00:00:00 --- log: started osdev/03.05.17 00:00:09 in pm i set idt. If i switch to rm and then again to pm, so i will need to set idt again? 00:01:00 erm... I don't think so 00:02:35 i've never written a demo before 00:05:23 maybe is better to write vga driver? 00:08:53 probably 00:11:57 --- quit: cmorris ("deconstructing computar") 00:19:40 --- quit: Perdix () 00:20:31 hrmm, i wonder how vga interface is setup in freebsd kerel 00:20:49 i'm hopping through code, figuring out how it works 00:26:27 --- quit: Soap` () 00:30:00 do you have a code which changes resolution? 00:31:49 just send something to int 10 00:31:56 preferably before you change to pmode 00:32:22 no no 00:32:54 i need something like outport(port, resoliution) 00:33:11 it can't be done through IO ports afaik 00:33:36 why? 00:34:22 video cards weren't designed that way 00:34:33 read up on the VESA and VGA specifications 00:47:03 --- join: wl (philipp@pD9E2D85E.dip.t-dialin.net) joined #osdev 00:51:48 int13h: Are you thinking about changing resolutions in protected mode? 00:52:26 yes 00:52:35 Well, you're in luck. 00:53:02 It's so terribly simple that you're going to be hideously annoyed by the time it's all over. 00:53:10 !!! 00:53:26 simple? 00:53:27 how? 00:53:37 http://yaxic.foo.is/docs/vesa_vbe3.pdf (I think that's the URL) 00:53:59 not found 00:54:03 darn.. 00:54:04 wait.. 00:55:03 /docs/ is missing 00:55:13 yeah 00:55:16 I see it now :) 00:55:27 but it's the vesa spec, isn't it? 00:55:40 Yup. 00:56:04 there 00:56:09 now the URL should work. 00:56:12 http://members.fortunecity.com/grkfn/VBE3.ZIP 00:56:53 Just a matter of copying a certain BIOS data area to a non-read-only area of memory and redefining a few things, then running a few built in functions.. 00:57:05 Fuck. 00:57:21 Icelandic final exam after one hour, and I can't get my yogurt open! 00:57:41 lol 00:58:24 There! 00:58:25 Finaly 00:58:50 I hate saturday morning exams. 01:00:04 you don't hate exams in general? 01:00:14 --- quit: air ("cria 0.2.8cvs15 -- http://cria.sf.net") 01:01:22 ReKleSS: Not really. More slight dispisal than pure hatred. 01:02:20 ah, ok 01:02:44 here's the thing. I was studying from 15:00 to 1:00 yesterday, and I didn't learn a bloody thing. 01:03:20 holy shit... 01:03:26 Well, okey, rather.. I didn't learn a bloody thing about the stuff thats to be tested from today. 01:03:57 I learned a great deal about a certain lady friend of mine.. but seriously, that's the last thing that should have been on my mind. *grmbl* 01:04:36 pfff 01:04:57 (I'm still a bit euphoric.. sorry. :) 01:05:02 lol 01:05:45 --- join: witten_ (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 01:06:16 smari, you don't understand me. i don't want to use bios interrupt in pm. 01:06:34 int13h: How do you want to do it then? 01:06:51 how do it bios? 01:07:05 int13h: Wave your magic wand of dereferencing and *pow*, you're suddenly in 1600x1200x32 @ 85Hz ? 01:07:07 --- quit: witten (Remote closed the connection) 01:07:14 --- nick: witten_ -> witten 01:07:23 no 01:07:36 hrmm 01:08:30 problem is that each video card is programmed differently maybe? you can go to VGA modes though i think 01:08:31 int13h: There are only so many ways of changing display modes. All cards support VGA, most support VESA, allmost all have their own superior interface documented in their relevant documentation. 01:09:11 VGA you can access via I/O ports. VESA via VBE/pm... 01:10:05 3dfx Banshee, for example, is done by reading addresses for a register table from the PCI Configuration Area, and tweaking them to your desire. 01:10:19 Most PCI cards probably work in a simelar way. 01:10:31 you offer me to use bios? 01:10:48 int13h: VBE/pm, not BIOS. 01:11:07 the VBE code segment is indeed located in the BIOS. 01:11:12 ..but that doesn't matter. 01:11:48 so anyway i have to use int 0x10? 01:12:03 you don't interrupt. 01:12:12 so how? 01:12:27 uhm, take a look at the doc 01:12:34 You just run the same code as you would have when you were in 16 bit mode - the code that the Real Mode Interrupt Vector pointed to. 01:13:19 you just do a normal "call" to that code. 01:13:24 read the doc. 01:17:46 wow, unit testing rocks 01:18:12 which is..? 01:18:34 writing a basic test for a single component or unit of your software 01:18:44 to run it through its paces and make sure it still does what it should do 01:18:53 great for flushing out new bugs that are introduced 01:19:38 indeed. Sounds clever. Any good way of doing this during OS design at lower levels? 01:20:14 sure 01:20:18 for instance.. 01:20:29 I was having all kinds of trouble with my physical memory allocator 01:20:34 couldn't track down where the bugs were 01:20:42 --- join: dax (dax@u212-239-165-213.adsl.pi.be) joined #osdev 01:20:59 (I really want to add support for a serial console on Yaxic as soon as possible.. have a few basic "poke" and "peek" style commands in a debug shell .. :) 01:21:02 so I wrote a function that just tried to allocate a bunch of memory over and over until there was no more left, and then it went through and freed all the memory that it just allocated 01:21:15 sounds simple, but it'll turn up all kinds of bugs if there are any 01:21:24 Smari: woot :) 01:21:59 As soon as these exams are done.. I'm going to take a good week at OS hacking. :) 01:22:32 heh 01:22:50 finals week is one of the best times to do coding 01:22:53 no pesky classes 01:22:57 I know.. :) 01:23:13 But it looks like I'm going to be unemployed this summer anyway.. :> 01:23:37 summer is likewise good coding time 01:24:37 Preferably get the PPC port up and running (my mother borrowed the screen from my IA32 test machine, so the only machine I can test on is a MPC823 machine with a LCD, but no keyboard...) and make memory allocation a worry of the past. 01:24:46 any one knows tutorial like switching from pm to 16 bit? 01:25:03 int13h: Why would you want to go backwards? 01:25:19 int13h: Just change the bit on cr0 01:25:45 and how about gdt,idt then? 01:26:02 just leave them. You're going to want them set up anyway f 01:26:03 Smari: memory allocation is fun :) 01:26:07 if you go back to PM. 01:27:14 smari: i set idt. if i switch to pm, then to rm, then i have to set idt again? 01:27:32 witten: I just can't get my mind around it. I've written SMP support, threading, virtual terminals, etc.. I've gone as far as I possibly can go without needing a memory manager.. because every time I've started, I've come closer to going nuts. 01:27:41 smari: i set idt. if i switch to pm, then to rm and then to pm. then i have to set idt again? 01:28:06 int13h: Don't have a clue. But why change to rm? 01:28:56 i don't understand you. Then how to use bios interrupt? 01:29:10 int13h: Read the bloody document. It explains it. 01:29:30 Smari: I'm surpised you've managed to get that far without allocating any memory! 01:29:31 witten: of course.. most of that which I've written doesn't even compile yet.. it just works theoradicly. ;) 01:29:40 Smari: hehe ok 01:29:58 Smari: well memory management is fairly difficult to do well 01:30:02 lol 01:30:08 Smari: what kind of management are you looking to do? 01:31:06 witten: Don't really care at this point. Paged, 4KB granularity, a free-list and a used-list,.. as far as anything else goes, I don't care. 01:31:35 just as long as I can get MM out of the way until I've reached a point where I can runtime debug the memory management process. :) 01:31:39 Smari: well either start simple and write an inefficient one that just works, or go all out and make a "modern" allocator 01:31:49 Smari: heh, good plan 01:31:54 well keep it simple then 01:32:17 witten: Yeah. That sounds like something I'd say. But.. but.. but.. HOW? 01:32:30 (and bear in mind that I'm going to school for a final in 15 minutes ;) 01:33:40 haha 01:33:43 i.e., got any good, clean, modular and simple code that I can look at? Almost all MM's I've seen are packed with all sorts of crud that doesn't really have anything to do with a MM. 01:34:04 Smari: well, you're welcome to look at my code: torsion.org 01:34:09 Okey.. 01:34:17 thanks. 01:34:18 but my current MM is an attempt to be a "modern" one and so it's efficient and not too simple 01:34:33 I'll check it out later. 01:34:37 if you're looking for simple, then just keep a statically-allocated array or something 01:34:44 no need to get fancy 01:34:54 indeed. 01:35:17 right now, I'm going to get my ass to school. 01:35:21 Be back later.. :) 01:35:25 goo luck! 01:35:36 thx 01:36:42 * witten just committed a metric crapload to cvs 01:42:13 'metric crapload'? 01:44:34 a lot 01:54:58 --- quit: witten ("rebooting to test new kernel") 02:06:45 --- join: witten (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 02:06:55 new kernel boots off a real floppy on real hardware! 02:07:18 heh 02:07:29 my kernel boots, but not on real hardware...... 