00:28:23 -!- GeekDude has changed nick to DrSinner.
00:28:37 -!- DrSinner has changed nick to GeekDude.
00:35:03 <HackEgo> [wiki] [[Hyperfunge]] http://esolangs.org/w/index.php?diff=43748&oldid=39856 * 69.165.212.148 * (+787)
00:36:59 <doesthiswork> I don't have ꙮ in my font but it is probably a nasal ingressive voiceless velar trill
00:51:19 -!- boily has joined.
00:56:17 <HackEgo> [wiki] [[Hyperfunge]] http://esolangs.org/w/index.php?diff=43749&oldid=43748 * 69.165.212.148 * (+238)
00:58:02 -!- singingboyo has joined.
00:58:55 -!- copumpkin has joined.
01:01:49 * boily pokes coppro in the haneman
01:06:24 <HackEgo> [wiki] [[Hyperfunge]] http://esolangs.org/w/index.php?diff=43750&oldid=43749 * 69.165.212.148 * (+338)
01:10:57 <HackEgo> [wiki] [[Hyperfunge]] http://esolangs.org/w/index.php?diff=43751&oldid=43750 * 69.165.212.148 * (-7)
01:11:40 -!- madbr has joined.
01:11:57 <madbr> oh man, I've created a monster
01:12:01 <madbr> http://esolangs.org/wiki/Hyperfunge
01:12:11 <madbr> hyperbolic geometry version of befunge
01:12:18 <madbr> it's aboslutely hideous :D
01:13:21 <ais523> a hyperbolic geometry befunge sounds excellently eso :-)
01:14:24 <ais523> I'm not sure I understand the directions in which you can execute code, though
01:14:41 <ais523> a picture would be helpful (that projects onto a euclidean space via the use of scaling)
01:16:44 <boily> madbr: hey! long time no see!
01:17:06 <boily> btw, you have no right creating hyperbolic esolangs on Sundays late at night.
01:19:04 <madbr> true, it would need a picture
01:19:31 <madbr> and yeah well... these things happen :o
01:20:04 <coppro> boily: ow that was rude
01:21:16 <madbr> I kinda wanted to incorporate the expansion properties of hyperbolic space into the language more but just gave up :D
01:23:11 <boily> we definitely need some kind of picture/example/synæsthetic experience for that.
01:23:44 <madbr> ok but not on sunday late at night :D
01:24:52 <boily> coppro: sorry. I only got tenpai on a closed chinitsu today. I couldn't achieve outrageousness.
01:27:27 <madbr> one idea I've had would've been to have a program with an infinitely tiled portion
01:28:15 <madbr> to store data independently from the stack you'd use the index of the tile respective to the non-tiled portion
01:30:11 <Phantom__Hoover> i feel like a square tiling would be better than a pentagonal one?
01:31:16 <Phantom__Hoover> you could translate befunge semantics a lot more naturally then
01:33:11 <mauris> madbr, is this the grid: https://upload.wikimedia.org/wikipedia/commons/4/44/Uniform_tiling_552-t1.png
01:34:21 <coppro> boily: it's ok. I haven't been doing well since that game
01:37:19 <mauris> i don't see where to draw the middle line
01:37:45 <madbr> I'm not sure it's the best grid for a hyperbolic fungeroid but I like the pentagonal thing going on
01:38:13 <madbr> mauris : the line at the left of the center pentagon looks like it'd work
01:39:16 <Phantom__Hoover> the problem with using pentagons is that you have no notion of opposite sides so you can't define a line on the tiling
01:40:07 <madbr> I guess it's more an edge
01:40:48 <madbr> it also has the catch that there are really two versions of each of the 5 directions you can move in
01:40:54 <Phantom__Hoover> if you're following an edge then that's the same as using the square tiling
01:41:59 <mauris> i feel like "sprialing out from the center" would be a more natural way to store the instructions?
01:42:01 <madbr> like, moving leftwards, you can either move leftwards in a way that undoes a right-up move, or in a way that undoes a right-down move...
01:42:23 <madbr> leading to the same pentagon except the IP is not in the same orientation
01:42:45 <madbr> mauris : yeah that's the other option
01:42:58 <madbr> mauris : or a sierpinsky type grid
01:45:15 <madbr> you can either start with a middle point (going to 5 edges and 5 corners), or a central equator with edges on both sides (the solution here), or a line of pentagons each with 1 edge leading downwards and 2 edges 1 corner upwards
01:47:21 <madbr> or a middle corner touching 4 pentagons
01:47:28 <mauris> you should draw some crude pictures about how the PC moves, mapping 1abcXYZdefghUVW etc.
01:47:50 <madbr> sure but tonight I can't be bothered :3
01:48:00 <mauris> maybe i have weak hyperbolic geometry skillz but this is basically impossible to read and imagine :(
01:48:20 <madbr> no hyperbolic geometry is impossible for everybody :o
01:48:27 <madbr> which is the point actually
01:49:25 <mauris> http://esolangs.org/wiki/Hi%5Cn help, i'm clicking through random pages and finding horrible languages
01:51:00 <boily> also, goodrjanight!
01:51:08 -!- boily has quit (Quit: CASSETTE CHICKEN).
01:51:47 <izabera> how did you even find that
01:52:54 <ais523> izabera: "random page", probably
01:53:07 <ais523> I invented the BF Derivative Random Page Game
01:53:14 <ais523> but stopped playing it because it was too depressing
01:53:35 <ais523> (the rules: keep pressing Random Page until you find a language you designed, your score is minus the number of BF derivatives you find on the way)
01:54:21 <ais523> hmm, Hi\n isn't as bad as some I've seen
01:54:32 -!- copumpkin has quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…).
