00:00:01 a multiplier of what? 00:00:16 Font size 00:00:32 are you assuming every typeface is monospace 00:00:57 anyway i think that definition of em is kind of outdated nowadays 00:01:50 -!- nooga_ has quit (Ping timeout: 252 seconds). 00:02:20 kmc: {{User:Zzo38/Userboxes/program famicom}} and {{User:Zzo38/Userboxes/text adventure game}} 00:02:39 -!- Phantom__Hoover has quit (Read error: Operation timed out). 00:02:46 Oops! There is prefer DVD over VHS, and VHS over DVD, and Blu-ray over DVD, but they forgot preferring DVD over Blu-ray. 00:03:40 Bike, I'm assuming that there's no reason for sizes to be an actual amount 00:07:15 -!- Phantom__Hoover has joined. 00:12:13 Sgeo: it could override another font-size rule on the same element 00:13:11 Sgeo: it is a multiplier of the font size. when you say "12 pt font", that means 1 em = 12 pt, where a pt is 1/72 of an inch 00:14:53 i believe, but i'm not sure, that font-size: 2em; and font-size: 200%; have the same effect 00:15:11 actually http://kyleschaeffer.com/user-experience/css-font-size-em-vs-px-vs-pt-vs/ says that they behave differently if the user activates the browser's font resizing feature 00:15:33 anyway CSS has both % and em because in other contexts they mean something quite different 00:16:10 "width: 50%" means '50% of the parent's size', "width: 0.5em" means '50% of a single em in the parent's font' 00:16:54 mostly though http://img.pandawhale.com/post-18529-Yes-mlkshk-2jkG.gif 00:17:23 People who made that gif probably aren't very good at CSS. 00:17:24 that window grubbing is kind of impressive 00:17:52 Lumpio-: lol 00:19:20 -!- Phantom__Hoover has quit (Ping timeout: 260 seconds). 00:20:55 -!- Phantom__Hoover has joined. 00:20:57 someone fix my connection 00:21:03 my girlfriend has perfected a procedure for turning free copies of the Wall Street Journal into food 00:21:10 by growing mushrooms on them 00:21:37 the mushrooms are tasty and do not appear to have poisoned me or shifted my political views to the right 00:22:21 i don't think the wsj can poison you kmc 00:22:46 i also suspect it would take some kind of massive rocket to shift your political views to the right 00:24:06 haha 00:24:24 or some money 00:25:28 what if i'm robbed by immigrants at knifepoint 00:26:19 what if the immigrants were right-wing 00:26:40 hmmmmmmmm 00:30:07 -!- DHeadshot has joined. 00:33:26 -!- Phantom___Hoover has joined. 00:33:27 -!- lahwran has changed nick to lahwran-. 00:33:28 -!- lahwran- has changed nick to lahwran. 00:33:31 for your next good deed, fix my connection 00:35:17 What if political views are modular? 00:36:24 Is htmldog.com a good site? 00:36:48 -!- Phantom__Hoover has quit (Read error: Connection reset by peer). 00:37:16 cursive on my machine isn't very cursive lol 00:44:29 You might also stumble across similar proprietary properties, such as -webkit-border-radius and -moz-border-radius which are for older, barely-used versions of Safari and Firefox respectively. Our carefully worded professional advice? Screw Сem. 00:44:34 ^^from htmldog 00:47:00 Well, you might use the -moz- properties if you are writing a XUL application, rather than a webpage. 00:48:47 http://25.media.tumblr.com/13b0bd5792b912e71aafc9b201ac0dba/tumblr_mllgbzm7xu1snfhwio1_1280.jpg 00:48:54 people like jerkcity right 00:49:56 yesssssss 00:50:22 yes 00:51:49 oh it's "jerkcity hd" http://jerkcityhd.tumblr.com/ 00:51:52 yes 00:52:09 Yes. 00:52:12 this is a ray of sunshine in my world 00:52:16 yes 00:52:16 kmc likes jerkcity. don't know about anyone else 00:52:24 yes 00:52:39 jerkcity is brilliant. who wouldn't like it 00:53:26 http://media.tumblr.com/f80afb234b91bd5d8bcaa8e2c0dccf17/tumblr_inline_mlr42hjCjn1qzv9v0.jpg 00:54:27 woodring's a cool guy, for someone who's psychotic 00:54:39 is he now 00:55:04 well he based a lot of his comic on hallucinations 00:55:43 welp 00:56:08 -!