00:00:10 what has the biggest number to do with that, though? 00:00:23 the biggest number is needed for $'s 00:00:30 oh, right 00:01:32 well, we can try with (m+1)n 00:03:41 actually hm 00:05:10 -!- Dovregubben has quit (Read error: Connection reset by peer). 00:06:39 (m-1) is enough for the $'s 00:07:04 (m-1)*n_$ + 2*n_: + n_rest 00:07:36 uh 00:07:47 I'm pretty sure that's wrong 00:08:00 simpel counter example: 00:08:08 oh hm 00:08:42 hum 00:09:09 * Arc_Koen begins to understand why Ftack is so useless 00:09:16 heh 00:09:55 it probably should be max(m-1, 1) 00:10:37 ahah 00:12:02 ok I have an idea 00:12:20 let me just check that it is really a counterexample 00:12:46 ok it's most definitely not 00:12:48 but it's fun 00:12:57 ~H~~~~~~~~~~~~~~~~~~~ 00:13:05 heh 00:13:36 -!- monqy has joined. 00:14:17 we have to define "(m-1)*n_$" to be 0 if there are no numbers though, I think 00:14:33 hm... 00:14:49 i guess if there are no numbers you are catting the moment you hit a $, so yeah 00:14:58 otherwise we could include a $ as the last element 00:15:36 +1 symbol, +1 step, -1 in the bound 00:16:45 hmm 00:16:54 (I'm not sure) 00:17:22 actually you can also have $0whatever 00:17:55 yes but that's +3 symbols, +1 step 00:18:01 oh 00:18:07 yeah, ok 00:18:35 so max(m-1, 0) with 0 even if m doesn't exist 00:18:52 -!- Dovregubben has joined. 00:19:17 -!- Nisstyre_ has quit (Read error: Connection reset by peer). 00:19:49 argh 00:19:59 well this is close to correct anyway, i think 00:20:08 hmm? 00:20:36 oh $0$ 00:20:45 haha 00:20:51 it works though 00:21:04 -!- Nisstyre_ has joined. 00:21:11 bound is 1, step is 1 00:21:14 yes, but we need the max 00:21:59 $1$ 00:22:27 i guess it works 00:22:49 note that only :~!( can be applied to themselves 00:22:58 yes i already did 00:23:19 and of those, : and ( give end states 00:23:20 I'm feeling so helpful right now 00:23:28 yup 00:23:34 while ~ and ! keep shrinking 00:24:08 ~ +1 step +1 symbol +1 to the bound 00:25:37 !x + 1 step either + 2 or +m to the bound 00:26:13 err, with a minimum of +1 00:26:19 ok, both are useless 00:30:55 btw, ./fueuec thuemorse still ends up on a segfault, even with catching empty queues 00:32:00 (and being a while loop) 00:32:00 and the suggestion to not use recursion? 00:32:45 oh, wait 00:33:37 fizzie seemed to have debugger proof that it overflowed the stack 00:34:19 yes 00:34:26 I guess I should learn to use debuggers 00:34:47 hey me too! 00:34:49 or maybe use an actual ide 00:35:12 TextEdit does have limitations 00:35:21 * oerjan is a vim guy 00:35:43 well, at least vim is related to programming 00:36:03 but i don't even have any of the advanced vim setup some people do 00:39:20 Arc_Koen, you can use TextEdit? 00:39:34 yes 00:39:39 why could I not? 00:39:43 * Arc_Koen is a mac user 00:39:44 ISTR it only writing RTFs, but I was young and stupid back then. 00:40:39 well, it's a text editor 00:40:45 like windows's blocnote 00:41:04 maybe sliiiightly more advanced as it can do some basic text formatting 00:41:16 but I'm definitely not using that 00:41:46 fun fact: it has an auto-spell check 00:41:59 if you don't disable it, it will correct what you're typing without asking 00:42:12 I wasn't aware of that, it messed with a lot of ocaml programs 00:42:47 like 'rec' which was systematically turned into 'ref', or the opposite 00:45:15 urrh 00:45:23 segfaulted on the alphabet 00:45:31 ok, this is a bug that should be corrected easily 00:49:42 -!- subleq has joined. 