00:00:00 --- log: started osdev/02.11.05 00:00:11 since i have a simple goal: single purpose OS/applications tightly bound 00:00:14 my job is simpler 00:00:18 yep 00:00:22 kind of like the danger os 00:00:30 strategies like giving each process a predetermined block of memory are feasible 00:00:31 it's as minimal as possible to run the java runtime 00:00:39 and the rest of the code is jav 00:00:40 i only have kernel threads 00:00:52 and right now all memory is shared and unprotected 00:00:56 since most of the real code is inside java, memory protection isn't a big deal 00:01:04 which is good because our ARM has no mmu 00:01:16 ic - why java? with such a small CPU, why not do it all in C? 00:01:23 because we have no mmu 00:01:26 ic 00:01:30 any bug would destabilize the system 00:01:36 and it's pretty fast, you'd be suprised 00:01:44 my method for that is to auto-generate all application code 00:01:54 that reduces bug counts and memeory issues a lot 00:01:56 and furthermore bytecode tends to be really small 00:02:01 ic 00:02:08 especially against RISC machine's native code 00:02:35 auto-generated code as lots of benefits - all my projects use this method now 00:02:38 one of the big reasons bytecode was all the rage 20 years ago or so in personal computers 00:02:44 (BASIC, forth, pcode, etc) 00:02:55 application code shouldnt be hand written, only the core system should be hand written 00:02:57 and they had to deal with stuff running all over the place 00:03:02 ic 00:03:47 what i do now is define applications in XML 00:03:52 still not gone to bed?!? 00:03:54 GO!!! 00:04:04 rico: soon, i promise 00:04:07 * geist thinks about that 00:04:15 * geist will dot hat 00:04:22 it's been a nice chat 00:04:30 no! 00:04:32 didn't get anything done tonight, but it was a nice chat 00:04:32 basically, i assume (1) all applications use a finite set a functionality and (2) access the same dataset 00:04:33 I didn't mean that 00:04:37 I want to chat now 00:04:40 please? 00:04:44 huh? 00:05:12 well, in my system there are only a finite set of things : buttons, tree views, lists, etcs 00:05:21 and only a finite set of data types and collections 00:05:28 so i define all the data screens in XML 00:05:34 and all the functional interfaces in XML 00:05:42 the core engine is the only code 00:05:54 it just parses the XML and executes the app 00:06:03 this reduces development time and bug count 00:06:16 because application developers basically are doing the same thing over and over 00:06:25 and deviance from the correct method is a bug 00:06:25 data driven layout 00:06:38 so i just have the core execute the XML 00:07:02 sort of similar to the byte-code idea, but theres no programming involved 00:07:13 just visual layout and data queries 00:07:39 well, I'm gonna head to bed 00:07:43 or at least sign off 00:07:44 ok, good night 00:07:50 good talking to ya 00:07:59 yeah 00:08:20 bye geist 00:10:06 night geist 00:14:14 --- quit: kemu (Connection timed out) 01:07:28 --- quit: adu () 01:08:02 --- join: adu (~andrew@dsl-64-130-166-225.telocity.com) joined #osdev 01:19:51 --- quit: pengo (Read error: 60 (Operation timed out)) 01:25:28 --- quit: trans (Read error: 110 (Connection timed out)) 01:26:45 hola? 01:26:48 anyone awake? 01:39:31 o 01:39:42 sorry 01:39:59 hi 01:43:07 what are you up to? 01:43:47 this is going to be one of those funny conversations with long turn-around time 01:44:10 high latency introduced by the time it takes to flip back and check this window :) 01:48:03 not much 01:48:18 im trying to fix some broken makefiles 01:48:20 joy :) 01:48:23 yey 01:48:30 i get really mad when makefiles dont work right 01:48:35 especially the depends 01:48:36 i made sure to keep the window visible this time :) 01:48:39 ;) 01:48:53 i try to keep makefiles small 01:48:56 i want it so if you touch any file, it knows exactly what to build and nothing extra 01:49:01 i like small projects too 01:49:03 yah 01:49:24 if you have 100 files, you probably have at least 5 projects 01:49:30 so i have this fancy auto-dep maker that makes the dependencies - but its sort of acting weird 01:49:38 well, i use a lot of directories 01:49:42 and make lots of libraries 01:49:47 thats good 01:50:01 then i make the top level make generate all the sub-makefiles 01:50:06 for consistency 01:50:24 i'm lame 01:50:25 thank god for gmake, though - regular make wouldnt do all the shit i want 01:50:26 ? 01:50:28 lame how? 01:50:32 i don't use make 01:50:38 ic 01:50:42 what do you use? 01:50:44 a script? 01:51:09 you just type gcc -o exe *.c 01:51:10 ? :) 01:51:28 i just write a shell script that does cc -c something.c; cc -c somethingelse.c; cc -o st something.o somethingelse.o; 01:51:36 ix 01:51:41 well, i get impatient 01:51:48 with make you can cut the build time down a lot 01:51:54 i know 01:52:09 i learned from this other guy a long time ago all the cool trick 01:52:14 i wouldn't use make for that tho 01:52:17 ever since then i was hooked 01:52:24 ? 01:52:29 i'd use either jam or perl's make 01:52:34 how come? 01:52:47 cool trick? 01:52:50 or tricks? 01:52:57 tricks 01:52:59 o ok 01:53:04 actually, i use only one or two cool tricks 01:53:15 the main one is using gcc -M 01:53:18 cuz there was this one make system made in perl and it was alot nicer to make makefile in 01:53:45 i think the coolest trick i know is gcc -S 01:53:51 -M parses a C file and outputs a list of the header files that it depends on 01:53:53 see the asm :) 01:53:56 yeah 01:54:01 i use that for debugging sometimes 01:54:04 not useful, but cool 01:54:08 the -M output goes to a sed script 01:54:12 that makes a Makefile rule 01:54:17 o 01:54:19 interesting 01:54:23 so i can touch any header files 01:54:29 and it knows exactly what to rebuild 01:54:34 hm 01:54:59 with that, if its working correctly, you never have to do a make clean to ensure its all correct 01:55:22 cool 01:55:26 but it requires gmake because gmake can construct files to include within itself 01:55:31 regular make doesnt really like that i think 01:55:31 you should look into jam, and that other one 01:55:33 just a sec 01:55:35 whats jam? 01:55:50 its the make system that apple's build tools use 01:55:50 * sleep- googles jam 01:55:52 ic 01:56:08 its like make but better somehow? is it open source? 01:57:11 not sure 01:57:20 i know theres a gnu one too 01:57:33 ic 01:57:37 is it java based? 