02:07:36 I get a triple fault when I try 02:07:53 that sucks :( 02:07:57 does it boot in an emulator? 02:08:01 bochs 02:08:14 haven't tried in vmware, because I can't be bothered finding and cracking it 02:08:14 cool, that's something 02:08:20 bochs is often a little more lax than real hardware 02:08:24 heh 02:13:35 * lodda is in #osdev on irc.freenode.net but noone is in there :( 02:14:11 split? 02:14:32 ??? 02:14:41 --- quit: int13h (Read error: 104 (Connection reset by peer)) 02:14:45 lodda: what are you on about? 02:14:47 nope :D 02:14:55 i'm on port 9001 there :D 02:15:02 idiot :p 02:15:09 heh 02:15:17 what's the port for? 02:15:31 Various buglets have been fixed with dancer-ircd, version 1.0.31+pre15 is now up. To try 02:15:34 the test server, go to your favorite freenode server, but use port 9001. Stop by #freenode if you get a chance. :) 02:15:39 from WALLOP msgs 02:15:47 ah 02:22:00 we should move #osdev temporary there :D 02:24:44 --- quit: trans (Read error: 110 (Connection timed out)) 02:25:03 people who join when the channel is kinda empty would freak out :D 02:37:23 noone interested? :( 02:38:49 argh... md5sum of a 16mb file at 206mhz... 02:39:13 --- join: Zenton (~vicente@8.Red-80-34-35.pooles.rima-tde.net) joined #osdev 02:41:23 night everyone 02:41:30 --- quit: ReKleSS ("cat /dev/bullshit > homework") 02:44:14 --- join: mur (murr@baana-62-165-186-206.phnet.fi) joined #osdev 02:44:51 @£$@£@£ 02:45:00 some money there 02:45:14 3 £, 3 net credits and 1 dollar 02:47:37 mur: irc.freenode.net port 9001 02:47:46 lodda? 02:48:29 :D 02:48:36 new irc server there 02:48:47 but i dont want to be banned 02:48:52 why banned?! 02:49:06 i am there and i'm kinda alive! 02:49:55 * witten just released torsion 0.0.16 02:50:11 hrm 02:50:16 --- quit: Dipp ("Client exiting") 02:51:47 the old one seems to work just fine :) 02:52:25 hrm 02:52:30 it is a test server 02:53:10 06:23 -!- WALLOP lilo: Hi all. Various buglets have been fixed with dancer-ircd, version 1.0.31+pre15 is now up. To try 02:53:13 the test server, go to your favorite freenode server, but use port 9001. Stop by #freenode if you get a chance. :) 02:53:24 do you believe me now? :D 02:54:09 okay 02:54:18 you'll be dead soon to bug ;) 03:01:23 --- quit: witten ("Client exiting") 03:02:13 --- join: Steve (Steve@pc4-grnk1-5-cust117.renf.cable.ntl.com) joined #osdev 03:03:45 oh no! Steve Urkle! ;D 03:04:02 Urcle 03:05:36 --- quit: Divine (Read error: 104 (Connection reset by peer)) 03:06:27 --- join: int13h (~no@calcium.nelte.lt) joined #osdev 03:09:55 --- quit: EtherNet- (Read error: 60 (Operation timed out)) 03:15:09 --- quit: Steve () 03:16:28 --- join: EtherNet- (~askbill@200.45.181.40) joined #osdev 03:38:17 --- join: gianluca (~kopeteuse@ppp-238-134.28-151.libero.it) joined #osdev 03:38:29 wow. 03:38:31 That sucked. 03:38:38 hi 03:39:06 hey 03:39:39 hey mur 03:40:39 anyone knows of sdl? 03:40:56 mur: Of course. 03:41:17 is there some library that can have multiple ttf fonts and woudl draw them antialiased? 03:42:54 and or at least one but antialiased 03:43:00 and is there some antialiasing libraries? :) 03:43:04 are 03:44:19 yeah.. 03:44:33 libSDL_ttf 03:44:49 uses Freetype 2, iirc.. 03:44:55 full antialiasing support. 03:46:01 * mur woudl like to have/make a multimedia cd software 03:46:35 for many platforms 03:48:33 smari: that pminfoblock and other stuff is very confusing 03:48:43 do you know about interplatform working? 03:48:48 do you have samples 03:48:49 ? 03:55:04 --- join: trans (ncuqow@fatwire-201-205.uniserve.ca) joined #osdev 03:59:03 transisotorrr!! 03:59:04 hey 04:03:09 mur: SDL works on just about anything. :) 04:04:28 how about playback of videos? 04:04:33 mpg or such 04:05:13 SDL_websrv - webserver in a library using SDL and SDL_net HAH! :P 04:05:46 SDL_ttf and SFont which is better? 04:06:45 SMPEG seems not to be supported but 3 os! 04:07:37 what os is libvorbis supported on? 04:08:38 SDL_ttf is good. 04:09:37 libvorbis works on Linux, Solaris, *BSD, HP-UX and IRIX as well as Wintendo.. if not more OS's 04:10:23 os it shoudl work on: win lin mac, osx be 04:10:42 be recommoded 04:10:43 yeah. 04:10:51 no 04:10:55 It definately works on OSX and Be. 04:10:56 does it work on those? 04:11:06 Dunno about classic MacOS though. 04:11:07 libvorbis deoes 04:11:08 `? 04:11:11 yup. 04:12:44 GNU people are pretty big on porting things. :) 04:13:02 and what about motion pictures? 04:13:05 with sdl 04:14:22 not tried that a lot.. but I'm sure you can do it. Xine and MPlayer have SDL output plugins. 04:14:43 sounds problematic 04:14:53 as i woudl like to embed everything to one executable 04:14:57 mm player 04:15:00 SDL_flic - a small C library that renders frames of FLI and FLC animation files (should work on all platforms) 04:15:11 well those are not so common formats anymore .P 04:15:27 not really, no. :> 04:20:17 Xiph tries to work literally everywhere. 04:22:59 --- quit: EtherNet- (Read error: 110 (Connection timed out)) 04:29:47 And it does. There's even a integer-only version of the Vorbis library for platforms without a floating point processor. :) 04:30:45 Yeah, but they don't have that for all the codecs. 04:30:52 Speex, for instance... 04:30:53 :> 04:31:22 Doesn't matter. It's a start. 04:33:26 Odin- you coudl contribute >:) 04:34:06 mur: If I had the skill, I could, yes. :p 04:35:13 Odin-: Like I've been saying for, what, two years now, GET the skill. Don't just wait around for it to pop up, because it never will. 04:36:48 Smari: Silent, you. I get distracted. 04:36:57 Odin-: Focus. 04:37:27 concentrate! you can use your brains to get things fly! contentracte! 04:39:02 you can even get things fried? 04:46:52 --- join: Dr_Evil-NB (marcus-nb@p508FE548.dip.t-dialin.net) joined #osdev 04:51:25 // put my number into a data buffer to prepare for sending to a remote host 04:51:25 char data[1024]; 04:51:26 Uint32 number=0xDEADBEEF; 04:51:28 hah >:) 04:52:51 --- quit: int13h () 04:57:38 mmm 04:57:50 DEADMUR 04:58:11 syntax error on line robert 04:58:18 DEADM00R 04:58:40 I'M NAKED! 04:59:19 file do you have webcam, Robert is hot! 04:59:35 :) 04:59:41 no have webcam :( 04:59:52 phew! ;) 05:00:50 * gianluca got webcam, but it's dressed :D 05:04:12 gianluca: Robert can use his imagination 05:06:05 --- quit: gianluca ("Using Kopete IRC Plugin") 05:08:31 ;) 05:08:34 lol 05:11:39 mmh 05:11:48 i drank nice Lycchee juice :D 05:12:11 and i've got some mango juice and some fruit juice i don't know here :DDD 05:31:45 blb 05:31:49 --- nick: mur -> mur__A 05:35:31 --- nick: mur__A -> mur 05:36:00 Smari! 05:36:05 mur! :) 05:36:08 a stupid question but i dont remember 05:36:12 do ihave to open a screen? 05:36:17 yes. 05:36:18 or can i just make console sdl app? 05:36:29 you can make a console SDL app, yes. 05:36:47 but you need to run SDL_Init() either which way.. 05:36:59 want to record phonemes for me, i coudl make speech sythetisator ;) 05:37:10 hehe, kúl. 05:37:28 * mur have made one already 05:37:29 hmm 05:37:36 perhaps i coudl mix the samples 05:37:39 and retry 05:40:37 speech synthesizer? cool 05:42:10 mur: your english rapidly deteriorated. why? 05:43:30 * file pokes people 05:44:01 * mur is recycling grammar :P 05:44:06 bbl 05:45:48 GRR 05:46:00 it's been raining for days in row :O 05:46:06 :(( 05:47:59 --- join: zwane_ (User-10531@modemcable204.207-203-24.mtl.mc.videotron.ca) joined #osdev 05:49:26 --- quit: trans (Read error: 110 (Connection timed out)) 05:54:20 --- quit: file (Read error: 60 (Operation timed out)) 06:20:09 --- join: pavlovskii (~zic@host217-44-184-5.range217-44.btcentralplus.com) joined #osdev 06:23:16 --- quit: zwane ("quit") 06:23:57 hi pavlovskii 06:25:53 hi lodda 06:25:54 --- quit: Boney ("Bye") 06:26:13 I'm on my Zaurus again 06:26:58 nice 06:27:41 can you come into internet with it with cell phone? 06:27:42 --- join: file (~intranet@mctn1-0912.nb.aliant.net) joined #osdev 06:28:13 lodda: Sure. 06:28:24 * file hugs his wireless connection 06:28:37 * kdehl pukes on his one 06:29:27 * lodda wants one 06:29:47 I want my new router so I don't have to mess with this proxy software... but soon, soon. 06:30:24 I had an ISP that was delivering Internet conenctivity via 802.11b. It worked as long as I was the only customer... 06:30:37 lol 06:31:02 hi 06:31:09 morning, debug_ 06:31:14 wth? 06:31:14 hi debug 06:31:15 gmorning 06:31:23 my connection just dived to 30% 06:31:23 lol 06:31:24 it is a wonderful day 06:31:38 bleh 06:31:44 file: Welcome to reality when it comes to wireless connections. :) 06:31:48 hehe 06:32:05 debug_: I just love when you say that. :) 06:32:10 * debug_ too 06:32:22 kdehl: how's stuff going? 06:32:24 la, la, la, la 06:34:23 ... 06:34:25 debug_: Stuff ain't going at all. Just sittig here gazeing at the wallpapers... 