00:49:45 hello 00:50:32 haha 00:50:57 `welcome subleq 00:50:58 empty program + ^D successfully prints an infinite amount of ? 00:51:01 subleq: 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.) 00:51:02 hello subleq 00:51:17 -!- copumpkin has changed nick to ChuckPonzi. 00:51:30 For the other kind of esoterica, try #esoteric on irc.dal.net, just incase 00:51:43 Or try Wikipedia. 00:52:08 as if someone named subleq 00:52:26 Just in csae. 00:52:27 monqy: i was being ironic 00:52:54 ^just in case 00:54:20 oerjan: funny bug... ./fueuec '--65' works as it should, but ./fueuec --print '--65' reveals it doesn't -- 65 65 0 A 0 (waits for input) 00:56:52 wut 00:56:59 fancy, uh? 00:57:39 so there's one (or more) function I've messed with which somehow prevents stuff to be sent back to queue 00:57:41 -!- nys has joined. 00:57:44 -!- monqy has quit (Quit: hello). 00:58:42 and that makes both ) and - destroy the next element 01:01:59 ok, now - and + interpret anything that's not a number as if it were a 0 01:05:42 ah, i see all the old folks are here 01:05:43 hi oerjan 01:06:06 hi subleq 01:07:09 Arc_Koen: something wrong with the use of matchwhat? 01:07:40 if you assume that is true you might get such an effect. 01:07:42 I guess 01:07:55 oh, of course 01:08:14 things would have gotten serious if I had used an union 01:08:40 s/serious/fancy 01:08:56 hi oklopol 01:09:30 urrrrh 01:09:32 haha 01:09:41 I used !*s as meaning "s is not empty" 01:09:56 OKAY 01:10:37 see, s != "" worked just fine 01:10:38 Arc_Koen: i suppose s[0] != '\0' is more readable, and probably compiles down to the same thing. 01:10:39 :p 01:10:43 heh 01:11:03 oh so matchwhat _always_ returned true? :P 01:11:30 yep 01:11:36 well i guess unless the queue is empty 01:11:47 no no, always 01:12:29 matchwhat is structured as "if (neither string nor queue are empty) {...} else if (string not empty) { false} else true" 01:12:55 i have a hunch that the entirety of matchwhat could be rewritten as a single expression 01:13:00 so, thinking all strings were empty, he always returned true 01:13:05 yes 01:13:12 in ocaml I wrote it another way 01:13:52 and could have wrote it many different other ways but heh one is enough 01:15:27 -!- impomatic has left. 01:16:41 +HH still transforms into 0 though 01:16:51 eek 01:16:59 paste the code? 01:18:09 http://sprunge.us/XfgY 01:18:21 http://sprunge.us/XfgY?c actually 01:23:27 -!- DHeadshot has quit (Read error: Connection reset by peer). 01:26:19 oerjan: adding printf("matchwhat: %s\n", (itsok?"TRUE":"FALSE")); in matchwhat shows matchwhat isn't even called 01:26:32 huh 01:27:28 so it's like I made a fancy error with !*s which wasn't given an occasion to mess around 01:27:30 If you make (s != "") in C it makes if the address does not match the address of "" so it is different. 01:28:09 haha 01:29:29 But !*s means if s points to a value zero so for string meaning, empty string. 01:29:38 Arc_Koen: does it get to if (op == '+') 01:29:44 ? 01:30:08 I'll add a printf but I guess it does, since + 3 4 gives 7 01:30:50 ok, uh, I agree this doesn't make any sense 01:31:31 -!- ChuckPonzi has changed nick to copumpkin. 01:31:37 kmc: FSVO "nicer" 01:31:41 ARRRH STUPID ME 01:32:01 I value hours not spent on a plane much more than hours spent on a plane being able to watch movies. 01:32:02 was using a different name for exec file when compiling and executing 01:32:15 ooooo 01:33:04 so basically I was using the last version which used that name, which probably was right after switching to !