01:57:40 cons! 01:57:43 thats it 01:57:52 cons? 01:57:53 http://www.gnu.org/directory/devel/cons.html 01:58:00 not java, perl 01:58:27 wow 01:58:31 it has all the features i like 01:58:32 thats cool 01:58:41 :) 01:58:46 i basically do most of those by extreme makefile hacking 01:58:49 glad i could help :) 01:59:04 the MD5 instead of timestamp seems so obvious now, but i never thought of it 01:59:45 hmmmm - the last release is may 2001 01:59:49 is it maintained anymore? 01:59:52 or just finished 01:59:56 dunno 02:00:20 it was in the back of my mind, i'm surprised i remember'd it was related to perl at all ... 02:00:38 hehe 02:00:49 its cool that its in perl -that would make it easier to alter 02:00:57 yup 02:00:58 and theres really no reason make has to be a c program 02:00:59 i like perl 02:01:09 ill check it out 02:01:10 but its pointers are just as hard to use as C's 02:01:21 i dont like developing in perl 02:01:26 why? 02:01:30 but its great for making tools 02:01:34 i do server stuff 02:01:36 ya 02:01:42 so i write everything in c++ 02:01:49 but perl is perfect for all the system glue 02:01:59 *nod* 02:02:08 ive had terrible experiences using perl for server tasks like web stuff 02:02:12 its a memory hog :) 02:02:47 if the person writing the code doesnt understand how and when perl allocates memory, its bound to screw up 02:03:11 but i cant fault it for what its good for: its the ultimate tool for system scripting 02:04:03 i think i wrote my first ever socket code in perl 02:04:10 i'm starting to think i should switch to lisp or scheme for my language 02:04:27 well, that probably depends on what you are making 02:04:45 i cant read lisp/scheme so i cant use it 02:04:53 cuz the stack and internal lists would be alot easier 02:04:57 my eyes cross when tring to read it 02:05:07 lol 02:05:11 oh, you mean the language you are implementing? 02:05:15 yes 02:05:24 yes, that would make life easier 02:05:48 if i made a language, im sort of embarassed to say, it would end up looking like java/c++ mixed together 02:05:52 not very pretty 02:05:54 * do[done] is back (gone 08:11:45) 02:05:55 --- nick: do[done] -> do 02:05:56 i could use car, cdr instead of d->val._head d->val._tail 02:06:00 ;) 02:06:09 hehe 02:06:09 im just totally fluent in c-like syntax 02:06:14 ever looked into ada? 02:06:17 nope 02:06:32 any language made by the government ... how can it be good? :) 02:06:34 its kinda like c++ and java mixed, but not really 02:06:40 ic 02:06:49 its very similar to the language used in Delphi 02:07:09 which is Object Pascal, and Ada is a superset of Pascal 02:07:32 i really dont know any weird languages like that - just things ive used for projects: c, c++, java, perl, python, etc 02:07:37 pascal a long time agpo 02:07:40 ic 02:07:52 and it wasn't made by the gvmt by the way, the contest that found the language was sponsored by the gvmt 02:07:57 ic 02:08:20 at least thats what i hear 02:08:23 even though c++ is ugly and broken, i still find it the easiest to develop in 02:08:40 heh 02:08:45 everything else feels wrong somehow 02:08:57 i understand 02:09:01 java a close second, but too limited 02:09:28 i wish i had free time to do more stuff like the os and language 02:09:34 sometimes i think most of what makes a language great is everyone using it, or having someone tell you how you won't be wasting your time by learning it 02:09:37 work is eating all my energy 02:09:42 i think thats why java cought on 02:09:46 hehe 02:09:59 java is pretty ok - ive written tons in java 02:10:01 not because it fit a need or was elegant, or fast, or cheap 02:10:08 the politic squabbling hurt it a lot 02:10:12 but because sun made people feel good about using it 02:10:49 im working on a project now to move my old companies codebase off of java 02:10:52 i think the fact that its still not std is killing it 02:11:02 its starting to piss off customers because of a lot of issues that we cant fix 02:11:04 heh 02:11:15 the client-side java is really hard to make work 02:11:21 we have to ditch that soon 02:11:43 M$ is kicking our ass on that one - they make sure its broken :) 02:11:56 --- nick: geist -> geist-sleep 02:12:07 how far a long is you lang? 02:12:20 M$ has C#, why would they want to provide Java? 02:12:37 C# actually seems cool, what little i know about it 02:12:46 sleep: in actuallity, i can push numbers, push variable names, add, show stack, and pop stack. 02:12:51 but its no good for me - i have to use linux based stuff and mono isnt there yet 02:13:01 ic 02:13:07 ive never made an interpreter 02:13:13 but ive fiddled with making a compiler 02:13:21 so i made a symbol table and stuff like that 02:13:38 C# is very cool, it has interfaces and classes, a distinction i never realized until i learned Objective-C 02:13:50 yeah 02:14:08 i like C++, because you can make the equivalent of interfaces, that also have implementation of some of the functions 02:14:18 that hybrid type class is impossible really in java 02:14:19 and its VM (CLI/CLR) can execute proper tail calls 02:14:48 --- join: trans (~trans@fatwire-202-64.uniserve.ca) joined #osdev 02:14:52 really? 02:14:53 once C# works well on linux, ill probably try it out 02:15:01 really what? 02:15:04 don't you mean an abstract class? 02:15:09 well 02:15:18 an interface in java 02:15:27 an abstract class would be a class with some functions, interfaces just have method signatures 02:15:31 right 02:15:37 but sometimes you want both 02:15:49 you can do that in C++? 02:15:54 an interface, but the boring helper functions that all of them need can be put there to be shared 02:16:29 my type system is still in limbo, not sure how to do it 02:16:32 in Java, since you (1) cannot put any code in an interface and (2) can only inherit from a single class 02:16:44 but i know it will make the distinction between interfaces and classes 02:16:45 you cant do what i want 02:16:51 i can't? 02:17:07 in C++ i do it with abstact classes and multiple-inheritence 02:17:17 ya 02:17:21 that works 02:17:22 its a minor complaint, but i seem to run into that pattern a lot 02:17:41 basically, i use multiple in heritence in C++ *only* to simulate interfaces 02:17:49 the thing i never understood about multiple-inheritance, is what it does when you have functions with the same signature 02:17:54 theres only a single "real" parent class 02:18:04 i know - thats creepy to me 02:18:10 thats why i have strict rules 02:18:14 you don't know? 02:18:21 i always have to look shit like that up 02:18:30 but I *never* write code that requires that shit to work correctly 02:18:46 you mean if it comes across two of the same kind, it could conceivably eat your hard drive? 