01:54:33 <ais523> HQ9+ does similar space better, though, I think
01:54:59 <ais523> that language appears to have been invented independently, though, and does one of HQ9+'s jokes better while missing the others
01:56:44 <ais523> oh dear, I've been clicking on random pages
01:56:58 <ais523> and this BF derivative is probably the second-worst I've seen: http://esolangs.org/wiki/And_then
01:57:10 <ais523> (the worst involved doing something that prevented loops working)
01:58:02 <mauris> this is certainly something
01:58:41 <ais523> really there's absolutely no reason for it to be based on brainfuck
02:06:08 -!- singingboyo has quit (Ping timeout: 272 seconds).
02:12:54 <HackEgo> [wiki] [[MagiStack]] M http://esolangs.org/w/index.php?diff=43752&oldid=43747 * Tripl3dogdare * (+0)
02:17:02 <mauris> i found an actually good language and wrote an interpreter!
02:19:52 <mauris> http://esolangs.org/wiki/Minimum -> http://lpaste.net/138992
02:19:57 -!- Phantom__Hoover has quit (Read error: Connection reset by peer).
02:20:31 <ais523> I like that use of absurd
02:21:08 <mauris> yes, especially absurd :: Void -> IO ()!
02:21:50 <mauris> and of course the parser failing for all finite strings. i'm proud
02:23:47 <madbr> I still have to figure out how to garbage collect one particularly crazy language I've come up with
02:24:18 <ais523> madbr: does the behaviour of the garbage collector change the language semantics?
02:25:12 <madbr> hmm, only if it changes which programs leak and which ones dont
02:25:17 <madbr> which... I think it does
02:25:42 <ais523> madbr: oh, my suggestion was going to be "just leak and use boehm-gc" if it doesn't change the semantics
02:25:59 <ais523> there are definitely languages, like Perl, where the GC behaviour is observable from inside the language itself though
02:26:37 -!- singingboyo has joined.
02:27:23 <madbr> the problem is that since the already computed data is immutable, computation must generate ever more data
02:27:49 <madbr> and it's hard to define when previously computed data becomes irrelevant and can be dumped
02:28:19 <oerjan> i don't like that eval.
02:29:21 <madbr> though I guess that due to the semantics of the language it can regenerate data that it dumped
02:29:31 <madbr> if it turns out it really needed it
02:29:38 <mauris> oerjan: how would you have written it?
02:30:17 <Jafet> ais523: referring to http://www.nntp.perl.org/group/perl.golf/2002/07/msg1289.html?
02:31:03 <ais523> Jafet: was actually refering to Scalar::Util::weaken; the description of Scalar::Util is as a place to put things that are effectively part of the base language, but aren't used enough to be in the language core
02:31:47 <ais523> but that thing is genius :-)
02:31:53 <ais523> pity it doesn't work on graphs with loops in
02:33:41 <HackEgo> [wiki] [[Minimum]] http://esolangs.org/w/index.php?diff=43753&oldid=13038 * Nooodl * (+657)
02:34:59 <HackEgo> [wiki] [[Minimum]] M http://esolangs.org/w/index.php?diff=43754&oldid=43753 * Nooodl * (+18) add a header for this i guess
02:40:48 <madbr> I guess it could just randomly dump data when reaching the data limit
02:41:12 <madbr> which means that with a real randomness source it would have to reach the solution...EVENTUALLY :3
02:43:48 <Jafet> On the other hand, it is also simultaneously possible that it never reaches the solution
02:44:49 <madbr> also I think the execution time would grow exponentially with each extra byte over the gc limit :o
02:45:13 -!- Wright has quit (Read error: Connection reset by peer).
02:45:24 -!- Wright has joined.
02:49:37 <oerjan> mauris: http://lpaste.net/138992#a138993
03:24:33 -!- Froo has joined.
03:27:15 -!- Froox has quit (Ping timeout: 246 seconds).
03:30:14 -!- Froo has quit (Quit: *bubbles away*).
03:37:56 -!- Frooxius has joined.
03:38:27 <Wright> is there a nickname for a group of three binary bits? I thought it was 'trit' but that's apparently a single ternary digit
03:38:58 -!- J_Arcane has joined.
03:39:02 <Wright> Ooh, wikipedia says it's a triad or triade
03:44:49 <myname> trit doesn't make sense. a bit would need to be two digits then
03:46:36 -!- callforjudgement has joined.
03:47:00 -!- ais523 has quit (Disconnected by services).
03:47:02 -!- callforjudgement has changed nick to ais523.
03:48:14 <madbr> "3/8ths of a byte"
03:48:58 <madbr> it's not a very popular bit grouping due to not fitting very well in bytes
03:49:52 <myname> octal is not very popular, either
03:53:34 <madbr> you could call it an "octal-digit"
04:04:53 -!- oerjan has quit (Quit: Nit).
04:15:49 -!- bb010g has joined.
04:16:50 <oren> octet, therefore tertet.
04:17:54 <oren> quatet would be 4 bits or one hex digit
04:18:41 <oren> pentet is one base 32 digit. sextet is one base64 digit
04:18:59 <oren> s/quatet/quartet/
04:19:33 <zzo38> I have used "nybble" before to mean one hex digit or four bits, this is also sometimes used elsewhere.
04:20:35 <zzo38> I use "octet" generally when defining a VM or something else that has memory cells larger than eight bits and are addressable by the longer cells instead of by octets.