*s 01:33:31 With that said, Virgin America's was probably the nicest domestic flight I've been on. 01:35:28 Arc_Koen: fancy 01:35:55 ok, it works with the wiki page's sample programs 01:36:08 yay 01:37:27 if you have any other remarks on the code... :-) 01:38:20 not haskelly enough ;P 01:38:37 haha 01:39:06 so hum, you have proven fueue was turing-complete? 01:39:15 essentially. 01:39:48 oerjan has even proven that norway is turing complete 01:39:59 erm 01:40:15 -!- pikhq_ has joined. 01:41:20 okay so up to now I was think those 46 46 46 111 117 116 32 111 102 32 115 116 97 99 107 33 10 things were supposed to be used by underload 01:41:31 :P 01:41:40 -!- pikhq has quit (Ping timeout: 248 seconds). 01:41:42 ^ul ! 01:41:43 ...out of stack! 01:41:51 yeah, I just had the idea to 01:41:52 wait 01:41:52 i just copied it from fungot's underload 01:41:52 oerjan: i already have 01:42:00 ok, ok 01:42:50 for a second I thought "^ul !" meant "find the last ascii-looking sequence of numbers written on this channel and translate it" 01:42:59 XD 01:43:36 i then had to add the initial [] when i realized that the translation of : would gobble up the first 46 before it could get printed. 01:44:02 since things happen slightly in parallel, more or less 01:45:24 what if they are two : in the underload program? 01:45:36 won't the first eat [] and the second 46? 01:46:10 shachaf: well, HEL-JFK and JFK-SFO are of comparable length 01:46:29 two consecutive underload functions aren't translated to two consecutive fueue subprograms. you need to use the AB translation to concatenate underload programs. 01:46:54 i'm not sure about the subjective difference between 6h and 8h flights 01:46:59 they both just register as "a long time" 01:47:04 so only one of the :'s will be running at a time. 01:47:18 shachaf: did you see http://www.ethanzuckerman.com/blog/2012/08/30/long-flights-a-somewhat-serious-business-idea/ 01:47:54 the ! translation also would gobble up the [], in a very literal fashion since it happens precisely at the spot where i use a fueue ! to remove the popped element 01:48:12 oh, ok 01:48:18 kmc: Hmm, I suppose that's true. 01:48:41 * shachaf is used to 10-12-hour international flights and 4-5 hours domestic flights. 01:48:51 Nope. 01:48:57 well, I'm not sure that's explicit on the wiki page, but I don't speak underload so i trust you 01:49:26 i didn't explain it much i guess :P 01:50:03 kmc: I find myself unable to concentrate on anything in flights. 01:50:06 i suppose i should add something about A and B representing nested underload programs 01:50:36 shachaf: why? 01:51:57 I don't know. I wish it wasn't the case. 01:52:12 I'm also rarely able to sleep. 01:53:28 nested? 01:53:34 "Conversation will be discouraged by a loud white noise machine that permeates the space, encouraging you to put on headphones and listen to whatever music you’ve brought with you." 01:53:43 I can also usually not get anything done while listening to music. 01:53:53 yeah, that's one of the dumber parts of this proposal 01:54:00 one of the commenters suggests that you rely on social norms to shut people up 01:54:07 like the library 01:54:42 Oh, and one of the most important of long flights is the loud baby in the row behind you. 01:54:54 * shachaf accidentally a lot of words these days. 01:56:40 Arc_Koen: i mean, to translate the underload program (:^):^ say, you translate A = (:^) B = :^ recursively and use the AB rule to combine them. 01:57:09 or you can split it as A = (:^): B = ^ , doesn't matter much 01:57:39 what if you split it as A = (:^):^ B = empty ? 