- Phantom___Hoover has quit (Ping timeout: 255 seconds). 00:58:26 -!- Phantom___Hoover has joined. 00:58:32 If you have hallucinations, but can reliably determine whether something is a hallucination or not, and you don't mind the hallucinations, are you 'psychotic'? 00:58:53 ...actually, I'm worse than that, come to think of it. 00:59:09 :> 00:59:14 I regularly have hallucinations where I can't determine that it's a hallucination during the hallucination. I do reliably work it out afterwards though. 00:59:25 what kind of hallucinations Sgeo? 00:59:30 -!- DHeadshot has quit (Ping timeout: 264 seconds). 00:59:38 Sgeo: yes 00:59:48 'psychosis' is a pretty general term. 01:00:14 -!- DHeadshot has joined. 01:00:58 Full sensory hallucinations. Strong enough that any real sensory input is barely noticable. 01:02:24 this had better not be a sarcastic reference to dreaming 01:02:24 if it's not: probably should see a doctor 01:03:27 * Sgeo thought it was obvious enough that you wouldn't need to make conditional statements like that 01:03:57 Although, hmm, some sensory input is strong enough to get through, usually disrupting the hallucination. 01:04:05 so you mean they're not? 01:04:27 I mean that yes, I was referring to dreaming. 01:04:55 with you sgeo it's not always clear 01:04:57 psychosis requires being awake. 01:06:09 -!- epicmonkey has joined. 01:06:59 i assumed you weren't being difficult 01:08:49 itt communicating poorly and then acting smug when misunderstood 01:09:10 -!- Phantom___Hoover has quit (Read error: Operation timed out). 01:09:23 kmc is that An XKCD Reference 01:09:27 yes 01:09:32 from Before It Was Cool 01:10:16 from when it gently poked fun at nerd tropes instead of just generating and copying them 01:11:58 -!- Phantom___Hoover has joined. 01:13:03 0 01:13:51 It just strikes me that harmless hallucinations that everyone has are considered 'normal', while if someone had additional odd but harmless hallucinations, that would be abnormal 01:14:14 it's quite common for mentally healthy people to have auditory hallucinations 01:14:33 there is a thing called the Hearing Voices Movement that involves such people (as well as unhealthy people who also hear voices) 01:15:35 Bike: I remember reading that autism was often misdiagnosed as schizophrenia due to insufficiently pedantic questions 01:15:49 people asking questions like "do you hear voices", which were then interpreted literally 01:16:00 "yes, when people talk nearby, I can hear their voiceтАж" 01:16:39 diagnosing mental illness is hard 01:17:25 https://en.wikipedia.org/wiki/Alogia#Example 01:18:11 -!- ais523 has quit. 01:27:50 -!- listofoptions has quit (Quit: Leaving). 01:29:10 -!- DHeadshot has quit (Ping timeout: 256 seconds). 01:30:52 -!- epicmonkey has quit (Ping timeout: 245 seconds). 01:54:11 Mental illnesses are tricky. 01:55:01 Also, I do not recall having a dream ever, so *suck it*. 01:55:05 I'm saner than all y'all. 01:55:37 I had a dream about being delusional this morning, what does that make me 01:56:58 yesterday i woke up sucking on a lemon 01:58:14 -!- Phantom__Hoover has joined. 01:58:25 Everything 01:58:38 i've lately started having fits of realising when i'm going to sleep that my thoughts make no fucking sense 01:59:03 http://en.wikipedia.org/wiki/The_Unsex%27d_Females also you all need to help me understand the eighteenth century because what 01:59:35 -!- Phantom___Hoover has quit (Ping timeout: 255 seconds). 01:59:43 http://en.wikipedia.org/wiki/File:Marywollstonecraft.jpg 01:59:51 my first thought was a distinct resemblance to mrs doyle 02:00:01 as in arthur conan? 02:00:10 no, as in 02:00:25 https://www.youtube.com/watch?v=Whbc5YJz7OU 02:00:32 i have been watching an awful lot of father ted of late 02:03:06 Phantom__Hoover, try lucid dreaming? 