:18:28 <pikhq> FreeFull: MBR is almost tolerable if you've got a small enough HD.
17:18:30 <elliott> It's, uh, currently looking at my MBR table.
17:18:34 <FreeFull> Gregor: How did it force you to?
17:18:35 <pikhq> FreeFull: And besides, this is new interestingness.
17:18:41 <elliott> But the problem is I want GRUB to see the partition I just made or something.
17:18:45 <elliott> FreeFull: mbr is limited to 2 tb
17:18:48 -!- micrypt has joined.
17:18:49 <pikhq> elliott: Erm. I dunno.
17:19:09 -!- brownies has joined.
17:19:20 <HackEgo> MICRYPT: 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.)
17:19:24 <FreeFull> elliott: Can't you just have the boot partition setup
17:19:30 <HackEgo> BrOwNiEs: 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.)
17:20:15 <pikhq> FreeFull: That doesn't help you if you cannot *represent partitions for the whole drive*.
17:20:21 <FreeFull> Like what was done in the olden days when the BIOSes didn't support booting partitions larger than 528MB
17:20:23 <shachaf> micrypt: They don't actually ever talk about esolangs in here.
17:20:31 <elliott> Oh nice, apparently tar doesn't save symlinks.
17:21:16 <pikhq> elliott: There's other file types tar doesn't preserve, though.
17:21:54 -!- elliott has quit (Remote host closed the connection).
17:22:27 <pikhq> Legacy tar only handles normal files, hard links, and soft links. Sane tar also preserves character specials, block specials, directories, and FIFOs.'
17:23:20 -!- elliott has joined.
17:23:33 <pikhq> You'll note it still doesn't handle named sockets.
17:23:55 <elliott> It's not meant to be possible to fiddle about with that many things related to your boot process in one go and still have it work first try.
17:24:03 <elliott> Maybe I messed up the conversion and it's still MBR???
17:24:35 <elliott> Okay, nope, gdisk says it's GPT.
17:24:58 <pikhq> I bet dmesg says what the kernel thinks it is.
17:25:02 <olsner> you can have both GPT and MBR at the same time
17:25:11 <elliott> olsner: Yeah, but gdisk says I just have the protective MBR thing.
17:25:29 <elliott> pikhq: dmesg | grep -i gpt gives nothing.
17:25:36 <FreeFull> olsner: Linux doesn't like that though
17:25:44 <elliott> But the same applies to | grep -i mbr.
17:25:46 <FreeFull> FreeBSD does but who uses FreeBSD? =P
17:25:47 <olsner> I remember trying to set MBR back up after accidentally the GPT
17:26:28 <elliott> FreeFull: Huh? Linux manages it fine, I thought.
17:26:58 <pikhq> FreeBSD does it a bit better than Linux, but Linux doesn't *utterly break* on hybrid MBR configs.
17:27:16 <pikhq> Linux only uses MBR or GPT, not both.
17:27:35 <elliott> Sort of upset that I didn't get any fancy breakage.
17:28:05 <elliott> Maybe I should enable [testing] for more fun.
17:28:06 <pikhq> Actually, from what I gather, the easiest way to check would be to see if /boot is a separate filesystem.
17:28:28 <olsner> yeah, partitioning and boot loader setup is so boring when it just works
17:28:43 <elliott> pikhq: Yeah, there's no /boot mounted, it's definitely GPT...
17:28:53 <olsner> how bad is it to use an SSD with the wrong sector size?
17:29:17 <elliott> olsner: I think it's just slightly slower.
17:29:34 <elliott> Oh hey, the ghc package got updated!!!!!!!!!!!!!!!!!!!
17:29:39 <elliott> Time to blow away /opt/ghc and break everything.
17:29:41 <FreeFull> elliott: It will show the GPT partitions, or the MBR partitions, but not both at the same time
17:30:12 <FreeFull> And which ones it shows depends on if you booted it from GPT or MBR
17:33:02 <elliott> Eurgh, the package only provides Cabal 1.16.0.
17:33:07 <elliott> So I guess upgrading probably won't work.
17:33:20 <elliott> FreeFull: That seems OK though, since in hybrid configurations you keep the two tables synchronised.
17:33:56 <pikhq> At least, to the extent possible.
17:34:22 <pikhq> (presumably if you're using a gigantic HD in a dual boot config, you don't try to keep track of your partition that goes over the 2TB boundary)
17:35:42 * elliott is just going to not install Windows until using a UEFI system.
17:36:05 <pikhq> Probably the best option.
17:37:10 <elliott> Not like the things I'd want a Windows partition for are viable on this machine, anyway... mostly games and this system is sort of crappy.
17:42:01 -!- brownies has left.
17:45:44 -!- micrypt has left.
18:07:02 <pikhq> It is, uh. Technically possible to use EFI with MBR.
18:07:24 <pikhq> Your MBR partition scheme just needs to have an EFI system partition and it "just works".
18:08:09 <pikhq> About the only benefit I can see here is if you wanted to image a giant bunch of disks with the same OS install and have it work identically on BIOS and EFI.
18:08:33 <pikhq> And, hypothetically speaking, it's an OS that doesn't boot from GPT on BIOS.
18:17:03 <pikhq> *cough*fuckyouwindows*cough*
18:17:43 -!- micrypt has joined.
18:26:12 <HackEgo> ＭＩＣＲＹＰＴ： ＷＥＬＣＯＭＥ ＴＯ ＴＨＥ ＩＮＴＥＲＮＡＴＩＯＮＡＬ ＨＵＢ ＦＯＲ ＥＳＯＴＥＲＩＣ ＰＲＯＧＲＡＭＭＩＮＧ ＬＡＮＧＵＡＧＥ ＤＥＳＩＧＮ ＡＮＤ ＤＥＰＬＯＹＭＥＮＴ！ ＦＯＲ ＭＯＲＥ ＩＮＦＯＲＭＡＴＩＯＮ， ＣＨＥＣＫ ＯＵＴ ＯＵＲ ＷＩＫＩ： ＨＴＴＰ：／／ＥＳＯＬＡＮＧ
18:27:42 <pikhq> God, zenkaku Roman.
18:28:32 <elliott> Hmm, I wish I could get rid of all these 32-bit libraries.
18:28:45 <elliott> I guess Dwarf Fortress is still 32-bit only, though.
18:28:49 <pikhq> Have you considering replacing all with musl?
18:29:07 <pikhq> Speaking of, I really need to help finish glibc ABI compat on musl one of these days.
18:29:30 <elliott> pikhq: I don't see how musl helps.
18:29:41 <pikhq> elliott: It doesn't directly.
18:29:47 <pikhq> But it makes me happier.
18:29:55 <pikhq> That's like helping.
18:30:15 <elliott> I sort of doubt musl's binary compatibility is good enough to use all of Arch's glibc-based binary packages with it.
18:30:28 <pikhq> Hence why I need to finish it.
18:30:47 <pikhq> As far as I'm aware, the single largest gaping hole is the lack of the _FORTIFY_SOURCE functions.
18:31:15 <elliott> What about those warnings you get about needing dynamically-loaded stuff if you statically link with glibc? Does musl have to do anything to support that kind of stuff?
18:31:24 <elliott> Not that the packages are statically-linked, but you know what I mean.
18:31:37 <pikhq> What those are is functions in glibc that call dlopen...
18:31:40 <pikhq> In musl, they don't.
18:33:27 <pikhq> Of course, static linked glibc stuff won't work without those...
18:33:36 <pikhq> Betcha they could be stubbed without much trouble though.
18:34:09 <pikhq> (mind, nobody, and I mean nobody, *actually static links against glibc*. It makes GHC's runtime look minimal.)
