00:00:25 FRACTRAN is a cool esolang 00:01:54 what im sayin 00:09:29 Godddammit dropbox stop being slow as pain 00:10:10 http://www.jerkcity.com/_jerkcity4523.html deuce explains the commutator 00:10:12 asdfasdfasdf the select files box did open I just dudn't see it 00:12:44 http://dl.dropboxusercontent.com/u/16240872/crossdomain/demo.htm 00:13:03 -!- sebbu has quit (Ping timeout: 245 seconds). 00:13:16 Hmm. "Origin http://dl.dropboxusercontent.com is not allowed by Access-Control-Allow-Origin." 00:13:32 Does that imply that farmingdale is in fact sending Access-Control-Allow-Origin data? 00:13:42 -!- sebbu has joined. 00:14:12 -!- sebbu has quit (Changing host). 00:14:12 -!- sebbu has joined. 00:15:06 Switched it to example.com 00:15:48 Sgeo: is the domain of a .js file actually the domain of the file including it? 00:16:24 It should be, I think. 00:16:29 * tswett nods. 00:16:34 Well, the domain of the global window object 00:16:41 Whatever it is 00:16:50 So, you guys know how Finnish is a cipher of Japanese? 00:16:59 German is likewise a cipher of French, right? 00:17:31 Certainly. 00:17:39 Oh, switching to example.com is dumb 00:17:45 common misconception, it's actually romansh it's a cipher of 00:19:03 Finnish is weird. 00:19:18 It's all like, "fingo ga hen da", only encrypted! 00:20:08 lol 00:21:08 http://snyfarvu.farmingdale.edu/~goldsj6/crossdomain/demo.htm 00:21:08 WAT. 00:21:14 goddammit dropbox 00:21:28 more like fartbox (a box you fart in) 00:21:41 Apparently dropbox supports CORS? 00:21:52 I need a packer sniffer 00:22:17 Sgeo: sweet blank page 00:22:19 Ok. I want two domains, neither of which support cors 00:22:26 Bike, look at Javascript console. 00:22:43 And at where resources are coming from 00:22:44 And source 00:23:39 the same combination on my luggage 00:25:03 It's retrieving it cross-domain 00:26:13 that counts as cross domain? 00:26:20 * pikhq feels like taking a risk; apt-get update 00:26:22 the script and the info are both dropboxusercontent. 00:26:39 That should count as cross-domain, although my boss disagrees. 00:27:04 why should it? 00:27:18 Because the page itself is on snyfarvu.farmingdale.edu 00:27:45 i mean, what's the securty flaw 00:28:17 Let's say gmail.com hosts some .js file. If I can trick the file into retrieving something interesting then calling my code, I now have something interesting. 00:28:28 -!- nooodl has quit (Ping timeout: 245 seconds). 00:33:34 Hey, so you guys know immibis? 00:34:35 Can I disable CORS somehow? 00:38:08 -!- Bike_ has joined. 00:38:56 -!- Bike has quit (Ping timeout: 256 seconds). 00:41:46 Warning: This method must not be called from JavaScript. 00:41:48 Hmm... 00:42:32 (init() on an XMLHttpRequest in Firefox) 00:43:37 -!- Bike_ has changed nick to Bike. 00:44:41 what 00:48:36 https://developer.mozilla.org/en-US/docs/DOM/XMLHttpRequest 00:49:16 Disable cors? Why 00:49:20 CORS is great! 00:50:11 type: Function? 00:51:27 I want to test what happens without the interference of CORS 00:51:52 You're trying to break same-origin? 00:52:00 Yes. 00:52:12 For .js files where the .js file itself comes from doesn't change the origin context 00:52:19 It's always set by the document. 00:52:26 Lumpio-, I am aware of this. I want to show my boss 00:52:37 He thinks it doesn't work like that? ¬u¬ 00:52:47 Just post a link to the specs or whatev 00:53:07 Specs suck when it comes to "Does this work on the web" 00:54:39 You know what the fun part is? It wouldn't really matter so much if cookies weren't sent with every request to a domain no matter who originated that request. 00:55:46 shachaf: "For example, http://hackage.haskell.org/trac/ghc/ticket/7634 ought to be a CVE" -- ezyang 00:55:51 go pro troll and request one 00:56:22 cve? 00:56:38 a number identifying a security vulnerability 00:57:12 http://en.wikipedia.org/wiki/Common_Vulnerabilities_and_Exposures 00:57:50 heh 01:02:00 -!- BillyZane has quit (Ping timeout: 260 seconds). 01:07:39 -!- Bike_ has joined. 01:08:54 also GHC still has loads of RWX pages :( 01:10:23 -!- Bike has quit (Ping timeout: 256 seconds). 01:10:56 -!- mnoqy has joined. 01:11:20 -!- Bike has joined. 01:14:21 -!- Bike_ has quit (Ping timeout: 256 seconds). 