02:03:44 it doesn't happen when i'm dreaming though 02:05:35 -!- augur has joined. 02:06:34 -!- augur_ has joined. 02:06:37 -!- augur has quit (Read error: Connection reset by peer). 02:09:33 -!- Phantom__Hoover has quit (Ping timeout: 240 seconds). 02:15:12 anyone here participating in the ludum dare? 02:18:37 -!- Jafet has joined. 02:29:32 At first when I put "This user knows the difference between scientific astronomy and superstitious astrology." on my userpage I thought I know the difference but actually I didn't. Now I do. 02:30:01 how do you know you know? 02:30:48 I don't. I won't necessarily know anything 100% certain. 02:30:58 Since everyone makes mistakes too, even scientists. 02:44:19 There's an Android app for Robozzle! 02:44:58 kmc, Bike, shachaf, you weren't here for the Robozzle addiction I think, you should play 02:45:05 hi Sgeo 02:45:07 what did i miss 02:45:09 what happened 02:45:10 help 02:45:21 http://robozzle.com/ 02:45:30 My server went down for mysterious reasons. 02:45:38 I have no idea what IRC channels I used to be in. 02:46:15 :( 02:46:18 wait why me 02:46:52 RoboZZle requires Silverlight 02:46:56 Thgeo 02:47:22 there's a js version 02:47:23 Write a clone of the software in other programming languages; I think I have done once in QBASIC 02:47:49 Bike, because you weren't here when Robozzle 02:48:02 `pastelogs robozzle 02:48:04 I should get a bag, I guess. 02:48:06 What's a good bag? 02:48:14 bag of holding 02:48:20 bag of tricks 02:48:22 It depends what you want to store in it. 02:48:40 kmc wins, with zzo38 in second place 02:48:45 I think Sgeo is just trying to trick me. 02:48:49 http://codu.org/projects/hackbot/fshg/index.cgi/raw-file/tip/paste/paste.410 02:53:01 kmc: have you figured out your excuse for moving to SF yet 02:53:12 my excuse is that my friends are there and also you can buy weed at the corner store 02:53:28 you can also get it delivered to you online i hear 02:53:43 or by delivery drone 02:53:56 yes you can use cryptocurrency to buy chinese synthetic hallucinogens through the onion router 02:54:02 because the future is a cyberpunk novel 02:54:38 no i mean a local san francisco thing 02:55:48 really 02:56:01 * shachaf looks 02:56:05 can i get large prime numbers delivered to me online 02:56:36 Cyberpunk has to have buying synthetic drugs with anonymous cryptocurrency IN SPACE kmc 02:56:54 large prime numbers are very rare [compared to most large numbers] 02:57:15 Like new bitcoins 02:57:49 how rare is your average large number 02:58:00 Very common. 02:58:13 I hear most numbers are large numbers. 02:58:46 I hear the average number is 0. 03:00:17 I just realized... now that I have money, I could buy an AW citizenship 03:01:26 Or a Worlds.com VP... but that would be unethical probably 03:01:32 *VIP 03:01:40 why? also what? 03:01:57 Worlds.com is a patent troll 03:02:00 kmc: ActiveWorlds. Think SecondLife but nineties. 03:02:36 It gets graphical updates sometimes.... kind of... 03:02:38 When I think SecondLife-but-nineties, I think SecondLife. 03:02:40 Still using RenderWare 03:02:43 Also "nineties" means "last decade". 03:02:52 And it's going to keep meaning that. 03:03:09 yes 03:03:12 Do I have to line my words up again? 03:03:19 remember when George Bush invaded Iraq in the nineties? 03:04:28 Remember when we partied like it was 1997 03:04:44 I think we did that in 1997, mostly. 03:05:38 OK, enough of that. 03:07:43 Use shorter sentences 03:12:25 kmc, have a wiki page about a game in ActiveWorlds that I loved 03:12:25 http://wiki.activeworlds.com/index.php?title=Mutation 03:14:11 You shouldn't neglect 03:14:12 polysyllabic measures 03:14:12 in your lines either. 03:14:21 -!- DHeadshot has joined. 03:14:41 monoids 03:16:01 тАЬActive Worlds Inc. supplies "White Label Solutions" to corporate clients and is able to deliver customized, turn-key 3D Universes complete with built-in tailored citizen registration systems and sales/marketing data-mining capabilities.