02:18:49 AHHHHHHH 02:18:58 hehe 02:19:06 :) 02:19:23 i dunno - but i figure, if i cant remember how its supposed to work, probably GCC doesnt do it right anyway 02:19:24 :) 02:19:29 that would be funny to read in the ISO C standard 02:19:57 i really love templates too, a lot of people hate them, but i love templates 02:20:14 "Conforming implementations of the programming language C SHOULD eat hard drives in the event that two identical method signatures are encountered." 02:20:33 hehe 02:20:38 i like them too, what i know of them 02:20:48 no, i think it should call each one round-robin 02:20:53 it would make my + function alot easier 02:21:06 hehe 02:21:55 i get creeped on some of the c++ weirdnesses - but with simple rules to follow things turn out ok 02:22:08 i hate exceptions in c++ and never use them 02:22:15 i use my own fake exceptions 02:22:22 they suck in java too 02:22:34 i never understood why they don't call them "errors" 02:22:35 without 'finally' like in python, i think exceptions are useless 02:22:38 haha 02:22:41 i dunno 02:23:03 and whats the difference between an exception and a continuation and a goto label 02:23:30 i use goto to emulate exceptions rather than use the real thing 02:23:41 good 02:23:46 that might sound lame, but it works ok and im not creeped out by it 02:23:51 the main thing is, finally 02:23:51 use the floor, not the carpet :) 02:23:54 :) 02:24:22 i like the idea of continuations in scheme 02:24:33 i think my macros could be re-written to use real exceptions, but until i understand how its implemented - now way 02:24:35 what that? 02:24:40 i dont know about continuations 02:24:42 scheme is actually a very nice language if you look deep into the semantics 02:25:07 ok i'll do some c code... 02:25:41 omg, my dog is humping the other dog - but their both female 02:25:55 int a,b,c; 02:25:55 funcall(); 02:25:55 cont: a = b+c; 02:26:13 "cont:" would be funcall's continuation 02:26:18 ok 02:26:30 basically where to go when somethings over 02:26:34 right 02:26:42 do you know what a tail call is? 02:27:00 but how is that different than placing code after the function? does that make it so no exception can jump past cont? 02:27:05 no, i dont know what a tail call is 02:27:57 according to the scheme spec its any expr that comes at the end of a body 02:28:06 ok 02:28:11 so the last thing in a code block? 02:28:19 yup 02:28:27 so whats special about a continuation? 02:28:47 why is that location labelled and what affect does it have? 02:29:35 when your doing recursive functions tho, you don't want to do all that yucky stuff when your done (ie exit 15 of the same function) so what you can do is (call-with-current-continuation name-of-fn) 02:29:52 they're usually not labeled 02:30:03 oh - cool 02:30:10 you can just move that stack back a bunch 02:30:15 to go back to where you were? 02:30:17 call-with-current-continuation is usually refered to as call/cc 02:30:36 and not to return, return, return, return? 02:31:02 i don't really know, i don't know that 02:31:11 continuations are first class objects in scheme 02:31:14 ok 02:31:30 it basically reduces some of the overhead of recursion? 02:31:32 (call/cc f) calls f with one arg, a continuation 02:31:34 that makes sense 02:31:57 so f would haveto take an argument like f(ret) = {.......... ret) 02:32:00 ack 02:32:03 bad punct 02:32:32 right, i sort of get it 02:32:32 you get the point, a continuation is a callable object passed to a function 02:32:44 see - heres my problem 02:32:49 ok 02:32:52 i leaned c and assembly 02:32:54 using DOS 02:32:55 right 02:33:01 before i took any classes 02:33:11 and if i cant understand the assembly 02:33:17 that would be produced by the code im looking at 02:33:24 i get scared 02:33:27 o 02:33:36 i need to grok the mapping from high to low levels 02:33:41 or i get worried :09 02:33:42 ok 02:33:58 in C++ most of the features seem clearly mapped to the machine 02:33:58 scheme is interpreted 02:34:07 so i cant "get it" 02:34:11 heh 02:34:12 s/cant/can/ 02:34:26 um i dunno 02:34:29 things like what you just showed me - take me time to understand :) 02:34:42 i think its because im not a theory guy - im a system level guy 02:34:49 if call/cc were to exist in c 02:35:08 so i was just thinking - it might alter the stack pointer 02:35:19 and jump to a label at the top where the function was first called 02:35:20 it would be like callcc(fn, cont); 02:35:26 when you hit the RET at the end 02:35:51 if i understand what happens to the stack pointer and the instruction pointer, i can totally understand 02:36:26 and fn would be a pointer to a function void fn(void *c) {... goto c;} 02:36:59 right - i think i understand 02:37:19 continuations are almost always used in association with a recursive function, correct? 02:37:48 and callcc(fn, cont); was used in another function fb, where fb(); cont: ... is where cont is 02:37:56 yes 02:38:04 ic - i think i totally get it now 02:38:31 how would YOU describe it in lamen's terms? 02:39:08 cool - it says in this page i just found that you can use continuations 02:39:21 huh? 02:39:23 to resume where an exception bailed - after fixing the problem 02:39:23 in what? 02:39:35 in scheme? 02:39:37 down you can restart where you left off 02:39:45 http://www.cs.utexas.edu/users/wilson/schintro/schintro_141.html 02:39:48 but this page 02:39:54 shows that i still dont get it 02:40:09 thanks for telling me about this - im going to have to read up on this 02:40:22 apparently theres alot to understand about continuations 02:40:42 if only i could read scheme.... 02:40:46 yah :) 02:40:54 heh 02:41:00 for now, ill just stick with my old fashioned exception macros ;) 02:41:33 heh ok :) 02:42:15 the main thing it seems to be, without getting into any of the details, is a way to jump around across functions and alter the stack pointer in various ways in order to (1) avoid useless overhead introduced by recursion, and (2) handle error cases in a more elegant fashion 02:42:50 thats the type of stuff thats natural in assembly, but difficult in c/c++ because of the atrifical structure introduced by the language syntax 02:43:38 like in C++ you cant jump to a label in another function, but thats fine in assembly 02:45:06 how long have you been writing code? 02:45:38 ever since i was 9 02:45:54 ic - i cant remember whenm, but i started around 5th grade with the atari800 02:46:05 heh 02:46:10 i started in QBasic 02:46:12 i didnt really get into until ago 18 or 19 with the PC and turbo c 02:46:23 c+asm is what really got me interested 02:46:31 i loved writing VGA hardware driver stuff 02:46:41 it was the first time programming felt real 02:46:42 yeah, when i learned ppc asm my eyes lit up 02:46:51 (gradually of course) 02:46:54 hehe 02:47:17 i remember trying to write a C program with no #includes 02:47:25 i disasm'ed pretty much all of MacOSX just after a month of owning the Beta release 02:47:26 because i wanted to do everything with my own code 02:47:31 thats how i discovered assembly 02:47:41 disasm'ing? 02:48:03 C program with no includes? 