00:15:10 1mpr0mp2 00:15:27 i mean 00:15:28 uh 00:15:37 if you like writing lots of HDL code 00:20:56 The only HDL codes I have written are FamicomHDL, although I have read Verilog codes for the Amber processor core and for other things. 00:22:14 what's FamicomHDL? 00:22:34 Hardware description for Famicom cartridges 00:22:51 It is actually a library in Haskell 00:23:14 16:22 Ohh xkcd 00:23:15 16:22 The monty python of nerds 00:23:15 16:22 :P 00:24:30 isn't that monty python 00:24:36 also http://xkcd.com/16/ 00:24:39 nrom = do { prgrom <- makeROM 0; chrrom <- makeROM 16384; connects $ zip cpuAddress (take 14 $ addressPins prgrom); connects $ zip cpuData (dataPins prgrom); connects $ zip ppuAddress (take 13 $ addressPins chrrom); ... and so on. 00:24:50 cool 00:25:30 war and peace, truly the crime and punishment of 19th-century russian novels 00:25:34 btw i have read neither of those 00:26:20 -!- Phantom__Hoover has joined. 00:26:41 Even though NROM is the simplest mapper (well, almost; there is a simpler one that has neither CHR ROM nor CHR RAM), you still need ten lines of code to program it. 00:27:51 And of course you also need to import (or define) functions such as makeROM, connects, cpuAddress, etc 00:28:59 There are other hardware description languages, I do not know a lot about them. I know a few things of Verilog since I have read the codes to implement the Amber processor core. 00:29:51 -!- Phantom_Hoover has quit (Ping timeout: 260 seconds). 00:33:31 -!- nooodl_ has quit (Ping timeout: 256 seconds). 00:34:37 -!- ais523 has quit (Ping timeout: 245 seconds). 01:02:57 -!- Sgeo_ has quit (Read error: Connection reset by peer). 01:04:07 -!- Sgeo_ has joined. 01:04:42 -!- Phantom__Hoover has quit (Quit: Leaving). 01:10:38 -!- Sgeo__ has joined. 01:13:31 kmc: do you know anything about linux networking 01:13:45 -!- Sgeo_ has quit (Ping timeout: 256 seconds). 01:15:06 -!- nooga has quit (Ping timeout: 264 seconds). 01:22:44 yes 01:22:51 that is a pretty vague question though 01:23:00 whatcha whatcha whatcha want 01:25:56 well i thought i figured out what i wanted 01:25:58 but then i became less sure 01:26:03 i realise it is vague 01:26:21 what i am wondering is how the hell do you communicate with tcp_diag from wizmode 01:26:38 kernel docs for /proc/net/tcp{,6} say that they are deprecated in favour of tcp_diag 01:26:45 but i have no idea how to use it and google is no help 01:26:50 maybe i should read the docs or something 01:27:10 ps what i am trying to do is figure out the UID of the owner of a TCP connection 01:27:37 i have no idea what that is 01:28:06 define that 01:28:14 tcp_diag or wizmode 01:28:25 er 01:28:28 s/wizmode/user mode/ 01:28:37 by which i mean 01:28:38 userspace 01:28:59 netstat -p will give you PIDs 01:29:25 right i guess that works 01:29:27 i am trying to do it programmatically 01:29:30 maybe i will look at netstat's source 01:29:44 or lsof 01:29:49 they are probably terrifying though 01:29:58 probably 01:29:58 sometimes when i want to see how a program does its magic, i use strace 01:30:06 i hear something about tcp_diag using netlink sockets 01:30:08 which uhhhhhh 01:30:10 strace++ kmc++ 01:30:11 buhhhhh 01:30:11 i don't know how to use netlink sockets 01:30:14 run away 01:30:18 i barely even know what netlink sockets are! 01:30:20 i don't wantto know 01:30:27 a trainwreck 01:30:28 s/wantto/want to/ 01:30:29 is what 01:31:05 strace: greater program or greatest program 01:31:25 the linux developers decided that there aren't enough ways for userspace to communicate with the kernel (you know, every system call and every file in procfs or sysfs