01:58:16 oh, wait, you must continue splitting until you have only one-symbol programs? 01:58:24 Yep. 01:58:31 ok :) 01:58:46 but you made me realize i actually forgot to represent the empty program in my translation 01:59:06 which is important, since you can have use for () which uses it recursively 01:59:28 shachaf: what about music without words 02:00:09 Sometimes it's OK... Sometimes it still doesn't work. 02:00:40 I should probably figure it out and collect some non-disruptive music. 02:04:52 Arc_Koen: expanded the wiki table 02:06:36 * oerjan OCD's on periods 02:07:44 Arc_Koen: oh and btw the subprograms must have matching parentheses, as always in underload, which is why the (A) rule is separate. 02:08:38 okay, but ( and ) are underload commands that are quite unrelated, are they? 02:08:53 I mean, the correct parenthesing is only an arbitrary rule, right? 02:09:10 no, they always match. (A) is more or less similar to fueue (A 02:09:20 or wait hm 02:09:25 fueue [A] 02:10:53 except that fueue [A] does nothing, while underload [A] moves itself from the running program to the stack 02:11:00 er 02:11:04 *underload (A) 02:11:56 did I mention it was 4am? 02:12:05 ...not that i recall. 02:12:40 well, it is, so i'll take off 02:12:45 good night then 02:12:53 thank you 02:14:22 zzo38: I think I corrected all bugs and memory leaks from fueue.c so i'd be thankful if you could replace it http://sprunge.us/iEAW 02:14:32 bye 02:14:39 -!- Arc_Koen has left. 02:20:00 -!- monqy has joined. 02:32:47 Review what I have written more of RogueVM; I have written a lot more by now. 02:43:28 kmc: "ANSI Windows code pages, and especially the code page 1252, were called that way since they were purportedly based on drafts submitted or intended for ANSI. However, ANSI and ISO have not standardized any of these code pages." 02:44:58 1252 at least is *compatible with* the ISO charset it gets confused with. 02:47:32 -!- Phantom_Hoover has quit (Read error: Connection reset by peer). 02:47:58 of course 03:02:42 i'm using gitit 03:02:49 i like that i can export any dumb page on my wiki as LaTeX or man page 03:06:53 or a S5 slideshow 03:16:48 -!- pikhq has joined. 03:16:53 -!- pikhq_ has quit (Ping timeout: 246 seconds). 03:34:04 -!- ogrom has joined. 03:46:48 -!- pikhq_ has joined. 03:47:09 -!- pikhq has quit (Ping timeout: 255 seconds). 03:51:47 -!- pikhq has joined. 03:52:13 -!- pikhq_ has quit (Ping timeout: 246 seconds). 04:11:07 -!- oerjan has quit (Quit: leaving). 04:32:34 -!- Concreto has joined. 04:57:23 OK, I wrote a lot more of RogueVM now: http://zzo38computer.cjb.net/roguevm/roguevm.tex http://zzo38computer.cjb.net/roguevm/roguevm.dvi Now I have many more instructions, more about the display, a complex numeric print mode, and various other things added or modified in case they were wrong or missing before. 05:05:08 Is there a simple way to make CTR mode suitable for disk encryption? 05:06:04 shachaf: I am going to sleep, but remind me about that question tomorrow. 05:06:20 @ask kmc Is there a simple way to make CTR mode suitable for disk encryption? 05:06:21 You can tell yourself! 05:06:23 @tell kmc shachaf: I am going to sleep, but remind me about that question tomorrow. 05:06:23 Consider it noted. 05:06:30 thachaf 05:06:30 kmc: You have 1 new message. '/msg lambdabot @messages' to read it. 05:06:36 bah 05:06:39 * kmc -> 05:06:40 "whoopse" -- kmc 05:06:50 kmc: Back to normal-people sleep schedule, are you? 05:07:00 @tell kmc 22:06 @ask kmc Is there a simple way to make CTR mode suitable for disk encryption? 