00:04:04 -!- cpressey has left (?). 00:16:57 -!- Pthing has quit ("Leaving"). 00:17:09 -!- Pthing has joined. 00:20:50 -!- Sgeo has joined. 00:22:56 -!- coppro has quit ("I am leaving. You are about to explode."). 00:24:39 Any non-regulars in here? 00:25:01 I took some pills for that 00:25:32 -!- coppro has joined. 00:25:33 lol 00:25:40 * coppro has quit ("I am leaving. You are about to explode.") 00:25:40 Any non-regulars in here? 00:25:40 I took some pills for that 00:25:40 * coppro (n=coppro@unaffiliated/coppro) has joined #esoteric 00:26:49 lol 00:26:53 I count as a regular now, don't I? 00:27:17 Sure ya do, pooppy! 00:27:41 * Sgeo is waiting for a particular person he invited from Reddit to talk about Robozzle 00:29:24 -!- Rygarb has joined. 00:31:07 -!- Rygarb has quit (Client Quit). 00:33:30 non-regular here 00:33:51 nonregular 00:35:26 * SimonRC goes to bed 00:35:34 Hi Ryg.. 00:35:41 Rygarb was the person I was waiting for 00:38:16 * Sgeo facepalms a few times 00:54:42 i'm an intermittent regular, really 00:54:50 actually, a hibernator 01:03:13 -!- madbr has joined. 01:09:30 Gregor: Many of us are like that, really. 01:12:01 http://filebin.ca/pfbmoy/maybewaltz.ogg <-- what happens when Gregor tries to write a waltz. 01:12:36 Gregor: do you do anything normal with your life? 01:12:44 Define "normal" 01:12:59 answer: he gets on all 3s 01:13:02 coppro: Do any of us? 01:13:28 pikhq: I'm increasingly of the opinion that none of us do 01:13:32 but I keep trying for that hope 01:26:03 Normal == perpendicular to tangent? 01:26:18 Sgeo: In some contexts, yes. 01:26:33 pikhq, what else can normal mean? 01:27:02 And I mean mathematically 01:28:53 normal distribution 01:53:34 -!- augur has changed nick to quantic1e. 01:53:38 -!- quantic1e has changed nick to augur. 01:58:46 -!- MissPiggy has quit. 02:13:49 where is soupdragon/fax? 02:13:52 :| 02:14:23 augur, just left 02:14:45 shes as misspiggy now? 02:15:29 iirc, yes 02:15:40 hm! 02:22:20 -!- oerjan has joined. 02:23:46 It's an FRCer! 02:23:58 Hail, oerjan. 02:24:16 'morn 02:31:33 oerjan's an FRCer? 02:32:09 MWAHAHAHA 02:32:38 Yes, oerjan's an FRCer and an occasional nomic player. Been a while since he was in Agora, though... 02:33:53 for a value of occasional ~ 5, 6 years or so 02:34:09 Hm, that implies oerjan is a _current_ FRCer? 02:34:13 I thought he was a former one? 02:34:46 you can sign out any time you want, but you can never leave 02:35:35 Sgeo: FRC doesn't remove players. They just consider you non-active. 02:37:56 * oerjan would be extremely surprised if the member list has been accurate since - sometime before he left, probably 02:38:12 I'd imagine it's not all that accurate. 02:38:30 I wasn't even aware there's a member list 02:38:35 *Formally* they never remove players, but since the list of players doesn't matter, lazy evaluation tends to take care of that. 02:38:51 coppro: I'm not sure there actually *is* a list. 02:39:06 i suppose that it also a possibility :D 02:39:30 * oerjan is pretty sure he kept such a list at one time. 02:39:47 Well, except in the sense that you can say "There exists a list that contains the players of FRC." 03:12:34 Hmmmmm 03:12:41 p 03:12:43 p 03:12:55 Actually, relistening to this waltz, thinking about it as game background music, it /could/ work. 03:24:21 -!- oerjan has quit ("leaving"). 03:27:33 -!- jpc has joined. 03:32:35 -!- Sgeo_ has joined. 03:39:27 -!- calamari_ has joined. 03:44:55 -!- calamari_ has changed nick to calamari. 03:45:16 Well, saw the PS3 hack... 03:45:20 That's freaking crazy. 03:50:24 -!- Sgeo has quit (Read error: 110 (Connection timed out)). 03:50:40 "And then I drop the memory bus low for 40 ns to avoid the cache writeback from the hypervisor" 03:50:41 -!- kar8nga has joined. 03:55:47 -!