01:14:27 -!- BillyZane has joined. 01:23:41 -!- augur has quit (Read error: Connection reset by peer). 01:25:01 -!- augur has joined. 01:25:28 I just got something that is similar to "goto" implemented in my language. ^_^ 01:25:49 i hear goto is considered harmful 01:25:49 Now if I get the "if" thingie implemented you might be able to do something with this confusing pile of shit :D 01:26:17 "considered harmful" considered harmful 01:26:25 Bike, 01:26:27 potato 01:26:35 Bike: what about asm goto 01:26:37 oh 01:27:12 Harm things? I hope so. 01:27:16 kmc, have you seen knuth's paper about goto, 's cool 01:27:21 no 01:27:23 potatos are great 01:27:28 Agreed 01:27:35 it must have sucked to live in europe before they got all the cool foods from the new world 01:27:39 "Structured programming with GO TO statements" 01:27:43 italian food with no tomatoes 01:27:45 direct response to djikstra 01:27:50 dijkstra? dutch 01:28:12 basically, in my language, everything in between the :'s have a numerical number..which is in order 01:28:14 what was europe's staple crop anyway 01:28:23 "i know shit" 01:28:36 You can now hop to any part of the program :D 01:29:05 oh, wheat, duh 01:30:07 apparently the most productive corn farms are in israel nowadays 01:30:11 then they got the potato in ireland and the population doubled 01:30:12 ~*~globalization~*~ 01:30:15 then it dropped by half >_< 01:30:45 disease or famine or 01:30:54 he's talking about the famine presumably 01:30:54 famine brought on by potato disease 01:30:59 since they were all like the same species 01:31:07 in which case you can blame the disease and also britain (THANKS ELLIOTT) 01:31:32 I love this channel XD 01:31:43 I'm so used to being kicked and banned for saying off-topic shit 01:31:54 You know that you can't talk about ubuntu in #ubuntu? 01:31:55 JesseH: how could would you say your language is at resource allocation in emergency conditions 01:32:01 "I think we are on the verge of discovering at last what programming languages should really be like." -- Donald Knuth, 1974 01:32:02 (now it's on-topic. you're welcome) 01:32:05 kmc: inorite 01:32:15 *how good would 01:32:16 no no i'm not complaining! 01:32:22 please continue to speak about potatos 01:32:27 that's good. complaints are off-topic. 01:32:31 lol 01:32:38 boil 'em, mash 'em, stick 'em in a stew 01:32:41 be happy, citizen 01:32:48 kmc, forgot the rest D: 01:33:17 Bike, this is a high level esoteric language 01:33:20 ah the british stole the best land in ireland for grazing cows 01:34:16 moo 01:35:00 🐄 01:35:07 (that's a cow, hth) 01:37:17 Isn't there a moo language? 01:37:32 There's the language that MOOs use 01:38:40 http://esolangs.org/wiki/COW 01:39:18 of course it's a brainfuck 01:39:34 kmc: is it MMIX 01:40:32 Blah, CORS considers POST of application/x-ww-form-urlencoded to be a "simple request" that it can just go ahead and send to a remote server without asking 01:40:39 (Because of course it could be done without CORS) 01:43:02 (Actually, hmm, not so sure) 01:46:28 Oh, I think if CORS isn't used browser still complains with a CORS sounding message 01:48:58 -!- NihilistDandy has quit (Ping timeout: 276 seconds). 01:49:48 Ta-da http://dl.dropboxusercontent.com/u/16240872/crossdomain/demo.htm 01:50:10 Might need to use Wireshark though to convince someone that snyfarvu.farmingdale.edu isn't explicitly denying it 01:51:40 this is a demonstration for work? 01:51:55 For a disagreement I have with my boss 01:51:58 Tangeltally related to work 01:52:02 *Tangentally 01:54:09 Hum. I think the server is actually SENDING the contents of the file, just the browser is refusing to tell the Javascript what it is. 01:54:22 * Sgeo wonders if there are any interesting implications of this. 01:59:17 -!- augur has quit (Remote host closed the connection). 01:59:55 -!- augur has joined. 02:05:35 -!- NihilistDandy has joined. 02:06:11 'Therefore, sites that prevent cross-site request forgery have nothing new to fear from HTTP access control.' 