04:20:49 <zzo38> So, that mean "octet" always means 8-bits regardless of the size of memory cells
04:23:30 <oren> yeah. and then we can also use hexdecet for unambigously a 16 bit word
04:24:08 <oren> bitrigintet is 32 bits
04:25:23 <oren> quattrisexagintet is a 64 bit unit
04:27:06 <oren> centioctovigintet is 128 bits
04:53:35 <ais523> myname: you feel the need to say that in /this/ channel? :-D
04:56:50 <oren> what's thon thon?
04:58:07 <myname> ais523: it's best suited here, isn't it?
05:00:37 <|f`-`|f> http://i.4cdn.org/wsg/1439678080638.webm
05:00:54 <|f`-`|f> Turn yourselves in peacefully terrorists
05:01:29 -!- Wright has quit (Ping timeout: 252 seconds).
05:02:07 -!- doesthiswork has quit (Quit: Leaving.).
05:10:50 -!- aretecode has quit (Ping timeout: 240 seconds).
05:10:58 <zzo38> The description of e-TeX says they added classes of marks since it seem like a good idea; however I have been able to do classes of marks without e-TeX, using only standard TeX operations.
05:44:46 -!- Froox has joined.
05:48:43 -!- Frooxius has quit (Ping timeout: 265 seconds).
05:49:45 -!- Froox has quit (Ping timeout: 255 seconds).
05:51:17 <ashl> is that chris morris
06:00:47 -!- x10A94 has joined.
06:02:40 <izabera> http://dl.acm.org/citation.cfm?id=275318 this is mildly interesting but not enough to pay 15$, can someone provide it for free? >.>
06:03:08 <zzo38> What is that one about?
06:03:53 <zzo38> I do not have a copy to provide for free anyways (as far as I know)
06:04:17 <Hoolootwo> is that English the natural language or English the computer language?
06:04:28 <izabera> i believe the natural language
06:04:57 <Hoolootwo> yeah natural language according to the references
06:06:16 <ais523> izabera: it's probably a bad idea to ask people to pirate academic papers for you, after what happened to aaron schwartz
06:06:37 <ais523> I'd recommend checking the author's/authors' website, though
06:06:47 <Hoolootwo> it's possible that someone is a part of the ACM
06:07:07 <ais523> often it has a preprint (basically, the same paper before proofreading/peer review), which the author's allowed to post publicly
06:07:11 <Hoolootwo> and has the ability to download copies and share maybe one copy, I haven't read any fine text but I could see that
06:07:45 <ais523> I can't remember how draconian the ACM is
06:08:03 <ais523> one of the major places has a guest access thing where you can look at (but not download) something like four papers per month
06:09:47 <Hoolootwo> go to 'Request Permissions' maybe, it appears as if that's for non-members
06:14:16 <Jafet> Actually, what happened to Schwartz was controversial because people are used to pirating academic papers.
06:14:45 <Jafet> This particular paper should be worth the $15 though. That's only $5 per page!
06:15:09 <izabera> that's 5$ per page more than what i'm going to spend <.<
06:16:51 <ais523> the thing that gets me about academic paper charging is that the charge for individual papers is always utterly overblown
06:16:59 <Jafet> “© 1998 ACM l-58113-030-9/98/0004 $3.50”
06:17:04 <ais523> if they price were reduced, there's a chance that people might actually spend it sometimes
06:17:21 <ais523> also the ACM's papers always have a price on the copyright line IIRC, and I have no idea what it refers to
06:23:23 <HackEgo> [wiki] [[???]] N http://esolangs.org/w/index.php?oldid=43755 * 50.170.122.255 * (+192) Created page with "??? is an esoteric programming language created by Stack Exchange users Alex A. and BrainSteel. It uses the punctuation within a literary work to perform operations akin to th..."
06:24:12 -!- mauris has quit (Ping timeout: 250 seconds).
06:24:43 <Jafet> We need a separate wiki to catalogue the rapidly developing field of brainfuck derivatives
06:28:02 <Sgeo> How about a brainfuck derivative with the exact syntax and semantics of Python
06:30:49 <HackEgo> [wiki] [[???]] http://esolangs.org/w/index.php?diff=43756&oldid=43755 * 50.170.122.255 * (+173)
06:33:07 -!- bb010g has quit (Quit: Connection closed for inactivity).
06:33:07 <HackEgo> [wiki] [[???]] http://esolangs.org/w/index.php?diff=43757&oldid=43756 * 50.170.122.255 * (+10)
06:34:23 <HackEgo> [wiki] [[MagiStack]] http://esolangs.org/w/index.php?diff=43758&oldid=43752 * Rdebath * (+72) Probably NOT TC
06:34:52 <HackEgo> [wiki] [[MagiStack]] http://esolangs.org/w/index.php?diff=43759&oldid=43758 * Rdebath * (+33)
06:40:34 <HackEgo> [wiki] [[MagiStack]] http://esolangs.org/w/index.php?diff=43760&oldid=43759 * Rdebath * (+5) Humm, OTOH branching doesn't look good enough for a PDA; is it ?
06:51:28 <HackEgo> [wiki] [[???]] http://esolangs.org/w/index.php?diff=43761&oldid=43757 * Rdebath * (+112) Definitely not TC
06:52:41 <HackEgo> [wiki] [[???]] http://esolangs.org/w/index.php?diff=43762&oldid=43761 * 50.170.122.255 * (+1155)
06:56:06 <HackEgo> [wiki] [[???]] http://esolangs.org/w/index.php?diff=43763&oldid=43762 * 50.170.122.255 * (+35)
07:05:13 <HackEgo> [wiki] [[Talk:???]] N http://esolangs.org/w/index.php?oldid=43764 * Martin Büttner * (+105) Created page with ""it is not Turing complete due to the loss of nested loops." I don't see why you can't have nested loops."
