00:01:04 -!- zzo38 has quit (Quit: killed to death). 00:01:30 -!- pingveno has joined. 00:05:54 -!- zzo38 has joined. 00:15:26 i found an interesting reddit comment on reverse engineering and (somewhat) haskell by copumpkin: http://www.reddit.com/r/haskell/comments/iqz0l/reverse_engineering_of_compiled_haskell/c264g5b 00:15:49 ohai 00:15:53 stalking.jpg 00:16:14 well i only noticed it was by copumpkin after i'd starting reading it 00:16:19 *started 00:16:20 :) 00:16:26 fine, stalking.png 00:17:37 I have re-read the panels around the end of act 4 00:18:22 It makes it obvious that it's refereing to it, but I'm still lost 00:18:46 what? 00:18:57 elliott, wrt Homestuck update 00:19:18 What about it 00:19:52 oerjan: I'd still be curious to feed hex-rays a ghc-compiled binary 00:19:54 pity I don't have hex-rays 00:20:11 I bet even function boundary analysis in basic IDA would fail, though 00:20:46 Sgeo, elliott is going to mock you about this, but he was completely lost by the ectobiology stuff when I read it with him. 00:20:47 I have found myself wanting to use a PHI command in C, although of course C doesn't have that. 00:20:56 And he was reading it for the second time. 00:26:30 Wait, why the heck is that the highest rated comment....? 00:26:40 this confirms it, YouTube commenting has now become a parody of itself 00:27:02 amazing 00:27:13 I've been suspecting it for a while, but not had solid proof 00:28:23 ais523: you _do_ know what the top voted reddit post ever is, don't you? 00:28:36 no, I don't 00:28:53 see http://www.reddit.com/r/all/top/?sort=top&t=all :) 00:29:12 heh 00:29:17 what does the icon next to it mean? 00:29:53 also, is the person who made it famous/special in any way that might lead to a post of theirs getting singled out? 00:30:19 ye 00:30:20 s 00:30:25 surely you know qgyh2? 00:30:37 he was the most famous user of old reddit 00:30:44 I don't, I don't really use reddit 00:30:49 I read proggit for the comments, but that's about it 00:30:53 I don't read homepage, for instance 00:31:35 elliott, was this before or after the test post? 00:32:22 Phantom_Hoover: much before 00:32:49 elliott, ah, so it was the power user crap Reddit prides itself on not having? 00:33:40 ais523: i'm not sure what that icon means, i _thought_ it meant a picture was missing but that's a self post so cannot have any... 00:34:09 Phantom_Hoover: no, qg was cool 00:34:35 elliott, well sure, but the fact that it was upvoted wasn't a random thing. 00:34:36 hm no looking at the frontpage the missing picture icon is an alien with a camera and a question mark 00:34:43 oerjan: it's something i've never seen before, probably added just for that post 00:34:43 oerjan: maybe that's why it's missing 00:34:49 Phantom_Hoover: Well sure, but come on, it's a one-off 00:36:16 ais523: but there other self posts which have no icon 00:37:26 http://www.reddit.com/static/self_default2.png 00:37:28 it seems 00:37:33 maybe using a question mark changes it:) 00:37:35 it :) 00:41:38 -!- zzo38 has quit (Remote host closed the connection). 00:43:15 oh i finally found one deep down in the frontpage which has it http://www.reddit.com/r/gaming/comments/is1s7/extralivesorg_is_trying_to_raise_8500_to_build_a/ 00:44:11 and moreover this one has the A+ icon for a self comment as well 00:44:34 while i've found a couple that have neither 00:44:44 *Aa+ 00:45:18 oh well maybe there's some obscure setting which does it 00:59:32 oerjan: you interested in reversing? 01:02:57 not very deeply, no 01:03:07 * Phantom_Hoover → sleep 01:03:12 -!- Phantom_Hoover has quit (Quit: Leaving). 01:07:21 -!- CakeProphet has quit (Ping timeout: 264 seconds). 01:18:05 -!- pumpkin has joined. 01:18:06 -!- pumpkin has quit (Changing host). 01:18:06 -!- pumpkin has joined. 01:18:36 -!- copumpkin has quit (Ping timeout: 276 seconds). 01:19:44 -!- pumpkin has changed nick to copumpkin. 01:34:25 -!