or debugfs) and so there should be a special socket type for communicating with the kernel as well 01:31:57 oh i forgot "every device file" and "every ioctl on every device file" 01:32:32 they couldn't even be like "here's a UNIX socket with the kernel on the other end" 01:32:38 kmc: okay well this sounds like a distinct regression from /proc/net/tcp, yes 01:32:42 it has to be a new socket type which behaves mostly like UNIX sockets 01:32:44 «Netlink was designed for and is used to transfer miscellaneous networking information between the Linux kernel space and user space processes.» 01:32:45 but not quite 01:32:54 Miscellaneous networking information. 01:33:00 That's pretty useful. 01:35:12 -!- hagb4rd has quit (Quit: hagb4rd). 01:35:38 kmc: hm 01:35:40 what package is netstat in 01:35:46 i don't have it here 01:38:58 Debian net-tools 01:40:50 kmc: "net-tools are deprecated and you should be using iproute2 instead: http://www.archlinux.org/news/deprecation-of-net-tools/" "So I think I have to use ss instead of netstat now. Thanks!" 01:40:51 joy 01:42:39 "This April marked the ten year anniversary of the last net-tools release." 01:44:46 -!- ais523 has joined. 01:46:09 let's rewrite everything again because it's shit 01:46:13 -!- Arc_Koen has quit (Quit: Arc_Koen). 01:46:22 okay so ss uses netlink 01:46:28 kill me now 01:46:41 you already said that once 01:46:47 am i obliged to kill you twice now 01:46:49 yes 01:47:10 will that make you into a zombie 01:47:15 yes 01:47:15 sendmsg(3, {msg_name(12)={sa_family=AF_NETLINK, pid=0, groups=00000000}, msg_iov(1)=[{"L\0\0\0\22\0\1\3@\342\1\0\0\0\0\0\2\0\0\0\0\0\0\0 01:47:15 \0\0\0\0\0\0\0\0"..., 76}], msg_controllen=0, msg_flags=0}, 0) = 76 01:47:21 kmc: so tell me about using netlink 01:50:11 hmm, I reckon that in messages that are clearly mostly out of the ASCII range 01:50:23 debuggers should escape every character, rather than just unprintable characters 01:50:32 because it almost certainly isn't ASCII, and consistent presentation is useful 01:51:15 I like the approach of "display as a string if it contains no unprintable characters, as a list of numbers otherwise 01:51:39 well if you have a long run of printable characters 01:51:41 you want to display as text 01:51:46 in case it's binary data with embedded text 01:52:31 make it an option 01:53:39 but I don't like options very much, and in that case it'd slow me down to have to set it 01:53:59 there's enough information for the computer to work it out automatically in this case 01:54:05 so I shouldn't have to tell it what to do 01:54:34 next you're going to tell me you're working on boehm GC 01:55:11 no, I'm not, is there something wrong with it that I'd need to fix? 01:55:41 no :P 01:55:52 just the "it looks like a duck, so it's a duck" argument is similar 01:56:20 I don't get it 01:56:31 oh, you mean how it looks for things that might be pointers 02:01:14 ais523: i think strace has that option 02:11:07 -!- Sgeo__ has quit (Ping timeout: 260 seconds). 02:27:06 okay seriously 02:27:08 how do you use netlink 02:27:14 i guess i want NETLINK_TCPDIAG 02:27:19 i wonder if there's any docs 02:28:16 http://git.kernel.org/?p=linux/kernel/git/shemminger/iproute2.git;a=blob;f=misc/ss.c;h=a0ab2e97cb8aea3f684d07b38df1cdc278c57bbe;hb=HEAD time to read this mess 02:28:52 oh jesus 02:30:00 ais523: hi 02:31:02 if you want to read 3000 lines of uncommented C code then I recommend you read Ksplice instead, it's a lot more interesting 02:32:04 kmc: but does it solve my problem 02:32:08 smoking salvia is like that TNG episode "The Inner Light" except that the dead civilization that possesses