02:06:39 Yes, but sites that are already vulnerable to CSRF just found being CSRFed slightly easier for the attacker to do.... which I guess isn't a big deal. 02:09:18 Hi can anyone look through the language overview and give me tips on what else the language needs? http://esolangs.org/wiki/Derplang 02:10:22 I just thought of an attack scenario due to CORS. 02:10:36 JesseH: 99 bottles 02:10:52 Bike, really? ill totally add that if you want ^^ 02:11:01 bike: will print out "99 bottles" 02:11:28 That's what CORS does. It sends a GET request. 02:11:28 as in 99 bottles of beer 02:11:35 This is why you don't ever make a GET request change anything. 02:11:41 You've been able to do that since the beginning of time anyways 02:11:41 Suppose you're on insecure wifi, and you're aware of it, so you decide that you won't visit youridiotbankthatdoesntusehttps.com 02:11:41 Bike, dont understand 02:11:51 With script elements, img elements, iframes, whatnot 02:12:02 bike's saying your language needs more beer 02:12:10 does not compute 02:12:11 "99 bottles of beer on the wall, 99 bottles of beer; take one down, pass it around, 98 bottles of beer on the wall! 98 bottles of beer on the wall, 98 bottles of beer...[etc]" 02:12:17 alt. something remotely interesting, like ackermann 02:12:22 But, on accident, you do visit evil.com, and someone working with evil.com is snooping your traffic. 02:12:23 hmm 02:12:35 I've heard the reference before 02:12:37 I forgot 02:12:38 I'm pretty sure that's a known attack vector, Sgeo 02:13:00 JesseH: http://esolangs.org/wiki/99_bottles_of_beer 02:13:01 evil.com sends a simple CORS request with credentials to the bank. 02:13:16 It doesn't even need to be a CORS request it can be a damn img tag 02:13:31 ....good point. 02:13:37 Nothing new there. 02:13:53 i don't think evil.com deserves this slander 02:14:04 Ohhhh Bike okay i understand now! 02:14:50 JesseH: also implement http://en.wikipedia.org/wiki/Ackermann_function 02:15:52 Alright thanks for the tips 02:17:05 "Important note: when responding to a credentialed request, server must specify a domain, and cannot use wild carding. " 02:17:32 I guess server could reflect Origin, but then what if it's cached 02:17:36 Bike, Does it have to get the first number, and put that in front of the string? 02:17:37 Oh, it being cached can be blocked 02:17:44 JesseH: does what have to what 02:17:50 oh wait 02:17:59 okay so basically 02:18:31 i have to write a program in my language that if you type something, it will manipulate that data and return something else 02:18:44 or well 02:18:57 I need an if statement in my language...I know that 02:19:01 i for one don't care about i/o 02:19:06 how you do it, rather 02:19:12 input already works 02:19:14 just have something compute ackermann somehow 02:19:18 Sgeo: Again, an img tag will always send credentials .-. 02:19:23 You can't even disable it. 02:19:26 Bike, alright 02:19:29 you don't 'need' an if statement per se 02:19:38 you can compute ackmerann in lambda calculus if you hate yourself sufficiently 02:19:48 Lambda calculus is fun 02:19:55 or more realistically use the µ-recursive formalism 02:20:02 Lumpio-, I'm saying what if I want to receive and reply to non-simple credentialed requests for some reason 02:20:04 "old skoooool" 02:21:08 One of the few languages I've actually bothered to implement was a language that gave you functions and a way to call them and then you could implement stuff like integers by hand. 02:21:25 Or well it gave you a way to define functions. None exist to begin with 02:25:34 ackerman function is a bit confusing. I may just be sleepy. 02:26:04 it's a "pretty cool function" 02:26:08 lern that math 02:26:12 I'm looking at other examples 02:26:56 How could I do this without if statement? 02:27:15 or well 02:27:32 How can I write this in a language without the ability to create functions? 02:27:49 you can't do much in an imperative language without control flow :-) 02:27:54 that's pretty vague 02:28:01 Bike: well have you seen the spec 02:28:03 * lifthrasiir reading a log 02:28:04 imagine doing it on a turing machine if you like 02:28:06 it has no "conditional jumps" 02:28:17 hmm 02:28:29 so no control flows and still trying to be Turing-complete? 