18:35:12 <pikhq> 675k for "int main(){}" is a bit much.
18:35:15 <elliott> I statically link against glibc sometimes.
18:35:45 <pikhq> That's 675k needed for glibc's *startup routines*.
18:36:14 <pikhq> And that's just the text segment.
18:36:43 <pikhq> The equivalent in musl is 1.5k for my current checkout.
18:36:46 <shachaf> 11:35 <shachaf> Zeno knew how to halve his cake and eat it too.
18:38:10 <pikhq> Went down a couple bytes when I did a git pull...
18:38:25 <olsner> I got 883kB for an empty main with whatever libc ubuntu has
18:38:42 <olsner> 854 with -Wl,--gc-sections
18:39:03 <pikhq> 650 B for a dynamic linked program.
18:40:14 <pikhq> olsner: I'm arbitrarily just measuring the text segment.
18:40:58 -!- epicmonkey has quit (Ping timeout: 245 seconds).
18:41:16 <pikhq> If you only look at the binary size proper, static linked musl is the smallest.
18:41:41 <olsner> there's no real reason that program should have anything more than an exit syscall in its text segment
18:41:58 <pikhq> olsner: Well, you still need the startup code.
18:42:22 <pikhq> How it works is _start runs, it sets things up for main, and only *then* does an exit syscall.
18:42:31 <pikhq> It's slightly more work than just _exit.
18:42:39 <pikhq> Still shouldn't be *that* much.
18:42:52 <olsner> yes, but even that is not necessary for this program
18:43:04 <pikhq> The linker can't magically know that main is a no-op.
18:43:56 <pikhq> All it knows is it needs the startup code and the program's objects to be glued together into a single binary.
18:44:10 -!- epicmonkey has joined.
18:44:13 <olsner> sure it can, but it's somewhat reasonable that the linker doesn't have that magic built in
18:45:03 <pikhq> s/somewhat/perfectly/
18:46:04 <pikhq> The linker is getting handed headered machine code. You need a disassembler and some sort of assembly logic analyzer to determine what main is doing, and if any bit of the startup code is necessary.
18:46:20 <pikhq> That is to say, you'd basically be shoving compiler infrastructure into the linker.
18:47:34 <pikhq> Does not quite work like this, actually.
18:48:11 <pikhq> Here's the linker's knowledge of LTO: it's passed a .so that should be loaded, and should have objects filtered through it.
18:48:23 <pikhq> And, also, that .so isn't even looking at the machine code blob in the objects.
18:48:39 <pikhq> It's looking at compiler IR that was shoved in the object files.
18:49:21 <pikhq> It's more-or-less just deferring optimization passes and code generation until all the translation units are available at once.
18:49:28 <elliott> i don't see much wrong with putting compiler logic in the linker, though
18:49:34 <elliott> if anything i'd rather there be less compiler logic in the compiler
18:56:55 <elliott> I wonder if Rust binaries are standalone enough to work as PID 1.
19:02:39 <HackEgo> /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: W: not found
19:05:08 -!- epicmonkey has quit (Ping timeout: 245 seconds).
19:07:53 <elliott> pikhq: Is poppler or mupdf better these days?
19:10:00 -!- FreeFull has quit (Quit: New tmux).
19:10:28 <Gregor> mupdf: The muppet defragmenter.
19:15:31 -!- FreeFull has joined.
19:19:17 <Sgeo> http://www.fileformat.info/info/unicode/char/1f648/index.htm
19:19:23 <Sgeo> Uh. The image they give seems to be wrong
19:21:41 <elliott> it's almost like the font they use doesn't have perfect unicode 6 coverage
19:25:06 <Sgeo> Aww, according to Wikipedia, Mayan symbols are not yet in Unicod
19:25:26 <Sgeo> Unicod. Universal fishing.
19:26:40 <Sgeo> Gregor, I don't count as being from #esoteric ? Or did I friend you?
19:26:56 <Sgeo> Also, I should figure out how to contact cpressey
19:27:23 <elliott> It's literally impossible to contact Chris.
19:28:06 <fizzie> Sgeo: There is, however, a tentative allocation for Maya hieroglyphs in the SMP roadmap: http://www.unicode.org/roadmaps/smp/
19:28:09 <Sgeo> I don't see an email address on his about page
19:29:08 <fizzie> Sgeo: I think it went something like so that you write "cpressey" in blood on a mirror at midnight, and then... something. Some chanting was involved.
19:29:49 <fizzie> Oh right, write it three times. Don't remember the other details.
19:32:05 <Sgeo> "And, because OpenZz thinks you are the input to a compiler, accumulating 10 of these errors will actually cause OpenZz to quit!"
19:32:19 <Sgeo> So you can have 9 syntax errors and OpenZz will not quit?
19:32:25 <Sgeo> That seems bizarre
19:33:19 <fizzie> I seem to recall some other compiler having a "more than K errors, giving up" rule too.
19:33:52 <fizzie> Though I don't know what "quit" means in that context.
19:34:31 <fizzie> I would've assumed "stops trying to make sense out of the file", as opposed to trying to report all errors.
19:34:38 <Sgeo> http://catseye.tc/projects/zzrk/README.markdown.html
19:36:03 <fizzie> Well, it sounds a bit like that, then.
19:37:04 <fizzie> Like doing -fmax-error=10 to gcc.
19:37:17 <fizzie> -fmax-errors=10, I mean.
19:37:57 <Sgeo> But surely an error is something that should make the compiler quit immediately?
19:38:49 <Sgeo> I don't know how often he comes in here, so don't know how long a @tell would take
19:42:59 -!- nooga has joined.
19:43:22 <fizzie> Sgeo: I think most compilers actually would instead read the whole file and try to report all errors at once, instead of quitting at the first, and that's what people actually want.
19:43:32 <elliott> Yes, gcc doesn't give up at the first error or anything.
19:43:33 <Sgeo> Ah, good point
19:43:42 <fizzie> There's the whole field of error recovery in parsing to make that possible.
19:45:20 <fizzie> But there was some tool with a finite default to '-fmax-errors'-like behaviour, I just can't remember which it was. The error message was mostly neutral, but there was a slight hint of disparaging the programmer.
19:45:39 <fizzie> The limit was something like 500 or 1000 errors, after all.
19:47:23 <fizzie> GCC just says "compilation terminated due to -fmax-errors=2." which is kind of boring.
19:47:26 <Sgeo> disparaging the programmer made me think INTERCAL, but I'm sure it would have a lower limit
19:49:09 -!- ais523 has joined.
19:49:17 <fizzie> No, this was some "serious" tool.
19:49:51 <fizzie> Oh, I'm probably just thinking of Valgrind's "go fix your program" message.
19:50:17 <fizzie> ==21573== More than 10000000 total errors detected. I'm not reporting any more.
19:50:20 <fizzie> ==21573== Final error counts will be inaccurate. Go fix your program!
19:50:30 <fizzie> That thing. It doesn't even quit.
19:50:59 <fizzie> Or at least mingling that message and the as-yet unidentified compiler.
19:51:45 <ais523> Valgrind's "go fix your program" is quite confusing
19:51:56 <ais523> given that 10 million errors can happen very quickly, and often stem from the same cause
19:52:05 <ais523> btw, I just received spam saying I'd been successfully unsubscribed
19:52:13 <Sgeo> "This seems difficult at least, and not really necessary if we're willing to forego writing "Hunt the Wumpus" in Burro"
19:52:15 <ais523> my guess is it's angling for replies asking wtf is going on
19:52:23 <ais523> which would confirm the address as genuine
19:52:31 <Sgeo> Aww, I wanted to write Hunt the Wumpus in Burro
19:52:36 <Sgeo> (note: not really)
19:52:37 <ais523> also, which one's Burro?