Picard is a race of ten-dimensional hexagons 02:36:43 good answer 02:42:21 IMO kmc should write this code for me 02:44:26 as your attorney i advise you to take a hit of some 20x salvia 02:44:40 perhaps it will fix your netlink sockets 02:45:02 ok 02:45:11 i hear salvia is awful though 02:45:30 I hear people react very differently to it 02:45:37 I tried sucking on the leaves for a bit and it was disgusting 02:45:41 it's pretty terrifying 02:45:41 and I had to spit them out 02:45:48 n.b. you should not heed my legal advice 02:45:53 i am not a real attorney 02:46:03 gasp 02:46:10 even though i do use latin abbreviations like "n.b." 02:46:32 you should spell them all out 02:46:36 nota bene! 02:47:05 sounds good 02:47:15 kmc: none of your advice so far has been legal though 02:47:18 kmc: Did you know the Hebrew letters for "P. S." are N B? 02:47:19 drugs are illegal! 02:47:36 As in נ"ב 02:47:42 shachaf: that sounds confusing. They should've designed hebrew better 02:47:48 actually is salvia still legal 02:47:58 yes, last I looked 02:48:06 I'm pretty sure sage is legal. 02:48:08 I bought a tube of it in boston a couple of years ago 02:48:17 shachaf: there are many varieties of salvia * 02:48:22 :P 02:48:32 Yes. 02:48:45 As long as any of them are legal, you can say that salvia is legal. 02:49:36 yeah always amused by people who go to home depot and buy the first plant that says "salvia" on the label and try to smoke it 02:49:55 i should grow some herbs 02:49:59 best way to have fresh herbs 02:50:08 Why would you smoke drugs? 02:50:09 smoke some parsley 02:50:10 oh and the oyster mushrooms may not be doomed after all 02:50:13 Smoking is disgusting. 03:00:06 http://www.spinics.net/lists/netdev/msg214494.html 03:00:07 okay 03:00:10 so tcpdiag is itself deprecated 03:00:13 and sockdiag is the latest thing 03:00:24 hahahahahahaha 03:00:33 let's rewrite everything again because it's shit 03:00:40 SURPRISE IT'S STILL SHIT 03:00:40 you sure you don't want to try smoking? 03:01:03 kmc: the only thing that isn't shit is @ 03:01:31 because @ is the fixed point of the "let's rewrite everything again because it's shit" operator 03:01:40 the... greatest fixed point 03:02:08 Greatest in what ordering? 03:02:14 the greatest ordering 03:02:43 Can't argue with that. 03:06:25 obviously not 03:18:44 -!- function has changed nick to trout. 03:25:14 apparently romney's victory rally is in Boston 03:41:46 o.O 03:42:04 Well he *was* Governor of the state. 03:42:06 you have to have a victory rally 03:42:09 even if you're going to lose 03:42:14 well, I hope I won't get to use those rotten eggs I had saved up for a rainy day 03:42:19 oh really? 03:44:00 @ping 03:44:01 pong 03:46:54 @ding 03:46:54 pong 03:51:12 -!- hagb4rd has joined. 04:04:58 elliott: Do you feel like fixing a bug in lens? 04:05:03 I bet you do. 04:05:54 what is the bug 04:06:25 https://github.com/ekmett/lens/issues/80 04:07:01 [19:42] shachaf will look into it later. 04:07:06 This is me looking into it. 04:07:34 it could use Bazaar 04:07:35 do that 04:07:58 Why don't you do it? 04:08:53 i don't have commit rights 04:09:07 I'll commit it for you! 04:09:12 Fine, I'll do it. 04:09:20 I don't quite like the holesOf solution. 04:11:22 I forget how sizeof works. 04:11:24 When does it require sizeof? 04:11:25 er. 04:11:27 When does it require parens? 04:11:31 Is "sizeof value" OK? 04:11:38 C? 04:11:42 Or does it need to be "sizeof(value)"? And is that actually "sizeof (value)"? 04:11:44 Yes, C. 04:11:49 It's "sizeof value" 04:11:55 sizeof isn't a function or function-alike. 04:12:02 ISTR you need parens sometimes. 04:12:07 Yes, for types. 