07:05:36 <HackEgo> [wiki] [[Talk:???]] http://esolangs.org/w/index.php?diff=43765&oldid=43764 * Martin Büttner * (+106)
07:06:42 <ais523> fwiw I agree with the talk page post
07:07:48 <HackEgo> [wiki] [[Talk:???]] http://esolangs.org/w/index.php?diff=43766&oldid=43765 * Ais523 * (+310) agree
07:09:36 <HackEgo> [wiki] [[Talk:???]] http://esolangs.org/w/index.php?diff=43767&oldid=43766 * Martin Büttner * (+193)
07:10:21 <HackEgo> [wiki] [[???]] http://esolangs.org/w/index.php?diff=43768&oldid=43763 * Ais523 * (+114) is TC; clarify what ' does
07:17:14 -!- Frooxius has joined.
07:21:24 -!- ais523 has quit (Ping timeout: 264 seconds).
07:35:07 -!- ais523 has joined.
07:59:50 <HackEgo> [wiki] [[Talk:???]] http://esolangs.org/w/index.php?diff=43769&oldid=43767 * Rdebath * (+250)
08:02:32 -!- zadock has joined.
08:05:49 <ais523> could someone here on Linux > 3.8 run "unshare -Ur true" as not-root and let me know if it produces an error? (this relates to weboflies 2)
08:06:40 <ais523> or, really, anywhere, but I thought I'd ask here because this channel a) is often responsive to bizarre requests, and b) is vaguely aware of weboflies
08:07:12 <izabera> unshare: unshare failed: Invalid argument
08:07:42 <izabera> with unshare from util-linux 2.26.2
08:08:02 <ais523> I'm confused, because that is a /different/ error from the one I'm getting
08:08:21 <ais523> and yet it should succeed, given that that's the simplest possible valid use of the -r option
08:08:30 <ais523> the reason I ask is that I suspect there might be a kernel bug
08:08:55 <ais523> unshare: write failed /proc/self/gid_map: Operation not permitted
08:09:15 <ais523> izabera: oh, probably you have unshare compiled out in your kernel
08:09:30 <ais523> err, unshare user namespaces
08:09:50 <ais523> that'd explain the symptoms you're seeing, it isn't very useful for me to know if the bug's fixed though
08:11:17 -!- AnotherTest has joined.
08:12:27 <izabera> i didn't compile my kernel
08:12:36 <izabera> and i honestly don't know much about this
08:13:32 <ais523> fair enough, I wasn't necessarily expecting people to understand what the command did (it is harmless, though)
08:14:01 <ais523> it can be surprising to see what it does without the "true" at the end (which makes it a no-op), if you aren't prepared for it, although it doesn't actually break security
08:16:18 -!- singingboyo has quit (Remote host closed the connection).
08:28:50 -!- Patashu has joined.
09:00:41 -!- FreeFull has quit (Quit: BBL today).
09:28:30 -!- Phantom_Hoover has joined.
09:37:23 <HackEgo> [wiki] [[Mov]] http://esolangs.org/w/index.php?diff=43770&oldid=43287 * 194.19.70.9 * (+10) mov is not 'esoteric' in any meaning of the word
09:55:36 -!- |f`-`|f has quit (Ping timeout: 264 seconds).
09:56:11 -!- |f`-`|f has joined.
09:57:26 -!- zadock has quit (Quit: Leaving).
10:00:48 -!- SopaXT has joined.
10:06:58 -!- shikhiniht has changed nick to shikhin.
10:10:05 -!- ais523 has quit (Ping timeout: 250 seconds).
10:22:07 <Taneb> Well, I have made progress today
10:24:07 <int-e> `learn Progress has been made today. It was invented by Taneb.
10:24:12 <HackEgo> Learned 'progres': Progress has been made today. It was invented by Taneb.
10:27:36 -!- Patashu has quit (Ping timeout: 246 seconds).
10:28:07 -!- Patashu has joined.
10:28:12 -!- boily has joined.
10:47:04 <HackEgo> Progress has been made today. It was invented by Taneb.
10:47:06 <HackEgo> Progress has been made today. It was invented by Taneb.
10:48:11 -!- ais523 has joined.
10:52:27 <boily> `` sed -i 's/the reals,/the reals, progress,/' wisdom/tanebvention
10:54:11 <Taneb> You know, I ought to be more careful about what I say just in case it becomes a tanebvention
10:55:46 <boily> the way I see it, everything already is a tanebvention. except those things which are explicitely not, therefore they are.
10:58:30 <boily> we're highlighting the most egregious ones, I think. like weetoflakes.
10:58:43 <Taneb> Which is the only one I actually invented
10:59:29 -!- AnotherTest has quit (Quit: ZNC - http://znc.in).
11:06:17 -!- ais523 has quit (Quit: lunch).
11:22:16 -!- boily has quit (Quit: SEDIMENTARY CHICKEN).
11:26:44 <int-e> damn, I missed the plural thing
11:27:07 <int-e> `rm wisdom/progres
11:27:33 <HackEgo> rm: cannot remove `/home/hackbot/hackbot.hg/multibot_cmds/env/.hg/store/data/canary.orig': Is a directory \ Done.
11:27:38 <HackEgo> Progress has been made today. It was invented by Taneb.
11:49:36 -!- SopaXT has quit (Remote host closed the connection).
11:57:11 -!- ais523 has joined.