19:52:45 <ais523> I'm sort-of losing track of all the esolangs we have nowadays
19:52:58 <Sgeo> Brainfuck-like where every program has an antiprogram that when concatenated to the program results in a no-op
19:53:12 <Sgeo> http://catseye.tc/projects/burro/doc/burro-1.0.html
19:53:47 <ais523> I remember talking to cpressey about it, because there was a mistake in it
19:54:31 <fizzie> JScript .NET has a syntax error "JS1197: Too many errors", but I'm sure I haven't used that.
19:55:28 <Sgeo> ais523, how do I contact cpressey, there's a mistake in his REDGREEN implementation, I think
19:56:02 <fizzie> Oh, and Apple MPW C has "Too many errors on one line (make fewer)".
19:57:40 <Sgeo> <bb'> ≡ <e> ≡ <>
19:57:52 <Sgeo> That line should NOT have confused me as much as it did, even if it was only for seconds
19:58:17 -!- kinoSi has quit (Read error: Connection reset by peer).
19:58:44 -!- kinoSi has joined.
20:01:19 <Sgeo> ais523, Burro question: Is it mandatory to have a / inside (), or can you have (e) and no / in there
20:02:01 <ais523> I think you'd need a / or a \ because otherwise it wouldn't make sense
20:02:24 <ais523> fizzie: I'm pretty sure I've used compilers that stop at a configurable number of errors, with "error: too many errors"
20:02:42 <ais523> I think it's because they have bad error recovery and errors after the first few don't actually make sense anyway
20:02:47 <ais523> or else that they fear a lot of duplicates
20:03:17 <fizzie> ais523: So am I, but I can't remember which ones they were. (GCC has an option for it too, but it is off by default.)
20:04:08 <ais523> hmm, Borland C++ 4, perhaps
20:04:13 <ais523> the compiler I grew up with
20:04:23 <Sgeo> ais523, is there a way to contact cpressey?
20:05:00 -!- Phantom_Hoover has joined.
20:07:14 -!- Phantom__Hoover has quit (Ping timeout: 255 seconds).
20:08:28 <ais523> Sgeo: I think he has an email address
20:08:48 * Sgeo is having trouble finding it
20:09:05 <ais523> does the email link on his userpage work?
20:09:12 <ais523> actually, probably not, given how often catseye moves
20:10:01 <Sgeo> Oh hey REDGREEN is on Github
20:10:06 <Sgeo> I could just do something there
20:10:27 <fizzie> They seemed to have identical commits.
20:10:47 <fizzie> It smells like some kind of an automatic thing to have both.
20:11:32 <ais523> you can commit to multiple repos
20:11:45 <ais523> the fact that they're different DVCSes shouldn't matter
20:12:32 <Sgeo> Do I just gratuitously add myself to the comment at the top of the file in my fork, or just make the change?
20:13:18 <fizzie> ais523: Oh, Bitbucket also supports Git? I thought they only did Mercurial.
20:14:20 <fizzie> In which case there would have needed to be some kind of a thing, or a person doing all commits twice with the same messages and so on.
20:15:37 <fizzie> By "thing" I mean something like that hg-git that people use.
20:16:53 -!- hagb4rd has quit (Quit: hagb4rd).
20:17:09 <Sgeo> "Sgeo opened this pull request in a few seconds
20:17:15 <Sgeo> You know, I should probably test it...
20:17:57 -!- ogrom has quit (Quit: Left).
20:18:09 -!- pikhq_ has joined.
20:18:28 -!- pikhq has quit (Ping timeout: 248 seconds).
20:19:45 <Sgeo> "Madison, a Term-Rewriting Proof-Checker"
20:19:56 <Sgeo> Misread that as time-travelling proof-checker
20:20:27 <Arc_Koen> "let's travel back in time to check if fermat really had a proof"
20:20:49 <Arc_Koen> (that's probably gonna be very disappointing)
20:21:46 -!- atriq has joined.
20:23:24 <Sgeo> fungot needs to be rewritten in Flobnar
20:23:25 <fungot> Sgeo: you'll be doing summer homework. *then* i can come up with a good solution
20:25:36 <Sgeo> Hmm. So Flobnar is more functional than Befunge?
20:25:51 <Sgeo> I'm trying to understand the distinction between a moving IP and "evaluate what's over there"
20:28:07 <atriq> "Evaluate what's over there" has a sort-of intrinsic "move back and do the next thing afterwards"
20:28:31 <Sgeo> The arithmatic stuff helps me understand
20:29:54 <fizzie> There is currently no text in this page. :/
20:30:10 <elliott> - This frequent relocation of our web address was done specifically to
20:30:11 <elliott> - please [[Phantom Hoover]].
20:30:21 <Sgeo> "(And it's not implemented yet, but never mind that.)"
20:30:31 <elliott> Phantom_Hoover: http://catseye.tc/node/Phantom%20Hoover.html
20:32:05 <Sgeo> Blah, I think the introspection and self-modification should perhaps use relative coordinates
20:33:15 -!- jiella has joined.
20:33:22 <atriq> And so mysterious!
20:33:37 <Phantom_Hoover> never did i think i would rub shoulders with douglas hofstadter
20:33:50 <Phantom_Hoover> i wonder if he'll lend me some of his harem of beautiful french women now
20:34:16 <Sgeo> Maybe I should have opened an issue than just issue a pull request
20:34:21 <atriq> Harems are overrated
20:35:46 <Sgeo> What is Befunge-93's :
20:36:12 <atriq> Duplicate, I think
20:36:58 <atriq> One of my friends hates me in her sleep
20:37:29 <fizzie> (At least that's always been my personal logic for it.)
20:38:46 <atriq> I'm kind of bored right now
20:39:07 <Phantom_Hoover> see, if you had a harem of beautiful french women you wouldn't be bored
20:39:12 <fizzie> !befunge "!olleh">:#,_@
20:39:42 <atriq> Phantom_Hoover, you're forgetting two things
20:39:54 <atriq> And I don't speak French
20:40:16 <atriq> Beyond "Ou est le W/C? Je voudrais une baguette!"
20:42:19 <olsner> hmm, I think I need new glasses
20:42:29 <Sgeo> I think that would make sense, but not sur
20:42:43 <fizzie> Yes. Well,it has a 0 on stack at @.
20:43:44 <Sgeo> But $ pops the stack
20:43:56 <fizzie> Yes, but only one of the 0s pushed by :.
20:44:00 -!- Phantom__Hoover has joined.
20:44:09 <Sgeo> Should the stack be considered to have an infinite number of 0s?
20:44:52 <fizzie> If you like. Poppig an empty stack yields 0.
20:45:38 <fizzie> The spec (at least 93) isn't explicit about ':' on empty, but pushing two 0s is I believe most common.
20:45:39 <Sgeo> The example cat program
20:45:42 <atriq> Phantom__Hoover, you've cloned yourself!
20:45:52 <Sgeo> Wouldn't that .. n/m
20:46:48 <fizzie> 98 seems to be explicit about :.
20:47:04 <fizzie> "pops a cell off the stack, then pushes it back onto the stack twice, duplicating it".
20:47:06 * Sgeo was talking Flobnar
20:47:20 <fizzie> Well, I haven't seen Flobnar.
20:47:27 <Sgeo> http://catseye.tc/projects/flobnar/doc/Flobnar.falderal.html
20:47:34 -!- Phantom_Hoover has quit (Ping timeout: 260 seconds).
20:50:08 -!- atriq has quit (Ping timeout: 245 seconds).
20:52:08 -!- atriq has joined.