02:48:07 yeah 02:48:21 i wanted to write a program that was all my own code 02:48:28 heh 02:48:30 i didnt like the idea of including something else 02:48:37 so i followed the h eader file 02:48:47 thats how i realized libc was there 02:48:53 the smallest include i've got was the stdlib with syscalls 02:48:58 the i found the libc source code (borland) 02:49:15 that included the assmebly source, so i started learning asm 02:49:20 and found a BIOS/DOS reference 02:49:27 wow 02:49:37 i cant figure out looking back on it why the fuck i did that 02:49:46 did what? 02:49:48 but i became fixated on learning it all 02:49:56 why not make a useful program 02:50:02 instead of re-writing libc crap :) 02:50:06 hehe 02:50:08 so i wrote some BIOS code 02:50:09 learning 02:50:13 then I realized the BIOS was actually code 02:50:16 just like libc 02:50:20 and validating that it was done efficiently 02:50:23 so i found a VGA reference 02:50:36 and i wrote directly to the VGA registers 02:50:53 thats when i finally felt like it was really all my own code 02:51:12 guess i was destined to be here in the osdev channel - a freak who wants to own all the code 02:51:24 LOL 02:51:38 ;) 02:51:50 after that i spent a lot of time writing VGA graphics code 02:51:52 that was fun 02:51:59 cool 02:52:07 i got really pissed at the DOS 64KB segment limits 02:52:27 did you find out about unix? 02:52:56 well, i had heard about it 02:53:08 then i started taking a programming class at the university near my house 02:53:12 and i got a unix account 02:53:25 i downloaded linux (SLS) distribution 02:53:29 yey! unix enters the picture :) 02:53:30 it was 40MB 02:53:38 it took a long long time 02:53:43 hehe 02:53:43 on my 14.4 modem 02:53:52 after i got all the floppies, i tried to install it 02:54:16 but i didnt understand the partition numbering started from 0, not 1 02:54:21 so i nuked my dos partition 02:54:27 omg 02:54:39 after that, i found out that linux was more fun anyway 02:54:42 so i stuck with it 02:54:46 hehe 02:54:59 the kernel was at version 0.9.x at the time 02:55:03 wow 02:55:12 and SLS was the best distro 02:55:16 I think thats what slackware copied 02:55:31 but the C compiler worked - thats all i cared about - and i got the modem working too 02:55:37 i loved X 02:55:43 so i was hooked from then on 02:56:05 i loved the feeling of having an actual password prompt for some reason - it felt more "real" than DOS 02:56:21 yeah 02:56:32 should i make my one-liners macros? 02:56:54 like: 02:56:55 bool dataIsa(data_t d, id_t type) { 02:56:56 return !strcmp(d->_type, type); 02:56:56 } 02:58:20 ? 02:58:47 i dont undestand exactly what you are asking 02:59:18 should i use #defines rather than functions... 02:59:27 umm - the compiler is real good at inlining 02:59:40 so i think the performance gain is minimal 02:59:50 and functions are safer and more readable i think 03:00:11 if youre just doing it for performance, dont bother with the macros 03:00:18 the compiler will inline often 03:00:23 i'm not using gcc tho 03:00:24 or is there some other reason? 03:00:29 what compiler? 03:00:33 apple's cc is "based" on gcc 03:00:43 ic - well, i have no idea about that 03:00:49 but i'm not sure what its inline abilities are 03:01:13 still, these days i only worry about really high-level performance issues 03:01:22 like the time complexity of the algorithm 03:01:34 ya 03:01:45 there usually other areas - structural - that improve a lot more than worrying about details like that 03:02:01 just make it easy to convert to a macro later if you decide its worth it 03:02:03 ok 03:02:18 thanx 03:02:38 and if you work with other developers, look through their code for all those bubble sorts and linear searches 03:02:52 bubble sorts? 03:03:13 usually, the performance issues are introduced by some developer who does something really silly 03:03:37 like i had one project where a guy did this: 03:03:52 int get_value_from_hashtable(hash h, key k) 03:03:52 { 03:04:06 for (v in get_keys(h)) 03:04:08 { 03:04:15 if (v == k) 03:04:17 { 03:04:23 return h->get(v); 03:04:26 } 03:04:29 } 03:04:53 does that make sense? 03:05:41 no 03:05:49 is that java? 03:05:54 it was actually in c 03:06:05 basically, a hash table lets you get a value by key 03:06:15 i know what a hash table is 03:06:28 instead of getting the key immediately, he got the list of keys 03:06:38 and iterated over the whole has until he found the key 03:06:46 and then returned the value by getting it 03:06:55 i have NO IDEA what he was thinking 03:07:12 it was in a core piece of code that slowed the whole product down to a crawl 03:07:21 and v and k are the same thing? hehe 03:07:24 yeah :) 03:07:28 thats funny 03:07:37 yeah, but some people get really confused about data structures 03:07:40 and do shit like that 03:07:48 in large projects, it becomes a nightmare 03:07:53 thats why i like small teams 03:07:57 4-5 people at most 03:07:58 ya i can imagine 03:08:26 anyways, a bubble sort is just a crappy sort method with really bad time complexity 03:08:41 but people do versions of that all the time - like for insertion into lists 03:08:43 worse than qsort? 03:08:49 qsort is good 03:09:42 thats why its built in to most standard toolkits 03:10:25 i've encountered qsort in so many places 03:10:39 its like pizza, its everywhere 03:10:49 yah 03:11:01 but i was pissed that java had no built in sort for the first few versions 03:11:08 that seemed like an oversight to me 03:11:08 do you know anything about tensors? 