04:12:14 You can think of it as the sizeof the cast, I guess. 04:12:17 "sizeof (int)" 04:12:21 Or something, I don't know. 04:12:33 Anyway values don't need parentheses. 04:13:53 C syntax is pretty wacky 04:14:05 there isn't that much of it, but there are some strange bits 04:14:16 "declaration follows use" is just dumb 04:15:55 kmc: Is it portable to do struct { struct nlmsghdr blah; struct my_actual_request req; }? And then use &thatstruct in an iovec. 04:16:02 Like, can't padding mess that up? 04:16:09 Or does Linux guarantee it's OK? 04:16:26 elliott: what are you doing !! 04:17:49 hey this lens library has some pretty nice documentation 04:18:06 elliott: if you want to be sure, use __attribute__((packed)) 04:18:26 yeah that 04:18:59 almost certainly not needed, but if nothing else, it documents that you have made this assumption 04:19:01 Can the TOGA computer be implemented using a 74163 IC? 04:20:18 kmc: it could be needed if the struct starts with a 64-bit integer 04:20:28 er, maybe not 04:20:40 only because nlmsghdr happens to be aligned though 04:20:44 i'm just suspecting that nlmsghdr... yeah 04:20:55 for more or less these reasons 04:21:00 actually, I guess nlmsghdr itself would be declared packed in that case 04:21:09 well, it wouldn't make a difference 04:21:16 * comex shuts up 04:22:22 suspecting that its size is a multiple of the largest alignment gcc will force on struct fields 04:22:25 but not sure 04:22:30 anyway yeah use 'packed' 04:24:39 Oh, are you reading the lens documentation? 04:24:49 yes 04:24:51 on github 04:25:08 People say it needs a lot more documentation. 04:25:23 oh 04:25:41 One day I'll know all the lens operators. 04:25:41 well it has more than "here are the functions and their extremely polymorphic types" 04:26:08 Most places it has "here are the functions and five different specializations of their extremely polymorphic types" 04:26:20 Also it uses doctest! 04:26:20 elliott: if you want to be sure, use __attribute__((packed)) 04:26:22 comex: that's so unportable :( 04:26:25 I thought that was a Python-only thing. 04:26:28 there's doctest for haskell now? 04:26:29 cool 04:26:44 elliott: eh, it works on GCC and every compiler that imitates GCC 04:26:48 comex: I guess the struct is guaranteed to work iff nlmsghdr works though 04:26:53 So those example are actual tests. 04:26:58 which is all the compilers you might use on Linux really 04:27:02 so I suppose I will skip doing the whole copying-memory nonsense to do it "properly" 04:27:55 elliott: pragma pack is more portable, i believe, in case you somehow need to compile it with MSVC 04:28:03 I would like to see an implementation of Checkout in some hardware description language. (Possibly also make a compiler to make a binary, which this implemented hardware then runs.) 04:29:17 comex: MSVC for linux 04:29:26 "a good idea" 04:30:12 did someone say a good idea 04:30:15 i love good ideas 04:30:31 monqy: do you collect them 04:30:45 yes 04:30:49 monqy: make it sadder 04:30:52 i feed them well and also let them drink 04:31:49 more like MICRO$HAFT VENEREAL COCKS 04:32:19 good one kmc 04:32:29 thachaf 04:32:29 the best one ?? 04:32:33 "yes" 04:32:53 __u8 pad; 04:33:00 high tech padding from linux kernel headers industry 04:33:30 /* Bytecode is sequence of 4 byte commands followed by variable arguments. 04:33:30 * All the commands identified by "code" are conditional jumps forward: 04:33:30 * to offset cc+"yes" or to offset cc+"no". "yes" is supposed to be 04:33:30 * length of the command and its arguments. 