20:52:39 <fizzie> Ooh, that division by zero is funky.
20:52:58 <Sgeo> "Enter Velo, a vaguely Ruby-inspired scripting language which unifies strings with code blocks, and scripts with object classes."
20:54:48 <Sgeo> Well, partly like Tcl
20:55:17 -!- sivoais has quit (Ping timeout: 246 seconds).
20:56:46 -!- sivoais has joined.
21:21:47 -!- iconmaster has quit (Ping timeout: 255 seconds).
21:50:16 <atriq> Help I'm answering things on Stack Overflow badly
21:51:28 <atriq> Hey, I've got the Tumbleweed badge on Ask Ubuntu
21:51:46 <atriq> And I still have this problem
21:53:09 <nooga> i've spent my evening eating beef and drinking portuguese wine
21:53:17 <Sgeo> nooga, in Tcl, Everything Is A String
21:54:07 <pikhq_> nooga: This includes code.
21:54:25 <atriq> In Windows Powershell Strings Are Not Actually Strings
21:54:36 <nooga> what's windows powershell?
21:54:56 <atriq> Thingy included with Windows 7
21:55:05 <atriq> Supposed to replace CMD.EXE or whatever
21:55:53 <nooga> microsoft finally noticed that unixy OSes had a serious command line shell for, llike, 40 years?
21:56:08 <nooga> and tried to copy the idea?
21:56:28 <olsner> that and more and wrong, I think
21:56:34 <atriq> Tried to copy the idea and wrote it thinking that attaching hidden information to EVERYTHING was a brilliant idea
21:56:43 <olsner> but I've never used powershell so I don't really know
21:56:54 <nooga> i uess nobody ever used it
21:56:57 -!- atriq has changed nick to awesome.
21:57:05 <nooga> even the, gah, windows sever powerusers
21:57:10 -!- awesome has changed nick to atriq.
21:57:28 <Lumpio-> I used to answer stuff on stack overflow but I got bored when I got all the privileges :I
21:57:36 <Lumpio-> There's still badges but they're meh hard
21:58:25 <pikhq_> Powershell manages to be more verbose than you reasonably want in a shell.
21:59:06 <atriq> Today I actually tried to climb a building
21:59:22 <Lumpio-> It's C# with a funny syntax.
22:00:13 <Lumpio-> new-object Searcher::$search(call-static-method File::OfType("txt"))
22:00:18 <Lumpio-> Or something like that I don't know
22:00:40 <nooga> what a spectacular fail
22:01:24 <pikhq_> BTW, because both POSIX and common systems permit some massive brain damage, "echo *.txt|grep foo" does not quite do what you think it does.
22:01:48 <ion> http://thebestpageintheuniverse.net/c.cgi?u=youre_not_a_nerd
22:02:40 <Lumpio-> Ok here's an actual example
22:03:04 <Lumpio-> Actually I don't know how to search .__.
22:03:10 <Lumpio-> But here's how to open a log: $AppLog = New-Object -TypeName System.Diagnostics.EventLog -ArgumentList Application
22:04:07 <ion> pikhq: echo *.txt|grep foo seems to do what i thought it does.
22:04:29 <pikhq_> ion: Consider the file -n.txt
22:04:49 <Lumpio-> Solution: don't name files like that
22:04:54 <pikhq_> Or perhaps the file \n.txt (in C string notation "\\n.txt")
22:05:48 <ion> pikhq: That behavior is what i think it has. It’s not what i’d *hope* it had, but that’s another matter. :-P
22:05:51 <pikhq_> Lumpio-: Solution, smack whoever thought echo should be anything but int main(int argc, char**argv){for(int i = 1; i < argc; i++)puts(argv[i]);}
22:06:19 <nooga> atriq: was it a portable toilet?
22:06:23 <nooga> or something bigger/
22:06:29 <elliott> pikhq_: that does not solve the bigger problem at all
22:06:30 <pikhq_> (which, incidentally, remains a POSIX compliant echo.)
22:06:39 <atriq> I didn't get very far, though
22:07:04 <pikhq_> elliott: Nah, but it's *less* brain-damaged at least.
22:07:26 <pikhq_> 'Course, second you add spaces to your filenames, it breaks like crazy.
22:07:29 * ion kills the gazebo
22:07:31 <atriq> I got a couple meters off the ground before I realised that the column I was ascending led to a horizontal wall thing
22:07:33 <elliott> well why fix echo when there is also every other command with options
22:07:36 <elliott> pikhq_: btw that is not a compliant echo
22:07:38 <atriq> nooga, a SHOPPING CENTRE
22:07:47 <pikhq_> Oh, dur, puts does append newlines after everything.
22:08:30 <atriq> No, I did not climb ion
22:08:36 <pikhq_> int main(int argc, char**argv){for(int i = 1; i < argc; i++)printf("%s",argv[i]);puts("");}
22:08:56 <atriq> It was a leisure and retail complex
22:08:59 <nooga> i meant an arbour, not gazebo
22:09:07 <nooga> i'm not even sure what gazebo is
22:09:28 <atriq> A gazebo is like a tent on stilts
22:09:36 <elliott> main = fmap unwords getArgs >>= putStrLn
22:09:37 <pikhq_> elliott: Complaint echo does not have options, so what do you mean, "every other command with options"?
22:09:48 <atriq> Except without walls
22:09:55 <pikhq_> "Implementations shall not support any options." -- POSIX
22:10:08 <elliott> well i mean "echo *.txt" not doing the right thing compared to globbing with any command that accepts options
22:10:10 -!- Phantom_Hoover has joined.
22:10:11 <elliott> like cat *.txt or whatever
22:10:18 <elliott> and don't complain about cat's options or i'll just find a command with options you like :P
22:10:25 <elliott> pikhq_: btw echo *.txt actually works perfectly
22:10:25 <atriq> No, I did not climb an arbor
22:10:30 <atriq> Why would you even IMPLY that
22:10:36 <elliott> (of course spaces still mess it up)
22:10:55 <nooga> once I climbed an arbor
22:10:58 <ion> nooga: http://www.comedycorner.org/90.html
22:11:10 <pikhq_> The above actually works. Perfectly.
22:11:16 <nooga> an I was slightly drunk and then i fell because i was afraid of falling
22:12:13 -!- Phantom__Hoover has quit (Ping timeout: 246 seconds).
22:12:16 <elliott> if "cat -- *.txt" is an ok solution then surely "echo -- *.txt" is too
22:12:22 <elliott> i.e. the problem isn't actually that echo has options there
22:12:34 <pikhq_> Except echo does not support options. Not even --
22:12:40 <ion> Isn’t echo officially deprecated in the POSIX spec?
22:12:53 <pikhq_> I'm reading it. It is not deprecated.
22:13:32 <Gregor> The problem with the POSIX echo spec is that it doesn't match any existing implementation.
22:13:49 <elliott> it matches the one pikhq_ will write in one IRC line in three seconds
22:14:00 <ion> “New applications are encouraged to use printf instead of echo.” – echo(1posix)
22:14:14 <pikhq_> Gregor: POSIX-but-not-XSI echo does map very well to sufficiently old UNIX.
22:14:20 <nooga> ion: yeah, i was the Eric before a looked up the goddamn gazebo in my dictionary
22:14:29 <pikhq_> ion: That is informative text, not a statement of deprecation.
22:14:49 <Gregor> ion: That's not the same thing as deprecation.
22:14:55 <nooga> i wonder how's echo in Plan9
22:15:10 <pikhq_> Also, the 1posix man pages are not the spec, but a derivative thereof. :P
22:15:27 <shachaf> elliott: Fun fact: It's impossible to write echo for GHC.