- myndzi\ has joined. 01:38:06 -!- myndzi has quit (Ping timeout: 276 seconds). 01:40:15 -!- Sgeo_ has joined. 01:43:33 -!- Sgeo has quit (Ping timeout: 255 seconds). 01:46:21 -!- pumpkin has joined. 01:46:46 -!- copumpkin has quit (Ping timeout: 240 seconds). 01:48:14 -!- copumpkin has joined. 01:48:14 -!- copumpkin has quit (Changing host). 01:48:14 -!- copumpkin has joined. 01:50:31 -!- pumpkin has quit (Ping timeout: 250 seconds). 01:56:36 -!- cheater_ has quit (Ping timeout: 255 seconds). 01:58:51 -!- Sgeo_ has quit (Ping timeout: 255 seconds). 02:03:29 -!- Sgeo has joined. 02:04:06 I need to write a patch for XChat 02:04:17 To stop its notifications from telling me that I disconnected 02:04:31 Or at least, limit it to only once in x amount of time. 02:07:19 stop disconnecting :P 02:08:49 -!- cheater_ has joined. 02:10:02 * Sgeo wonders if XChat scripts have access to that stuff 02:22:04 elliott, how much does prgmr HD space cost? 02:22:45 Um? 02:22:52 What do you mean/why do you want to know? 02:23:07 I want to know because I'm considering trying prgmr 02:23:20 For what 02:23:27 To play around with 02:23:54 That seems like a bad reason to spend at least ten dollars a month :-P 02:24:01 But it increases with the RAM. 02:24:20 elliott, so what VPSes are more suited to just playing around with? 02:25:00 You have a Linux box with an IP already, don't you? 02:25:16 A VPS is for when you have actual services that you want to serve twenty-four seven :P 02:25:22 I mean, your money, but... 02:25:34 (Well, probably my dad's money) 02:25:36 You're not gonna find anything more than a few dollars cheaper than prgmr and what you do find will be a lot worse. 02:26:16 I want toy services that I want to serve 24/7 02:27:27 Your usecase sounds rather badly thought-out 02:27:30 Or rather, not 02:27:42 (thought out, that is.) 02:28:36 IRC bots, maybe some Second Life stuff, wanting to play with web stuff without being restricted in my choice of language 02:28:51 (Web stuff that other people can actually use) 02:29:31 You could just leave your laptop on, then it costs [fraction of electrical bill] rather than whatever prgmr does:P 02:29:33 does :P 02:29:37 But sure, prgmr would fit for that I guess. 02:55:33 -!- cheater_ has quit (Ping timeout: 255 seconds). 02:59:26 * Sgeo vaguely wonders what happened to a certain DJ that shares his name 03:00:06 well we found it confusing with several people named the same thing, so we killed him. 03:00:18 it was a fair coin toss, though 03:00:20 actually, we found it confusing with several people named the same thing, so we killed you 03:00:30 and are not sure why your client is still sending messages 03:00:53 ais523: ...are you sure you remember the right coin result 03:01:22 no 03:01:35 hm, awkward 03:01:47 -!- zzo38 has joined. 03:05:59 Actually, each of oerjan and ais523 was tasked with killing one of you if the coin flip went that way, but they both misread the coin and killed both of you. 03:06:06 Still, the ambiguity was resolved. 03:08:12 -!- cheater_ has joined. 03:09:16 oh well 03:09:31 @dice 1d2 03:09:32 1d2 => 1 03:10:03 -!- BeholdMyGlory has quit (Remote host closed the connection). 03:26:39 -!- azaq23 has quit (Ping timeout: 276 seconds). 03:43:51 -!- elliott has quit (Remote host closed the connection). 03:44:33 -!- elliott has joined. 03:44:40 -!- elliott has quit (Changing host). 03:44:40 -!- elliott has joined. 03:59:48 -!- elliott has quit (Remote host closed the connection). 03:59:51 -!- elliott_ has joined. 04:22:55 My brother made up his character's name by now. He named his character "Also". 04:23:42 X-D 04:28:09 zzo38 is part of this group, and also his brother 04:28:46 Now I should *also* put it into the computer. 04:30:39 also puns can easily fall flat if you don't get the grammar just right 04:45:58 http://jeffkatz.typepad.com/.a/6a0120a721c2d7970b0133ed630c7b970b-800wi Delicious, delicious heart attack. 