12:30:34 <ais523> hmm, now a different set of people is online
12:31:14 <ais523> <ais523> could someone here on Linux > 3.8 run "unshare -Ur true" as not-root and let me know if it produces an error? (this relates to weboflies 2)
12:31:30 <ais523> one person already tried but I think the relevant kernel config option was disabled
12:33:15 <APic> ais523: unshare: unshare failed: Die Operation ist nicht erlaubt
12:33:34 <ais523> hmm, I guess it isn't enabled on your kernel either :-(
12:33:41 <APic> Linux Ant 4.1.0-1-amd64 #1 SMP Debian 4.1.3-1 (2015-08-03) x86_64 GNU/Linux
12:34:59 <ais523> the relevant feature is user namespaces
12:35:12 <ais523> but they seem to be broken with respect to GIDs (UIDs work correctly)
12:35:20 <ais523> either that or both me and unshare(1) are using them incorrectly
12:35:54 <APic> On my UserModeLinux-VM it fails too
12:36:28 <ais523> `run unshare --version
12:36:29 <HackEgo> unshare: unrecognized option '--version' \ \ Usage: \ unshare [options] <program> [args...] \ \ Options: \ -h, --help usage information (this) \ -m, --mount unshare mounts namespace \ -u, --uts unshare UTS namespace (hostname etc) \ -i, --ipc unshare System V IPC namespace \ -n, --net unshare network nam
12:36:38 <ais523> I can believe it would fail in UML
12:36:45 <HackEgo> unshare: invalid option -- 'U' \ \ Usage: \ unshare [options] <program> [args...] \ \ Options: \ -h, --help usage information (this) \ -m, --mount unshare mounts namespace \ -u, --uts unshare UTS namespace (hostname etc) \ -i, --ipc unshare System V IPC namespace \ -n, --net unshare network namespace \
12:36:53 <ais523> oh wow, that's an old version of unshare
12:37:05 <ais523> that doesn't support "--version", for some reason
12:37:11 <HackEgo> Linux umlbox 3.13.0-umlbox #1 Wed Jan 29 12:56:45 UTC 2014 x86_64 GNU/Linux
12:39:14 -!- Patashu has quit (Ping timeout: 260 seconds).
12:42:15 <Taneb> ais523, I get invalid option U
12:42:25 <Taneb> But mine doesn't have --version either
12:42:33 <Taneb> This is Ubuntu 14.04
12:42:46 <ais523> right, I think that predates Linux 3.8?
12:42:54 <ais523> or at least the matching util-linux release
12:43:09 <ais523> it probably took a while for new Linux functionality to be implemented in util-linux
12:43:46 <ais523> hmm, I hope that distros don't disable user namespaces by default
12:43:50 <ais523> or nobody will be able to run weboflies
12:44:42 -!- ProofTechnique has joined.
12:44:42 <Taneb> ais523, invalid argument on Arch with kernel 4.0.5-ARCH
12:44:54 <Taneb> What is web o' flies?
12:45:04 <ais523> Taneb: formerly known as the Secret Project
12:45:51 <ais523> could you do "cat /lib/modules/$(uname -r)/build/.config | grep USER_NS"
12:46:21 <ais523> and, umm, it's quite hard to explain what it does, but it involves intercepting system calls and replacing them with different ones
12:46:44 <ProofTechnique> Taneb: http://stealthcatproductions.blogspot.com/2008/11/web-o-flies.html HTHeth
12:47:19 <ais523> ProofTechnique: I doubt it's been discussed much outside #esoteric
12:47:24 <Taneb> ais523, file does not exist
12:47:30 <ais523> I only came here for help with it because it's a) ridiculous, b) hard to explain what you're doing
12:47:41 <Taneb> This is a server I don't control
12:47:45 <ais523> I take it /proc/config.gz doesn't exist either?
12:47:56 <Taneb> No, that does exist
12:48:07 <ais523> zgrep for USER_NS in that, then
12:48:22 <ais523> (the two files should have the same content, just one is usermode and one is kernelmode)
12:48:33 <Taneb> # CONFIG_USER_NS is not set
12:48:40 <ais523> that's the option that I'm trying to test
12:48:48 <ais523> but apparently it's disabled in whatever distro that server's using
12:50:07 <Taneb> I think my laptop (with Ubuntu LTS) has it enabled
12:50:14 <Taneb> Just an old version of unshare
12:50:20 <ais523> I can believe Arch, on the basis that multiple people have reported the same outcome with kernel versions in the 4s
12:52:12 <ais523> so it has to be some rolling-release distro that's widely used
12:55:46 -!- ais523 has quit.
12:57:03 -!- ais523 has joined.
13:00:50 -!- ais523 has quit (Client Quit).
13:01:19 -!- ais523 has joined.
13:02:28 <int-e> same here with Arch
13:02:46 <ais523> hmm, I guess this is going to be really hard to test then
13:02:56 <ais523> also, I guess unshare(1) isn't widely used
13:03:07 <ais523> I didn't see any relevant results from a web search on the error message
13:03:17 * int-e switched to a kvm host for lambdabot to get namespace support
13:04:14 <int-e> (the evaluation part runs as a seaparate user and with unshare(CLONE_NEWIPC | CLONE_NEWNET | CLONE_NEWNS) )
13:07:09 <ais523> hmm, secure boot may end up preventing people turning all these great kernel features on
13:07:32 <ais523> on the other hand, it may force distros to turn them all on on the basis that people literally can't compile the kernel themself any more
13:08:06 -!- doesthiswork has joined.
13:08:21 <int-e> Well, Docker is becoming quite popular and relies on namespace support. Also systemd has plans in that direction, I believe.
13:08:49 <int-e> I'm actually surprised that Arch doesn't enable namespaces by default.
13:13:07 -!- oerjan has joined.
13:13:11 <int-e> https://bugs.archlinux.org/task/36969?project=1 tells a bit of the story; apparently the namespace feature is still suffering from holes.