- kar8nga has quit (Remote closed the connection). 04:13:33 pikhq: sounds hardcore, what does the hack mean exactly, I mean can it be patched by sony? 04:16:05 zeotrope: By releasing a CPU without a cache. 04:16:32 (that is to say "not only no but fuck no") 04:17:23 time to get a ps3 then? 04:17:45 Get a fat one; the slims don't have Linux support, and he runs the code as a Linux module. 04:17:54 Might be possible otherwise, but it's a PITA. 04:18:02 :( I'm sure they are hard to track down 04:18:09 Requires an arbitrary code execution bug in any extant PS3 program. 04:18:11 No, not really. 04:18:23 Even the PS2-compatible ones aren't too hard to find. 04:18:28 * pikhq picked one up a couple weeks ago 04:18:57 gonna wait and see what they do with it 04:19:14 * zeotrope is totally broke 04:19:25 Peek and poke implies "just about anything". 04:19:58 -!- oraqol has joined. 04:21:07 -!- oraqol has left (?). 04:23:01 his description of the exploit totally went over my head, gotta read up on electronics one of these days 04:28:46 Most of it's software, really. 04:29:23 The only bit of hardware is screwing up the memory bus to prevent the hypervisor from doing the cache writeback right... 04:38:11 -!- Pthing has quit (Remote closed the connection). 05:16:09 -!- Sgeo has joined. 05:30:16 -!- Asztal has quit (Read error: 110 (Connection timed out)). 05:33:19 -!- Sgeo_ has quit (Read error: 110 (Connection timed out)). 05:34:32 http://filebin.ca/qtqhva/maybewaltzpointdos.ogg (Now with 100% more Gregor!) 05:36:44 where'd that filename come from? 05:36:55 Maybe Waltz Point Two->Es 05:37:07 ah 05:43:19 Any opinions on the second half? Other than "the part that's clearly ripped off from Pachelbel is clearly ripped of from Pachelbel" 05:44:36 INSUFFICIENTLY STOLEN FROM PACHELBEL. 05:44:45 STICK HIS ENTIRE BODY OF WORK IN A SINGLE PIECE. 05:44:53 :P 05:44:59 But make it a waltz? 05:45:11 Make it 4 minutes and 33 seconds. 05:45:37 I sampled that and overlayed it verbatim on top of this. 05:45:40 So you can have the most unique performance of Cage's 4'33". 05:45:45 (Trimmed of course) 06:38:30 -!- madbr has quit ("Radiateur"). 06:39:19 -!- MizardX has joined. 06:43:38 -!- calamari has quit ("Leaving"). 07:02:56 -!- tombom has joined. 07:48:45 -!- oerjan has joined. 07:50:30 -!- tombom has quit ("Leaving"). 07:59:59 -!- clog has quit (ended). 08:00:00 -!- clog has joined. 08:05:24 -!- FireFly has joined. 08:15:44 -!- MigoMipo has joined. 08:34:10 -!- oerjan has quit ("leaving"). 08:54:08 -!- MizardX has quit (kubrick.freenode.net irc.freenode.net). 08:54:08 -!- oklo has quit (kubrick.freenode.net irc.freenode.net). 08:54:41 -!- MizardX has joined. 08:54:41 -!- oklo has joined. 09:00:24 -!- jpc has quit ("I will do anything (almost) for a new router."). 09:03:06 -!- MigoMipo has quit (Remote closed the connection). 09:13:37 -!- GreaseMonkey has quit ("HydraIRC -> http://www.hydrairc.org <- Nobody cares enough to cybersquat it"). 09:49:31 -!- FireFly has quit ("Leaving"). 10:29:13 -!- lament has quit (Read error: 60 (Operation timed out)). 11:10:02 -!- MizardX- has joined. 11:26:09 -!- MizardX has quit (Read error: 110 (Connection timed out)). 11:26:28 -!- MizardX- has changed nick to MizardX. 11:47:55 Gregor: sorta weird ending 12:12:51 -!- oklo has quit (kubrick.freenode.net irc.freenode.net). 12:13:19 -!- oklo has joined. 12:51:06 -!- MizardX- has joined. 12:54:54 -!- MizardX has quit (Read error: 60 (Operation timed out)). 12:55:03 -!- MizardX- has changed nick to MizardX. 13:06:43 why don't text editors have call stacks 13:07:28 if i want to fix a function, then return back to where i was, i have to use my own memory 13:38:57 -!- fizzie has quit (kubrick.freenode.net irc.freenode.net). 13:38:57 -!- EgoBot has quit (kubrick.freenode.net irc.freenode.net). 