04:47:04 I am American and I approve of this murder weapon. 04:52:19 -!- elliott_ has quit (Remote host closed the connection). 04:52:25 -!- elliott has joined. 04:57:23 I've made a couple of those 04:58:14 and you're still alive? 04:58:41 he didn't say he ate them 05:00:16 you'd inhale the cholesterol 05:01:56 -!- oerjan has quit (Quit: Good night). 05:05:22 they were delicious 05:05:28 but I did feel rather gross after eating them 05:05:46 Clearly copumpkin is as deceased as I am. 05:08:25 oh wow java generics are crappy 05:09:16 oh wow water is wet 05:09:33 :) 05:09:41 but they're, like, _really_ crappy :O 05:10:47 but water is, like, _really_ wet :O 05:10:54 like, completely drenching 05:11:45 perhaps wetness is the extent to which a thing is saturated with water 05:12:27 does that mean water is the wettest thing of all? 05:12:35 it does 05:12:36 oh, phew, they're not _quite_ as bad as I thought I'd just proved 05:12:52 but wetness might relate to liquids, which would complicate things 05:12:59 since not all liquids are water 05:13:51 it also brings up the question of whether ice is wet 05:14:22 man, that's deep 05:14:32 So, in wetness, is water a necessary element? is liquid state a necessary element? 05:14:48 Can you have something wet which is neither water nor liquid 05:16:09 Is humid air wet? Is mercury? 05:16:20 It's a natural word, so its probably ambiguous at best. 05:16:22 I am showing off because gregor is here 05:16:57 ok case closed 05:17:15 Not capitalizing my name or using punctuation isn't particularly impressive showing off ... :P 05:17:24 I didn't know which way to go 05:17:34 When Particle Man is in water, does he get wet, or does the water get him instead? 05:30:17 ais523: are you there? 05:30:21 In Soviet Russia, wet gets YOU!! 05:30:32 elliott: yes 05:30:40 ais523: haskell/java bridge using jni 05:30:44 it's rare for me to leave the computer on and unattended overnight 05:30:49 therapy is available down the hall. 05:30:53 sorry for the nightmares. 05:30:54 in the daytime, it's possible my boss has contacted me or something like that 05:31:03 you can only avoid reading that line for so long, ais523. 05:31:09 elliott: that's no more insane than haskell/imperative language, isn't it? 05:31:23 ais523: you're saying that involving Java in something doesn't make it more of a terrible idea? 05:31:34 that would be some kind of Act of Maximum Terribleness 05:31:36 elliott: only marginally 05:31:38 Haskell/JavaScript bridge 8-D 05:31:44 I don't consider Java some sort of great, all-encompassing evil 05:31:49 like, killing a baby with Java isn't really worse than just killing a baby, I suppose 05:31:49 just an ordinarily evil 05:31:53 *ordinary 05:31:59 almost mundane, in fact 05:32:00 ais523: OK, but as a language, it sure doesn't like talking to other people :P 05:32:07 Of course Java is mundane, it's just /strange/ and mundane 05:32:25 Java constantly surprises me, and I'm hardly inexperienced with C and Python and other such "boring languages" 05:32:36 really? 05:32:38 it doesn't surprise me much 05:32:39 Gregor, I think I have use for that 05:32:50 ais523: that may be cause for concern :-) 05:32:56 * Sgeo abruptly gets shot 05:32:59 ais523: but yes, it honestly does 05:33:02 it's designed exactly like I expect it to be designed 05:33:18 ais523: you expect type erasure? 05:33:21 more abstraction layers than are particularly sane, and a standard library that tries to do everything itself 05:33:45 and yes, but partly because I know that Java's template system is a last-minute hack that was retrofitted to the language 05:33:55 that's unfair 05:33:56 and the name is also weird 05:33:57 that's not an "expectation" 05:34:00 that's cheating :-) 05:34:02 ais523: you expect the rather strange anonymous subclass syntax? 05:34:19 elliott: it's saner than what C++ would come up with for the same thing 05:34:30 that means nothing! 