06:35:06 ok 06:35:11 any exams left? 06:35:15 Nope. 06:35:33 so you're free? 06:35:50 have you found any summer job? 06:35:52 No. There'll be courses this summer for us that don't want to work... :) 06:35:57 ok 06:35:59 smart move 06:36:03 hehe 06:36:08 how about you? 06:36:08 * file disco dances 06:36:23 * debug_ pours up another glass of coke 06:36:40 I need to make supper then go pay for stuff 06:36:59 food is good. 06:37:09 or rather dinner 06:37:13 yes 06:37:16 ok 06:37:30 I need a girl. Anyone got one for me? 06:37:43 kdehl: :) 06:37:54 why are you not in #love anymore? 06:37:55 Just some desperate bitch. All I need. 06:38:00 kdehl: Yeah. But you're paying. 06:38:00 kdehl: depends on how much you pay for it 06:38:06 hahaha 06:38:08 :) 06:38:19 Smari, lodda: You tell me. :) 06:38:37 1,000,000,000 $ 06:38:40 I don't pay for sex. I just go without it. 06:38:42 really good girl 06:39:35 lodda: "Good woman not worth her weight in gold" 06:39:43 debug_: I'm coming back. When the IPv6 servers work. 06:40:01 lodda: Sure. When and where? 06:40:10 :) 06:40:18 Shit. 06:40:31 There's nothing as unproductive as a geek in love. 06:40:47 Smari: You are so right. 06:40:47 Smari: Yes, there is. A suit in love. 06:41:03 Odin-: Bullshit. 06:41:07 Suits don't count. 06:41:14 * debug_ is more productive when in love 06:41:14 * file wishes his parents would depart 06:41:14 (The suits being so unproductive anyway that it's difficult to go farther.) 06:41:29 debug_: Depends on the status of the other party, I guess. 06:41:31 * kdehl doesn't fall in love anymore.. 06:41:38 * mur is not productive at all when in love, otherwise he is hyperactive 06:41:54 mur: Same goes for me. 06:41:59 I am very productive, I come up with romantic tricks and stuff like that. stuff thta I wouldn't have the energy to come up with otherwise 06:41:59 indeed. 06:42:26 Not knowing the status of the other party towards oneself is worse than knowing it's a definate negative status. 06:43:13 I though about how a book about Logic would be like, if it was written by a woman. 06:43:21 true = false, sometimes 06:43:22 and so on 06:43:47 Heh. 06:43:52 a computer based on such "logic" would not even be probabalistic, it would be totally undeterministic 06:43:59 Don't apply boolean algebra... 06:44:09 girlean algebra 06:44:19 Precisely. 06:44:37 --- join: coredump (~ben@c-24-128-146-98.ne.client2.attbi.com) joined #osdev 06:45:10 kdehl får du studentbidrag om du studerar sommartid? och inte? 06:45:20 yawn 06:45:24 * geist gets up 06:46:24 mur: Japp. Man får studeimedel för sommarkurser. Annars skulle jag aldrig läsa. :) 06:46:25 gmorning geist 06:46:41 morning 06:46:48 it is a wonderful day. 06:46:57 studeimedel? 06:47:05 morning geist 06:47:27 mur: Pengar från CSN. :) 06:47:33 hi geist. 06:47:35 får man bostads hjälp och studentpengar om man studeran inte sommartid? 06:47:38 CSN? 06:47:40 hiya kids 06:47:42 up early again 06:47:51 waiting for the coffee place to open 06:47:53 pfff 06:48:06 go to bed again =P 06:49:01 mur: Eh, va? Man får inga pengar om man inte läser... 06:51:03 --- quit: pavlovskii (Remote closed the connection) 06:51:17 --- join: pavlovski (~tim@host217-44-184-5.range217-44.btcentralplus.com) joined #osdev 06:52:01 heya pav 06:52:02 hey papaa! 06:52:17 hey geist, mur 06:52:44 well, I'm gone. 06:52:52 bye Smari 06:52:57 no, you're still here 06:53:07 You know me too well. ;) 06:53:12 Smari: Where to? :> 06:53:13 But I'm leaving momentarily. 06:53:17 heh, I knew there was a computer here that was still running 06:53:18 Odin-: Out with Gummi Foss. 06:53:23 but it was so quiet I couldn't pinpoint it 06:53:28 it was this old ibook 06:53:54 now, gone. 06:54:02 cya'all 06:54:08 later 06:54:10 cya Smari 06:55:05 what program can I write? 06:55:26 --- quit: Dr_Evil-NB (Read error: 104 (Connection reset by peer)) 06:56:23 * geist wonders how useful it would be to expose a transaction model to all kernel functions 06:56:35 ie, every thing you can do with the kernel is part of a transaction that you define and can back out, etc 06:56:50 certainly fs level stuff beyond the metadata updates would be interesting 06:57:20 ie, start transaction 1, create file a, write to file a, create file b, decide you want to undo it all, abort transaction 1 06:57:31 greetings geist 06:57:37 it would be useful in a distributed environment, where the reliability of subcomponents is an issue 06:57:52 not sure how useful taht would be at the threading and vm level 06:57:59 and how friggin hard it would be 06:58:06 yeah, it would be useful in a distributed environment 06:58:13 --- join: Dr_Evil-NB (marcus-nb@p508FE548.dip.t-dialin.net) joined #osdev 06:58:35 VMS does that for filesystem stuff, right? 06:58:39 what about: 06:58:45 yes and no 06:58:52 create a process, Sleep(10000), undo process creation 06:59:00 vms had a log based filesystem 06:59:14 you couldn't, unless you recorded everything that process did 06:59:20 which isn't the same as a transaction based fs. a log based fs is an extension of a journalled fs 06:59:20 --- join: nullify (~nullify@138.89.11.39) joined #osdev 06:59:35 pavlovski: right, and how can you 'undo' the sleep. you cant get the time back 06:59:56 related to this, I always wanted the ability to undo single-stepping in a debugger 07:00:00 i.e. run a program backwards 07:00:13 BSD had lfs as well, though i dont think it was ever widely used 07:00:24 it's alive in netbsd again 07:00:27 log based fs's are pretty interesting really, I'm suprised no one has resurrect development on it 07:00:36 is it 'alive' or just 'runs' 07:00:48 there are people who work on it again 07:00:53 oh interesting 07:00:58 I've not tried it myself though 07:01:03 though in this day and age they're better off starting off 07:01:17 lfs was a good first attempt, but there are lots of things they can do better 07:01:26 er starting over I mean 07:01:43 you get pretty close to the disk's real transfer rate 07:01:45 using lfs 07:01:49 yep 07:02:01 and it lets you do stuff like take snapshots of files 07:02:06 vms let you do that with spiralog 07:02:12 * debug_ is going to have an install/test session trying out different OSes some time 07:02:17 you could do an atomic duplication of a tree of files at once 07:02:21 copy them basically 07:02:49 since vms has version numbers on files, the log based system meshed with that perfectly 07:03:25 since log based fs's dont overwrite blocks in a file, they just write a new one and make the old blocks obsolete 07:03:43 if I write a filesystem driver with write support, I'll definitely not do it without soft updates (or some equivalent) 07:03:56 just write a journal 07:04:02 a simple block based journal is pretty simple 07:04:07 ok 07:04:18 I'm not talking about inventing a new fs 07:04:33 soft updates is an interesting strategy, and basically accomplishes the same thing but approaches the problem from a different angle 07:04:46 yes 07:04:48 though all in all a really good journal implementation is pretty slick 07:04:56 like NTFS for example, it has a very cool journal 07:05:05 way beyond the 07:05:13 'new' stuff thats hitting linux 07:05:51 or the naive implementation that beos did 07:06:00 back in a minute, gotta get coffee. it opened 07:14:10 --- join: trans (namnke@fatwire-201-205.uniserve.ca) joined #osdev 07:25:32 wow, the conversation totally died 07:25:46 though now that I look at it it was mostly me talking anyway 07:26:26 at least you've got your coffee 07:26:35 that's true 07:27:32 hehe 07:28:09 i guess in the time you needed to get the coffee you could have made the coffee yourself 07:28:53 you bet 07:29:03 it's not the same though 07:29:05 --- join: wl_ (philipp@p50865DCB.dip.t-dialin.net) joined #osdev 07:31:12 geist: :) 07:31:16 * debug_ was taking a shower 07:31:31 how sexy 07:31:34 yes 07:31:37 * debug_ is all wet 07:31:40 hm 07:31:45 how about towel? 07:31:45 now, back to os discussion 07:31:45 :) 07:31:50 mur: EOD 07:31:57 how bout those spinlocks, eh? 07:32:01 :) 07:32:42 geist: I don't think you want to log every syscall 07:32:52 it would simply be too much data 07:33:09 uh what? 07:33:30 transaction recording 07:33:48 I'm not talking about my shower anymore 07:34:04 --- quit: notten (Read error: 60 (Operation timed out)) 07:34:49 oh that, yeah 07:35:02 nah, just trying to start a real conversation for a change 07:35:06 lol 07:35:14 still, it's an interesting concept 07:35:17 spinlocks... 07:35:19 yup! 07:35:29 at the fs level might be interesting 07:35:42 and possibly at some network level 07:35:58 damnit, my laptop has a terribly noisy headphone jack 07:36:02 * debug_ is dreaming about clusters 07:36:44 oh all those nice inodes 07:36:55 assigning moore and mooore to use 07:37:02 taking away reservations and assigning more 07:38:05 mmmm Mango Juice 07:38:13 better than coffee =P 07:38:17 * pavlovski contemplates installing NT 4 on his dual PPro machine 07:38:26 would require installing a hard disk and CD-ROM drive though 07:38:43 --- join: zwane (User-10531@modemcable204.207-203-24.mtl.mc.videotron.ca) joined #osdev 07:38:44 pavlovski: which version? 