13:38:58 -!- Gregor has quit (kubrick.freenode.net irc.freenode.net). 13:39:09 -!- EgoBot has joined. 13:39:29 -!- Gregor has joined. 13:56:36 -!- MissPiggy has joined. 14:08:23 -!- bsmntbombdood has quit (Read error: 110 (Connection timed out)). 14:28:31 -!- MizardX- has joined. 14:28:39 oklo: It's unended. 14:41:40 http://www.mail-archive.com/gcc@gcc.gnu.org/msg49113.html Huh. 14:44:19 -!- MizardX has quit (Connection timed out). 14:44:27 -!- MizardX- has changed nick to MizardX. 14:45:34 -!- MissPiggy has quit (Read error: 104 (Connection reset by peer)). 14:45:59 -!- MissPiggy has joined. 14:48:44 -!- cpressey has joined. 15:02:13 -!- Pthing has joined. 15:45:35 Help! I promised myself I wouldn't solve a RoboZZle puzzle this morning, and I am 15:45:42 It's a pathetically easy one though 15:49:05 * Sgeo puts in a trivial shorter solution to one he solved a while ago 15:53:07 Having all my past solutions available to me is a big help 15:59:26 If Rygarb comes in, say hi. E's the redditor I've been talking about, and hasn't used non-IM chat before 16:01:03 * MissPiggy did some robozzle today 16:01:36 Why am I not getting ready to go to school? 16:04:12 robozzle does that to you 16:05:18 It's not RoboZZle 16:05:24 RoboZZle's fault right now 16:05:30 I'm going to school later than usual, so 16:07:10 It's messing with my mind, I think 16:07:35 just how much have you been playing it o_o 16:10:23 I meant the "not going to school at the time I'm used to" is messing with my mind 16:10:37 oh heh 16:10:48 ais523 / scarf: i have important matters to discuss. 16:12:01 ais523 / scarf: this time, actual provable progress, not just refined ideas. 16:13:31 Chrome needs a Firefox tab extension, analogous to IE Tab 16:14:26 -!- FireFly has joined. 16:16:56 Sgeo: why? 16:17:15 There are websites that work in Firefox but not Chrome 16:17:22 My school requires use of one such site 16:17:55 is that because the websites use browser sniffing? or what exactly 16:18:27 the IE tab makes sense because IE is so god damn broken 16:19:03 i guess that might be interesting to the public as well: an append function in clue has been compiled, although there's no parser yet 16:19:06 http://www.vjn.fi/pb/p666553643.txt 16:19:13 compiles, although parsing has to be done manually 16:19:31 zeotrope, it also says it doesn't work with Safari, so maybe it's an issue with Webkit 16:20:07 it's an append function programmed by giving suitable examples of how it should work 16:20:57 what are the numbers? 16:20:59 compiles in 0.03 seconds without any sort of optimization (purely brute force search) 16:21:03 MissPiggy: they are just numbers 16:21:25 how do you read it? 16:21:34 examples of how append should work, "[2,5,7],[3,6,8] -> [2,5,7,3,6,8];" means "given the lists [2,5,7] and [3,6,8], output [2,5,7,3,6,8]" 16:21:51 this produces a function that appends any two lists 16:21:58 wow how does it work? 16:22:00 after compilation 16:22:11 it works by finding a clever set of examples 16:22:35 basically while the actual functions and computation is completely brute force searched, the recurrence isn't, you give that. 16:22:45 ah you have actually divided it into base and recursive cases 16:22:56 yes 16:22:57 and 16:23:07 i've also linked to each recursive case a subcase, stuff it should recurse to 16:23:37 so you never have to "recurse into darkness", just use cons, car and cdr randomly 16:24:18 this is the only way i've found that makes the concept implementable without having a strongly optimizing compiler from the start 16:24:28 okay 16:24:46 does it use types to guide the search? 