05:34:46 elliott: well, you come up with a better syntax 05:34:59 ais523: Why, when I could come up with a better language? 05:34:59 The C++ template syntax is bad because it uses < > as delimiters even though they also mean less than and greater than signs. 05:35:01 I think you're just not thinking from the Java point of view 05:35:05 It's strange to be able to do that with shortcut syntax in the first place 05:35:16 And certainly not what I'd expect from the Most Boring Orthodox OOP Language Ever 05:35:17 The syntax isn't what's bothersome, although I haven't used Java much. But how do you not have lambdas? 05:35:20 Java is not orthodox IMO 05:35:20 not really, if everything is a class or an object, then what's a new method? a class 05:35:33 so you need to be able to make throwaway classes like that 05:35:40 and the syntax is much the same as defining an ordinary class 05:35:49 except that you put the class definition where the class name should go 05:36:04 Why can't anonymous functions just be syntax sugar for a method of an anonymous class, then? 05:36:06 * elliott steps away from ais523, curls up in a ball, and cries himself to sleep. 05:36:10 Erm, hmm 05:36:23 elliott: I note that C does the same thing, with anonymous structs 05:36:37 struct {int i; float f;} x; 05:36:38 ais523: I wish you were around when Java actually surprised me so I could tell you why 05:36:47 what happens in IDEA, stays in IDEA 05:41:51 Sgeo: Java 8 is actually adding lambdas. 05:42:17 what syntax, incidentally? 05:42:25 pikhq, yay! Now how about all the APIs that don't take lambdas? 05:42:39 Or are they using Runnable or something, which, I _think_ stuff takes? 05:42:47 well, they're just sugar for a five-liner that creates an anonymous inner class 05:43:11 Runnable is sort-of the defacto standard for those if you're just using them as a callback, but taking no arguments and returning void is a bit crippling 05:43:21 Ah 05:43:34 Haven't used much Java 05:43:37 (not /fatal/, but you don't want to have to use stupid workarounds) 05:43:52 The planned syntax is #{ arglist -> code } 05:44:18 /now/ I'm surprised 05:44:22 mostly at the -> 05:44:28 that's such an unjavaish operator 05:45:19 And you can only close over final or "effectively final" variables... 05:45:40 * elliott mostly at the # 05:45:52 i never expect to see that outside of a literal/comment in java 05:46:33 well, it's for consistency with method pointers, isn't it? 05:46:40 pikhq: that restriction's been around for ages 05:46:49 it's to save having to distinguish between current and original value 05:46:55 Effectively final? 05:47:18 Sgeo: Basically, can be declared final without causing a compilation error. 05:47:31 Ah, cool 05:47:46 (Well, cool's the wrong word. But ... smarter than... an unsmart compiler?) 05:48:01 Anyways. It's basically syntactic sugar around anonymous classes. 05:48:18 I suppose it's to make idiots think about whether they mean {final T x = this.x; return new C() {C(){setT(x);} } } or {final T finalThis = this; return new C() {C(){setThingToAccessToGetT(finalThis);} } } 05:48:36 I probably have made minor syntax errors there, because Java more or less requires an IDE to do the boilerplate for you 05:48:41 it's almost as bad as VHDL 05:48:50 Well. Anonymous classes of a single method. 05:49:07 -!- zzo38 has quit (Remote host closed the connection). 05:49:15 yep 05:49:23 And the type gets inferred. 05:49:45 elliott: you'll love this syntax, then (proposed, I'm not sure it was accepted): Math#sin for what is basically #{ x -> Math.sin(x) } 05:49:59 pikhq: type inference? in /Java/? now I'm shocked 05:50:10 ais523: It's about as advanced as C++0x's. 05:50:21 and how advanced is that? 05:50:25 ais523: heh 05:50:49 C++0x is more insane than my mind can easily grasp 05:50:59 I couldn't even follow the move constructor debate 05:51:05 ais523: Pretty much, if you're trying to use #{ } where only a class with a single method is expected, it figures out that you want to use that type. Otherwise it's a compile error. 