тАЭ 03:16:46 -!- Koen_ has quit (Quit: Koen_). 03:17:29 On the one hand, AW's update process is annoying, on the other, it's better than it used to be 03:18:12 Used to prompt that it needs to be updated. New version starts, still not up to date, prompt again 03:18:14 and again 03:18:25 Now it only prompts the first time 03:19:46 -!- jconn has joined. 03:27:33 -!- DHeadshot has quit (Ping timeout: 248 seconds). 05:21:06 -!- conehead has quit (Ping timeout: 252 seconds). 05:29:02 b^a^l^l^s 05:29:20 mod p 05:29:28 the magic equation behind RSA 05:29:48 > sin (pi / 10) * 2 05:29:50 0.6180339887498948 05:30:00 i didn't know that 05:31:26 idgi 05:31:36 that's phi 05:31:42 well, phi - 1. the other phi 05:33:51 phi, the fuckiest number 05:35:06 -!- conehead has joined. 05:36:18 -!- ogrom has joined. 05:44:49 -!- nooodl has quit (Ping timeout: 258 seconds). 06:04:41 -!- trout has quit (Read error: Operation timed out). 06:05:35 -!- variable has joined. 06:08:39 -!- variable has quit (Excess Flood). 06:09:37 -!- variable has joined. 06:12:23 -!- sebbu has joined. 06:13:01 -!- sebbu has quit (Changing host). 06:13:01 -!- sebbu has joined. 06:15:36 -!- sebbu2 has quit (Ping timeout: 256 seconds). 06:23:01 At the instant the Omega Point is reached, life will have gained control of all matter and forces not only in a single universe, but in all universes whose existence is logically possible; life will have spread into all spatial regions in all universes which could logically exist, and will have stored an infinite amount of information, including all bits of knowledge which it is logically possible to know. 06:23:53 i was so pissed when they even mentioned that shit on some science TV show 06:25:06 I...what? 06:26:05 tipler's fucked and de chardin's a teleological fuck and fuck 06:28:56 looks p.fucked, yeah 06:30:11 damn fucked 06:30:15 d. fucked 06:44:25 -!- Bike has quit (Quit: leeps). 06:52:46 -!- kmc has set topic: KEYCAP PICTURE INSERT MODE | Soap with a prize inside | http://underhanded.xcott.com/?page_id=5 | http://codu.org/logs/_esoteric/. 06:57:37 -!- conehead has quit (Quit: Computer has gone to sleep.). 07:07:23 gamma goblins (part 2) 07:15:15 -!- btiffin has joined. 07:18:33 single character read interpret programming tool (toy), script. BF would be a script. The mother BFing script. 07:24:07 -!- lahwran has changed nick to lahwran-. 07:24:09 -!- lahwran- has changed nick to lahwran. 07:37:25 -!- zzo38 has quit (Remote host closed the connection). 07:38:34 -!- Zerker has joined. 07:41:48 there is a crack in everything, that's how the light gets in 07:44:39 Only the cracked see the light 07:46:29 a crack you can't see, and when the wind blows through it, it makes no sound... 07:49:38 -!- Zerker has quit (Quit: Colloquy for iPad - Timeout (10 minutes)). 08:04:53 -!- mnoqy has quit (Quit: hello). 08:14:05 -!- ThatOtherPerson has joined. 08:33:53 -!- epicmonkey has joined. 08:52:28 http://lcamtuf.blogspot.com/2011/03/other-reason-to-beware-of.html 08:52:59 -!- Zerker has joined. 08:54:42 -!- nooga has joined. 09:03:13 тПОтГг 09:19:16 -!- epicmonkey has quit (Ping timeout: 260 seconds). 09:29:20 -!- btiffin has left. 09:41:03 It interesting how when the fire alarm goes off we all just ignore it. 09:41:06 *It's 09:45:28 -!- ThatOtherPerson has quit (Ping timeout: 245 seconds). 09:52:22 Interesting indeed. 09:52:42 hi Jafet 09:53:03 haf 09:56:45 Jafet: Any puns for us today? 10:01:57 Perhaps you can find one at http://heasarc.nasa.gov/docs/xte/learning_center/discover_0606.html 10:16:49 -!- ThatOtherPerson has joined. 10:32:08 -!- ThatOtherPerson has quit (Ping timeout: 245 seconds). 10:34:39 -!- ThatOtherPerson has joined. 10:34:55 -!- Zerker has quit (Quit: Colloquy for iPad - Timeout (10 minutes)). 