05:07:01 Consider it noted. 05:12:02 -!- nys has quit (Ping timeout: 246 seconds). 05:13:30 -!- Vorpal has joined. 05:25:24 -!- Vorpal has quit (Ping timeout: 276 seconds). 05:29:35 -!- Concreto has quit (Quit: who gives a shit?). 05:53:09 -!- impomatic has joined. 06:22:31 -!- Lumpio- has quit (Ping timeout: 245 seconds). 06:32:01 -!- zzo38 has quit (Remote host closed the connection). 06:33:26 -!- ogrom has quit (Quit: Left). 06:48:46 -!- asiekierka has joined. 06:49:54 -!- Lumpio- has joined. 07:12:17 -!- derdon has joined. 07:16:22 @tell kmc Also, can you easily turn a stream cipher into a block cipher? (Also, some other things.) 07:16:22 Consider it noted. 07:16:52 * shachaf will save conversation for tomorrow. 07:27:59 Has anyone used Joy? What do you think of it? 07:48:27 -!- nooga has joined. 07:48:54 -!- AnotherTest has joined. 08:03:10 -!- Vorpal has joined. 08:38:53 -!- nooga has quit (Ping timeout: 260 seconds). 09:22:38 -!- atriq has joined. 09:25:15 @messages? 09:25:15 Sorry, no messages today. 09:25:26 > hi "atriq" 09:25:28 hi atriq 09:25:32 @src hi 09:25:32 Source not found. Maybe if you used more than just two fingers... 09:26:19 @tell Arc_Koen Remarks: You have another memory leak: consider ![...] or $n[...] -- in both, you throw away the block Token with deletetop(), but it doesn't free the Queue data in it; yet the only pointer to that data is lost. See http://sprunge.us/WiHK 09:26:19 Consider it noted. 09:28:03 "Two Fingers" is a brand of tequila. I don't know where the name comes from. 09:29:10 I ought to upload my Fueue interpreter somewhere 09:31:23 Is there a way I can pipe something directly to the clipboard? 09:31:40 foo > /dev/clipboard or something like that 09:31:58 > hi "atriq" 09:32:00 hatriq 09:32:01 > hi "elliott" 09:32:03 helliott 09:32:04 > hi "oerjan" 09:32:06 hoerjan 09:32:07 > hi "kmc" 09:32:08 h 09:32:10 Whoops. 09:32:11 Oh no! 09:32:14 > hi "keegan" 09:32:16 heegan 09:32:16 atriq: Depends on which clipboard you're talking about. For the X clipboards, there are several command-line tools. 09:32:30 'xclip' is one, if you happen to have it installed. 09:32:59 It uses the primary selection by default, and the CLIPBOARD selection if you ask it to. 09:33:38 "foo | xclip" will let the middle-mouse-paste paste the output of foo, for example. 09:33:49 * shachaf uses xsel 09:33:57 ...Because it usually seems to be installed. 09:34:08 It's not, here. 09:34:12 Though neither is xclip. 09:34:47 It looks very similar, except with slightly more bells on it. 09:41:47 @tell Arc_Koen Remarks: I'm still a bit confused by is_empty not taking a Queue*; the "const Queue q" parameter declaration is especially confusing, since the 'const' can't really do anything too meaningful; it's passed a new Queue by value, so from the viewpoint of the caller, it couldn't modify it anyway. 09:41:47 Consider it noted. 09:42:30 @tell Arc_Koen Remarks: In main, printf("Too many arguments.\n", argc); should either lose the "argc", or add a corresponding "%d" in the format. (The '-Wall' gcc flag would've noticed this.) 09:42:30 Consider it noted. 09:43:07 Aaargh 09:43:20 Why is getting syntax highlighting for Haskell on Tumblr so hard 09:46:41 -!- MoALTz has joined. 09:47:56 atriq: Perhaps you should start a more Haskell-friendly competitor, called Humblr? 09:53:46 I don't think anyone questions that some of us have or haven't done that. 10:17:37 there is no integer between 1 and 2! 