05:51:07 C++0x is a beautiful, crystalline trainwreck 05:51:12 despite knowing early C++ (really early, before namespaces were invented) 05:51:21 it is the Best language and, therefore, a completely useless piece of shit 05:51:30 *Really* naive. 05:51:36 can anyone here explain move constructors? 05:51:40 but it lead to C++ sudoku! 05:51:47 my guess is no, because they're basically inexplicable 05:51:55 but perhaps someone here is insane enough to grasp them 05:52:05 I mean, I understand the intention and what they're meant to encourage 05:52:25 I just don't get a) why that's considered a desirable feature for a language, or b) any of the millions of details needed for them to work correctly 05:52:42 summarise? opening a tab will crash flash 05:52:44 (this is not a joke) 05:52:49 they also make perfectly innocent-looking code do ridiculous things 05:52:55 (flash /really/ hates pages loading or closing in Chrome for some reason) 05:53:01 elliott: I can't easily, because I don't understand it really 05:53:06 well, 05:53:07 I mean, I understand the intention and what they're meant to encourage 05:53:08 this part 05:53:14 you know what a copy constructor is, right? 05:53:25 it's like that, except you destroy the thing being copied, because it isn't needed any more 05:53:26 yes, it's constructing a new object with just an existing one as an argument 05:53:28 and copying all its fields over 05:53:31 but it's really hard to get right 05:53:33 ais523: wow 05:53:34 ais523: linear typing 05:53:42 ais523: but... at runtime 05:53:50 or wait 05:53:52 is it uniqueness typing 05:53:56 whatever Clean uses for IO, anyway 05:54:08 it's reminiscent of linear typing, at least 05:54:14 The *reason* for move constructors is so that the compiler doesn't have to actually do deep copies for its temporary values. 05:54:18 except that linear typing's "legitimate ways to cheat" are rather unlike C++'s 05:54:33 pikhq: why can't it just use the original value? 05:55:11 mutability? 05:55:20 ais523: Because things tend to get passed *by value* to relevant functions. 05:55:32 but yes, the reason for anything insane in C++ tends to be "performance, even though you'd just use a struct if you _really_ cared" 05:55:32 pikhq: oh right, and C++ doesn't hide calling convention stuff 05:55:42 heh 05:55:47 elliott: that's because C++'s market is very specific nowadays 05:56:08 it's people who care massively about performance, but who can't use C for various reasons, mostly library support 05:56:09 ais523: games programming, application developers (who don't care about the new features), and "specialist" areas? like embedded/realtime stuff 05:56:27 (I don't include needing to use an object-oriented view of things as a reason to use C++, as C is better at that) 05:56:38 `addquote (I don't include needing to use an object-oriented view of things as a reason to use C++, as C is better at that) 05:56:39 512) (I don't include needing to use an object-oriented view of things as a reason to use C++, as C is better at that) 05:56:45 yay, a milestone 05:56:52 e.g. foo = bar + baz + qux; would have to create a value for "baz + qux", to pass it to operator+ for adding bar to get the final value... 05:56:55 i'll delete most of them later :-) 05:57:06 I, uh, think? 05:58:02 pikhq: unless operator+ took references as arguments 05:58:09 which is the standard for arithmetic operators, isn't it? 05:58:23 Oh, right. To *work around that*. 05:58:38 ais523: haha, nice 05:58:43 is there ever really a reason to pass by value in C++? 05:58:44 C++'s semantics are far too complicated. 05:58:45 pikhq: yes 05:58:53 elliott: I dunno, but people *do it*. 