10:42:58 -!- ThatOtherPerson has quit (Ping timeout: 245 seconds). 11:18:16 -!- Nisstyre has quit (Read error: Connection reset by peer). 11:18:22 -!- Phantom_Hoover has joined. 11:32:18 -!- DHeadshot has joined. 12:58:26 -!- DHeadshot has quit (Ping timeout: 252 seconds). 13:16:39 -!- ThatOtherPerson has joined. 13:16:58 -!- copumpkin has quit (Ping timeout: 252 seconds). 13:17:35 -!- copumpkin has joined. 13:36:40 -!- DHeadshot has joined. 13:39:11 Oh Taneb oh Taneb, where art thou Taneb? 14:16:34 i'd say he's taneb most places 14:17:05 Really? He actually uses the name "Taneb" at school and such? 14:17:29 I would think he'd just use "Nathan" 14:18:14 ah, no 14:18:18 nathan's his internet name 14:18:38 but then why does he use "Taneb" on the internet? 14:19:18 he's very trusting 14:20:09 So what is the point of him even having an internet name? 14:21:06 to throw any pursuers off the trail 14:21:32 Okay, so I'll amend my question: 14:22:49 Oh Nathan who is really Taneb, where is thy current location? What coordinates art thou currently located at? When is thy estimated time of arrival to the nearest IRC client? 14:36:00 Heretofore doth ye antiquarian preten┼┐ions make airs 14:47:57 -!- nooodl_ has joined. 14:51:35 hm TIL this is a valid C declaration: int (*x)[10]; 14:52:06 is that just an array of ten pointers to ints, or 14:52:13 or... a pointer to a pointer to ints? 14:53:11 * FireFly guesses the latter 14:54:15 a pointer to an array, i.e. just a pointer really 14:55:54 yeah it's a pointer to an array 14:57:36 [3633262.162782] [BLOCK] IN=eth0 OUT= MAC=b8:27:eb:3e:2e:5c:58:98:35:41:33:0e:08:00:45:00:00:4a:7d:06:00:00:6d:11:ca:00 SRC=85.181.239.22 DST=192.168.1.40 LEN=74 TOS=0x00 PREC=0x00 TTL=109 ID=32006 PROTO=UDP SPT=51703 DPT=6802 LEN=54 14:57:36 [3633834.476985] [BLO\x19 14:57:36 [3634184.787933] [BLOCK] IN=eth0 OUT= MAC=b8:27:eb:3e:2e:5c:)J] 8:35:41:33:0e:08:00:45:00:00:4a:72:4b:00:00:27:11:7d:9b SRC=61.58.164.178 DST=192.168.1.40 LEN=74 TOS=0x00 PREC=0x00 TTL=39 ID=29259 PROTO=UDP SPT=11388 DPT=6802 LEN=54 14:57:40 Hm that does NOT look good 14:57:46 look at those weird chars in dmesg 14:57:49 what the hell is going on 14:57:52 hacked 14:58:10 probably not though 14:58:20 dmesg is read from a ring buffer 14:58:32 kmc, doubtful, anyway the machine is acting as a firewall, so I do expect a lot of iptables messages. 14:58:43 so I guess it overwrote part of an old message 14:58:51 but i don't know why weird bytes 14:59:02 kmc, seems strange it would be the last three lines of dmesg too 14:59:05 it is my RPi 14:59:05 if you can afford to take the machine down, you could memtest it 14:59:07 ah 14:59:16 can't memtest a RPi 14:59:18 afaik 14:59:29 Not with memtest86+ 14:59:42 sure 15:00:02 there might be a similar tool for ARM but it would have to deal with the bajillion different ARM platforms 15:00:21 perhaps the Linux kernel itself can run a test 15:00:21 kmc, also it is a PITA to check what happened, since it is headless. Would need to locate that HDMI->DVI cable and so on 15:00:25 mm 15:00:43 i blame the cosmic microwaves 15:00:58 anyway arrays aren't 'just pointers really'; they decay to pointers in certain contexts, but act differently in others 15:01:13 if you have int (*x)[10]; int **y; then I can think of at least 2 significant differences 15:01:30 elliott, possibly, but doubtful 15:01:33 one is that sizeof(*x) != sizeof(*y) 15:02:09 the other is that *x == &(*x) but *y != &(*y) 15:02:19 kmc: well I didn't realise you could even have a pointer to an array 15:02:22 which is why I thought it might decay 15:02:44 yeah me either, or anyway I didn't know the syntax for it until just now 15:02:49 by reading http://cm.bell-labs.com/cm/cs/who/dmr/chist.html 15:02:52 which is an interesting article 15:06:22 Btw, function pointer syntax: WHY 15:06:43 You also can't point x and y at the same kind of things, and (char*)(x+1) - (char*)x != (char*)(y+1) - (char*)y. 15:06:48 Function pointers are even worse in C++ (like most stuff) 15:08:04 kmc, so int (*x)[10]; is specifically an array of length 10? Hm 15:08:11 it's a pointer to an array of length 10 15:08:15 well yes 15:08:29 will the compiler warn on (*x)[11] then? 15:08:38 how do you mean? 15:09:01 I mean when accessing element 11 of the length 10 array 15:09:01 It will certainly warn if you try to point it at an int[11]. 