07:38:48 * geist loves itunes music store 07:38:50 * geist buys more stuff 07:38:57 file: which version of NT 4? 07:38:59 erm... 07:39:02 this is *precisely* what I always wanted 07:39:11 to be able to buy albums and songs off the internet 07:39:11 geist: I want Linux support! 07:39:47 --- quit: zwane_ ("Client Exiting") 07:39:51 heh 07:40:03 who _buys_ music in the internet? 07:40:38 I do 07:41:10 this was no question i wanted an answer to ;P 07:41:25 well I should go to the bank... then the post office... 07:41:41 * geist buys a paul simon album 07:41:53 * pavlovski plays Combat Flight Simulator 07:42:07 * lodda plays FlightGear 07:42:19 --- quit: nullify ("Client exiting") 07:43:13 * file ponders 07:45:22 --- quit: wl (Read error: 110 (Connection timed out)) 07:47:55 --- join: xeon (~xeon@cs6668188-147.austin.rr.com) joined #osdev 07:47:58 Hey all... 07:48:02 hi xeon 07:48:15 I'm surprised GNU doesn't already have a project named ghost. 07:49:45 * xeon thinks he'll make one and somehow fit a recursive acronym comtainging OS... 07:49:54 hello 07:51:21 transistor! 07:51:27 hey mur! 07:51:32 supsup? 07:51:33 what up? 07:51:36 hahaha 07:51:38 not much 07:51:49 you? 07:52:19 my random quote is "i don't think they are very nutritious" 07:52:27 (headlamps) 07:52:38 =) 07:53:05 ??? 07:53:18 well 07:53:19 eleven 07:53:22 it has been raining all day 07:53:22 i was the loudest? i never one spoke in capitals 07:53:31 that sucks 07:53:32 twelve little elf! 07:53:37 its raining here too 07:53:44 2 days in row already :( 07:53:44 trans: what are you talking about, you're always in caps 07:53:49 I thought you knew that 07:53:53 ? 07:53:57 maybe your irc client is screwed up 07:54:06 am i in caps now? 07:54:09 he' 07:54:09 no 07:54:10 er 07:54:10 :) 07:54:14 trans was never in caps LOL :p 07:54:18 mur: thanks mur 07:54:26 way to screw that one up 07:54:31 geist: any rough idea of a time? 07:54:31 ;) 07:54:41 oh i see 07:54:42 =) 07:55:02 oh no! geist is becoming osdev idler, he already talks a bit nonsense! :O 07:55:13 oh its !'s 07:55:21 * geist hacks on the ppc port some more 07:55:21 that makes sense! 07:55:51 *hack hack hack* 07:56:27 --- quit: zwane (Read error: 60 (Operation timed out)) 07:57:49 ... 07:58:02 so like ummm yeah 07:58:06 sweet I might not have to go to the bank... 07:58:18 * trans is bored and he just woke up 15 min ago =) 07:58:29 what's the time there? 07:58:40 7:45 07:58:43 =) 07:58:46 ah 07:58:50 * trans woke up too early too 07:58:50 my bank card's protective covering is coming off and I don't feel like having it getting jammed in the ATM today 07:58:54 8am here 07:58:55 i thought it was 12:00 or 13:00 07:58:56 :) 07:59:21 5pm here :D 07:59:22 --- quit: elvstone (Read error: 60 (Operation timed out)) 07:59:29 i have until 18:00 to kill 07:59:30 brb 07:59:31 --- quit: xeon (" The IRC Client of the Gods! - HydraIRC -> http://hydrairc.sourceforge.net <-") 07:59:58 oO 5pm that means TV 08:00:09 gods irc? 08:00:13 yay 08:00:16 well 08:00:21 who is a god in here? except xeon 08:00:30 * trans gets to smoke up with a girl who thinks he's cute tonight!!!!!!!!! 08:00:32 =) 08:00:41 ooh 08:00:43 girls 08:00:43 lodda: huh? 08:00:50 uhm 08:00:56 read the quit msg 08:00:57 couldn't contain myself any longer 08:01:01 of xeon 08:01:28 oh 08:01:42 trans: does that girl exist and does it have a brain? :D 08:01:48 yes 08:02:03 a rl girl!!! 08:02:03 do you like her then? 08:02:04 omg 08:02:17 i only met her once a year ago but yes 08:02:25 >:D 08:02:28 dont screw up then 08:02:29 :) 08:02:30 bbl anyway 08:02:37 hahaha 08:02:58 stand in front of [m ]ir(c|ror) and practise 08:02:58 i'm sure i will but hopefully it wont negatively affect the situation =) 08:03:00 bbl 08:03:50 argh, why the hell does this fail? 08:03:54 nah i'll be good and stoned anyways so i have a partial excuse =) 08:04:02 I allocate a new pagetable, zero it out and the system immediately fails 08:04:08 but as far as I can tell that memory isn't already used 08:04:39 --- join: zwane (User-10531@modemcable204.207-203-24.mtl.mc.videotron.ca) joined #osdev 08:04:44 and it can't be that I'm reallocating part of the old pagetable, because I cleared the register that points to it and flushed all of the TLBs 08:05:00 it's a big page table, maybe I should try a smaller one and see if that magically works 08:05:13 since this machine has 384MB, it's creating one that's 4MB big 08:05:23 * mur has splendid business idea 08:06:49 compareforlove.com - you upload pictures of yourself and the another one, it prompts please wait page and then outputs "you match together! probablility %i%%", rnd(i) /*60 to 100 % */ 08:06:57 of course you need to buy credits first :) 08:07:09 argh, ppc code is big 08:07:15 damn risc machines 08:08:10 hmmmm... 08:15:15 --- quit: Dr_Evil-NB ("Vision[0.9.6-0203]: i've been blurred!") 08:17:02 --- join: Uroboros (~Uroboros@ptd-24-198-195-129.maine.rr.com) joined #osdev 08:33:55 mur: lol, evil :) 08:34:34 --- quit: file (Remote closed the connection) 08:35:22 --- join: file (~jcolp@mctn1-0912.nb.aliant.net) joined #osdev 08:35:44 --- quit: file (Client Quit) 08:35:51 --- join: xeon (~xeon@cs6668188-147.austin.rr.com) joined #osdev 08:35:52 There. 08:36:05 No more sudden disconnections anymore (hopefully). 08:36:11 I upgraded my drivers :P 08:36:29 --- join: file (~jcolp@mctn1-0912.nb.aliant.net) joined #osdev 08:37:31 Slow or fast its way too costly to get that stuff. 08:37:37 err wrong window 08:37:42 hah 08:38:07 --- quit: file (Killed (NickServ (Ghost: jcolp_!~jcolp@mctn1-0731.nb.aliant.net))) 08:38:16 --- join: file (~jcolp@mctn1-0731.nb.aliant.net) joined #osdev 08:38:26 --- quit: file (Client Quit) 08:40:45 --- join: file (~jcolp@mctn1-1527.nb.aliant.net) joined #osdev 08:46:17 --- join: EtherNet- (~askbill@200.45.181.106) joined #osdev 08:46:34 --- nick: EtherNet- -> EtherNet 08:51:31 brb... uninstalling hard drive 08:51:32 --- quit: pavlovski ("Client exiting") 08:51:46 --- quit: trans (Read error: 110 (Connection timed out)) 09:01:43 --- part: Uroboros left #osdev 09:12:45 --- join: redblue (star@ppp037.216-96-207.sherb.mt.videotron.ca) joined #osdev 09:15:30 --- quit: wl_ ("Quit") 09:29:18 Is anyone here forced to use -std=c99? 09:29:53 my boss threatens to fire me if I don't use it ;> 09:38:45 --- join: bono (~bono@modemcable027.101-200-24.mtl.mc.videotron.ca) joined #osdev 09:39:38 yay 09:39:40 back 09:40:49 gwdwrfwf 09:43:35 I can't use inline or long long (which sucks if you want to support POSIX) 09:45:02 xeon: what's your problem? 09:49:32 --- nick: mur -> mur__A 09:51:08 --- join: pavlovski (~tim@host217-44-184-5.range217-44.btcentralplus.com) joined #osdev 09:55:38 does anybody know how is the bios's option to disable the onboard video card and use an agp one ? 09:56:02 heh 09:56:10 depends on your bios 09:56:12 this is different for each BIOS 09:56:24 wb pavlovski 09:56:27 "Guess" is the answer :) 09:56:29 lodda: hi 09:56:31 pavlovski: ok 09:56:39 I'm now down a hard drive and a CD writer :( 09:56:50 but I now have a hard drive and a CD writer in my dual PPro machine 09:56:52 :) 09:56:56 but I can't install NT :( 09:57:05 :( 09:57:21 use something else instead :) 09:57:39 the BIOS supports CD-ROM booting, but the boot loader says it can't find NTLDR 09:58:18 make the boot disks 09:58:20 floppies that is 09:58:28 Windows XP won't let me 09:58:30 3 floppies for Nt 4 09:58:34 you're supposed to run I386\WINNT /OX 09:58:46 boot a dos boot disk with cdrom drivers 09:58:48 but XP tries to be clever and stop you from installing NT 4 on top of itself 09:58:57 brb 09:59:02 * EtherNet is away: trying to flash a bios 09:59:02 http://www.bootdisk.com/ 09:59:10 cool 09:59:18 then I can run the DOS WINNT.EXE 09:59:29 yeah, actually that's an even better idea 09:59:40 but if you do, make sure you start smartdrv.exe first 09:59:46 it'll make a massive difference in the copy speed 09:59:58 if you can find a bootdisk here with smartdrv on it 09:59:59 does anyone here know m68k OSes that run on old macs (apart from netbsd/linux and macos) ? 10:00:17 I could just download DOS 6.22 10:00:31 I was disappointed to see that MS had dropped NT 3.51 from MSDN 10:00:38 that would have been cooler (= more retro) than NT 4 10:00:43 pavlovski: normally if I had access to the hard drive ahead of time I'd format a FAT partition, copy the contents of the NT4 i386/ dir to the disk 10:00:54 then boot into dos with a floppy, start smartdrv, and run winnt.exe 10:01:01 drat 10:01:05 should have thought of that 10:01:18 I've installed NT probably at least 1000 times 10:01:19 would have been much faster to copy the contents of the CD onto this hard disk from XP before I moved it 10:01:21 seriously 10:01:31 hardcore 10:01:37 I worked at Compaq for a few years as QA 10:01:46 3 years of installing NT each day... :D 10:01:48 hey 10:01:52 "Windows 2000 And NT4 Bootdisk Sets " 10:01:53 and my job was to basically install NT onto servers with various hardware configs 10:02:01 in about 5 different languages 10:02:06 pavlovski: oh yeah, and you can do that too 10:02:16 that boots directly into the nt setup 10:02:18 even in chinese? :D 10:02:33 lodda: nah, european languages 10:02:38 :/ 10:03:19 ever see Windows 3.2? 