02:28:30 closest it gets are "repeat this thing `n` times" and "unconditional jump" 02:28:31 I can add conditions >:P 02:28:42 lifthrasiir: well obv. it's not tc 02:28:54 That's the part I'm working on 02:28:59 conditions are so mainstream 02:28:59 Seeing what I need for it to be turing complete 02:29:05 mnoqy, my point :P 02:29:27 well, if it can compute the ackermann function it's turing complete, probably, so there's a goal 02:29:34 mnoqy: of course it is not TC but sometimes very obscure feature can turn out to be a "TC-able control flow" 02:29:38 Bike: um are you sure 02:29:38 In my language, to go on forever, printing a number that gets bigger each time you do this... 02:29:48 mnoqy: "probably"!! 02:29:56 ah, "probably" 02:30:20 though can you make tcness out of composing ackermanns or some shit, iunno 02:30:33 ackermanns is total sooooooooo 02:30:42 JesseH: I can't get a gist of your language (from the past log), can you describe a bit more? :) 02:31:12 nu:x:0:ad:y:x:1:ou:y:gt:4 02:31:13 oh right, you need to halt at some point, durr. 02:31:21 let me test that code 02:32:01 well that just prints 1 forever :/ 02:32:25 Ill need to fix ^_^ 02:32:53 ah... it had a link 02:32:58 * lifthrasiir reading the spec 02:33:03 lifthrasiir, sorry im super thinking 02:33:13 brains cpu is all on one thing right now 02:33:20 what's cool is those "conjecturally tc" languages where you've got tc iff someones math conjecture holds 02:33:21 You can force close or wait 02:33:38 mnoqy: more interesting than 'something something goldman's' i hope 02:33:48 i think there are at least two? 02:33:52 probably one is that 02:33:53 or whatever the prime one is? how do i always forget it 02:33:56 Goldbach 02:33:59 JesseH: yes, it is definitely not TC, meaning that it is not on the same level as other esolangs and normal languages (at least conceptually) 02:34:06 we should stop naming people after elements 02:34:27 http://esolangs.org/wiki/Oozlybub_and_Murphy this is the goldbach one, thank's cpressey 02:34:35 forget where the other one is 02:34:42 Bike: and hydrogen should be renamed to unium 02:34:51 infinitely long identifiers, nice 02:35:00 lifthrasiir, turing completeness is my goal 02:35:47 there are several ways to add turing completeness 02:35:51 add? implement. 02:36:37 I think ill make the if statement only check if the two variables are == 02:36:45 That will make things interesting. 02:36:54 but in general you need i) unbounded code repetition or loop and ii) unbounded storage that can be read and written arbitrarily (though it can be also arbitrarily slow) 02:38:37 imo, jesseh should learn ski calculus and then make something interesting 02:38:49 ski free 02:38:51 I just may 02:39:08 imo make something sub-tc and interesting 02:39:16 ki calculus 02:39:27 make something which is tc iff the riemann hypothesis holds 02:39:36 mnoqy: total function programming FTW 02:39:37 that sounds painful. 02:39:48 lifthrasiir: ye 02:40:11 ill making something that is only sometimes tc 02:40:12 "One lesser known aspect of the game deals with the numerous tree stumps found around the slopes of SkiFree. Users have found that when skied over backwards they transform into a small spotty mushroom.... The mushrooms may illustrate an extrapolation of Plato's theory of ideas in which the physical form of something may not reveal its true identity and only its representation in our particular plane of existence." 02:40:20 "sometimes"? 02:40:23 lifthrasiir, yep 02:40:26 kmc: nice 02:40:51 everytime the interpreter is run, it throws a random, and if it is whatever value, it will use the code that makes it turing complete 02:40:53 if not, it wont 02:40:54 :-) 02:41:02 i should make an esolang based on dynamics instead of all this "discreteness" crap. first, i will learn to math 02:41:09 lol 02:41:10 that reminds me of... 02:41:29 CHI... I don't remember the name right now 02:41:41 CHIQRSX9+, okay 02:41:50 reminds me of neumann's paper on computing neural networks with faulty components, except suxx0rz 02:42:11 which has a simple instruction X which supposedly makes the entire language TC 02:42:34 the reference implementation invokes Perl for that purpose 02:43:29 #!