15:09:11 which I presume you do by that syntax? 15:09:17 Probably not. 15:09:19 Or does x[11] work directly 15:09:23 'warning: initialization from incompatible pointer type [enabled by default]' 15:09:23 Why does declaring the length there matter then? 15:09:30 Hm I guess that 15:09:41 Vorpal: Because x++ will advance it by ten ints. 15:09:52 Oh, awesome 15:10:03 and sizeof will factor in the array size 15:10:04 That might be useful 15:10:09 x = malloc(sizeof(*x)); 15:10:22 If you have an array of array of 10 ints, you need an int (*)[10] to point to one "row". 15:10:24 You could work on RGB data really easily with this 15:10:45 (3 or 4, not 10 then obviously) 15:11:03 http://sprunge.us/hJaD -- from my candide backscroll. 15:11:13 Someone was asking about a very similar thing on ##c recently. 15:11:25 Vorpal: the syntax for declaring function pointers is meant to mimic the syntax for using them (as with other declarations in C) 15:11:29 it's definitely bad though 15:11:37 and i think declaration-follows-use is a bad rule 15:11:49 It's not *that* bad until you have a function that returns a pointer to a function. 15:12:17 void *call_cc(void *(*)(void (*)(void *) __attribute__((noreturn)))); 15:12:17 The argument types of the returned function pointer end up very far from where you'd normally expect a return type to be. 15:12:21 you can make it arbitrarily bad 15:12:49 declare call_cc as function (pointer to function (pointer to function (pointer to void) returning void) returning pointer to void) returning pointer to void 15:12:55 C is the worst imo 15:13:00 fizzie, I never even tried that without typedefs... 15:13:13 elliott, worse than C++? 15:13:48 well at least C++ has things 15:13:48 even if you can write or read types like that, you're really better off pretending you can't 15:13:50 Vorpal: int (*foo(int))(char); is a function that takes an int, and returns a pointer which takes a char and returns an int. 15:14:14 The return type -- int (*)(char) -- gets split up pretty badly there. 15:14:22 fizzie, ouch 15:14:41 :/ 15:14:47 yeah typedefs are called for 15:15:06 btw why does C have typedefs and not just use macros for that purpose? 15:15:10 Also it all gets worse in C++ as usual. Pointers to member functions and so on. 15:16:06 How does a pointer to a virtual member function even work? What if there is virtual diamond inheritance mixed into it? 15:16:08 kmc: You couldn't (easily) use a macro for that returns-a-pointer-to-a-function, since you can't form "pointer to T" just by appending a *, always; but you can if T is a typedef. 15:16:23 right 15:16:52 va_arg I think requires that the type is specified in such a way that you can turn it into a pointer like that. 15:17:17 Vorpal: There's a fascinating article about the implementation of pointers to member functions when it comes to virtual functions. 15:17:27 fizzie, oh? Where? 15:17:39 Vorpal: I seem to recall (on some implementation) it involving 4-, 8-, 12- and maybe even 16-byte pointers. 15:17:45 Vorpal: I'll see if I can re-find it. 15:18:08 fizzie, 16 bytes on a 64 or 32-bit system? 15:18:30 it would be utterly silly on a 32-bit system, not quite as bad (relatively speaking) on a 64-bit system 15:18:34 Vorpal: 32-bit. 15:18:40 ouch ouch ouch 15:18:57 Vorpal: It might have been http://www.codeproject.com/Articles/7150/Member-Function-Pointers-and-the-Fastest-Possible or something linking to it / linked from it. At least that one mentions the different sizes. 