10:03:25 that was Chinese 10:05:32 crap 10:05:39 got 80% of the way through the *server* boot disk 10:06:43 I liked NT4, but not it's lack of FAT32 support 10:06:56 why not use NTFS? 10:08:23 pavlovski: I don't remember why I didn't use NTFS 10:08:32 I think because I couldn't boot with an MSDOS boot disk and read it 10:10:14 hmm 10:10:22 I keep my C: drive formatted as FAT16 for that reason 10:10:37 it's only 2GB but it's useful for old DOS games that won't run under XP 10:11:04 well I use Linux now, so bleh to it. 10:14:05 w00t! 10:14:13 "Setup is inspecting your computer's hardware configuration" 10:16:53 wonder why it's only detected 1 CPU 10:17:06 --- join: trans (tcepgd@fatwire-201-205.uniserve.ca) joined #osdev 10:17:46 --- join: Steve (Steve@pc4-grnk1-5-cust117.renf.cable.ntl.com) joined #osdev 10:17:49 pavlovski: setup only sees one cpu, it's a uniprocessor kernel 10:17:50 dont worry 10:18:21 hmm 10:18:32 it said 1 CPU, 64MB, MultiProcessor kernel 10:18:36 we'll see 10:20:09 yeah it always says that 10:20:22 maybe it's the smp kernel but it's overridden to only see the first one 10:20:35 pff, why? 10:20:36 but yeah, it *always* only sees one of them during setup, do not fret 10:21:05 ok, it's detected my "MPS Multiprocessor PC" 10:21:10 maybe since you haven't installed a custom HAL and there are some streange setups that wont smp right without some tweaking 10:21:18 --- nick: mur__A -> mur 10:21:18 so they dont want the setup program not booting 10:21:18 and also the fact that I haven't got a mouse connected :) 10:21:23 so they do the safe thing 10:21:39 but they want a smp kernel because it has the code to find the other cpus so that they can detect it 10:21:43 that's my theory 10:21:56 that's pretty good reasoning 10:22:33 * geist goes to take a shower 10:22:37 I'm icky 10:22:39 well, but they could at least detect them but not bootstrap the other procs 10:25:06 lodda: I think that's just a UI point 10:26:52 some users could get afraid and bring their workstations back, because they don't have several processors in their peecees ;D 10:27:03 (well, they'd think that) 10:27:56 or they could wait for about 2 minutes, when they read that NT has after all detected their MP system 10:29:09 if they are impatient... 10:30:10 or stupid :p 10:30:27 hmm 10:30:33 no, i want to blame MS 10:30:44 ther user's always right 10:30:47 or wrong 10:30:55 stupid people aren't allowed to use NT 4 10:31:00 there's a clause in the EULA 10:31:00 pfff 10:31:32 so you are basicly breaking with the law? :PP 10:33:48 my CD-ROM drive sounds like a lawnmower 10:34:01 here goes... NT 4 booting! 10:34:09 * file cheers 10:34:26 geist: you're right... "2 System Processors" 10:35:07 * lodda feels being ignored :'( 10:37:00 Anybody know what new Windows version MS is thinking of releasing in the near future? 10:37:18 I've only heard of Longhorn and Windows Server 2003, thats all 10:37:25 have no idea how they're related though. 10:37:36 Windows Server 2003 is already out 10:37:55 is Windows XP Media Center Edition out? 10:38:27 I think so 10:39:10 Windows Server is nothing new, I know... I want leap like Windows 2000 => Windows XP kinda leap, like something innovative... 10:39:38 sell ToriX 10:39:42 to ms 10:39:48 well wait for Longhorn then 10:41:21 lodda: Ahahaha, I don't think they want a bootloader 10:41:51 hmm 10:43:34 --- join: zwane_ (User-10531@modemcable204.207-203-24.mtl.mc.videotron.ca) joined #osdev 10:44:23 --- quit: Steve (Read error: 110 (Connection timed out)) 10:45:07 --- quit: zwane (Read error: 110 (Connection timed out)) 10:52:10 --- quit: sayke (Read error: 104 (Connection reset by peer)) 10:54:24 lol 10:54:31 "Congratulations to Mr. Ville Mikael Turjanmaa" 10:54:35 "Congratulation to you, concerning your initiative, and thanks for dowing this exploration. 10:54:35 We all suffer from the Microsoft WinOS, which causes TIMELOST to millions of Users every Day. 10:54:35 " 11:00:39 --- join: file[outside] (~intranet@mctn1-1527.nb.aliant.net) joined #osdev 11:02:13 so yup 11:04:00 yup 11:05:42 * file[outside] is on the front bench of his home 11:08:37 --- join: witten (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 11:15:58 --- nick: mur -> mur_A 11:24:39 --- join: Dr_Evil-NB (marcus-nb@p508FFD0B.dip.t-dialin.net) joined #osdev 11:25:34 hm 11:26:39 --- join: SLACKo (~SLACKo@62.114.151.0) joined #osdev 11:31:17 --- nick: mur_A -> mur 11:31:47 --- nick: Dr_Evil-NB -> Dr_evil 11:32:19 --- nick: Dr_evil -> Dr_Evil 11:38:59 hey SLACKo SLACKo SLACKo! 11:39:55 mur mur mur ! 11:42:57 sup=? 11:56:55 nm 11:57:00 playing with stuff 11:57:02 you? 11:58:20 redoing political test 11:58:21 :) 11:58:32 www.politicalcompass.org 11:58:38 --- join: nutrageous (~Ishq@frm-64-4-102-202.access.ntelos.net) joined #osdev 11:58:40 --- quit: Ishq (Read error: 54 (Connection reset by peer)) 11:58:55 --- nick: nutrageous -> Ishq 12:06:03 >:) 12:09:10 trans http://www.digitalronin.f2s.com/politicalcompass/images/enParties.gif 12:14:56 --- quit: trans (Read error: 110 (Connection timed out)) 12:15:11 --- join: zwanem (User-10531@modemcable204.207-203-24.mtl.mc.videotron.ca) joined #osdev 12:15:29 --- quit: zwane_ (Read error: 110 (Connection timed out)) 12:20:55 damnit 12:24:19 --- join: green_ (~green@adsl-64-160-53-55.dsl.snfc21.pacbell.net) joined #osdev 12:27:36 --- nick: mur -> mur__A 12:29:44 --- join: Divine (~john@12-246-112-182.client.attbi.com) joined #osdev 12:33:54 --- join: revanthn (lsh@202.9.183.131) joined #osdev 12:34:47 --- join: nutrageous (~Ishq@frm-64-4-102-202.access.ntelos.net) joined #osdev 12:35:02 --- quit: Ishq (Killed (NickServ (Ghost: nutrageous!~Ishq@frm-64-4-102-202.access.ntelos.net))) 12:35:12 --- nick: nutrageous -> Ishq 12:44:24 --- quit: witten ("Client exiting") 12:49:06 --- nick: mur__A -> mur 13:04:54 --- quit: file (Killed (NickServ (Ghost: jcolp_!~jcolp@mctn1-1322.nb.aliant.net))) 13:05:09 jf. 13:06:34 --- quit: redblue (Read error: 104 (Connection reset by peer)) 13:07:48 wow the difference between -O0 and -O1 is huge. 13:08:31 wow the difference between dax and daxie is huge. 13:08:48 as in 39 vs 14 seconds :O 13:08:57 oh 13:20:29 --- join: air (~brand@12-210-162-121.client.attbi.com) joined #osdev 13:20:44 hola amigas 13:21:00 Hola, freak. 13:21:48 lol 13:21:52 hi air 13:23:50 --- quit: file[outside] (Read error: 113 (No route to host)) 13:25:52 crush just keeps getting better 13:26:08 (old a) ; previous value of variable 'a' 13:26:17 heh 13:26:20 (org a) ; original value of 'a' 13:26:24 air: http://www.ucl.ac.uk/~ucapare/outline.htm 13:26:32 (sum a) ; total of all values assigned to 'a' 13:26:33 sounds nice air 13:26:52 (length a) ; number of assignments to 'a' 13:27:11 not sure about the length name 13:27:39 Anybody got a good full, scalable, fixed-width, unicode font that can be read with minimal problems? 13:27:45 stole length from an m$ research paper 13:27:52 hehe 13:28:00 Smari: There doesn't exist a single full unicode font. 13:28:07 Okey. 13:28:10 air: Why would someone need that, air? 13:28:17 xeon: need what? 13:28:24 Woulden't that introduce a lot of complexity? 13:28:27 Anybody got a good ALMOST full, scalable, fixed-width, unicode font that can be read with minimal problems? (I hate you Herbert ;) 13:28:29 xeon: nope 13:28:47 Odin-: What are you working on anyway? 13:28:56 Smari: Nothing, really. 13:29:17 Smari: Might get back to messing in Emacs; I just arrived. 13:29:20 for(a = list_head; a != NULL; a = a->next) free(a); 13:29:32 Odin-: Jakk. 13:29:43 Odin-: Write that friggin SDL YBV implementation. 13:29:44 :) 13:29:46 'a' is freed and then for must execute a=a->next; 13:29:58 from freed memory 13:30:05 air: It mgihtl become one of those things hated like goto... 13:30:15 Smari: Hm. Maybe. 13:30:26 so u must explicitly keep old values 13:30:36 Odin-: I'll take that as a yes and expect results tomorrow. 13:30:46 each variable is going to be 5 virtual variables then... 13:30:46 Smari: No can do. 13:31:02 for(a = list_head; a != NULL;){ old = a->next; free(a); a = old; } 13:31:16 Smari: Tomorrow, I'll be working like a maniac, trying to (not) fail at math. 13:31:18 Odin-: I didn't ask you if it were possible. I told you to perform the impossible. Now move! Time is a wasted! 13:31:25 air: i like it :D 13:31:43 --- quit: revanthn () 13:31:43 (damn, I'd probably make a nasty CEO. ;) 13:32:02 and what happens when someone wants to do (old (old a)) ? 