22:15:32 <shachaf> A compliant echo, that is.
22:15:54 <ion> As an answer to your question: feel free to.
22:16:26 <pikhq_> Gregor: It appears to be POSIXly correct.
22:16:43 <pikhq_> Though it does handle "-n" as the first operand.
22:16:59 <ion> How about with -rtsopts=none?
22:17:22 <shachaf> ion: z: RTS options are disabled. Link with -rtsopts to enable them.
22:17:22 <pikhq_> No translations, though, so it's not XSI at all.
22:17:30 <ion> shachaf: duh
22:18:44 <pikhq_> Also: int main(int argc, char**argv){for(int i = 1; i < argc; i++)printf("%s%c",argv[i],i==argc-1?'\n':' ');}
22:20:15 <shachaf> pikhq_: That doesn't print a newline for an empty command line.
22:20:17 <elliott> pikhq_: fails to handle printf failing
22:22:44 <Arc_Koen> I wonder what would happen if Kipple was working with queues instead of stacks
22:23:48 <shachaf> elliott: You forgot to mention the undefined behavior bit where it doesn't return 0.
22:24:05 <olsner> main doesn't have to return
22:24:20 <shachaf> Isn't that a C++-only thing?
22:24:22 <pikhq_> shachaf: In C99, if you reach the end of main you return 0.
22:24:25 <pikhq_> elliott: Fine. int main(int argc, char**argv){for(int i = 1; i < argc; i++)if(printf("%s%s",argv[i],i!=argc-1?" ":"")<0)return 1;return putchar('\n') != EOF;}
22:24:38 <shachaf> Or maybe "things that make any sense" land.
22:24:50 <elliott> pikhq_: makes assumptions as to what a return value of 1 means
22:24:58 <olsner> also known as "ancient C" land?
22:24:59 <elliott> perhaps you want EXIT_FAILURE?
22:25:17 <pikhq_> elliott: No. POSIX echo returns some value greater than 0 on an error.
22:25:32 <pikhq_> elliott: It does not have to have semantics other than being greater than 0.
22:26:29 <elliott> pikhq_: what if argv[i] is NULL? can that ever happen?
22:27:40 <shachaf> It even handles the argc == 0 case.
22:28:03 <fizzie> shachaf: No, it's not. It's initialized to 1 even if argc <= 1.
22:28:26 <Sgeo> "and even if Chris doesn't respond to your pull request for three or four years, at least your fork will be out there and available to anyone else who might be interested."
22:28:30 <fizzie> Oh, it was inside the loop.
22:28:31 <Sgeo> Hmm, is that particularly likely?
22:28:46 <elliott> Sgeo: yes. chris hates pull requests. and the web.
22:28:59 <elliott> he quit IRC because he refuses to use any other protocol but ssh and gopher.
22:29:04 <elliott> this is also why he cannot be contacted
22:29:40 <elliott> pikhq_: anyway I guess that might be valid
22:29:48 <shachaf> I bet it's still not compliant.
22:30:03 <elliott> it seems fairly airtight though
22:30:10 <elliott> i was tryign to think of overflow behaviours with argc but that seems ok too
22:30:23 <shachaf> Wait, which echo specification are we talking about here?
22:30:38 <Sgeo> All of them. Simultaneously.
22:30:53 <olsner> POSIX echo is boring though, judging by the echo(1posix) man page
22:30:55 <pikhq_> http://pubs.opengroup.org/onlinepubs/9699919799/utilities/echo.html
22:31:17 <shachaf> Oh, part of that spec is XSI.
22:31:18 <pikhq_> -n as the first operand is implementation-defined, as are instances of backslash in the operands.
22:31:32 <pikhq_> And I define them to be "they are treated as ordinary strings".
22:31:40 <elliott> pikhq_: looks like you have to deal with locales
22:31:45 <elliott> given ENVIRONMENT VARIABLES
22:32:25 <fizzie> Not if you have no diagnostic messages.
22:32:34 <pikhq_> elliott: I do so correctly, by having no diagnostic messages.
22:32:45 <shachaf> "Determine the locale for the interpretation of sequences of bytes of text data as characters (for example, single-byte as opposed to multi-byte characters in arguments)."
22:32:54 <elliott> shachaf: yes but that's XSI :(
22:33:13 <elliott> seems like this is more likely to be an invalid C program than non-compliant with the POSIX spec
22:33:29 <shachaf> Well, you need #include <stdio.h>
22:33:33 <shachaf> But I guess that was implied.
22:33:37 <pikhq_> I was taking that as read.
22:42:27 <pikhq_> int main(int a,char**b){int i=1;for(;i<a&&printf("%s%s",argv[i],i!=b-1?" ":"");i++);return i==b&&putchar('\n')!=EOF;}
22:44:20 <elliott> is continuing to printf after an error ok
22:44:38 <olsner> are you allowed to execvp("/bin/echo", argv)?
22:44:51 <pikhq_> olsner: No, because /bin does not need to exist.
22:44:59 <shachaf> I'm pretty sure that won't compile because argv isn't in scope.
22:45:39 <pikhq_> /bin/echo is not guaranteed to exist, nor is /bin.
22:46:12 <elliott> int main(int a,char**b){while(*b)printf("%s%c",*b,*b++?' ':'\n')>0||exit(1);}
22:46:46 <pikhq_> I'm pretty sure the argv vector is not NULL terminated.
22:47:02 <shachaf> argv[argc] == NULL is in the C spec.
22:47:14 <pikhq_> olsner: It is strictly conformant for all POSIX utilities to be shell builtins, and for the shell itself to not actually exist in the filesystem at all.
22:47:46 <shachaf> elliott: exit() isn't golf if you count the #include
22:47:54 <olsner> pikhq_: ah, that kills any exec-based solution
22:48:14 <Sgeo> I wonder if I should learn how to use yoob
22:48:24 <olsner> elliott: that only works if the shell is in the filesystem?
22:48:37 <olsner> or are you always able to exec sh?
22:49:00 <shachaf> elliott: Wait, what's your correct program?
22:49:10 <elliott> int main(int a,char**b){while(*b)printf("%s%c",*b,*++b?' ':'\n')>0||exit(1);}
22:49:25 <elliott> technically both mine and pikhq_'s programs are wrong
22:49:29 <elliott> because printf can print only some output sometimes
22:50:19 <shachaf> elliott: Yours is wrong because it prints argv[0]
22:50:59 <pikhq_> According to POSIX, all of the standard utilities must be executable via exec(), even if they are implemented as shell builtins.
22:51:06 <shachaf> Also doesn't handle the no argument case.
22:51:21 <pikhq_> Except the "special" builtins, which needn't be.
22:51:30 <shachaf> elliott: Oh boy, when you run your code with argc==0, it prints the environment!
22:51:44 <shachaf> (On Linux/glibc/etc., that is. It's undefined behavior, of course.)
22:51:55 <shachaf> Also if you fix the thing that makes it not even compile.
22:52:08 <pikhq_> So, it's valid to have the C compiler as a shell builtin, but a POSIX system must execute that shell builtin on exec("cc")
22:52:10 <elliott> is '\n' guaranteed to equal 10
22:52:37 <pikhq_> There are no charset guarantees beyond what characters must exist.
22:52:44 <elliott> shachaf: what makes it not compile
22:52:50 <shachaf> pikhq_: What about '\0' == 0?
22:52:54 <elliott> pikhq_: that's not true. 'a' to 'z' are required to be adjacent IIRC
22:53:08 <elliott> and i think '0'+n has to work?
22:53:09 <elliott> shachaf: is that not valid?
22:53:22 <shachaf> Instead of figuring it out I replaced it with an if().