05:58:56 elliott: you pass integers and smaller primitive types by value, generally 05:59:19 because if you pass them by reference, the compiler has to find some memory to store them in in order to pass a pointer to them, which is pretty convoluted 05:59:43 ais523: well, yes 05:59:46 I mean non-primitive types 06:00:01 most languages don't even have a /way/ to specify that you want a by-value passing 06:00:10 (as opposed to by-reference) 06:00:15 nowadays, I think value is faster for even structures up to 64 bits wide, with modern compilers and calling conventions 06:00:25 (although this is an implementation detail in the case of immutable objects) 06:00:31 ais523: not many structures are less than sixty four bits wide... 06:00:48 elliott: they are if you go up to Java levels of enterpriseyness 06:00:49 (I would say that int is usually used even when char would do fine, for futureproofing and not-having-to-think-about-it reasons) 06:01:10 if you have nothing but a vtable, for instance 06:01:16 and the vtable is stored as a pointer 06:01:22 a pointer to a pointer would just be ridiculous 06:01:28 clearly 06:01:40 elliott: well, in the minds of C++ users 06:01:40 so what you're saying is, C++ needs opaque typedefs :) 06:01:43 (Haskell newtype) 06:01:53 Badly. 06:01:54 in Haskell, those desugar to the contained value 06:01:58 so there's zero runtime overhead 06:02:29 elliott: gah, never suggest a new feature for C++, even as a joke 06:02:33 haha 06:02:34 the consequences aren't pretty 06:02:50 I'm sad that concepts weren't accepted, those were really what C++ template metaprogramming needed 06:03:00 (he thinks I'm joking) 06:03:00 something wasn't accepted? seriously? 06:03:06 you don't know about concepts? 06:03:14 they were half the size of C++ itself 06:03:17 not if the word has a specific technical meaning in the context of C++ 06:03:18 (hyperbole) 06:03:24 and required a massive gcc fork to implement 06:03:29 presumably to prove it was even humanly possible 06:03:29 aha 06:03:37 ais523: take a skim of http://en.wikipedia.org/wiki/Concepts_(C%2B%2B) 06:03:41 In the pending C++0x revision of the C++ programming language, concepts and the related notion of axioms were a proposed extension to C++'s template system, designed to improve compiler diagnostics and to allow programmers to codify in the program some formal properties of templates that they write. Incorporating these limited formal specifications into the program (in addition to improving code clarity) can guide some compiler optimizations, and 06:03:41 can potentially help improve program reliability through the use of formal verification tools to check that the implementation and specification actually match. 06:03:51 it was kind of like.... typeclasses 06:04:00 except... used as a sort of formal verification system 06:04:02 presumably it's someone trying to embed a different language into an existing one? 06:04:13 ais523: Wikipedia lists no prior art 06:04:19 I suggest a quick read of the article, it's truly jaw-droppingly C++ 06:04:23 apparently too C++ for the committee 06:04:23 hmm 06:04:44 it reminds me of a random reddit poster saying that incomprehensible template error messages were the compile-time version of a segfault 06:04:50 presumably, that's an attempt to statically prevent segfaults 06:04:59 or, metastatically 06:05:15 I really hope they were powerful enough to make a big C++ subset --> C++ template compiler easy-ish 06:05:28 then you could propose to the committee to specify C++ templates based on the C++ semantics itself 06:05:34 thus shortening the report and allowing for further extension 06:05:41 like, say, allowing every C++ feature to be used at the type level 06:06:04 the Java-like solution to this would be to allow templates to extend each other 06:06:08 does the C++ committee take the same crap that the html5 committee does? 06:06:16 it wouldn't /help/, but would be more javaish 06:06:26 itidus20: same crap as in what? 06:06:30 oops i actually meant XML 06:06:40 Ah. Most likely. 