04:33:30 */ 04:33:32 oh good grief 04:33:43 what is this 04:33:49 bytecode for what 04:33:53 bytecode 04:33:58 for 04:33:58 the code of bytes 04:33:59 inet_diag 04:34:05 why does this involve bytecode 04:34:43 Should I read _Purely Functional Data Structures_? 04:34:46 yeah 04:34:57 It looks good. 04:35:01 it feels good 04:35:15 theegan 04:37:47 why does that involve bytecode... 04:38:10 monqy: depends what `that' is 04:38:19 whatever involves bytecode 04:38:29 why does bytecode involve bytecode 04:38:32 in the grim linux of the far future there is only bytecode 04:39:14 the children of the generation of the future: they're just bytecode??? humankind is dead. 04:39:48 monqy: im already bytecode 04:40:27 hows it like being subhuman 04:41:22 monqy: superhuman 04:42:22 hows it like being wrong 04:42:56 monqy: how's it like being racist against bytecode :/ 04:43:25 racist against every kind of code 04:43:27 except unicode 04:56:19 kmc: okay apparently I actually have to send bytecode over netlink to find out the pid that owns a socket 04:56:38 so 04:56:39 that's a thing 04:57:20 shachaf++ 04:58:08 stupid future 05:01:38 this request code is going to be like 05:01:40 60 lines 05:01:49 and that doesn't even count parsing the response 05:02:35 what are you even doign 05:04:33 -!- augur has quit (Remote host closed the connection). 05:04:49 monqy: you dont want to know.......... 05:05:30 Is ais523 on? 05:06:36 * pikhq feels too awesome 05:09:31 zzo38: yes 05:09:32 ais523: sorry 05:09:41 kmc: q 05:09:42 monqy: p 05:09:50 * pikhq feels far too awesome 05:10:35 that's a dangerous feeling 05:10:44 elliott: ? 05:10:47 Potentially 05:11:21 pikhq: oh? 05:11:28 I love that feeling 05:11:31 what did you accomplish? 05:11:39 Girlfriend get 05:11:39 monqy: yes. 05:12:29 lol 05:13:21 pikhq: congrats! 05:14:11 Perhaps a tiny bit weird; hard to be happy at people IRL 'bout it. Thus, here. 05:14:21 nah, makes sense 05:14:35 people get upset that you're gloating if you do it IRL 05:15:02 q 05:15:40 hi 05:23:39 coppro: Bah 05:24:32 ? 06:13:16 It is time to take out the trash, and we need to stop supporting pay basis distros such as communist Redhat. They are charging us for the stuff thats spose to be free. 06:13:59 We need only juche distros. 06:14:52 shachaf: please tell me you are not serious 06:15:36 I'm completely serious, coppro 06:15:51 I know that most of you would rather run Windows than Redhat. This is my opinion again, but personally Redhat runs like crap even on newer systems. Stock package management and static binaries may save time, but in the long run they lower your peak perormance and also make the system unstable, and as useless as a used condom. Even recompiling the kernel on Redhat causes problmes. Their modern init scripts interfere with some newer kernel ... 06:15:57 ... operations. It takes as much time to get Redhat working right as Gentoo. 06:16:05 oh 06:16:06 now it makes sense 06:16:10 is this funroll-loops 06:16:50 yes 06:16:56 goddamn freeloading redhat, always taking our stuff and never giving back 06:17:43 i'd just like to say, there are so many uses for a used condom 06:24:01 -!- Sgeo has joined. 06:28:09 There is the chess variant "Chaotic Chess", which is not really quite as chaotic as they say it is. Each piece, after move, change into another kind of piece of your choice, subject to restriction cannot change into the pair Pawn-King, Knight-Queen, Bishop-Rook. Win if your opponent has no king. 06:33:25 -!- FreeFull has quit (Quit: Bye). 06:38:38 -!- augur has joined. 