10:22:05 std::vector { 1, 3, 2} 10:22:14 itidus21: now there is 10:22:29 gasp 10:22:29 * insert = between > and { 10:23:02 (and a name of choice of course) 10:23:53 AnotherTest: Can you do that nowadays? I remember a common complaint about std::vector and such that you can't use the {a,b,c} initializers with them. 10:24:16 fizzie: yes, that's C++11's initialize list feature 10:24:21 Funky. 10:24:24 *initializer 10:24:53 std::initializer_list I think 10:24:58 Yes, I see. 10:25:02 and yes, that was very very annoying 10:29:33 i always admired the DATA statement in basic 10:31:14 mroman: is the source code of stlisp available somewhere? 10:55:04 -!- kinoSi has quit (Read error: Connection reset by peer). 10:55:31 -!- kinoSi has joined. 11:06:45 AnotherTest: No. 11:06:50 The source code has gone missing. 11:10:25 As I'm not a person who uses backups. 11:18:50 -!- Phantom_Hoover has joined. 11:35:03 "You need me on your staff, because I'm a man who thinks." 11:35:58 -!- nooga has joined. 11:40:15 -!- atriq has quit (Quit: Leaving). 11:42:34 -!- Arc_Koen has joined. 11:43:23 hello 11:43:23 Arc_Koen: You have 3 new messages. '/msg lambdabot @messages' to read them. 11:43:28 @messages 11:43:28 fizzie said 2h 17m 9s ago: Remarks: You have another memory leak: consider ![...] or $n[...] -- in both, you throw away the block Token with deletetop(), but it doesn't free the Queue data in it; 11:43:28 yet the only pointer to that data is lost. See http://sprunge.us/WiHK 11:43:28 fizzie said 2h 1m 41s ago: Remarks: I'm still a bit confused by is_empty not taking a Queue*; the "const Queue q" parameter declaration is especially confusing, since the 'const' can't really do 11:43:28 anything too meaningful; it's passed a new Queue by value, so from the viewpoint of the caller, it couldn't modify it anyway. 11:43:28 fizzie said 2h 58s ago: Remarks: In main, printf("Too many arguments.\n", argc); should either lose the "argc", or add a corresponding "%d" in the format. (The '-Wall' gcc flag would've noticed this. 11:43:30 ) 11:44:02 http://www.reddit.com/r/guns/comments/z55a3/has_there_been_more_highprofile_gun_crime_lately/c61pen4 11:44:08 Ladies and gentlemen, /r/guns. 11:48:31 I didn't say those things with that many newlines in the middle. 11:53:24 So handing out a gun to everyone doesn't work out quite as expected? 11:57:22 fizzie: do you think I should change is_empty(Queue q) to is_empty(const Queue *q)? 11:58:26 I mean, there is not really any argument why to use a pointer here; it would be like having something like is_empty(const char **s) to test emptiness of a string 11:58:43 or is_zero(const int *n) 11:59:10 "and how soon before they have a drug where side effects may include rectal ventriliquism. if your asshole starts talking call a doctor.. or get friends over cos it's gonna be a fun night" 12:24:32 hi shachaf 12:24:32 kmc: You have 3 new messages. '/msg lambdabot @messages' to read them. 12:24:39 against my better judgement, i am awake 12:24:43 kmc: Just as I was about to go to sleep. :-( 12:24:50 @tell oerjan ./fueuec '~!~)): [[48 [)):] [~!~)):] ~~) !][49 [~!~)):] [)):] )~]]' 11010011001011010010110011010011001011001... '110' seems a weird start for the thue morse sequence... 12:24:50 Consider it noted. 12:24:53 Did you sleep tonight yet? 12:24:58 yes 12:25:07 Oh. You just woke up early? 12:25:10 but my alarm went off by accident at 08:00 12:25:16 Doing better than I am, then. 12:25:22 and i haven't been able to get back to sleep 12:25:27 I wish my alarm could just wake me up. That would be great. 