06:06:47 i do not know entirely what I mean.. but I hear the ISO is a joke 06:07:07 elliott: OK, the axioms addition is very C++ish indeed 06:07:11 actually, it reminds me a bit of C99 06:07:37 hmm, I see the kind of feel you mean 06:07:42 it's just... on a C++ scale 06:08:12 yep, that's a good description 06:08:34 hmm, this reminds me a bit of listening to a piece of computer game music and trying to figure out what series it comes from by the style 06:08:35 (sorry for fake foreign accent) It is much the feeling that dependancy on C/C++ undermines potential of esoteric languages. 06:08:42 only with programming languages 06:09:00 that's not fake foreign accent, that's just academicspeak 06:09:05 itidus20: C in particular has a view of the universe that was low-level once, but nowadays is an abstraction inversion 06:09:06 or maybe the latter is an absorption of the former 06:09:25 but it's hard to do better 06:09:32 Well, there is danger in dependance. 06:09:35 ais523: I have a feeling the C committee went that way after someone came up to them with a copy of the Lisp Machine ZETA C compiler, and said "explain this". 06:10:11 I've got annoyed with gcc-bf, because I just run into abstraction inversions everywhere, and it saddens me 06:10:24 abstraction inversions? 06:10:33 coppro: java 06:10:38 Wow, coppro asked the question I was going to ask 06:10:38 hmm, that made sense when i typed it 06:10:51 Sgeo: thus proving that neither of you can use Google (you're not ais523, you have no excuse) 06:10:52 A dependance on c/c++ means that GNU can be led along by the c standard. 06:10:55 coppro: basically, when a language provides high-level concepts only, so if someone wants to use low-level concepts, they have to implement them in terms of the high-level concepts 06:10:58 itidus20: well, no 06:11:00 GNU code in GNU C 06:11:06 which are again implemented in terms of low-level concepts 06:11:07 which doesn't advance automatically with the C standard 06:11:07 ais523: ah, yes 06:11:08 do they define their own standards? 06:11:15 itidus20: --std=gnu99 06:11:16 itidus20: they refer to published ones 06:11:17 ais523: e.g.? 06:11:17 which can't change 06:11:23 hmmmm 06:11:29 ok its probably safe 06:11:31 pikhq: the classic example is synchronization in ADA, but most people, including me, don't know how it works 06:11:32 you can change what the latest ISO standard for C is, but you can't change what ISO C90 is 06:11:34 *Ada 06:11:37 but you can see the danger here right? 06:11:44 itidus20: it's just the trust issue. 06:11:50 in the end, you have to (depend|trust) someone. 06:11:55 ais523: No, I mean "in C". 06:12:00 if google and microsoft lobby the standard 06:12:07 ais523: I know what abstraction inversion is, and can think of a few really good examples. 06:12:22 Most obviously, implementing locking primitives in Erlang on top of its messaging. :) 06:12:26 pikhq: hmm... vectorised instructions, perhaps 06:12:43 that's not quite what I mean, though 06:12:48 let me think 06:12:57 they should make a language called stallman 06:13:16 :D every program needs to have the name stallman in the filename to compile 06:13:20 (so you can do locking-based multithreading on top of message-passing multithreading on top of locking-based multithreading!) 06:13:50 smacks my knee with jocularity 06:14:29 itidus20, ...why? No offense, really, but that sounds boring 06:14:56 pikhq: fetch/modify/store's a decent example, I think 06:15:06 in C, all movement of data into and out of caches is implicit 06:15:35 so you have to do silly things normally used only by Vorpal, and weird unusual libraries, or else confusing code that doesn't really mean what it looks like it means, to get your memory access patterns to what you want 06:15:49 Will anyone be mad at me if I strap my hash output language into a sucky thing like BF? 06:15:54