22:53:24 <atriq> NAME FOUR FICTIONAL JAKES
22:53:27 <pikhq_> shachaf: You're handing a numeric literal there.
22:53:55 <pikhq_> shachaf: Which of course works. '\0' is (char)0
22:54:16 <shachaf> OK, but are there guarantees like 'a' != 0?
22:54:27 <pikhq_> elliott: *Very* certain '0'+n doesn't have to work.
22:54:36 <pikhq_> shachaf: Don't think so.
22:54:43 <elliott> pikhq_: i think you are wrong
22:54:45 <pikhq_> elliott: Heck, that's not even true of EBCDIC systems.
22:54:53 <elliott> I know there are some ordering guarantees
22:55:12 <pikhq_> Nor is 'a' through 'z' adjacency...
22:55:31 <shachaf> To specify an integer character constant containing
22:55:31 <shachaf> the two characters whose values are 0x12 and '3', the construction
22:55:31 <shachaf> '\0223' may be used, since a hexadecimal escape sequence is terminated
22:55:31 <shachaf> only by a non-hexadecimal character. (The value of this two-character
22:55:32 <shachaf> integer character constant is implementation-defined also.)
22:55:36 <shachaf> I didn't know that was in C89.
22:55:45 <shachaf> Implementation-defined, but still.
22:55:49 <pikhq_> Oh, 0 through 9 should be adjacent.
22:56:12 <pikhq_> There are no *other* ordering guarantees in POSIX.
22:56:49 <pikhq_> That is the only character with a fixed value.
22:58:35 <shachaf> pikhq_: What about the guarantee that if you order a copy of the POSIX spec, they'll send it to you?
22:58:39 <shachaf> I should hope that's a guarantee.
22:58:48 <pikhq_> (POSIX volume 1, 6.1 "Portable Character Set")
22:58:57 <pikhq_> shachaf: I don't think that's *in* POSIX.
22:59:07 <shachaf> It's still a POSIX guarantee, though.
22:59:18 <pikhq_> I said "guarantees in POSIX".
22:59:25 <olsner> ordering the spec should be undefined behavior
22:59:26 <Sgeo> Oozlybub and Murphy is great so far
22:59:27 <pikhq_> Not guarantees *about* POSIX.
23:02:29 <Sgeo> TRIVIA PORTION OF SHOW
23:02:29 <Sgeo> WHO WAS IT FAMOUS MAN THAT SAID THIS?
23:02:29 <Sgeo> A) RONALD REAGAN
23:02:29 <Sgeo> B) RONALD REAGAN
23:02:29 <Sgeo> B) RONALD STEWART
23:02:41 <Sgeo> ^^ in the Oozlybub and Murphy docs
23:02:55 <shachaf> Sgeo: What does that have to do with Clojure?
23:03:02 <olsner> the correct answer is D) CAPTAIN PICARD
23:03:29 <elliott> Sgeo: do you not "get" cpressey
23:03:55 <shachaf> elliott: WHAT IF!!! i do not "get" cpressey :'(
23:04:03 <shachaf> I don't even know who that is.
23:04:11 <elliott> shachaf: you don't get anything
23:04:13 <Sgeo> BURN THE BLASPHEMER
23:04:15 <elliott> cpressey is the only esolanger worth a damn
23:04:21 <elliott> well ais523 and oerjan too
23:04:27 <olsner> shachaf: http://www.catseye.tc/projects/oozlybub-and-murphy/doc/oozlybub-and-murphy.html
23:05:12 <olsner> hmm, that reminds me, I should start working on yonguilexiphonaugh some day
23:05:48 <ais523> elliott: cpressey is more prolific than me, and oerjan is more profound
23:06:17 <ais523> and AmBF/AmbiF (it has two official names) has really got me annoyed, because I'm still not sure if it's TC
23:06:23 <ais523> the only thing it's missing is arbitrary effect at arbitrary point
23:06:28 <elliott> I guess fizzie is good too for fungot
23:06:28 <fungot> elliott: did they make it easier to use than a keyboard.
23:06:29 <ais523> and you know how annoying /those/ things are to prove TC
23:06:43 <elliott> but every esolanger beyond that is unnecessary
23:07:00 <shachaf> elliott: It's a good thing I'm not an esolanger, isn't it?
23:07:28 <elliott> ais523: Is arbitrary effect at an arbitrary point even well-defined?
23:07:31 <elliott> I've always been suspicious of it.
23:07:34 <Sgeo> arbitrary effect at arbitrary point is what's needed for quineness, right? Or, well, providing arbitrary effect at arbitrary point is necessary for a proof of the existence of a quine to hold, in addition to TC
23:07:45 <Sgeo> Am I understanding correctly?
23:07:57 <ais523> but it's a good explanation to use for "I've been able to create two bignum counters in this with increment, decrement, and zero test and I'm /still/ not sure it's TC"
23:08:16 <ais523> elliott: I basically consider the TCness rules as not a mathematical definition but a heuristic
23:09:22 <ais523> elliott: anyway, this nondeterministic BF variant is definitely TC if you add conditional goto
23:09:26 <ais523> I'd explain to oerjan but he isn't here
23:09:51 <elliott> also the word I was thinking when I said "unnecessary" was actually "inessential"
23:10:06 <ais523> elliott: it's like BF, with a tape infinite both ways and bignum elements; and + does - instead half the time, and < does > instead half the time (and vice versa)
23:10:13 <ais523> (i.e. + and - are the same, and < and > are the same)
23:10:24 <elliott> ais523: is this on the wiki?
23:10:31 <ais523> I should probably add it some time but I'm lazy
23:10:42 <ais523> anyway, the name is /pronounced/ like "am BF" or "ambi-F"
23:10:48 <ais523> and it is spelled in such a way that you can't tell which it is, which is impossible
23:10:52 <ais523> so I'm having issues titling the page
23:11:07 <shachaf> ais523: What's the point of the "vice versa"?
23:11:12 <olsner> at least it's not spelled ///
23:11:45 <elliott> ais523: make it so that [ does ] half the time too
23:11:55 <shachaf> elliott++ makes a good point.
23:12:05 <ais523> shachaf: oh, I did get rid of , and .
23:12:07 <ais523> just forgot to tell anyone
23:12:14 <ais523> and that would make the thing even harder to deal with
23:12:23 <elliott> it doesn't harm semantics, right?
23:12:51 <shachaf> You can keep . as long as it does , half the time.
23:13:09 <ais523> but you can't distinguish between n and -n
23:13:26 <ais523> and above 2, can't distinguish between two numbers if they have the same parity
23:13:33 <ais523> so you couldn't actually construct anything coherent to output
23:13:43 <ais523> at least, can't non-destructively distinguish
23:13:49 <ais523> and any distinction would be probabilistic as-is
23:13:52 <elliott> ais523: make it normalise the output to what's observable, then
23:13:58 <elliott> take abs as first step, obviously
23:14:13 <ais523> you should probably make it output bits, actually
23:14:17 <ais523> just caring about 0 and non-0
23:14:18 -!- jiella has quit (Quit: Leaving.).
23:14:25 <ais523> you can construct 0 and ±1 trivially
23:14:34 <shachaf> ais523: Just base it on boof
23:14:36 <elliott> IMO use * and @ for +- and <>, I think some other derivative does that
23:14:37 <ais523> there's some doubt as to whether you can reliably construct ±2, and if it's useful to do that
23:14:52 <ais523> I guess we should work out the IPA for it
23:14:54 <ais523> and use that in the title
23:14:57 <shachaf> elliott: Why would you use @ for <>?
23:16:24 -!- atriq has quit (Quit: Leaving).
23:16:35 <shachaf> Does æ mean the thing which is pronounced as "ea"?