13:32:21 xeon: dont work that way 13:32:25 air: So isn't that true? for every variable there will be five virtual variables? 13:32:38 old takes a variable symbol, the inner old would return a value 13:32:44 Smari: No. You'd make a demonic CEO. 13:32:45 That could get very bloated, couldent it? 13:32:56 hrm 13:33:12 Odin-: Or that. 13:33:16 xeon: no 13:33:56 xeon: ORg would only store the oginial value IF it is used on a variable, not for all vars all the time 13:34:12 xeon: OLD only stores the old value right before it is used 13:34:36 xeon: right before the last set before it is used 13:34:52 xeon: the sum and length would only be kept if they are used 13:35:18 xeon: it only adds overhead when u USE em, and they replace what u would have to explicitly do 13:36:16 that free was bad example 13:36:19 --- join: revanthn (evvwsu@202.9.183.169) joined #osdev 13:36:56 since crush doesnt have a free 13:37:58 what does it have? a gc? 13:38:02 ya 13:38:13 whedid you put that in? 13:38:20 and how often does it run? 13:38:53 it runs in its own thread 13:39:14 low priority unless memory is exhausted 13:39:40 its priority will increase as memory fills up 13:40:50 --- join: reduser3 (~blah@user-12hdn4q.cable.mindspring.com) joined #osdev 13:42:39 so, every variables assignment is actually two and possibly three if it's an original assignment? 13:43:00 --- join: trans (wkswqg@fatwire-201-205.uniserve.ca) joined #osdev 13:43:15 no 13:43:55 you need to store the old value, right? 13:44:27 ``IF'' there is an (org a) in the function ``THEN'' the top of the function will have a hidden (set old-value-of-a a) 13:44:52 i put the `` and '' in there just for ``YOU'' :) 13:45:25 --- join: mrMister (~andri@ti122110a080-0648.bb.online.no) joined #osdev 13:45:58 uhm 13:45:58 wtf 13:46:10 (set a: 1) ... (set old-value-of-a: a) (set a 2) ... (old a) ... 13:46:15 er 13:46:20 somehow i think people should stop doing those idiot jokes 13:46:36 in the ``IF'', s/old-value-of-a/org-value-of-a:/ 13:46:55 dax: which? 13:47:21 i don't believe those "girls" on msn who are so-called friends/sisters of "someobody you know" that have a crush on you 13:47:23 the (set ???-value-of-a: a) code is implicit and not typed by the programmer 13:47:31 that just sounds like a bloody idiot practical joke 13:48:40 hum this time i got "her" GSM number 13:54:23 you ARE an idiot if u use msn 13:56:09 --- join: dh (~gfafgawrg@213-35-251-83-dsl.kvm.estpak.ee) joined #osdev 14:06:22 air: and you are arrogant, but do i complain? :P 14:07:03 u just did :) 14:08:38 --- join: keyhack (keyhack@216-190-244-176.nrp3.mon.ny.frontiernet.net) joined #osdev 14:09:20 hehe 14:10:22 why do people always write the wrong things with capital letters, in example what air just did; he should write YOU big 14:10:44 s/big/with capital letters 14:11:51 is that soo difficult? *g* 14:13:11 --- quit: keyhack (Client Quit) 14:13:15 pff 14:14:06 * lodda is bored 14:14:37 lodda do you know css? 14:14:45 --- join: witten (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 14:14:50 i had to make sure he saw it as ARE and not aren't :) 14:14:50 mur: a little 14:14:57 pff 14:14:58 * dax admires his raytracer 14:15:11 * air admires his language :) 14:15:18 can someone tell me how i can have 1 pixel wide solid colour borders in table? 14:15:30 between cells not around 14:15:36 air: i would prefer using capitals for "are" too, instead of "you", as suggested by lodda 14:15:39 air: sure, everyone would expect you to give a good comment on msn 14:15:51 no 14:16:01 air, dax anyone other knows 14:16:02 ? 14:16:02 heh 14:16:25 mur: html? 14:16:26 it's like... "nobody expected to see a negative comment on msn in this chatbox" 14:16:37 (cfr. spannish inquisition) 14:16:38 you said those girls were idiots, but air wanted to say that YOU are if you even use msn 14:16:56 writing are with capital letters makes no sense to me 14:17:04 * witten writes a disk cleaner 14:17:15 air html or preferably css 14:17:18 both will do 14:17:20 lodda: thats because yer german, yer cars make no sense to us americans 14:17:25 :p 14:17:33 because you americans are weirdos 14:17:43 lodda: wtf would anyone in their right mind wanna drive a friggen egg???? 14:17:53 pff 14:18:01 :) 14:18:26 * dax renders a 20480 by 15360 image 14:18:31 air can you help? 14:18:39 mur: ya 14:18:50 what is the border argument in css anyhow (forgotten) :/ 14:18:50 dax: heh 14:18:58 dax: what do you render? 14:19:19 lodda: a big central sphere with 4 spheres touching it, reflective surfaces 14:19:28 1 ambient and 3 point lights 14:19:33 hrm 14:19:43 gogo 32x supersampling. 14:19:43 mur: border:1 14:19:43 how long does it take? 14:19:51 nobody knows 14:19:54 air: woudlnt help 14:19:55 not 1px? 14:19:55 border-color:#000000; 14:20:02 ooh 14:20:03 --- join: HeavyJoost_ (~HeavyJoos@a213-84-139-110.adsl.xs4all.nl) joined #osdev 14:20:03 --- quit: HeavyJoost (Read error: 104 (Connection reset by peer)) 14:20:10 mur: border-width:1;? 14:20:27 mur: border-style:solid; 14:20:42 --- nick: HeavyJoost_ -> HeavyJoost 14:21:01 hmm memory usage of the raytracer seems to be ok 14:21:08 border-width made it visible 14:21:13 using a bit less than 2mb 14:21:15 but what about cell's borders? 14:21:18 hmm 14:21:19 oh wait 14:21:21 sec 14:21:22 k this is going to take a bit too long lol 14:22:38 tomorrow i'll try to spec a scene/shader language 14:23:07 have fun :D 14:23:17 and i have to clean the code a bit 14:23:23 --- nick: reduser3 -> sh1nta 14:23:41 and figure out a way to get it into multiple threads (to take advantage of smp), and maybe a way to split it over multiple pcs 14:24:09 hm 14:24:26 and i seriously have to work on the speed lol it's BAD. 14:26:47 mur: http://www.w3.org/TR/REC-CSS2/tables.html 14:27:48 ugh okay 14:28:33 thanks 14:34:37 --- part: sh1nta left #osdev 14:34:49 --- part: revanthn left #osdev 14:40:57 mur: figure it out? 14:41:06 yes 14:41:11 how? 14:41:15 good you gave me this link 14:41:29 i want to do xml + css documents 14:42:10 TR#row1 { border-top: 3px solid blue } etc 14:43:40 --- join: revanthn (uice@202.9.183.235) joined #osdev 14:45:36 air, btw border: 1px solid #505050; 14:45:53 ya 14:46:04 perhaps i coudl somehow specify TR and TD inside TABLE.foo to follow some pattern hmm 14:46:15 TD and TR with not class specified, that is 14:46:24 table { 14:46:24 border: 3px dashed #ff0000; 14:46:24 border-spacing: 10; 14:46:24 padding: 20; 14:46:24 } 14:46:25 tr { 14:46:27 border: 3px solid green; 14:46:29 } 14:46:38 tr is used in other tables too 14:46:39 bum 14:46:51 my tr isnt working 14:46:52 TABLE.foo {..} 14:47:11 s/tr/td/ and it works 14:47:18 hm 14:49:30 Arg! Damn `` and '' ! 14:49:49 (Sorry, I was getting my hair cut + taking shower) 14:50:10 :) 14:52:18 --- join: thin (~thin@198.162.21.158) joined #osdev 15:01:07 --- quit: Dr_Evil ("Vision[0.9.6-0203]: i've been blurred!") 15:06:11 night 15:12:43 why? 15:12:49 day 15:12:56 WHY!?!!?! 15:12:58 :P 15:17:29 ARG! 15:17:33 Tunes stands for ``Tunes is a Useful, Not Expedient, System''. 15:18:57 ya 15:19:34 xeon: what about it? 15:27:58 xeon is learning to read 15:28:05 encourage him 15:28:27 it uses `` and '' 15:28:28 --- join: keyhack (keyhack@216-190-244-176.nrp3.mon.ny.frontiernet.net) joined #osdev 15:28:55 and? 15:29:06 I don't get Wiki's... I mean, can't someone just delete all the work? 15:29:16 geist: It's just annoying. 15:29:18 There 15:29:22 There's no purpose to it 15:29:23 xeon: sure, but someone else can restore it 15:29:32 They could have just used " and " 15:30:10 thin: Yes, but what if someone JUST got through writing a huge book before it was backed up? 15:30:27 it's backed up immediately after you input it 15:30:43 AH 15:30:46 i see 15:30:48 xeon: in microsoft word, the quotation marks point in different directions depending on where the quotation marks are (beginning or end) 15:30:49 xeon: where did they use ``and ''? 15:31:01 Tunes stands for ``Tunes is a Useful, Not Expedient, System''. 15:31:33 xeon: what page uses ``and ''? 15:31:37 --- quit: keyhack () 15:32:15 --- quit: revanthn () 15:32:29 Quotation marks are not the same as the second symbol. 15:32:33 http://www.eleves.ens.fr:8080/home/madore/computers/tunes.html 15:32:46 (A second being a part of a degree.) 15:33:37 xeon: wtf u reading that site? 15:34:36 air: Umm, yea, you told me to 15:34:40 --- quit: mrMister (Read error: 60 (Operation timed out)) 15:35:05 xeon: look at hostname 15:35:08 Wait, I thought Tunes meant Tunes is a Useful, Nevertheless Expedient System 15:35:15 xeon: it does 15:35:18 it means both 15:35:44 how bout it means 'name isn't important really so we came up with a clever one' 15:35:45 xeon: www.tunes.org/cliki/ 15:35:48 the page yer reading is OLD 15:36:03 cliki.tunes.org 15:36:38 Wow, good point... 