03:11:11 nope 03:11:15 n/m 03:11:15 i forgot what those are 03:11:42 what is 'tensor theory'? that rings a bell 03:11:47 i thought i knew 03:11:58 they're used in differential geometry 03:12:11 i think they're a collection of points 03:12:17 ic - like attractors 03:12:19 or something 03:12:29 dunno 03:12:35 i think i encountered discussion of that when we were working on clustering algorithms 03:12:43 for clustering data points 03:12:51 i've read lots of equations with tensors in them, but not much about what a tensor IS 03:12:56 i liked working on data mining software, that was a lot of fund 03:13:12 fund? 03:13:15 or fun? 03:13:17 meant 'fun' 03:13:20 ;) 03:13:31 hehe, were you paid a lot for that project? 03:13:35 :) 03:13:37 not really 03:13:42 but i learned a lot 03:13:46 thats cool 03:13:50 we did credit card fraud detection 03:13:53 that was a fun project 03:14:13 we had a system that scanned old credit transaction data and learned what types of transactions were stolen cards 03:14:23 then we scanned new data to find suspected stolen cards 03:14:53 it was fun because we had to make the system very scalable - it had to process 15K transactions a second 03:15:23 we coudlnt even afford to buy a machine like the customer had, so we had to just hope it would work on 32 processors 03:16:19 so how did you end up making it learn? 03:16:28 we used classification and regression trees 03:16:43 brb 03:16:50 a learning method that gererates tree models by recursive entropy based splits 03:16:50 k 03:19:50 recrusive entropy? 03:20:06 well 03:20:11 i dont pretend to undestand it all 03:20:13 but basically 03:20:17 you sort all the data 03:20:21 on each attribute 03:20:32 travel accross each sorted array 03:20:37 and try to find the single cut 03:20:46 that maximizes the "difference" between both sides 03:20:52 using some kind of entropy calculation 03:20:58 after the first cut 03:21:03 oh ok 03:21:05 you recursively do this to each subset 03:21:13 until you get to some place to stop 03:21:19 then you usually prune back a bit 03:21:41 this results in really expensive off-line model generation, but really fast run-time scoring of incoming data 03:21:57 because you just run the object down the tree until you hit a leaf - bingo thats the score 03:22:15 we made a system that used hundreds of trees in parallel for better results 03:22:34 later we found out that was called "boosting and bagging" 03:22:46 heh cool 03:22:48 i didnt undestand the math, but the over all algorithm was really cool 03:23:05 and it actually worked - it worked about as well as their existing neural net based system 03:23:14 wow 03:23:34 but it had the advantage that the trees were readable - you could see "why" it scored the transaction as posssibly stolen 03:23:44 have you ever made a genetic algorithm? 03:23:54 yes - we did a little bit of that too 03:23:59 but not much 03:23:59 cool 03:24:09 the GA was to pick some parameters for the system 03:24:15 but it never got put into production 03:24:29 o well 03:24:32 GAs seem good for determinig parameters for other algorithms 03:24:44 i think GA's should be builtin like qsort 03:24:47 hehe 03:25:03 GAs are underrated - they are far more useful than people realize 03:25:06 just pass a pointer to the judgeEffectiveness() function 03:25:09 hehe :) 03:25:13 yup 03:25:19 er 03:25:25 have you read Koza's books? 03:25:36 pass the judgeEffectiveness() function to GA() 03:25:41 oops 03:25:46 Koza? 03:25:46 no 03:25:54 hes the main GA researcher 03:26:02 ic 03:26:06 i saw him once at the game developers conference 03:26:11 i learned about them on alife.org 03:26:14 one of his students has a game AI company 03:26:16 yeah 03:26:26 i learned about GAs at ARtificial life 3 03:26:31 in sante fe 03:26:39 a conference? 03:26:41 yeah 03:27:04 there were some really cool people there - one dude modelling ant colonies that were amazing 03:27:18 another guy had fractal plants that grew 03:27:24 all kinds of interesting stuff 03:27:27 cool 03:27:38 yeah - it was sort of a mis-matched group - but all interesting 03:27:42 fractal plants? 03:28:02 --- join: Javanx (~javanx@213.45.18.12) joined #osdev 03:28:06 yeah, you could tweak parameters to make ferns, or different things 03:28:19 very artistic 03:28:36 wow 03:29:21 i have Steven Wolfram's new book on CA's 03:29:25 but i think the main machine learning methods havent improved much - at least not stuff that ive heard of 03:29:42 MARS was the last interesting thing i heard about and thats not even much better than anything else 03:29:44 --- join: SLACKo (~SLACKo@65.199.129.129) joined #osdev 03:30:01 i had a few ideas that might work 03:30:09 yeah? 03:30:13 so did ree, but he hasn't been here in awhile 03:30:21 i remember that guy - what was he working on? 03:30:27 he's been thinking of making a universe in a computer 03:30:29 hi folks,anyone know how to configure paging under linux like in solaris? 03:30:31 cool 03:30:42 slacko: what do you want to tweak? 03:30:47 paging? proly not 03:31:10 like to set MINFREE etc.. 03:31:23 so if it droped below that value 03:31:25 minfree is what? the minimum free pages in memory? 03:31:33 it'll page out 03:31:47 id look around in /proc 03:31:55 if its not there you probably cant tweak it 03:32:00 without reading the kernel source 03:32:27 ic 03:32:32 k thx 03:33:07 sleep: i've thought about using lists to express the building blocks of a thought 03:33:29 sleep: i've thought about using lex/yacc in conjunction with ALICE-bot 03:33:30 adu: you are treading into dangerous territory here :) 03:33:58 why? 03:34:04 adu: i believe such things are only possible through emergence - you would need to start at a lower level 03:34:14 so does ree 03:34:33 i just have this attitude that cognitive computation requires a fine granularity as the base 03:34:55 and modelling anything at the high level wont produce good results 03:34:59 just a hunch 03:35:04 yeah 03:35:26 but go on, tell me more 03:35:29 i also particularly like the base-4 system 03:35:42 (i think DNA/RNA is trying to give us a hint) 03:35:59 interesting... 03:36:10 also most of the mathmatical constants are less than 4 03:36:28 log 10 = 3..., pi = 3.14, e = 2.7... 03:37:01 ic 03:37:08 just when you said that i thought 03:37:15 that space and time are four dimensional 03:37:21 perhaps thats there reason 03:37:23 i don't really intend on improving AI tho 03:37:33 s/there/the/ 03:37:39 i like to concentrate on HCI / UI 03:37:50 ic 03:38:01 from a philosphical standpoint you mean 03:38:25 and the 4th spacial dimension has some interesting properties, toboot 03:38:43 no, just ease, and pritiness 03:38:45 my friend katherine 03:38:53 when she was really upset would cry on my shoulder 03:38:59 and say "thank god for linear time" 03:39:06 my ideal interface would be a single, large color touchscreen 03:39:07 "thank god this will pass" :) 03:39:19 lol 03:39:20 ok 03:39:37 what would this interface be like? 03:39:53 it would use my widget system 03:39:59 * do is away: at university 03:40:04 --- nick: do -> do[done] 03:40:33 which doesn't have any silliness 03:41:01 like tabs and option groups and popup menus and combo boxes all being different 03:41:04 thats just silly 03:41:43 how do you mean different? 