23:17:08 <ais523> shachaf: it's basically what English people do when trying to pronounce short "a"
23:17:37 <Arc_Koen> so i'm trying to run the example brainfuck program used on the Dbfi page to test the self-interpreter's speeds
23:17:47 <shachaf> Wait, "ambiguous" isn't pronounced like "and".
23:17:59 <ais523> are according to Wiktionary
23:18:03 <Arc_Koen> and I'm trying to run it in a brainfuck interpreter written in thue, itself ran in a thue interpreter in java
23:18:04 <ais523> and I pronounce them the same way
23:18:09 <ais523> how do they differ in your language?
23:18:16 <Arc_Koen> and well it's been running for almost two minutes now
23:18:21 <ais523> Arc_Koen: you can expect that to take a while
23:18:22 <shachaf> You = wiktionary = wrong :-(
23:18:35 <ais523> thue has an inherent O(n) slowdown
23:18:38 <Arc_Koen> I have no idea what that brainfuck program was supposed to do
23:18:41 <ais523> shachaf: so how do they differ in your accent?
23:18:48 <Arc_Koen> ais523: I thought this slowdown could easily be reduce
23:18:50 <shachaf> ais523: I'd pronounce the 'a' in "ambiguous" like the 'u' in "up".
23:19:11 <Arc_Koen> by making a thue interpreter that, instead of selectng rules randomly, would "sort" them
23:19:19 <shachaf> It might not quite be like that, but it's certainly not like "and".
23:19:23 <ais523> although pretty much any vowel can be pronounced as anything if you go looking around for weird accents
23:19:32 <ais523> Arc_Koen: you still have to see where the rules match
23:19:35 <Arc_Koen> that is, always trying to apply rules from the top of the rule list, and always placing the last rule used on top of the list
23:19:38 <shachaf> I might pronounce it that was for a long word like "ambiguity".
23:19:45 <olsner> I've said ambiguous in so many ways now that I've forgotten the right way to say it (if I ever knew any)
23:20:04 <shachaf> ais523: I might just be wrong.
23:20:08 <Arc_Koen> writing a system to try first where the last change has been made might be slightly more complicated
23:20:14 <olsner> (took about three tries to get there)
23:20:17 <ais523> Arc_Koen: so basically each interpreter step is proportional to the size of the internal state
23:20:21 <ais523> that's what causes the O(n) slowdown
23:20:26 <ais523> Thutu has the same problem for the same reason
23:20:27 <Arc_Koen> but still, the sorting rules thing would make it much faster I think
23:20:32 <ais523> despite running rules in deterministic order
23:20:53 <olsner> is that the best that can be done, or is that just because the thue interpreters are not very good?
23:21:07 <ais523> olsner: it's possible to optimise it in many practically useful special cases
23:21:10 <ais523> but not in general, IIRC
23:21:17 <ais523> Arc_Koen: sure, benchmarks rarely do
23:21:22 <ais523> unless you're running Crysis or something
23:21:25 <Arc_Koen> well technically it tried to print the number 11001010
23:21:48 <ais523> (note: I don't have a BF version of Crysis to check; Lost Kingdoms is normally considered the closest substitute)
23:21:51 <Arc_Koen> but I only implemented the result of the . operator for ascii values 10 and 32-126
23:21:56 <shachaf> ais523: Do you pronounce "ambiguous" like "America"?
23:22:03 <ais523> has anyone here actually completed Lost Kingdoms, btw?
23:22:24 <Arc_Koen> the thue interpreter in java seems to be pretty slow too
23:22:34 <shachaf> So you pronounce "america" with "æ"?
23:22:40 <shachaf> And you pronounce "and" with "æ"?
23:22:40 <Arc_Koen> (it allows me to see the internal state)
23:23:00 <ais523> I only really have one sort of short "a" for the starts of words
23:23:07 <Arc_Koen> but it does allow escaped characters in replacement rules, which is nice
23:23:46 <shachaf> ais523: You pronounce "and" like "america"?
23:23:59 <elliott> and an america start with the same vowerl
23:24:15 <ais523> shachaf: only the first vowel
23:24:18 <shachaf> "vowerl" is the British spelling of "vowel", right?
23:24:37 <ais523> shachaf: "er" for "e" in "vowel" is quite a common accent in the UK
23:24:43 <ais523> although I'd typically pronounce it as "vowl"
23:25:23 <shachaf> elliott: Just like "colour", right?
23:26:55 <ais523> huh, New York just banned soda
23:27:04 <copumpkin> I thought it only banned large ones
23:27:24 <Arc_Koen> what do you mean "banned soda"?
23:27:46 <ais523> basically shops aren't allowed to sell more than 16oz at a time
23:27:51 <shachaf> copumpkin: Rec is pretty nifty.
23:27:55 <ais523> sadly, I'm British, and as such am not entirely sure how much 16oz is
23:28:00 <ais523> how much is that in metric?
23:28:01 <lambdabot> newtype Rec a = InR { outR :: Rec a -> a }
23:28:04 <Arc_Koen> wasn't that "unless the customer asks"?
23:28:10 <copumpkin> shachaf: oh, the non-termination type
23:28:19 <shachaf> elliott: Can you use newtype P = InP { outP :: forall a. P -> a } to write Y?
23:28:21 <olsner> I thought 1 oz was about 30 grams
23:28:27 <shachaf> copumpkin: Yes. It's Curry's paradox.
23:28:32 <ais523> Conversion result: 473.2 mL (milliliters), assuming oz is US fluid ounces.
23:28:46 <Arc_Koen> I thought they just banned the "if the customer doesn't say how much they want, then they want a huge one" policy
23:28:49 <ais523> 350ml is about standard for a small bottle of drink in the UK
23:28:54 <elliott> shachaf: hey how do I define P in lambdabot
23:29:03 <ais523> which corresponds to 12oz
23:29:06 <Arc_Koen> for the record our previous president made up a tax on soda
23:29:12 <olsner> 50cl is the normal soda bottle size here
23:29:34 <elliott> @let newtype P = InP { outP :: forall a. P -> a }
23:29:36 <ais523> olsner: I've seen 50cl bottles too, but they're rarely used for soda
23:29:45 <shachaf> elliott: @let doesn't work that way.
23:29:48 <ais523> (although they'd be marked 500ml in the UK)
23:29:53 <elliott> @let newtype Test = Test Int
23:29:56 <ais523> bottles are 500ml indeed
23:29:57 <elliott> @let newtype Test = Test { test :: Int }
23:30:00 <ais523> but cans are more common
23:30:11 <ais523> mostly I see fruit juice in the 500ml ones rather than soda, though
23:30:18 <copumpkin> http://upload.wikimedia.org/wikipedia/commons/2/21/Double_big_gulp.jpg
23:30:24 <Sgeo> "Because using a built-in operation store in this way saturates it, it cannot be used again. Typically you want to make a copy of the store first, and use that, leaving the built-in store unmodified."
23:30:39 <Sgeo> I can imagine how bad things would be if a mainstream language was like that
23:30:43 <Sgeo> http://catseye.tc/projects/xoomonk/doc/Xoomonk.falderal.html
23:30:44 <shachaf> copumpkin: Pft. That cup is still smaller than the person holding it.
23:30:47 <olsner> according to this article the ban is for restaurants, "concession stands" and "other eateries"
23:30:59 <Sgeo> I'm sure badly written libraries and programs would saturate important stores
23:31:06 <ais523> anyway: fun fact: at the University, I can buy a 500ml bottle of carbonated water for 95p
23:31:13 <ais523> and I can buy a 2l = 2000ml bottle for 55p
23:31:24 <ais523> even from the same shop, I think
23:31:33 <ais523> I'm not 100% sure of the logic behind this
23:32:07 <shachaf> ais523: «'Only you MUST eat them both, if you buy two,' said the Sheep.»