15:19:09 "Most bizarrely, in Visual C++, a member function pointer might be 4, 8, 12, or 16 bytes long, depending on the nature of the class it's associated with, and depending on what compiler settings are used!" 15:19:22 Hm 15:19:25 There's a table of them. 15:19:34 one (GCC) trick i learned from ksplice is to use typeof like so: 15:19:34 extern const typeof(int (*)(void)) ksplice_call_pre_apply[] 15:19:55 C's syntax is pretty weird sometimes 15:20:07 `int typedef (*bar)(int);` is also nice 15:20:13 тАЛ/home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: int: not found 15:20:13 (or not so nice) 15:20:21 HackEgo: huh 15:20:22 instead of, uh, extern const int (*ksplice_call_pre_apply[])(void) ? 15:20:23 hush* 15:20:39 typeof makes the function pointer type into something that can be composed in a syntactically normal way 15:21:03 also people might be amused by the 'templated data structure library' here: https://github.com/CentOS/ksplice/blob/master/objcommon.h 15:21:03 Yeah, in that it's kind of like an anonymous typedef. :p 15:22:30 "Current versions of the GNU compiler use a strange and tricky optimization. It observes that, for virtual inheritance, you have to look up the vtable in order to get the voffset required to calculate the this pointer. While you're doing that, you might as well store the function pointer in the vtable. By doing this, they combine the m_func_address and m_vtable_index fields into one, and they ... 15:22:37 ... distinguish between them by ensuring that function pointers always point to even addresses but vtable indices are always odd --" 15:25:00 -falloyed-rims 15:29:08 -!- DHeadshot has quit (Ping timeout: 255 seconds). 15:29:38 -!- ogrom has quit (Quit: Left). 15:32:01 -!- Koen_ has joined. 15:34:17 fizzie: O____O 15:34:26 that sounds like... that sounds like sbcl's tagging, almost 15:34:37 or for that matter the ARM thing with thumb function calls 15:37:31 -!- quintopi1 has changed nick to quintopia. 15:37:35 -!- quintopia has quit (Changing host). 15:37:35 -!- quintopia has joined. 15:43:13 kmc: nice cpp 15:44:19 hey, why isn't Robozzle Joust a thing yet 15:45:44 `int typedef (*bar)(int);` is also nice <-- that does something? 15:45:58 typedef in the middle? What does that mean 15:48:25 Fiora: heh, yeah 15:49:16 pointer tagging comes up a lot 15:49:28 if you give programmers some unused bits, we will find a way to use them 15:49:44 -!- DHeadshot has joined. 15:50:13 i find it really amusing that the whole 'canonical address' thing on AMD64 exists explicitly to prevent people from using 'unused' address bits that may become used later 15:50:37 it's kind of too bad really 15:52:03 kmc: that reminds me of that thing on win32 I remember reading 15:52:21 i think maybe they should have let the OS disable those checks; some platforms don't care about binary compatibility 15:52:30 that microsoft couldn't enable /LARGEADDRESSAWARE by default because many windows programs would use the top address bit for something, because Windows guaranteed the kernel reserved the top 2GB 15:52:34 (at the time, at least) 15:52:49 hey does anyone here have any Go experience 15:52:50 so 32-bit programs are still limited to 2GB address space by default, even on 64-bit 15:53:02 heh! 15:53:40 i wonder if there are Linux/i386 programs that do similar things 15:53:48 they would break running as compat processes on amd64 15:53:52 Fiora, what about the /3GB switch though? 