13:13:29 <ais523> this does not surprise me
13:13:43 <ais523> being able to run a command as non-root and gain /all/ capabilities is crazy
13:13:50 <ais523> even if most of them don't actually do anything
13:14:04 <HackEgo> [wiki] [[Special:Log/newusers]] create * Vioz- * New user account
13:15:04 <int-e> Oh, I mixed this up. I need namespaces (CONFIG_*_NS), but not for arbitrary users (CONFIG_USER_NS). So never mind.
13:16:03 -!- ais523 has quit.
13:16:34 <int-e> (lambdabot uses a suid program to do the namespace setup)
13:16:34 <fizzie> ais523: Reports from here: no '-U' in unshare at work desktop (distribution derived from Ubuntu 14.04); "unshare failed: Operation not permitted" on home desktop (Debian something); "unshare failed: Invalid argument" on VPS (Debian 8.1).
13:16:44 <fizzie> Aw, just a little bit late.
13:17:04 -!- ais523 has joined.
13:17:18 <fizzie> ais523: Reports from here: no '-U' in unshare at work desktop (distribution derived from Ubuntu 14.04); "unshare failed: Operation not permitted" on home desktop (Debian something); "unshare failed: Invalid argument" on VPS (Debian 8.1).
13:17:54 <ais523> huh, that EPERM on unshare is interesting (the second one)
13:18:22 <ais523> must have hit the narrow version range where it was root-only, or the somewhat less narrow range where certain distros restricted it to root only as a mitigation for potential security holes
13:18:35 <ais523> on my Ubuntu 15.04, the unshare works, and setting uid_map works, but setting gid_map doesn't
13:18:45 <ais523> which is so bizarrely asymmetric that it makes me suspect a bug
13:18:55 -!- madbr has quit (Quit: Pics or it didn't happen).
13:19:49 <HackEgo> [wiki] [[Stuck]] N http://esolangs.org/w/index.php?oldid=43771 * Vioz- * (+218) initial submission.
13:19:55 <fizzie> It's got 4.1.0-1-amd64 installed too, but I haven't had a chance to reboot lately.
13:26:44 -!- ais523 has quit.
13:26:53 -!- ais523 has joined.
13:28:56 <ais523> aha, I think I found the cause of my problem
13:29:03 -!- `^_^v has joined.
13:29:08 <ais523> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=87c31b39abcb6fb6bd7d111200c9627a594bf6a9 talks about "/proc/self/setgroups" which I've never heard of
13:29:25 <ais523> and it isn't documented
13:29:29 <ais523> but it seems very relevant
13:29:36 <ais523> also I love the security hole it's there to work around
13:30:07 <ais523> (you can have group permission lower than other permission on a file, thus dropping group membership via gid_map can give you higher perms on a file)
13:36:30 -!- SopaXT has joined.
13:36:47 <ais523> drwxrwxr-x 4 root root 4096 Aug 16 16:04 test
13:37:15 <ais523> int-e: thanks for the bug report link, I found out what I was missing by chasing links from there
13:37:23 <ais523> the solution is to write "deny" to /proc/self/setgroups
13:38:52 <ais523> and here it is, my unshare -mUr replacement: http://sprunge.us/GRST
13:42:01 <Taneb> Hang on, what have you done
13:42:32 <ais523> Taneb: have you tried running it yet?
13:43:29 <ais523> it's not weboflies itself, just a test file to play around with some of the features it needs
13:43:32 <ais523> and it doesn't need root
13:44:18 <ais523> what it does is /gives/ you root, but a restricted sort of root that isn't actually able to do anything malicious (in theory)
13:46:06 <ais523> I've been trying to mount filesystems using it
13:46:19 <ais523> but loop mounts don't work because I can't read or write any of the loop devices
13:46:27 <ais523> bind mounts work, mounting a new tmpfs works
13:46:37 <ais523> mounting a physical block device doesn't work for obvious reasons
13:48:10 <fizzie> I think I did run across some context where group permissions were being used to "exclude" a group (as in, more bits for other than group), but it seemed highly nonstandard.
13:49:59 <ais523> cygwin has an article about implementing permissions like that on Windows
13:50:23 <ais523> it turns out that it's possible, but the configuration you have to use is so unusual that even just viewing the permissions in Explorer changes them to something semantically different
13:50:37 <fizzie> I think you linked to it before.
13:51:22 <ais523> I can't remember where it is though
13:51:45 <ais523> probably here?: http://cygwin.com/cygwin-ug-net/ntsec.html
13:52:05 <fizzie> Under the "File permissions" heading.
13:53:09 <Taneb> ais523, how risky is this exploit?
13:53:56 <ais523> Taneb: it's not an exploit, it's an intended feature
13:54:25 <Taneb> Being able to become root is an intended feature?
13:54:25 <ais523> although based on the discussion in this channel, it's one with more security holes in than you'd expect from a typical intended Linux feature
13:55:15 <ais523> because you can't do anything you couldn't do with your own account
13:55:26 <ais523> at least, shouldn't be able to
13:55:50 <ais523> also you can't interact with any user /but/ root (which is secretly your own account), e.g. su won't work
13:56:09 <Taneb> So it just makes me look like root or something?
13:56:55 <ais523> it's sort-of like kernel fakeroot
13:57:09 <ais523> however it does let you do various things you can't normally do as a user, that are meant to be safe when sandboxed
13:57:57 <Taneb> Can I chroot to a directory I can't otherwise access?
13:58:12 <fizzie> ais523: Out of curiosity, if you "rootify" yourself like that, can you connect to a Unix domain socket in the non-unshared network namespace, and if you can, what would using SO_PASSCRED/SCM_CREDENTIALS report to the other end?