06:56:20 * Sgeo noticies the elixir language 07:02:17 hi whats that 07:05:39 http://elixir-lang.org 07:05:49 hi :') 07:06:01 Language that runs on the Erlang VM. Syntax looks vaguely Rubyish, but it has macros 07:06:19 Sgeo: is it better than clojure? is it better than tcl 07:06:23 is it better than erlang 07:06:24 is it better than factor 07:06:30 is it better than common lisp 07:06:54 is it better than.... 07:06:55 @ 07:07:04 lambdabot: @ 07:07:04 Maybe you meant: . ? @ activity activity-full admin all-dicts arr ask b52s babel bf bid botsnack brain bug check choice-add choose clear-messages compose devils dice dict dict-help djinn djinn-add 07:07:04 djinn-clr djinn-del djinn-env djinn-names djinn-ver do docs dummy easton echo elements elite eval fact fact-cons fact-delete fact-set fact-snoc fact-update faq farber flush foldoc forget fortune 07:07:04 fptools free freshname ft gazetteer get-shapr ghc girl19 google googleit gsite gwiki hackage help hitchcock hoogle hoogle+ id ignore index instances instances-importing irc-connect jargon join karma 07:07:04 karma+ karma- karma-all keal kind learn leave let list listall listchans listmodules listservers localtime localtime-reply lojban map messages messages? more msg nazi-off nazi-on nixon oeis offline 07:07:04 oldwiki palomer part paste ping pl pl-resume pointful pointless pointy poll-add poll-close poll-list poll-remove poll-result poll-show pretty print-notices protontorpedo purge-notices quit quote rc 07:07:06 read reconnect remember repoint run shootout show slap smack source spell spell-all src tell thank you thanks thx ticker time todo todo-add todo-delete topic-cons topic-init topic-null topic-snoc 07:07:08 topic-tail topic-tell type undefine undo unlambda unmtl unpf unpl unpointless uptime url v vera version vote web1913 what where where+ wiki wn world02 yarr yhjulwwiefzojcbxybbruweejw yow 07:07:28 @yhjulwwiefzojcbxybbruweejw 07:07:28 "\"\\\"\\\\\\\"\\\\\\" 07:07:28 @yow 07:07:28 Couldn't find fortune file 07:07:30 I've always been curious about Erlang's approach to things, but the syntax does bother me, although I think the final straw was when the online book I was reading said that hot-swapping was a thing to try to avoid and is difficult 07:07:48 well, that was enlightening. 07:07:49 I think my only use-case for hot-swapping is to make development faster. 07:08:16 @get-shapr 07:08:16 shapr!! 07:08:24 can we not do the lambdabot: @ thing 07:08:48 these bots are made for spammin' 07:08:57 isn't erlang syntax just prolog syntax taken too far 07:09:10 oh no 07:09:14 how far is too far 07:09:15 can it be vaguely more productive spam than lambdabot: @ 07:09:18 because that is really annoying 07:09:21 thanks 07:09:21 I'm going to venture a guess that Elixir is better syntax-wise than Ruby 07:09:32 whoa, wild guess man 07:09:35 Sgeo: are you a venture capitalist 07:12:57 Sgeo: what makes you say that. are you going to learn elixir now. what makes you like elixir. 07:13:15 is elixir the new clojure. 07:14:50 exciting world of new languages and doing(?) things(??) with(????) them(???????) 07:15:48 doing things with, n. telling #esoteric all about 07:15:51 wait 07:15:53 that's not a noun 07:15:53 that's a lot of question marks monqy 07:15:54 fuck dictionaries 07:15:58 don't use up your lifetime supplies 07:16:26 watching from a distance, community involvement(???) 07:16:54 is there an elixir reddit you can teach about how elixir works 07:16:55 "lambdabot: @" "Maybe you meant: @" Yeah, maybe. 07:17:16 Elixir restores both your HP and MP to full. 07:27:46 -!- ais523 has quit. 07:40:56 ... 07:41:04 I think Elixir's "records" are mutable. 07:41:05 -!- lambdabot has quit (Remote host closed the connection). 07:41:23 oh gosh 07:41:27 mutable state in 2012?? how unfashionable 07:41:36