15:36:38 Now the whole situation is absurd: there is no conceptual distinction between memory and disk - both are used to store information, and the only valid difference is that the memory is faster so it acts with respect to the disk as a cache acts with respect to memory. Now a programmer will not usually choose to control the cache explicitely so the same should be true of the memory with respect to the 15:36:38 disk: both should be accessible as an (essentially infinite) storage medium for data. 15:36:51 air: I know, I'm going through the whole 'course' 15:39:02 do u have a problem with that or u just dumping it for geist? :) 15:39:30 I was showing it to you. 15:39:35 why? 15:39:48 I'm impressed 15:39:57 This design philosophy is winning me over 15:40:23 brix is the fast version of tunes 15:40:49 screw you all. flos is the best os 15:40:54 thin: :) 15:40:58 flos = fuxing l33t OS 15:41:15 and it's coded in forth! and uses the humane interface with ZIP! 15:41:24 and in an exokernel, etc etc etc 15:41:41 heh 15:42:12 u know exokernels suck 15:42:29 heh 15:42:39 air: whaddya mean? 15:43:15 thin: heh 15:43:28 hm 15:46:55 Mach gave exokernels a bad name because they fucked up 15:48:07 The ONLY thing I dislike about this design is the need to force every program to be OSS 15:48:24 Mach isn't an exokernel, if I'm not mistaken. 15:48:34 It's a microkernel 15:49:47 air: What if two functions have the same name? 15:50:04 xeon: depends on which version of mach you're talking about 15:50:19 older ones were more pure ukernel, then later they started pulling lots of stuff into the kernel 15:50:25 xeon: that's right, forgot 15:50:32 xeon: mach gave microkernels a bad name :P 15:50:36 it's still much more modular than you'd think 15:50:49 xeon: crush allows multiple functions with same names and different interfaces 15:52:02 air: What do you mean by that? What if two interface functions conflict? 15:52:52 xeon: but primitive types belong to 'special' classes ::INT, ::FLOAT, ::NULL, ::SYMBOL so u cant have (defun foo [a:int]...) and (defun foo [a:uint]...) cuz they both belong to ::INT 15:54:10 air: So files, functions, and variables are all the same thing, in one big database? 15:54:16 umm, so you allow function overloading then 15:54:36 xeon: but u can have (defun foo [a:int]...) and (defun foo [a:int b:int]...) 15:55:49 xeon: no 15:56:02 Oh :( that would be cool though. 15:56:27 xeon: variables arent stored in any database 15:56:44 xeon: functions are stored in libraries, the interfaces are stored in the crush db 15:56:51 xeon: files are stored in lots of other dbs 15:58:14 --- join: kyelewis (~w00t@203.221.42.60) joined #osdev 16:01:29 --- quit: SLACKo (Remote closed the connection) 16:01:35 heyas 16:02:04 hayas 16:02:59 lol 16:03:14 * kyelewis is looking at stats for may 16:04:30 air: What if they all WERE the same? 16:05:03 so, my NT4 install didn't work 16:05:09 ooh - my windows machine has been up 2 days, 16 hours, 17 minutes and 21 seconds - maybe this uptime will beat my last uptime record of 3 days :P 16:05:14 I reckon one of my CPUs is kaput 16:05:15 You could do namespace:file-or-function-or-variable-here to access ANYTHING! 16:05:24 xeon: huh? 16:05:37 xeon: files DO NOT have names 16:05:44 all functions, variables, and files were the same 16:05:45 it would get, say, half-way through drawing a piece of text, and hang 16:05:50 crush has modules (namespaces) 16:06:01 YES! that's what I mean 16:06:08 kinda 16:06:25 xeon: http://brix-os.sf.net/?p0=info&p1=crush&p2=modules 16:06:27 How would you represent a, say, browser? 16:06:52 xeon: huh? 16:07:16 xeon: u should wait until i commit the new crush doc changes and then read it 16:07:43 What I mean is represent how to launch browser 16:07:50 launch a browser* 16:08:06 Under your "file system" 16:08:15 u would have a browser module that contains function to parse html and render it 16:08:43 and u have a window object that contains widgets and code snippets that call the functions in the browser module 16:09:05 Wow... 16:09:28 If these ideas work, the job market for programmers will become shit :P 16:09:37 Are you sure you want that on your head? 16:09:43 the browser window object is bound to the text mime type, along with the normal text viewer window 16:10:36 the url loader thingy will open the html file using the browser window but u can click in titlebar to change window objects 16:11:28 is a true document-centric OS 16:11:34 its .. 16:12:01 wtf would brix kill off programmers? 16:12:08 er programming jobs 16:12:20 xeon: go learn something about economics 16:12:29 it's okay to kill off the job market for programmers 16:12:40 in fact, it's a GOOD thing for the economy 16:12:46 change = good 16:12:50 heh 16:13:19 same with farmers, stop subsidizing farmers = better for the economy 16:13:24 maybe thats why geist hates brix, he's gonna get fired 16:13:36 :) 16:13:48 --- join: ReKleSS (~ReKleSS@c17565.mckinn1.vic.optusnet.com.au) joined #osdev 16:20:04 geist: http://www.geekculture.com/joyoftech/joyarchives/001_300/243.html 16:20:45 geist: don't worry about why i gave you the url 16:20:53 air: Why don't you get rid of types once and for all ? :P 16:28:32 no 16:28:47 i have implicit declarations, thats good enuf 16:28:50 http://www.geekculture.com/joyoftech/joyarchives/436.html 16:29:57 function/macro templates + implicit declarations give crush both static and dynamic typing benefits 16:30:08 --- join: keller` (~wing0@200.103.137.104) joined #osdev 16:34:30 leeeeaving 16:36:53 --- quit: mur ("MURR!") 16:52:11 --- quit: kyelewis (calvino.freenode.net irc.freenode.net) 16:52:11 --- quit: Zenton (calvino.freenode.net irc.freenode.net) 16:52:11 --- quit: TheStar (calvino.freenode.net irc.freenode.net) 16:52:11 --- quit: mors (calvino.freenode.net irc.freenode.net) 16:52:11 --- quit: gab (calvino.freenode.net irc.freenode.net) 16:52:11 --- quit: lodda (calvino.freenode.net irc.freenode.net) 16:52:11 --- quit: CLxyz (calvino.freenode.net irc.freenode.net) 16:52:20 whats the regular expression for .* that doesnt contain a space? 16:52:22 --- join: kyelewis (~w00t@203.221.42.60) joined #osdev 16:52:22 --- join: Zenton (~vicente@8.Red-80-34-35.pooles.rima-tde.net) joined #osdev 16:52:22 --- join: TheStar (~star@nsw39-adsl-046.tpgi.com.au) joined #osdev 16:52:22 --- join: CLxyz (~clxyz@00-01-03-d2-a4-07.bconnected.net) joined #osdev 16:52:22 --- join: gab (~prfalken@gaia.chx-labs.org) joined #osdev 16:52:22 --- join: mors (~skywalker@64.104.136.140) joined #osdev 16:52:22 --- join: lodda (~playgroun@h138n2fls31o965.telia.com) joined #osdev 16:52:31 --- quit: thin ("nooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo") 16:53:08 --- quit: Odin- (calvino.freenode.net irc.freenode.net) 16:53:37 --- join: Odin- (~sbkhh@adsl-2-216.du.snerpa.is) joined #osdev 17:13:46 --- quit: Odin- (calvino.freenode.net irc.freenode.net) 17:14:06 --- quit: pavlovski ("Client exiting") 17:15:17 --- join: Odin- (~sbkhh@adsl-2-216.du.snerpa.is) joined #osdev 17:22:38 --- quit: trans (Read error: 60 (Operation timed out)) 17:25:59 --- join: zwane_ (User-10532@modemcable204.207-203-24.mtl.mc.videotron.ca) joined #osdev 17:30:55 --- quit: zwanem (Read error: 110 (Connection timed out)) 17:39:28 --- quit: zwane_ (Read error: 60 (Operation timed out)) 17:42:13 --- join: redblue (star@ppp095.216-96-207.sherb.mt.videotron.ca) joined #osdev 17:46:50 --- join: zwane_ (User-10532@modemcable204.207-203-24.mtl.mc.videotron.ca) joined #osdev 17:55:57 --- join: file (intranet@mctn1-0890.nb.aliant.net) joined #osdev 17:57:48 --- join: Uroboros (~Uroboros@ptd-24-198-195-129.maine.rr.com) joined #osdev 18:02:09 --- quit: xeon (Read error: 104 (Connection reset by peer)) 18:06:44 --- join: zwane (User-10532@modemcable204.207-203-24.mtl.mc.videotron.ca) joined #osdev 18:07:44 --- join: Perdix (~robre@adsl-32-193-94.bna.bellsouth.net) joined #osdev 19:02:25 --- join: trans (urfptq@fatwire-201-205.uniserve.ca) joined #osdev 19:08:08 --- join: Boney (~paul@dsl-203-113-192-218.VIC.netspace.net.au) joined #osdev 19:26:02 --- quit: Uroboros ("Leaving") 19:26:12 --- join: air- (~brand@12-210-162-121.client.attbi.com) joined #osdev 19:26:13 --- quit: air (Read error: 104 (Connection reset by peer)) 19:26:24 --- nick: air- -> air 19:36:58 --- quit: Zenton ("Client Exiting") 19:37:18 --- join: air- (~brand@12-210-162-121.client.attbi.com) joined #osdev 19:37:35 --- quit: air (Read error: 104 (Connection reset by peer)) 19:40:21 --- join: air-- (~brand@12-210-162-121.client.attbi.com) joined #osdev 19:40:21 --- quit: air- (Read error: 104 (Connection reset by peer)) 19:43:22 --- quit: trans (Read error: 110 (Connection timed out)) 19:53:52 --- quit: redblue (Read error: 104 (Connection reset by peer)) 19:53:59 --- nick: air-- -> air 19:56:08 --- join: air- (~brand@12-210-162-121.client.attbi.com) joined #osdev 19:56:08 --- quit: air (Read error: 104 (Connection reset by peer)) 19:56:31 Make up your MIND!!! Sezhe air! 19:56:35 :) 19:57:15 heh 19:57:17 --- nick: air- -> air 19:57:30 damn server keeps disconnecting me 19:57:56 hm 19:58:25 air: You won't have that problem if you visit #hackers :) (DONT GO IN THERE! K-LINE CHANNEL) 19:58:39 lol 19:59:33 TheStar: who do u think set the onjoin message for this channel? 