13:58:39 <ais523> fizzie: I remember reading the answer to this but can't remember what it was
13:58:42 <ais523> let me try to find it again
13:59:20 <ais523> fizzie: you can, the only UID you can send is root (because it's the only one you can access), and the other end sees your own actual UID
14:01:09 -!- nycs has joined.
14:02:48 -!- `^_^v has quit (Ping timeout: 264 seconds).
14:04:07 <ais523> Taneb: anyway, you are getting just a tiny taste of what WEB OF LIES does
14:04:22 <ais523> basically, imagine this sort of shenanigans going on except for everything
14:04:37 -!- SopaXT has quit (Remote host closed the connection).
14:05:10 <Taneb> So, nothing is the user it seems it is, or something?
14:05:25 <ais523> name had to come from somewhere
14:05:36 <ais523> but the hope is that everything the user sees is actually internally consistent
14:06:44 <ais523> I should probably not play around with nsfakeroot as much as I do
14:06:59 <ais523> maybe some day I'll /actually/ have root but mistake it for just another nsfakeroot
14:08:19 -!- SopaXT has joined.
14:10:32 <Jafet> That's when you wake up from the dream into another dream
14:10:40 <fizzie> What's Debian's fakeroot built on, LD_PRELOAD trickery on the syscall wrappers?
14:11:04 <ais523> fizzie: yes, it's pretty fragile
14:12:30 -!- atrapado has joined.
14:16:01 -!- `^_^ has joined.
14:17:47 -!- Lorenzo64 has joined.
14:18:38 -!- nycs has quit (Ping timeout: 265 seconds).
14:19:04 -!- ais523 has quit.
14:26:55 -!- doesthiswork has quit (Quit: Leaving.).
14:28:43 -!- j-bot has quit (Ping timeout: 252 seconds).
15:21:28 -!- Frooxius has quit (Ping timeout: 265 seconds).
15:35:11 -!- SopaXT has quit (Read error: Connection reset by peer).
15:36:26 -!- ProofTechnique has quit (Ping timeout: 240 seconds).
15:37:05 -!- mauris has joined.
15:39:20 -!- GoToTell has joined.
15:39:43 -!- Wright has joined.
15:57:33 -!- mauris_ has joined.
16:00:48 -!- mauris has quit (Ping timeout: 246 seconds).
16:06:36 <HackEgo> [wiki] [[Stuck]] http://esolangs.org/w/index.php?diff=43772&oldid=43771 * 72.38.29.19 * (+1632) added commands for math function.
16:06:55 -!- Wright has quit (Ping timeout: 244 seconds).
16:07:36 -!- mauris__ has joined.
16:07:40 <HackEgo> [wiki] [[Stuck]] http://esolangs.org/w/index.php?diff=43773&oldid=43772 * Vioz- * (+0) /* Overview */
16:10:32 -!- mauris_ has quit (Ping timeout: 244 seconds).
16:10:46 -!- oerjan has quit (Quit: Temporarily Unstuck).
16:15:36 -!- J_Arcane has quit (Ping timeout: 265 seconds).
16:30:32 -!- Wright has joined.
16:30:49 <HackEgo> [wiki] [[Stuck]] http://esolangs.org/w/index.php?diff=43774&oldid=43773 * Vioz- * (+1838) /* Overview */
16:32:16 <HackEgo> [wiki] [[Stuck]] http://esolangs.org/w/index.php?diff=43775&oldid=43774 * Vioz- * (+98)
16:33:29 <HackEgo> [wiki] [[Stuck]] http://esolangs.org/w/index.php?diff=43776&oldid=43775 * Vioz- * (+119) /* Overview */
16:37:57 -!- ProofTechnique has joined.
16:41:00 -!- GoToTell has quit (Quit: Want to be different? Try HydraIRC -> http://www.hydrairc.com <-).
16:46:15 -!- mauris__ has changed nick to mauris.
16:46:27 -!- mauris has quit (Changing host).
16:46:27 -!- mauris has joined.
16:50:55 -!- J_Arcane has joined.
17:10:43 <HackEgo> [wiki] [[MagiStack]] http://esolangs.org/w/index.php?diff=43777&oldid=43760 * Tripl3dogdare * (+910)
17:11:10 -!- ProofTechnique has quit (Ping timeout: 240 seconds).
17:14:54 <HackEgo> [wiki] [[MagiStack]] M http://esolangs.org/w/index.php?diff=43778&oldid=43777 * Tripl3dogdare * (+177)
17:15:25 <HackEgo> [wiki] [[MagiStack]] M http://esolangs.org/w/index.php?diff=43779&oldid=43778 * Tripl3dogdare * (-46)
17:38:36 <HackEgo> [wiki] [[MagiStack]] M http://esolangs.org/w/index.php?diff=43780&oldid=43779 * Tripl3dogdare * (-5)
17:41:37 -!- Lorenzo64 has quit (Read error: Connection reset by peer).
18:23:46 -!- Frooxius has joined.
18:26:22 -!- FreeFull has joined.
18:28:07 <tswett> The way I'm writing this operating system kernel, it'll allow you to do something hilarious.
18:28:17 <tswett> So, right now I'm implementing dynamic memory allocation.
18:29:02 <tswett> Dynamic memory allocators are going to be objects. You could, in principle, have a bunch of them, each with its own heap.
18:29:37 <tswett> In practice, I don't see needing more than one.
18:30:27 <tswett> Memory that's been allocated isn't actually associated with the heap any more, in any way.
18:30:58 <tswett> And what this means is that if you have multiple dynamic memory allocators, you can allocate memory from one and "free" the memory using another one.