03:42:24 i think it should be a system-wide choice as to whether you want your "one-of" choices to be in a popup list or in a always visible list 03:42:38 they're all "one-of" widgets 03:42:42 ic 03:43:00 all the disparate ways of viewing collections is silly and confusing 03:43:06 tabs just happen to have extra code that switches pannels 03:43:15 ic 03:43:27 but i like tabs for navigation 03:43:47 then you can set your system to have always-visible lists 03:43:47 its a good way to see the high level tree, while looking at a leaf 03:44:34 would you really compute differently if you had to choose your panel from a popup? 03:44:56 umm 03:45:13 not really i guess 03:45:17 YEY 03:45:31 but navigation is important - users should be able to see things that are available without searching menus 03:45:32 menus suck 03:45:40 i have a whole specification for my widget system somewhere 03:46:00 menus save valuable screen space 03:46:39 ic - but users have a hard time finding stuff in menus - i make all my apps tabbed views 03:46:48 so they can see whats available 03:46:56 i use a toolbar like outlook on the left 03:46:59 for major levels 03:47:05 and a tab collection across the top 03:47:09 for minir levels 03:47:12 minor 03:47:17 thats why it should be customizable then, so you can have your tabs 03:47:20 :) 03:47:38 hmm - id like to see your spec sometime 03:47:46 this is stuff i deal with all the time - usability of the gui 03:47:51 it used to be on my webpage 03:48:05 but apple started charging for iTools membership 03:48:10 *bastards* 03:49:15 how annoying 03:49:30 well - at least all my gui code is abstractly described 03:49:37 theres no gui *code* in my software 03:49:42 well for the love of 03:49:43 only XML-based descriptions 03:49:57 i also thought of alternative uses of dragging, and text selection, collumnar selection rather than linear selection, and scrolling by proximity to edge rather than requireing scroll bars 03:50:06 so changing it out to act the way you are decribing would just require altering the rendering engine 03:50:19 XUL 03:50:20 --- quit: file () 03:50:25 XUL? 03:50:48 XUL is mozilla's gui XML language 03:50:55 ic 03:51:08 ive never seen that, but mine is designed to be really simple 03:51:19 and it doesnt enforce much of anything 03:51:36 so the same files produce many outputs - theres a GTK version, an HTML version and a Java version 03:51:38 from the same source 03:52:05 i also use an XML description of the functional interface to the product 03:52:40 that produces the binding to the GUI, as well as the scripting API - like SOAP/XML-RPC - that interface is auto-generated from the XML functional interface description 03:53:16 the whole goal is to never write code for any of the applications except for functions that plug into the back-end that actually implement stuff 03:53:56 its great - it reduced our GUI java code to 1/10th the amount of lines of XML 03:54:53 whats XUL like? 03:57:18 --- join: file (proxy@mctn1-2922.nb.aliant.net) joined #osdev 03:58:12 re file 03:58:17 thanks 03:58:21 ADSL went down for a few minutes 03:58:24 and now I must leave 03:58:51 k 03:58:53 adios 03:59:43 --- quit: adu (Read error: 104 (Connection reset by peer)) 04:13:37 --- join: miro (~miro@picknicker.codeon.de) joined #osdev 04:15:48 Hey miro 04:16:04 --- quit: sleep- ("...time for work :(") 04:16:57 --- join: lynx (~root@pD9E63A0A.dip.t-dialin.net) joined #osdev 04:17:07 woo 04:17:24 Mister r00t! 04:17:50 dun use irc as root sum skriptkid can hack u w sub7!!! 04:19:01 --- join: _Bradd_ (~bbobak@d57-52-229.home.cgocable.net) joined #osdev 04:22:09 --- quit: trans (Read error: 60 (Operation timed out)) 04:24:49 --- quit: SLACKo ("[x]chat") 04:32:34 --- join: eks (~eks@h24-82-197-140.wp.shawcable.net) joined #osdev 04:32:46 konnichi wa, eks-san 04:33:15 konnichi wa, sampai 04:33:39 cornichon ! 04:33:42 watakushi wa sensei desu ;) 04:33:52 lol 04:34:43 * oink wants thinkgeek's sound bug 04:34:58 sampai = ? 04:35:13 oink : soundbug is afaik a german product 04:35:20 the assistant of a sensei 04:35:25 ah 04:35:26 ok 04:35:50 lynx: really ? 04:35:59 good news :) 04:36:24 * oink looks at computeruniverse.net 04:38:12 yay! 04:38:14 they've got it 04:38:24 (^_^) 04:38:35 48 EUR :o 04:39:39 cheaper at thinkgeek.... 04:39:52 how much is it there? 04:40:09 28, 24 or something 04:40:20 dont forget taxes/chipping you have at thinkgoat. 04:41:09 yeah, i'm not gonna buy there 04:41:59 look at geizhals first. 04:42:39 --- part: file left #osdev 04:42:48 that's what i'm doing 04:42:53 they're all at the same price 04:43:22 45e chapest 04:43:25 yes 04:43:26 sorry 04:43:56 http://www.0xfi.com/oslib/topx.html <-- good site for os info :) 04:49:13 the site is MEGA slow 04:52:40 omg 04:52:42 fn hos 04:52:56 --- join: do (~green@194.85.84.244) joined #osdev 05:00:28 fn orim 05:04:00 hrm 05:04:01 time 05:04:01 for 05:04:03 my 05:04:06 table turns 05:29:39 --- join: Zenton (~vicente@8.Red-80-34-35.pooles.rima-tde.net) joined #osdev 05:35:58 --- quit: _Bradd_ (Read error: 104 (Connection reset by peer)) 05:36:53 --- join: _Bradd_ (~bbobak@d57-52-229.home.cgocable.net) joined #osdev 05:41:06 --- join: eclips (~kemu@150.43-200-80.adsl.skynet.be) joined #osdev 05:41:26 --- nick: eclips -> kemu 05:56:44 --- quit: eks ("going to work") 06:00:16 --- join: trans (~trans@fatwire-202-64.uniserve.ca) joined #osdev 06:00:38 --- join: corsairk8 (~Stealthy@pc-62-31-1-150-cr.blueyonder.co.uk) joined #osdev 06:04:03 --- quit: lynx ("BitchX-1.0c19 -- just do it.") 06:13:43 --- join: revanthn (~revanthn@203.129.127.18) joined #osdev 06:15:38 --- quit: miro (Remote closed the connection) 06:22:44 --- join: brandtd (~brandtd@pool-08-129.pix.aea9.k12.ia.us) joined #osdev 06:23:15 --- join: lodda (~lothar@p508FE45B.dip.t-dialin.net) joined #osdev 06:23:17 hoi! 06:23:25 yo 06:24:06 question for you all: why, with bochs, is it taking forever to load even the grub menu? 06:24:40 brandtd:it's cuz your CPU is low speed 06:24:55 but it will eventually load, right? 