16:25:05 no. but obviously it will. 16:25:17 wow I am surprised this works without being type directed 16:25:31 well, it uses types in the sense that if a function fails, the result is ignored, but i don't use type for any sort of optimization yet 16:25:58 there are many (natural) cases where types completely remove the combinatorial explosion 16:26:29 in fact, i think once you can introduce your own types, this might actually become "practical", by which i mean you might be able to program like an actual program 16:27:12 what about strange recursion patterns? 16:27:27 like what about zipping two lists, or splitting one list into even and odd parts? 16:27:46 you can have an arbitrary number of examples, and an arbitrary number of subcases 16:28:01 and an arbitrary number of branches, although all functions can only branch once, in the beginning 16:28:03 AAAAAAAAAAAAAAARGH 16:28:07 my monitor died 16:28:11 using laptop atm 16:28:27 seems like the background light on my desktop is dead 16:28:36 also no warranty any longer 16:28:37 there are many subtleties to this, and i'm planning to write a comprehensive spec, which might actually happen now that i've seen this really works 16:28:54 how do the examples come into it? 16:29:02 come into what? 16:29:13 I mean, could you just as easily replace them with a generator and checkable specification 16:29:19 into the program generation 16:30:11 the idea is you give just enough examples to find the right function. there's a separate sort of example that can be used for testing the result 16:30:34 (which i haven't implemented yet, but obviously it's just a few lines) 16:31:34 do you mean could i get rid of examples, and instead have specifications for how functions should work abstractly, like some sorta assert? 16:32:17 -!- BeholdMyGlory has joined. 16:32:56 {:. [2,5,7],[3,6,8] -> [2,5,7,3,6,8]; 16:32:56 : [5,7],[3,6,8] -> [5,7,3,6,8] 16:33:14 this is really the most interesting part, the rest is just filler 16:33:47 i don't know how a generator would guide the recurrence exactly 16:39:23 I was just thinking if you could generate inputs and outputs then check they are correct 16:39:52 you could, but you could only use that for testing, not actually making new functions 16:42:00 well at least with this system 16:42:37 in clue, i do not plan to make the language understand what objects are smaller than others in the sense of inductive construction of objects 16:42:58 that is, a human will know [1,2] should not recurse to [1,2,3] usually 16:43:03 but to [1] 16:43:09 because [1] is smaller than [1,2] 16:43:51 this is made general by making a new type that's constructed inductively in some other way, and adding an isomorphism between the two types 16:44:01 but that all is for another project 16:44:46 or, rather, that one trivial idea i just mentioned. 16:47:30 * Sgeo needs to get ready to go to school 16:47:58 when do you have school exactly? 16:48:52 i usually start getting ready 5 minutes after i have to leave 16:48:59 -!- oklo has changed nick to okloNEVERMORE. 16:49:07 wait that was too random 16:49:09 -!- okloNEVERMORE has changed nick to oklopol. 16:50:30 3:30 local time. To make it there in a timely fashion, I want to leave the house by maybe 12:20 16:50:48 in half an hour? 16:50:55 And I'll get there an hour or so early, assuming no bus screwups 16:51:07 Yes. But I still haven't showered, shaved 16:51:13 I think I'm going to skip shaving today 16:51:20 yeah shaving is for girls 16:51:24 just let it live 16:52:29 Not shaving is awesome. 16:52:55 My monthly shaving ritual allows me to use a single disposable razor for ~six months. A pack lasts me years. 16:53:29 I (as of earlier this week) became addicted to shaving every day 16:53:34 Now, because of time concerns, I can't 16:53:49 Time concerns completely caused by me being online right now 16:53:57 I've yet to shave :D 16:54:03 Maximum money saving 16:54:18 put physics into your prorgam oklopol and get a unfied theory output 16:54:43 * Sgeo plucks out an eye and gives it to MissPiggy 16:55:41 yeah i just need a few examples of how particles move 16:56:48 Eeep, it's late 16:57:01 man, just chill... 