23:32:13 <Arc_Koen> for a second there I thought 'p' meant 'pound'
23:33:41 <ais523> Arc_Koen: that'd be massively expensive
23:33:54 <ais523> and a dollar is usually somewhere between 60p and 70p
23:34:05 <ais523> although it's probably somewhere really different right now because the world economy is in a mess
23:34:31 <ais523> it's prety much where you'd expect it to be
23:35:04 <olsner> I somehow read p as cent
23:35:19 <olsner> I guess I forgot how england and america are actually different countries that just happen to speak the same language
23:35:30 <ais523> shachaf: which country's currency is that?
23:35:45 <ais523> olsner: and also there's a distinction between England and the UK
23:35:49 <ais523> which many Americans don't seem to realise at all
23:36:00 <ais523> (some of them catch on when you mention the existence of Scotland)
23:36:40 <olsner> ais523: yeah, I will never learn how that distinction works
23:36:57 <Arc_Koen> well my internal money converter is still using the values from my childhood: 1$ ~= 10f (french francs), 1€ = 6.56f, 1 swiss franc = 4f, and 1$ is slightly less than 1€
23:37:13 <shachaf> ais523: To be fair, even people who live in the UK are confused on the finer points.
23:37:19 <shachaf> ais523: Including elliott.
23:37:42 -!- monqy has joined.
23:37:54 <shachaf> elliott: Remember when I confused you by talking about the Isle of Man or something?
23:37:56 <ais523> shachaf: you can start a mindboggling flamewar by asking whether England is a country or not
23:38:05 <ais523> mindboggling because nobody's entirely sure of which side of the argument they're on
23:38:14 <shachaf> ais523: I'm with Rosencrantz.
23:38:24 <Arc_Koen> (and when thinking about it you get "10 is slightly less than 7" but those rates might come from different times so it's ok)
23:38:49 <shachaf> GUIL (leaping up): What a shambles! We're just not getting anywhere.
23:38:49 <shachaf> ROS (mournfully): Not even England. I don't believe in it anyway.
23:38:50 <shachaf> GUIL: Just a conspiracy of cartographers, you mean?
23:39:29 <copumpkin> I suppose Rosencrantz probably is too then
23:39:33 <ais523> I think the latest terminology is that the UK is a nation, or a state, and that England, Scotland, Wales are countries, and screw all the smaller constituents of the UK because classifying Northern Ireland is a political minefield and the other places are too small to bother with
23:39:36 <shachaf> copumpkin: Did you see R&G Are Dead?
23:39:41 <ais523> but I'm not convinced I agree with it
23:39:51 <copumpkin> I read it at some point and then watched a video of a performance
23:39:51 <shachaf> copumpkin: Well, he's in a box. Take that as you will.
23:40:54 <elliott> ais523: UK is a kingdom, obviously
23:41:04 <Arc_Koen> ais523: and where do you place Great Britain in all that?
23:41:05 <shachaf> ROS: We might as well be dead. Do you think death could possibly be a boat?
23:41:05 <shachaf> GUIL: No, no, no... Death is... not. Death isn't. You take my meaning. Death is the ultimate negative. Not-being. You can't not-be on a boat.
23:41:08 <shachaf> ROS: I've frequently not been on boats.
23:41:11 <shachaf> GUIL: No, no, no - what you've been is not on boats.
23:41:20 <shachaf> Arc_Koen: Great Britain has already been placed.
23:41:25 <shachaf> It's pretty solid where it is.
23:41:53 <Arc_Koen> shachaf: I was referring to ais523: I think the latest terminology is that the UK is a nation, or a state, and that England, Scotland, Wales are countries, and screw all the smaller constituents of the UK because classifying Northern Ireland is a political minefield and the other places are too small to bother with
23:41:54 <ais523> Arc_Koen: Great Britain is reasonably simple, it's a geographical term referring to an island
23:41:57 <ais523> it's a reasonably large island
23:42:05 <ais523> and it contains England, and most of Scotland and Wales
23:42:10 <olsner> "The 14 British Overseas Territories are under the jurisdiction of the United Kingdom, though they do not form part of it."
23:42:58 <nooga> what's the english name for that dash in ń character?
23:43:01 <shachaf> What about the Underseas Territories?
23:43:10 <Arc_Koen> so the UK is a state that contains countries... and the US are a country that contains states
23:43:17 <olsner> shachaf: they are secret and don't officially exist
23:43:29 <ais523> we don't have acutes in our own language, but we have a word for them so we can talk about them in other languages
23:43:40 <Jafet> They're at Lantis.
23:43:56 <ais523> nooga: we're not really used to seeing them on n, though
23:44:06 <ais523> mostly we use them to talk about French because the French are closest
23:44:09 <Arc_Koen> I'm beginning to learn dutch and they have acutes, though they seem to be saving them for onomatopoeias
23:44:13 <ais523> and we used to repeatedly have wars with them
23:45:01 <nooga> like... give me a second
23:45:10 <shachaf> There should be a modifier for silent letters.
23:45:23 <Arc_Koen> I'm pretty sure we have that in french
23:45:32 <olsner> shachaf: silent letters should simply be removed
23:45:36 <Jafet> That's so cliché, ais
23:46:15 <Arc_Koen> so ń alone is a whole syllable?
23:46:24 <Jafet> Silent letters are what they read in silent films
23:47:00 <nooga> it's platal nasal consonant
23:47:11 <nooga> it just sounds like ni in night
23:47:55 <olsner> maybe the sound isn't used in english at all
23:47:58 <Arc_Koen> "ghhh"? that's a weird way to pronounce an n
23:48:50 <nooga> http://en.wikipedia.org/wiki/%C5%83
23:49:24 <Arc_Koen> Ń (minuscule: ń) is a letter formed by putting an acute accent over the letter N.
23:49:45 <Arc_Koen> I'm so glad wikipedia can teach me those things every day
23:50:05 <olsner> seems it's the same sound as ñ in spanish
23:50:11 <nooga> http://en.wikipedia.org/wiki/File:Pl-kwiecie%C5%84.ogg
23:50:18 <nooga> olsner: yep, pretty much
23:50:37 -!- Phantom__Hoover has joined.
23:50:48 <nooga> ó is pronounced like Polish u, which is like english oo
23:51:31 <nooga> i think ś and ć do not exist in english
23:53:25 <nooga> we've also got ł,ż,ź,ą,ę and ł
23:53:46 -!- Phantom_Hoover has quit (Ping timeout: 246 seconds).
23:54:01 <olsner> don't forget rz, sz and cz
23:54:12 <Arc_Koen> mind me if I go off-topic for just one second? - do you know of a brainfuck program I could feed my interpreter with to check for potential bugs? (instructions +-><[] are supposed to be working properly, instruction . works if the current cell is either 10 or 32-126, and instruction , is considered a no-op)
23:55:10 <nooga> brzeszczot -> knife edge
23:56:16 <nooga> chrząszcz brzmi w trzczcinie w strzebrzeszynie, to remind the most known tongue breaker
23:56:36 <olsner> let me guess, that's polish for "hi"?
23:56:50 <nooga> no, hi is cześć ;)
23:56:50 <Arc_Koen> hrm aren't we supposed to throw some vowels into that before even considering trying to pronounce it?
23:57:20 <Jafet> http://www.muppetlabs.com/~breadbox/bf/quine.b.txt
23:58:32 -!- kinoSi has quit (Read error: Connection reset by peer).
23:59:01 -!- kinoSi has joined.