06:25:20 cuz it loaded linux+lilo a lot faster than it's taking just grub to load 06:25:42 bochs is still sucking down cpu time, btw 06:43:30 --- quit: brandtd ("Client Exiting") 06:47:20 --- quit: revanthn (Success) 06:50:59 --- join: quantis (quantis@130.88.157.186) joined #osdev 06:51:04 ello 06:55:01 --- join: sleep- (~ivan@68.22.197.230) joined #osdev 06:55:25 lo sleep 07:01:20 --- quit: trans (Read error: 110 (Connection timed out)) 07:01:54 --- join: bono (~bono@host-ip20.8D.com) joined #osdev 07:02:10 --- quit: gab (Read error: 54 (Connection reset by peer)) 07:07:26 --- join: gab (~prfalken@gaia.chx-labs.org) joined #osdev 07:09:13 * mrd wonders if it's actually possible to kill someone with silence 07:10:57 * Robert shouts in mrd's ear. 07:11:23 * oink looks at Robert 07:11:48 Hey oink :) 07:12:00 :) 07:12:01 helo 07:14:39 * Robert pets his dear little pig from France. 07:15:35 --- quit: witten (Remote closed the connection) 07:15:35 --- join: witten_ (~witten@adsl-gte-la-216-86-199-140.mminternet.com) joined #osdev 07:21:07 L'Etat, c'est moi. 07:23:40 mon cul ouais 07:23:46 et la marmotte, elle met le chocolat dans le papier alu ? 07:24:02 eh...could you translate into an human lang? 07:24:09 ;) 07:25:26 into english...but is english=human? 07:25:44 depends on the english speaker 07:25:53 unfourtanately i don't speak any french... 07:27:10 only "je ne comprepa" or so and "L'Etat, c'est moi" and "fou c'est moi" 07:27:18 (or so) 07:27:55 * oink can't talk 07:27:59 * oink is eating fried chicken 07:28:03 xD 07:28:09 mmm 07:32:50 --- quit: quantis (Read error: 54 (Connection reset by peer)) 07:38:02 --- join: quantis (quantis@130.88.157.186) joined #osdev 07:39:58 --- quit: kemu () 07:43:48 --- join: eirikn (eirik@odin.eirikn.net) joined #osdev 07:47:12 morning all 07:50:01 Morning. 07:50:34 mOorning 07:51:29 g'morning 07:51:33 --- quit: malenfant (Read error: 104 (Connection reset by peer)) 07:54:53 what are you all up to? 07:55:01 * sleep- is fightin code-writers block 07:55:21 im wrting up some documentation for a new kind of compiler im writting :) 07:55:25 well hope to write 07:55:30 im crap at coding me thinks 07:55:59 hehe - why do you say that? :) 07:56:14 why so self-critical? 07:56:35 i have problems passing variables in c, took me two hours to work out :( 07:56:55 well pointers, havent done any proper coding for about 18 months :( 07:57:13 ic - itll come back 07:57:17 write some graphics code 07:57:25 thats a good way to become familiar with pointers real quick 07:57:32 :) 07:57:43 yeah ill be doing that soon hopefully 07:57:51 but ive got to get my maths up to scratch 07:58:12 whats this 'new kind of compiler'? 07:58:20 ive spent the last two years doing pretty much bugger all, and ive forgotten just about any skill that seems to be of practival use :( 07:59:08 im writting a system which instead of automatically outputing to a object format, you can actually embed instructions in it to build the object format 07:59:08 ;) 07:59:19 ? 07:59:23 explain further 08:00:25 look at the elf format, it contains a list of all the objects, its automatically produced by the compiler, with my sustem youd have a schema file, which says at byte to therell be a point to main, and also be able to buil tables 08:00:32 its a bit like tex for object formats 08:00:50 hmmm still not a very good explaniation 08:01:30 well 08:01:38 have you looked at linker scripts with GNU ld? 08:01:48 nope 08:01:50 ld lets you define the exact output format manually 08:02:02 the default is elf, but you can lay it out however you want 08:02:06 might have a look at that then 08:02:09 i use this to link my kernel 08:02:22 to set the boot sector to the corrrect place, etc. 08:02:34 everyones always thought of whatever i think of first :( 08:02:37 ;) 08:02:37 tisnt fair 08:02:40 i hate the world 08:02:42 :) 08:02:47 that was done to support other formats 08:03:04 its really at the linker stage, not the compiler stage 08:03:11 well there goes my backup 3rd year project down the toilet then 08:03:16 haha :) 08:03:27 sorry to disappoint you 08:03:37 this was combined the idea was to have it all output to assembler, then fasm would just assemble it straight 08:04:18 whats fasm? 08:04:48 im mean , i figure its an assembler, but who made it 08:04:49 ? 08:06:15 its the flat asmebler, its freeware, cant rember the homepage (i just go to google whenever it need a newe copy), its like a really cut down assembler, all it does it assemble the instructions as it reads them in the source file and not a lot else, hence its pretty flexable 08:06:29 sounds very riscy 08:06:35 ? 08:07:22 sllep: i dont think the ld scripts do quite what i was plaaning 08:08:15 --- join: amon-re (ddefreyne@D5E080CE.kabel.telenet.be) joined #osdev 08:15:49 my sys would generate a lot more metadata, like you could tag each fucntion in a different way, the idea was to have a ada/pascal type sysntax for describing the logic, cause i believe it to be clearer, and then to have all the logic code encapsulated in c/tex like statements that would generate the metadata 08:17:55 quantis 08:18:02 ello 08:18:07 me not understands sentences longer than 5 words 08:18:19 me 8 means 08:18:27 ? 08:18:36 lol, nm 08:18:53 sorry im pretty crap at writting and describing things 08:19:12 english is my first language as well 08:19:27 doh 08:19:30 --- join: dax_ (~you@u212-239-192-73.adsl.pi.be) joined #osdev 08:19:35 Heh 08:19:36 Hey dax 08:19:42 HEY MISTARH 08:19:48 Mister H? 08:20:04 Hidden? Heath? Homosexual? Hetarded? 08:20:16 --- nick: dax_ -> dax 08:20:16 hey dax 08:20:19 :) 08:20:43 homosexual would be the correct one for amon 08:21:04 moo would be the correct one for dax 08:21:12 yes. 08:21:27 Daxecca Heedron 08:21:27 :o 08:21:29 * dax is downloading a damned good CD 08:21:39 * amon-re buys his music 08:21:42 looks like everyones friends here then :) 08:21:49 dax: is your server up? 08:21:51 * Robert downloads ir. 08:21:53 it 08:21:53 no 08:22:00 (His music) 08:22:08 powerfailures every 2 secs... 08:22:17 that's bad 08:22:28 yea 08:23:44 dax, like 08:23:55 found a couple of nice new bands to listen too 08:23:56 to* 08:23:57 *** signoff 08:24:03 *** joined blabhlabhlah 08:24:12 that about 150 times in a minute? 08:24:19 well no 08:24:24 my pc doesn't boot that fast 08:24:34 Ah 08:25:03 THIS ALBUM IS GOODXORZ. 