12:26:44 what does it do instead? 12:27:15 Well, it beeps. 12:27:21 Occasionally I wake up. 12:29:23 shachaf: can you use an ESSIV for the nonce part of the counter 12:29:26 isn't that enough? 12:30:06 ESSIV is "generate an IV based on the sector number"? 12:30:29 based on the sector number and the hash of the key 12:30:30 yeah 12:30:54 i guess this is not secure though 12:30:57 That's very broken since you can overwrite old blocks. 12:31:00 yeah 12:31:13 x ^ y will give you the xor of the plaintexts. 12:31:15 you end up reusing the keystream otput from CTR 12:31:17 maybe the internet should shut down overnight 12:31:58 whereas reusing the IV for CBC is not as much of a problem 12:32:03 but still bad 12:32:23 You get the property that "someone can see if what you wrote to a block has ever been there before". 12:32:35 Which is undesirable in normal uses of CBC but maybe doesn't matter in disk encryption. 12:32:56 (By the way, I don't see the advantage of thinking of the "nonce part" and the "counter part" of the CTR input, rather than just a "counter that starts at a random value".) 12:33:16 right 12:33:22 (Alternatively, ciphertext_i = ENC(key, iv + i) ^ plaintext_i) 12:33:51 did i tell you i saw a cold boot attack on a spy show on TV? :) 12:34:53 No. 12:35:01 I didn't know that was called a cold boot attack. 12:36:02 I,I called boot attack 12:40:30 i guess it's a pun 12:40:47 05:40 puns aren't worth much. 12:41:08 i meant "cold boot" 12:41:23 I know. 12:41:26 ok 12:41:42 Is that a fictional spy show or a nonfictional spy show? 12:41:43 should i give an informal talk at MIT (a "SIPB cluedump") on how to screw up using cryptography? 12:41:56 the show is real; it is a work of fiction 12:41:57 Yes! 12:41:57 ;P 12:42:02 the show is Burn Notice 12:42:15 it is a fairly ridiculous show but lots of fun to watch 12:42:27 Also you should come up with a bunch of CTF challenges that get cryptography wrong. 12:42:33 And then we can make a #esoteric CTF out of it. 12:42:38 wish i could make nelhage and gdb do that 12:42:39 Since Stripe probably won't do it. 12:42:42 i think i won't have time 12:43:00 -!- Phantom__Hoover has joined. 12:43:06 Hmph. 12:43:14 I'll do some if you do some! 12:43:24 also it's annoying to run a CTF 12:43:26 * shachaf wonders whether making good CTF problems is really hard. 12:43:29 because people try to hack your shit 12:44:01 did you see https://blog.gregbrockman.com/2012/08/system-design-stripe-capture-the-flag/ 12:44:06 i don't know why this wasn't on the stripe blog 12:44:24 Yes. 12:44:32 He talked about it at the CTF event. 12:44:50 Arc_Koen: Sure there is: large structs are conventionally passed as pointers in C, to cut down the amount of necessary copying. (Admittedly a compiler can sometimes optimize. But not without inlining, while still following ABI rules.) 12:44:56 Apparently several of the levels had alternate solutions that they didn't think of. 12:45:16 -!- Phantom_Hoover has quit (Ping timeout: 246 seconds). 12:45:23 fizzie: a queue is a struct made of one int and two pointers 12:45:24 For example level 5 was solved by nearly everyone using an unintentional bug, it seems? 12:46:45 which bug? 12:46:47 Arc_Koen: Yeah; anything larger than a pointer quite often counts as large. Anyway, it's a convention, doesn't need to make all that much sense. 12:47:03 Apparently you were supposed to make a response which consisted entirely of a string matching the regexp 12:47:13 /somethingAUTHORIZEDsomething$/ 12:47:15