16:57:31 Bye 16:57:40 byes 17:03:46 -!- Asztal has joined. 17:12:33 -!- tombom has joined. 17:19:56 Deewiant, btw mycology tests o even if y claims it isn't supported. This seems wrong to me 17:21:26 -!- oerjan has joined. 17:22:54 -!- cpressey has quit ("Leaving."). 17:23:48 AnMaster: Didn't we go over this and it says UNDEF now? 17:23:57 Or was that a separate issue 17:24:02 Deewiant, well the issue would be that it *tries* it at all 17:24:12 if i want to fix a function, then return back to where i was, i have to use my own memory 17:24:17 or rather, could potentially be 17:24:17 vim has a tag stack 17:24:23 AnMaster: And like I said we went over this 17:24:30 oerjan: cool. does anyone ever use it? 17:24:42 Getting it to not try it is nontrivial and I'm not in the mood for something nontrivial :-P 17:24:51 Deewiant, what if you do like the _POSIX_WHATEVER* define on openbsd for mmap? basically openbsd implements mmap() but not all details of it 17:25:03 heck if i know, except the help system seems based on it 17:25:05 thus it doesn't define the posix define for supporting it 17:25:16 Deewiant, see my point? 17:25:30 i understand for programs you need to preprocess the file first 17:25:52 AnMaster: No, not at all. What is it? 17:25:54 oerjan, also about tags. do you mean ctags? 17:26:10 i think so 17:26:29 Deewiant, that not claiming that it has o doesn't imply that it doesn't have a partly working o 17:26:54 Argh, triple negative 17:27:08 :D 17:27:13 Deewiant: are you on drugs? 17:27:14 Deewiant, thus if you implement it differently but claims you don't have o, you could get BAD 17:27:26 oklopol: Not to my knowledge 17:28:04 Deewiant: was just wondering about your not liking nontrivialities atm 17:28:15 Deewiant, for example you could claim not to have o, but have an o that doesn't perhaps support the text mode 17:28:22 i'm sure you usually *love* triple negatives 17:28:23 or only supports text mode 17:28:24 AnMaster: But then you would know that you have a half-implemented o and thus could disregard the error. 17:28:32 Deewiant, true 17:28:39 oklopol: No, actually I don't usually not hate them 17:29:04 Deewiant, that sounded awkward grammatically 17:29:07 AnMaster: Besides, having a half-implemented o like that is probably not very likely :-P 17:29:17 no it didn't 17:29:27 Deewiant, well yeah "no text mode" is more likely than "no binary mode" 17:29:35 your mom sounded awkward grammatically in bed last night 17:29:38 I mean in general 17:29:52 Why would you have a half-implemented o if you're going to report that you don't support it at all 17:30:00 Deewiant, well cfunge used to have no text mode for a long time iirc, mycology doesn't test that afaik 17:30:47 It does test it but because i ignores spaces it's impossible to test it automatically, so it asks the user to verify 17:31:09 Deewiant, hm what about binary mode for i 17:31:45 Still ignores spaces 17:32:13 Deewiant, also no it doesn't test it, since I'm implementing o in efunge atm, and text mode throws an exception. However it passes mycology just fine 17:32:44 Okay, so it just prints a message then 17:33:16 AnMaster: hm, vim can also support emacs style tag files 17:33:21 Deewiant, yeah: 17:33:23 "Can't test o in linear text mode: i ignores spaces, no way to know from within standard Funge-98 whether they are output to file." 17:33:37 oerjan, I don't use etags either 17:35:01 well those are the styles i have heard about (because vim supports them, probably) 17:35:14 oerjan, well yeah I don't use *tags... 17:35:27 well i don't either 17:35:45 anyway what oklopol wanted sounds closer to bookmarks 17:36:01 which I believe almost all "more advanced than notepad" editors support 17:36:25 you mark a line then you can jump to it 17:36:31 why don't text editors have call stacks <-- the line before the one i quoted 17:37:25 and [ce]tags is essentially using a program to create bookmarks for all functions in a program file, afaik 17:37:42 yeah but I prefer a specific line usually 17:44:29 -!