15:53:59 /3GB? 15:54:17 Fiora, it is a boot.ini option that switches the split to 1 GB for the kernel, 3 for the app 15:54:20 ohhhh. I see. that's the parameter required on 32-bit to make the OS fit itself in 1GB, and leave 3GB for the app 15:54:27 yes 15:54:39 I think if I rmeember right on x86_32 to make large address aware work, you need both the app compiled with /LARGEADDRESSAWARE *and* the /3GB startup flag 15:54:58 but on 64-bit it's just largeaddressaware, I think. since the kernel has all the address space ver 15:55:01 *ever 15:55:06 Fiora, the guys at work who are still on 32-bit XP use it to be able to use incremental linking in Visual Studio. Kind of funny really. 15:55:15 -!- DHeadshot has quit (Remote host closed the connection). 15:55:38 (yeah, the ilk file gets THAT big, boost is scary) 15:56:15 kmc: did linux always enforce the 2/2 split before or something, like windows did? 15:56:26 I'd imagine on linux you probably couldn't have relied on that as much but I wouldn't know (?) 15:56:37 the address model is a compile time kernel config parameter 15:56:40 it's usually 3GB / 1GB 15:56:48 people shouldn't have relied on it, but they might have 15:57:01 actually the kernel itself uses the top page or so of addresses for non-pointer meanings too 15:57:24 functions that normally return a pointer will return ERR_PTR(-ENOMEM) or whatever on errors 15:57:35 ooh 15:57:46 The concept of the kernel having memory pages in the application address space is kind of silly really. It is an artifact of the architecture though 15:57:51 oh, so that's just like doing (void*)(-err) 15:57:57 Fiora: yeah I think so 15:58:12 Vorpal: yeah, it's hard to get good performance on x86 without it (except on some very recent chips?) 15:58:16 http://www.reddit.com/r/programming/comments/1d7g28/hacking_v8_yield_and_asyncawait_in_nodejs/ 15:58:28 kmc, oh? How did they fix that? 15:58:36 This sort of thing, if it gets standardized and into node.js, may make me actually want to try node.js 15:58:51 I think they added a PCID system? 15:59:06 IIRC S/370 allows using 16 or so address spaces per process 15:59:13 Sgeo, why link reddit rather than the original source? That just means an extra click for me 15:59:21 yeah, there's something like a TLB tag 15:59:22 Fiora, PCID? 15:59:25 "Another long overdue improvement to the page tables is the Processor Context ID (PCID). The PCID is a field in each TLB entry that associates a given page to a process. Previously, IntelтАЩs TLB could only contain entries from a single process and whenever the CR3 register was written (e.g. a context switch), the TLB was flushed. 15:59:30 The PCID lets pages from different processes safely inhabit the TLB together, so that CR3 writes no longer flush the TLB. Whenever a process tries to access a page in memory, the PCID is checked to determine whether the page is actually mapped into the processтАЩ address space; 15:59:33 ah 15:59:35 if the PCID does not match then a TLB miss occurred. This is very much analogous to IntelтАЩs VPID, which enables the TLB to contain pages from different virtual machines and avoid TLB flushes during VM transitions." 15:59:39 (sorry for the spam) 15:59:48 Because there are useful links in the Reddit thread too 16:00:02 My Scheme thing I wrote (I think it was maybe for some course?) used the whole low byte of 64-bit values as tag bits; that way (on x86-64) you can work easily on the tags just by looking at al/bl/..., and it's a simple matter of a sign-extending shift of 8 bits to turn it back to a pointer. (Of course it'll fail if virtual addresses are ever extended to full 64 bits.) 16:00:06 Fiora, nice 16:00:23 -!- Bike has joined. 16:00:29 say nice to Intel maybe? :P though apparently it came because of VM-related things 16:00:31 why would it fail, if you're using the /low/ byte? 16:00:47 fizzie: mmap everything yourself then 16:01:02 kmc, he is shifting the entire address up to store stuff in the low byte 16:01:05 oh 16:01:08 kmc: Because when you shift a pointer left by 8 to make room for that byte, the top 8 bits fall off. 16:01:11