19:59:33 TheStar: btw, they now warn u before u get the kline 20:01:00 Cool - I wasn't warned last time (a while ago) 20:01:23 Oh, and it was probably you! :) 20:02:44 --- quit: CLxyz (Read error: 60 (Operation timed out)) 20:04:31 i have not told anyone to go to those channels, geist did tho 20:04:39 --- join: xeon (~xeon@cs6668188-147.austin.rr.com) joined #osdev 20:05:07 ehy all 20:10:49 --- join: air- (~brand@12-210-162-121.client.attbi.com) joined #osdev 20:10:50 --- quit: air (Read error: 104 (Connection reset by peer)) 20:11:53 * air- is gonna kick himself if he doesn't stop disconnecting 20:15:04 * TheStar watches air kick himself... 20:20:58 --- join: xeon` (~xeon@cs6668188-147.austin.rr.com) joined #osdev 20:21:28 lol 20:22:48 air: Is the user going to be allowed to use device driver functions in "programs" ? 20:23:59 no 20:24:31 crush has 4 different levels that it assigns to functions 20:24:51 each level can call functions in its level and the level above it 20:25:35 Why only 4? 20:25:39 crush, system, lib and user 20:25:57 crush is crush coded crush code? 20:26:13 drivers would be in the system level and u need a layer (lib) between them and the user functions 20:27:08 the crush level is just a super system level 20:27:30 for defining functions that libs can call like ASM 20:27:40 --- quit: witten ("Client exiting") 20:27:43 er cant call 20:29:22 Interesting... 20:29:30 hmm 20:29:31 --- quit: xeon (Read error: 60 (Operation timed out)) 20:29:31 --- nick: xeon` -> xeon 20:29:34 Oh! I remember a question. 20:29:37 maybe i should use dual levels 20:29:43 in and out 20:30:01 What ever happened to fprem? 20:30:50 same thing that happens to 90% of the oss projects :) 20:31:03 Oh ok, it's still alive... 20:31:06 Thank God. 20:31:11 I thought it had died out... 20:31:17 well 20:31:38 I was beginning to get worried since no changes to the CVS have been made in months 20:31:55 Let's just say that 90% of oss projects stay "just barely alive"... 20:32:08 Almost a year actually 20:32:59 if the project has space on sourceforge it will live forever, or until sf has a bad drive and no one backed up the project 20:41:44 geist: I fixed the motherboard!!!!!!!!!1 20:42:27 Soldering iron? 20:43:10 the bios 20:43:24 --- quit: keller` ("gtg") 20:55:45 ?!? 20:55:48 How does this work: 20:55:48 static insn_def_t insn_1byte[256] = { 20:55:48 [0x3C] /* CMP AL, Ib */ 20:55:48 { { INSN_CMP_R8L_IB, INSN_CMP_R8L_IB }, 20:55:49 { INSNOP_IB, INSNOP_IB }, 20:55:50 { NDX_AL, NDX_AL } }, 20:56:03 How do they put that [0x3C] in there? 21:01:31 huh? 21:04:46 THat's what the source says 21:04:55 what does the [0x3C] do? 21:04:55 I 21:05:03 I've never seen anything like that 21:05:28 i have accessors for normal reading/writing objects (set a 1) and (set b a) and a second kind of accessors for doing other stuff like (my-array index) which returns the array element at index 21:05:40 what do i call the second type of accessors 21:06:37 element perhaps? 21:06:44 (element index) 21:06:50 and no, parameterized accessor isnt correct because u call em with no parameters, (my-array) could return element 0 or something 21:07:28 no, my-array is a variable 21:07:49 im just looking for a name to refer to that type of accessors, nothing u would type in yer code 21:08:12 --- join: trans (wdcmfp@fatwire-201-205.uniserve.ca) joined #osdev 21:08:57 air 21:09:34 13. Compounds = Functions on Indices 21:09:38 > ("hello" 2) 21:09:53 compound accessors? 21:34:28 * EtherNet is away: sleeping 21:46:16 AH 21:46:39 new crush docs are online 21:52:13 --- quit: trans (Read error: 110 (Connection timed out)) 21:52:50 --- quit: air- (Read error: 113 (No route to host)) 21:52:50 --- join: air-- (~brand@12-210-162-121.client.attbi.com) joined #osdev 21:53:30 can anyone tell me 21:53:35 the new hyperthreading intel processors 21:53:41 do they count as SMP or UP? 21:53:54 because they seem to be capable of SMP, but there's only one chip, so....? 21:54:41 SMP 21:54:47 ok 21:54:52 so which linux kernel would be used? 21:56:13 smp 21:56:28 hrm, ok 21:56:37 is it really smp tho? 21:56:59 do u program it as smp or as ht? 21:57:00 that's what I've been thinking.... 21:57:03 ht? 21:57:04 as far as the s/w is concerned, yes 21:57:09 HyperThreading 21:57:21 hehe... and do SBU's still work? :p 21:57:47 green_: u sure? 21:57:58 pretty sure 21:58:00 (standard bash units, from linux from scratch. Defined as the time it takes to compile the bash shell; anything's compile time can be measured in SBU's) 21:58:38 (and SBU's don't work right on SMP processors) 21:58:41 green_: u've looked at specs? 21:58:56 no 21:59:01 actually, there's a section on hyperthreading in the intel developer manual 21:59:04 I'll look latyer 21:59:06 *later 21:59:26 i'm seeing ht support in 2.4.17 21:59:39 !!! 21:59:52 was hyper threading even around 2.4.17 was released???? 22:00:14 i dont think so 22:00:32 i think they added support before it came out 22:00:34 Air: I think you should reimplement a Crush as a Scheme derivative 22:00:38 :P 22:00:40 i know its probably not possible to write code that doesnt stall the pipeline but im sure some code does a pretty good job at not stalling it and if u treated it as smp then the process running on the second core wouldnt get much execution time 22:01:01 This is from a ht machine... 22:01:24 damn copy and paste no work. hang on. 22:01:25 xeon: scheme uses cons, crush doesnt and that makes it faster 22:01:48 cons? 22:02:03 u know how C allocates local variables on the stack? 22:02:14 Yea 22:02:18 Linux build.tokyo.redhat.com 2.4.18-14smp #1 SMP Wed 4 12:34:47 EDT 2002 i686 i686 i386 GNU/Linux 22:02:23 lisp allocates cons cells from the heap for ALL variables 22:02:37 so.. that's a 2.4.18 smp kernel on a dual xeon 22:02:48 green_: You called? 22:02:56 no :-) 22:03:17 cat /proc/cpuinfo lists 4 processors. 22:03:22 a cons cell has two pointers car and cdr, one points to the object, the other is kinda like a list next pointer 22:04:05 yeah 22:04:07 I see. 22:04:12 Ok, I gotta go. 22:04:21 I guess I'll see you all on the flip side... 22:04:51 --- quit: xeon (Read error: 104 (Connection reset by peer)) 22:04:53 is HT just like a super pipeline or what 22:05:31 how much of the chip is actually duplicated on die? 22:05:35 ht puts two register sets on one die but the registers share the same execution units 22:05:53 tahts it? 22:06:17 and it adds a more complex instruction scheduler that interleaves the processes thru the pipeline 22:06:35 --- join: Hutch (charles@on-tor-blr-a58-02-846.look.ca) joined #osdev 22:06:58 when it detects that thread 1 will stall it starts filling the pipeline with thread 2 22:08:03 oh so this is getting into out of order execution 22:08:18 branch prediction is only going to be in the 64bit processor irhgt? 22:10:15 huh? 22:10:26 all processors have branch prediction 22:10:38 cept for really old ones 22:11:35 --- join: sayke (~abuse@sttldslgw29poolF193.sttl.uswest.net) joined #osdev 22:16:36 i thought they were doing something differently though 22:21:05 some chips let you give prediction hints, others do static prediction. 22:21:29 mips64 has a branch-likely instruction, for instance. 22:21:43 and there are hints bits for p4. 22:38:16 --- quit: Perdix () 22:44:00 --- quit: bono ("Lost terminal") 22:45:43 --- quit: air-- (Read error: 104 (Connection reset by peer)) 22:47:49 --- join: air-- (~brand@12-210-162-121.client.attbi.com) joined #osdev 22:55:40 --- join: Xerroz (~a@66-214-121-247.hes-cres.charterpipeline.net) joined #osdev 23:02:54 --- quit: Hutch (Read error: 104 (Connection reset by peer)) 23:03:54 --- join: Hutch (charles@on-tor-blr-a58-02-846.look.ca) joined #osdev 23:09:59 --- join: trans (shzlea@fatwire-201-205.uniserve.ca) joined #osdev 23:14:14 --- join: elvstone (~elvis@as3-2-3.sgp.lk.bonet.se) joined #osdev 23:19:07 does anyone think Paging is essintial to an operating system? 23:19:19 yes 23:19:28 because the only other option is segmentation 23:19:33 oh wait, no 23:19:37 a modern OS, yes 23:19:39 flat model... 23:20:23 i was just reading in my 386 DX manual about paging, i think its essential, however i dont fully understand it yet. i suppose i need to read more 23:20:36 erm... osdev.neopages.net has a few good tutorials 23:21:53 and I suggest finding a newer manual 23:22:22 i have the pentium 4 manuals, but i like the older ones better 23:22:36 insured compatiblity with older processors 23:43:15 --- join: CLxyz (CLxyz@00-01-03-d2-a4-07.bconnected.net) joined #osdev 23:59:59 --- log: ended osdev/03.05.17