/bin/perl \n #X :P 02:46:25 http://25.media.tumblr.com/12868568fcce5b6a70804b519de198f3/tumblr_mm480oyaRw1qjyqixo1_1280.jpg h t h 02:50:16 I'm going to bed. Tomorrow...turring completeness is among us. 03:04:08 Bike, i choose to interpret this as a savaging of dinosaur comics 03:04:20 good interpretation 03:06:14 meanwhile: john carmack didn't come up with the magic inverse square root constant? if i was the hero-worshipping type i'd be shattered 03:08:43 who did? 03:09:10 What's wrong with allowing a webpage to read arbitrary domains' GETable-without-cookies stuff? 03:09:23 In theory, someone could just put up a server somewhere that allows that indirectly 03:10:07 maybe the real problems is just cookies in general eh 03:10:31 Yes. 03:10:51 They need to be scrapped and a new mechanism developed. 03:12:00 Sgeo: cached content 03:12:12 Sgeo: imagine I requested page A with cookies and it is now cached 03:12:30 if an arbitrary page could read that, they may get sensative data 03:12:46 Well, if you're not sending the cookies, it deserves to be considered different in the cache. 03:13:00 sometimes 03:13:01 also, even if there's no private data involved 03:13:08 they can measure how long it takes for the request to come back 03:13:14 to see if you've visited the website recently 03:13:47 even without cookies, they can use: "did this image come back fast enough" 03:13:53 etc 03:14:16 Sgeo: also, some (silly) websites don't have indepotent GET requests 03:14:33 that should be a shooting offence 03:14:58 constant, said websites are already screwed 03:15:17 Sgeo: it turns a misconfiguration or miscoding application bug into a security issue 03:15:45 It was already a security issue. 03:15:53 hrm? 03:16:00 img tag with src, or attacker's proxy somewhere 03:16:07 true 03:16:12 Sgeo: poor example 03:16:34 years ago I used to know all this stuff, havn't done web application pen testing for a long time though 03:17:06 hm that reminds me of reading in a ph'p book that used extract on $_GET or $_POST i forget which (both?)..... i forget the book too 03:17:24 http://php.net/manual/en/function.extract.php what a cute language 03:17:58 -!- Phantom_Hoover has quit (Read error: Connection reset by peer). 03:18:05 Oh, yeah, that used to be infamous. 03:18:19 But uh... I guess they still let you do it, huh. 03:18:40 oh right, it's register_globals 03:18:41 register_globals is horrible too 03:18:47 ah i remember that too 03:20:02 "superglobals" 03:29:30 You know what would be interesting? A browser marketing itself (and trying to be) secure by sacrificing some backwards-compatibility 03:40:32 ΡΗΡ 03:41:36 lol 03:42:56 what do i need to know about heap corruption exploits 03:43:15 i get this idea that you can overwrite block list pointers and use that to trick the allocator into writing a chosen value to a chosen address 03:43:18 Sgeo: that happens all the time, e.g. Firefox disabled some CSS features for a while for security reasons 03:43:27 or you can just clobber other app data on the heap as desired 03:43:39 that sounds pretty cool kmc 03:44:03 :visited, or does CSS have more security issues than that? 03:44:50 http://www.phrack.com/issues.html?issue=57&id=8 is a really sophisticated heap attack that I never understood in full 03:45:00 we should just scrap the web and use project xanadu 03:48:18 this splitter thing seems like it should be somewhere that isn't sudo. 03:51:25 huh, 8 byte alignment... 03:52:26 "Luckily for the attacker who smashes the heap in order to execute arbitrary code, the GNU C Library does not activate these error detection mechanisms" sweet 03:55:44 "It may be possible, although hard to do reliably in a real-world attack, for a malicious web site to execute arbitrary machine code in the context of the web browser" some good links here kmc 03:56:01 kmc: Where did he say that? 04:02:55 who say what 04:11:22 lol 04:11:27 apparently I have moderator powers on stackoverflow now 04:11:33 I haven't done anything on that site in like 2 years 04:11:41 ban everyone 04:11:44 ban everything 04:12:45 ban all non-PHP questions 04:14:32 kmc: ezyang 04:16:26 oh, on zephyr 04:17:26 Oh. 04:19:07 why does C have calloc for zeroed arrays but nothing special for uninitialized arrays? like, the way calloc takes a unit and all. 04:20:46 Why do people expect any kind of internal consistency from C 04:21:29 kmc: As far as I understand the state of the art in MD5 collisions isn't good enough to exploit that. :-( 04:21:32 Is it at least more consistent than the web 'standards'? 