18:31:12 <tswett> The result is that this section of memory will actually be moved between the heaps.
18:32:57 <tswett> It's like a library system that doesn't care where you return books.
18:33:21 <tswett> And libraries gain and lose books when people return them to different libraries.
18:51:00 -!- TieSoul has joined.
18:57:25 -!- Frooxius has quit (Quit: *bubbles away*).
19:08:51 -!- bb010g has joined.
19:08:57 -!- AnotherTest has joined.
19:18:50 -!- mauris has quit (Read error: Connection timed out).
19:19:28 -!- mauris has joined.
19:31:23 -!- hppavilion[1] has joined.
19:52:16 <HackEgo> [wiki] [[Stuck]] http://esolangs.org/w/index.php?diff=43781&oldid=43776 * Vioz- * (+4096)
19:57:09 -!- hppavilion[1] has quit (Quit: Page closed).
19:58:31 -!- mauris has quit (Read error: Connection timed out).
19:59:26 -!- mauris has joined.
20:05:02 -!- _256Q has joined.
20:15:21 -!- AnotherTest has quit (Quit: ZNC - http://znc.in).
20:16:32 -!- AnotherTest has joined.
20:17:24 -!- TieSoul has quit (Remote host closed the connection).
20:18:41 -!- MoALTz has quit (Quit: Leaving).
20:27:07 -!- oerjan has joined.
20:59:58 <HackEgo> [wiki] [[Special:Log/newusers]] create * DigitalCannon * New user account
21:03:01 <oerjan> the wiki seems to be under attack
21:03:23 -!- Patashu has joined.
21:07:48 -!- atriq has joined.
21:12:06 -!- atriq has quit (Ping timeout: 250 seconds).
21:14:24 -!- x10A94 has quit (Read error: Connection reset by peer).
21:17:23 <oerjan> fizzie: with a cannon hth
21:18:57 <oerjan> edwardk: i think you would be the person to know the answer to this http://stackoverflow.com/questions/32059754/are-there-useful-applications-for-the-divisible-type-class
21:22:49 <oerjan> hm Control.Lens doesn't seem to use it, what is this?
21:23:30 -!- `^_^ has quit (Ping timeout: 246 seconds).
21:28:18 <HackEgo> [wiki] [[Niblet]] N http://esolangs.org/w/index.php?oldid=43782 * DigitalCannon * (+3297) Created page with "'''niblet''' is an [[esoteric programming language]] where nibbles are used to store data. Cat: ,<.<[[ Hello World: ++++<<<<v< ---<<<v<. -<v< ----<<<<v<. ----<<<<<v<..."
21:28:20 <edwardk> oerjan: bookmarked to answer
21:28:25 <edwardk> oerjan: discrimination uses it heavily
21:29:52 <HackEgo> [wiki] [[Language list]] M http://esolangs.org/w/index.php?diff=43783&oldid=43743 * DigitalCannon * (+13)
21:35:20 -!- Patashu has quit (Ping timeout: 260 seconds).
21:40:36 <oerjan> edwardk: hm the new fancy hackage source browsing seems to break anchors http://hackage.haskell.org/package/discrimination-0.1/docs/src/Data-Discrimination-Grouping.html#Grouping
21:42:00 -!- ProofTechnique has joined.
21:43:32 <edwardk> oerjan: you should mention that to hvr
21:43:44 <edwardk> i think he was responsible in some way for the hackage source browsing
21:47:54 -!- ProofTechnique has quit (Quit: My Mac has gone to sleep. ZZZzzz…).
21:50:07 <fizzie> oerjan: That's pretty strange, the anchor on the "class Grouping a ..." line is <a name="Group"> -- the same as for "newtype Group", so it's there twice.
21:51:03 <fizzie> (Ditto for "class Grouping1".)
21:51:31 <oerjan> there was a github issue about duplicated anchors, maybe this is known then.
21:52:43 <oerjan> except it was supposed to be fixed before that package was uploaded
21:53:38 <fizzie> If you mean https://github.com/haskell/hackage-server/issues/319 that seems to talk only about the line number anchors.
21:54:01 * oerjan doesn't have a github account.
21:54:15 <fizzie> The hscolour 1.21 fix it mentions just says "ensure that line-number anchors do not reset in literate code fragments".
21:54:31 <fizzie> But there's a 1.23 change "bugfix for anchor-generation for instance decls".
21:55:30 <fizzie> It's not exactly an instance declaration, though, so maybe that's not related.
21:56:16 <HackEgo> [wiki] [[MagiStack]] M http://esolangs.org/w/index.php?diff=43784&oldid=43780 * Tripl3dogdare * (+103) /* Hello, world! */
21:56:45 <fizzie> If you're feeling really enthusiastic, maybe you could run the source through the newest hscolour and see if it's fixed or not.
21:57:18 <fizzie> That's reasonable. I wouldn't, either.
21:58:11 <oerjan> oh hscolour uses darcs, no wonder i couldn't find it
22:02:03 <oerjan> fizzie: that hscolour version is also older than the discrimination version
22:02:18 -!- AnotherTest has quit (Quit: ZNC - http://znc.in).
22:05:41 -!- mauris has quit (Read error: Connection timed out).
22:06:28 -!- mauris has joined.
22:08:37 <HackEgo> [wiki] [[ResPlicate]] http://esolangs.org/w/index.php?diff=43785&oldid=43638 * Stalem * (+2951) Conjecture and interpreter source
22:16:47 -!- atrapado has quit (Quit: Leaving).
22:39:49 -!- boily has joined.
22:57:45 -!- boily has quit (Quit: STRUT CHICKEN).
23:48:16 -!- _256Q has quit (Read error: Connection reset by peer).