- zeotrope has changed nick to zoetrope. 17:45:01 -!- zoetrope has quit ("leaving"). 17:49:36 bookmarks? what if i'm in function f, and want to fix f, then return to f? 17:50:04 well okay, obviously i could still use the bookmark for f. still, that doesn't solve the problem 17:50:10 in fact it has nothing to do with it 17:50:23 i don't mind searching for f, i do mind trying to remember where i was 17:50:31 well having to try 17:53:36 -!- augur has quit (Read error: 113 (No route to host)). 17:54:22 oh, it seems vim has a stack (or list) for marks as well 17:54:57 or any jump 17:55:10 * Sgeo missed the bus he wanted to take 17:55:48 so where are you now? 17:55:52 At home 17:55:52 another bus? 17:56:05 1:20 then? 17:56:14 1:09 17:56:26 is this the c++ thing 17:56:37 Um, I have that class later in the day 17:56:43 First is a Psychology class 17:56:52 ...phychology? 17:57:01 is this high school 17:57:05 i mean 17:57:06 that other thing 17:57:37 or actually i guess high school 17:58:03 This is college 17:58:24 are you 17 or something? 17:58:32 20 17:58:36 ...what? :P 17:58:43 hmm 17:58:56 I'm taking a mix of early and late courses 17:59:07 >.> 17:59:14 i think i've misremembered your age before as well, i have a recall of you saying you're 15 a few years ago 17:59:55 I doubt I would have shared my age 5 years ago 17:59:56 my brain doesn't like change 18:00:03 i wasn't here 5 years ago 18:00:29 Anyway, off to catch a predit.. bus 18:00:37 Deewiant, how should o work with respect to values greater than 255? 18:01:29 Deewiant, either I can do % 256 or I can treat it as an unicode codepoint. The latter means it may reflect if it isn't a valid codepoint (for example, it is a number reserved as a surrogate pair or such) 18:01:46 UNDEF 18:01:56 Deewiant, how does ccbi handle it btw? 18:02:17 % 256 18:02:48 hm 18:03:10 probably a better idea, you might want to output non char data 18:05:48 -!- augur has joined. 18:17:13 Deewiant, oh btw you are missing a test for binary o 18:17:21 you check that no more than expected columns are included 18:17:26 but you are not doing the same for rows 18:17:40 took me a while to track down why some other things weren't working 18:17:45 since I had that exact bug 18:19:37 IIRC, the condition for valid unicode codepoint x is: ((x >= 0 AND x < 0xD8000) OR (x >= 0xE000 AND x < 0x10FFFF)) AND (x & 0xFFFE != 0xFFFE). 18:19:43 -!- cpressey has joined. 18:23:29 Ilari, actually it need to be 1) valid 2) representable in UTF-8 18:23:38 anyway I'm treating it as binary % 256 now 18:23:48 UTF-8 is a binary encoding, it can represent anything. 18:23:58 Deewiant, eh? 18:24:06 sure? 18:24:17 iirc there are some of the higher planes or something 18:24:31 or was that UTF-16= 18:24:34 s/=/?/ 18:24:52 The standard limits it to 0x10ffff (formerly 0xffffffff) but the idea of UTF-8 can be extended to any value 18:26:15 All those codepoints can be represented by UTF-8 (unicode only goes up to plane 16 (17th plane)). 4-byte UTF-8 would go up to plane 31. 6-byte UTF-8 (older form) would go up to plane 32767. 18:26:36 -!- augur has quit (Read error: 110 (Connection timed out)). 18:26:48 BMP is plane 0. 18:28:01 Ilari, what is the "AND (x & 0xFFFE != 0xFFFE)" bit about 18:28:36 All planes have two final code points which are assigned to noncharacters 18:28:45 ah, but why= 18:28:47 s/=/?