04:21:44 -!- NihilistDandy has quit (Ping timeout: 260 seconds). 04:21:53 because it's a widely used language that bla bla god forbid i ask a historical question without being greeted by a chorus of "it just sucks" 04:22:25 IT SUCKS 04:22:31 thx 04:22:36 This is a bizarre misprint. 04:23:04 just suck it 04:23:05 The first half of this page is taken from the first half of a page a dozen pages later. 04:23:25 ouch. 04:23:29 tjat sounds like a misprint alright 04:23:29 Up to the new chapter heading. 04:23:32 -!- NihilistDandy has joined. 04:23:56 But the page number isn't. 04:24:31 should i sue the publisher 04:24:40 Who's the publisher? 04:24:52 Dell 04:25:48 they publish books? 04:26:50 By book I meant laptop catalog. 04:26:55 ( https://en.wikipedia.org/wiki/Dell_Publishing ) 04:27:20 http://en.wikipedia.org/wiki/IKEA_Catalogue 04:27:39 The American Gun Mystery 04:28:35 Each edition of the catalogue takes about 10 months to develop from concept to final product.[3] 04:28:47 Is the concept, uh, different every year? 04:29:05 Is a baby really different to another baby? 04:52:56 i love that this old sudo exploit comes from utterly mundane string manipulation in logging 04:53:05 and the more recent one that was some obscure debug code 04:53:28 am i naïve to think that that sort of thing shouldn't be part of the sudo code anyway 04:53:50 that's the thing, even if you have some good reason to write your app in C, it probably only applies to 5% of the code and not to the 95% of boring error handling, argument processing, etc. 04:54:04 like shouldn't that be in a library or something 04:54:21 even in a kernel, most of the code (by SLoC) is not doing anything particularly performance sensitive or low-level 04:55:04 -!- carado has quit (Ping timeout: 246 seconds). 04:57:35 are you implying we might be well-served by writing applications out of multiple interacting modules, the implementation considerations of which don't impinge on each other's implementation considerations without good reason 04:57:39 or is that just me 04:58:36 We're all just doomed 04:58:39 yeah 04:58:51 or we're all just doomed. 04:58:53 some combination of using multiple languages, and using languages that are safe by default but have these low-level capabilities 04:59:17 Now of course, those languages are implemented 04:59:18 in C 04:59:21 Haskell is a good example of a language that's safe by default, and suitable for systems programming (maybe not kernels so much, but like, high-preformance network servers) 04:59:36 this is part of why I get annoyed that everyone wanks about Haskell being sooooo mathematical and pure and not about the cool real world stuff it can do 04:59:47 Jafet: sure, but you end up with a smaller trusted base 04:59:54 imo fuck mathematical purity (hiiii elliottttt) 05:00:11 ghc is like 10MB? 05:00:25 That's a lot smaller than a linux system full of programs, is the point 05:00:33 the GHC RTS is 50,000 lines of C and assembly code 05:00:35 it's scary 05:00:36 also is that the whole thing or just the C part 05:00:49 but not as scary as every app being written separately in C 05:01:12 also, we hope that kernel and programming language development can attract a higher caliber of programmers than applications development 05:01:18 The trusted base also contains the code generator, bytestring, and if lens is still using unsafePerformIO somewhere 05:01:31 and we can justify expending more effort per line of code, because those core projects are used widely 05:01:44 That's already true though 05:01:48 i mean, this battle is largely won, most software is written in Ruby and JavaScript now 05:01:53 could be better but could be worse 05:02:10 unfortunately the web has its own awful security problems 05:02:25 The most amusing javascript attack is the heap fengshui thing 05:02:40 the most amusing web security thing is billion laughs. 05:02:58 Where they load a javascript program to attack the javascript vm 05:03:39