/ 18:28:58 Something wrong with your shift key today? :-P 18:28:59 BOM? 18:29:28 Deewiant, I'm on my laptop, slightly different key placement. I use my desktop more 18:29:30 BOM for the first plane, doesn't seem to make much sense for the others though 18:29:41 reason for this is that the backlight died in desktop monitor this evening 18:29:49 so I have to buy a new one tomorrow or something 18:30:00 which is highly inconvenient 18:30:13 Deewiant, well yeah 18:30:27 Deewiant, does BOM manage to detect PDP endianness btw 18:30:54 There is no such thing as PDP endianness for UTF-16 or UTF-32 18:30:54 Should work if its UTF-32... 18:31:12 Deewiant, there should be 18:34:31 -!- augur has joined. 18:46:34 UTF-32 BOM in PDP-endianess would be '00 00 FE FF'. Little endian would be 'FE FF 00 00' and big endian '00 00 FF FE'. 18:48:26 UTF-32 BOM would contain sufficient information about byte order for decoding any valid unicode codepoint correctly, no matter what the byte order. 18:49:24 The tricks would only work if there are no characters above plane 255, but Unicode doesn't allow that anyway. 18:52:57 For UTF-16, PDP endianess and little endian are the same. 18:54:58 and UTF-8 has no endianness, no matter what Microsoft thinkgs 18:55:00 *think 18:55:02 *thinks 19:03:58 coppro, indeed I know that 19:04:25 I know. I just wanted to complain about MSFT 19:04:35 Ilari, what about PDP but with the first two bytes exchanged? XD 19:04:59 you said "no matter what the byte order." but maybe you meant "existing byte order" 19:05:13 coppro, why the FT btw 19:05:23 FT? 19:05:27 in MSFT 19:05:40 oh, that's their ticker symbol 19:05:59 yeah but why 19:06:16 because? 19:06:19 meh 19:21:55 "No matter what byte order" includes all 24 possible byte orders for UTF-32. 19:25:01 how is abcd distinguishable from bacd 19:25:10 you get double zero leading the file either way 19:25:34 Would be if the BOM were FEFF0102 or some such. 19:25:42 yeah 19:25:51 ... Oh, except that'd overlap with other chars. 19:26:07 FEFFFCFD might not. 19:26:17 it would too 19:26:32 coppro: How so? 19:26:47 Is FC or FD the top of the private usage range? 19:27:04 U+FFFD is � 19:27:14 UTF32. 19:27:34 what about it? 19:27:55 It won't consider FFFDFCFE or FFFDFEFC as "U+FFFD". 19:27:57 oh 19:27:59 right 19:28:02 yeah, that would be safe then 19:28:06 It'll consider it as U+FFFDFCFE or U+FFFDFEFC. 19:28:06 since you're never even in a valid plane 19:28:21 but then it wouldn't be a legal character 19:28:33 Oh well. 19:28:56 The only point of the BOM is to indicate the byte order. It's really just metadata for the UTF. 19:29:01 yeah 19:33:27 hm I should read stdarg.h 19:33:30 it should be fun 19:34:14 argh it uses __builtins 19:45:43 typedef __builtin_va_list __gnuc_va_list; 19:45:45 wtf 19:45:57 a intrinsic type? 19:45:59 an* 19:46:45 * AnMaster prods pikhq 19:47:25 AnMaster: Yeah, GNU C has a lot of additional intrinsics. 19:47:40 pikhq, I thought they were all function-like 19:47:40 Particularly for stdarg.h 19:47:59 but a type like one is messy 19:48:27 Nope. Good number of intrinsic types. 19:48:35 pikhq, other ones? 19:48:36 like? 19:50:17 Vector types. 19:50:21 __label__ 19:50:26 pikhq, that one is the only case of typedef and __builtin on the same line in /usr/include, /usr/lib/gcc/*/*/include 19:50:42 pikhq, aren't they made using __attribute__s iirc? 19:51:00 Oh, right. The vector types are __attribute__s on int. 19:51:12 like float foo __attribute__((something)) 19:51:17 pikhq, not float? 19:51:54 float is also valid. 19:52:02 That's a float vector rather than an int vector. 19:52:23 yeah 19:52:29 double should be valid too 19:52:55 probably char/short/long as well 19:53:07 and long long I guess 19:54:16 There's the complex types. 19:54:42 _Complex int = 3i; 19:54:54 Erm. 19:54:57 _Complex int foo = 3i; 19:55:14