08:25:07 Hmm, I was just going to ask, "does your WHOLE COMPUTER crash when there's a power failure, not only your IRC app? 08:25:17 <- less intelligent person 08:25:17 (In The Eyes Of The Lord by 100 demons) 08:25:20 :o 08:25:45 Now that was funny 08:25:46 08:26:16 --- join: trans (~trans@fatwire-202-64.uniserve.ca) joined #osdev 08:27:47 got some music for our maths teacher amon-re 08:27:53 Abstrakt Algebra 08:27:54 :) 08:28:00 Boards of Canada - Music is Math? 08:28:08 Boards of Canada - a is to b as b is to c? 08:28:30 Boards of Canada - the smallest weird number? 08:28:36 no 08:28:40 ;( 08:28:43 all songs by Abstrakt Algebra 08:28:43 OGG'ed? 08:28:57 no... just large mp3s :( 08:29:22 wtf 08:29:28 got to make homework..bbl.. 08:29:29 their english is weird 08:29:42 ic 08:29:51 "Abstrakt" 08:29:54 You point your finger, i break it you are 08:29:55 lol 08:30:34 I wonder why they didn't call it "appstrakd allgheibra" 08:30:41 i should vote today, but im too lazy to go out 08:30:58 where you liveb? 08:31:19 mmm 08:31:21 --- quit: lodda ("Client Exiting") 08:31:27 GOOD MUSIC EVENING MB 08:31:34 (mother isn't home... heh) 08:31:39 neighbours are though 08:32:11 wonder if they mind deathmetal blasting with volume at 75%... 08:32:44 play it to them long enough and they might come to appricate its finer points 08:32:51 * sleep- loves death metal 08:33:07 * amon-re loves Aphex Twin 08:33:11 haven't played anything loud yet 08:33:22 i got a couple of nice "bass test" songs too 08:33:24 * quantis more of a nu-metaler 08:33:30 cryptopsy 08:33:35 come to daddy..... 08:33:46 * dax pets his B&W speakers 08:33:53 Black&White? 08:33:53 600W each :) 08:34:03 Bloody&W00tage? 08:34:04 www.bwspeakers.com <-- ph34r 08:34:58 Polynomial C owns 08:34:59 just admit it... bowers & wilkins ownz 08:35:32 i want a pair of Prestiges... Model Nautilus ones :) 08:37:43 Hahahah, "pinkfloid" instead of "pingflood", owned 08:38:49 hmm 08:38:56 those speakers aren't that expensive 08:39:05 only £17500 each 08:48:12 --- join: acme (acme@pD9E634F0.dip.t-dialin.net) joined #osdev 08:50:10 --- quit: amon-re ("kthxbye") 08:54:44 --- quit: quantis () 09:03:01 --- quit: sleep- ("Leaving") 09:03:56 HUMP. 09:08:29 --- join: stormbind_ (~stormbind@p50835F6C.dip.t-dialin.net) joined #osdev 09:08:43 --- quit: stormbind (Read error: 110 (Connection timed out)) 09:09:52 --- join: tirloni (gpt@aline.bs2.com.br) joined #osdev 09:12:55 hmm 09:13:05 i told you guys i was going to lay down osdev for a while 09:13:09 i don't think i can do that 09:13:31 the osdev-virus is attack me again... can't... resist... the... urge... to... finish... daxos 09:17:00 --- join: lynx (~root@p50808E69.dip.t-dialin.net) joined #osdev 09:17:18 --- quit: acme (Read error: 54 (Connection reset by peer)) 09:17:28 huhu 09:21:24 :)) 09:22:58 WOO 09:40:06 --- quit: bono ("Lost terminal") 09:43:06 --- quit: do ("bbl") 09:53:17 --- join: eks (~eks@h66-38-248-194.gtconnect.net) joined #osdev 09:53:23 --- quit: eks (Client Quit) 09:55:58 --- quit: trans (Read error: 110 (Connection timed out)) 09:57:22 --- join: HeavyJoost (~HeavyJoos@a213-84-139-110.adsl.xs4all.nl) joined #osdev 09:59:10 --- join: sberla84 (~ma@212.141.138.200) joined #osdev 09:59:50 --- quit: sberla84 (Client Quit) 10:10:00 nice speakers dax 10:10:03 ;) 10:11:59 * do[done] is back (gone 06:32:02) 10:12:02 --- nick: do[done] -> do 10:18:55 --- quit: do ("Quit") 10:22:44 --- join: do (~green@194.85.84.244) joined #osdev 10:37:42 http://slashdot.org/comments.pl?sid=44177&cid=4598605 10:43:04 --- join: Liesbeth (Simon@11.144-201-80.adsl.skynet.be) joined #osdev 10:56:22 Hoi 11:00:44 Hey 11:00:59 --- quit: eirikn ("Segmentation fault") 11:01:34 hey honey 11:05:17 --- join: lodda (~lothar@p508FD9D0.dip.t-dialin.net) joined #osdev 11:05:25 hallo 11:05:27 konban wa, lodda-san 11:05:59 what does san mean? i only know chan and kun 11:06:47 was ist chan und kun? 11:06:57 san heisst... herr oder fräulein 11:07:07 kann glaube ich auch lehrling sein 11:07:26 FRÄULEIN LODDA 11:07:37 lynx-chan heisst, glaube ich, in etwa lynx-freundin und lynx-kun in etwa lynx-freund oder so... 11:07:56 ah 11:07:56 ok 11:13:14 ok ok 11:13:21 ihr seid alle langweilig 11:14:00 Ja. 11:14:10 Aber du gefällst uns! 11:14:25 Bleib bei mir, lynx! 11:15:30 hrm 11:15:32 maybe later 11:15:38 would i be allowed to sleep in your bed? 11:15:49 * lynx goes to watch portishead live on DVD 11:16:29 :P 11:16:38 * Robert watches pothead live on IRC 11:17:13 BAH 11:17:22 talking aout dax´ new gf ? 11:17:49 new ? 11:18:13 i was his former one :( 11:18:21 cough new cough 11:18:36 been together with her for 1 year & 1 month already 11:19:32 poor boy 11:20:25 1 year and one month ago I had never been with a girl. 11:20:26 Liesbeth :( 11:20:29 (I still haven't FYI :P) 11:20:38 Robert : hrm... me neither 11:20:39 hehe 11:20:43 way to go kiddies 11:20:46 i think i am the oldest , how hrm... 11:20:49 "mature"... 11:20:54 :D 11:21:10 At least you're no stupid script kiddie :)) 11:21:13 *hint* keep away from women... better for your health *hint* 11:21:18 Robert :PPP 11:21:19 Hehe. 11:21:30 dax : wherefore? 11:22:01 uhm 11:22:03 he just wants to annoy me, isn't it spoiled brat! 11:22:22 well when your gf tries to bodyslam you a few times each week... 11:22:26 Liesbeth : there is always a place free for you @ me 11:22:27 * dax runs 11:22:36 * dax hits lynx 11:22:46 ... 11:22:50 *ouch* 11:22:56 where did you hit me? 11:23:08 heyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy that was because he was making fun of me! 11:23:16 he needed to be punished 11:23:19 a bit 11:23:20 hmm 11:23:22 beat him! 11:23:22 maybe 11:23:25 he 11:23:26 h 11:23:28 or.. 11:23:30 you don't need to tell her that 11:23:31 bite him! 11:23:32 --- join: trans (~trans@fatwire-202-64.uniserve.ca) joined #osdev 11:23:36 ugh 11:23:39 * lynx loves biting ppl 11:23:42 just not down there 11:23:48 where? 11:23:59 last time i was drunken i bit 3 ppl showing me their red bellies and shoulders in the mroning... 11:24:11 lol 11:24:13 ... 11:24:20 lynx: how old are you btw? 11:24:28 19, afair 11:24:44 omfg 11:24:47 that's just.... 11:24:50 ancient 11:25:00 stfu 11:25:05 LYNX YOU'RE ODL! 11:25:13 at least i have some percent of old-school puting in my venes 11:25:25 or was it alcohol? 11:25:32 alcohol prolly 11:25:34 you'll be an uncle in a few days dax 11:25:36 dax is oooold 11:25:37 beh 11:25:41 rub it in. 11:25:51 * Liesbeth rubs it in 11:26:14 --- join: huntrckr (~huntrckr@myr53-01-p144.gt.saix.net) joined #osdev 11:26:19 dax : you have a sister :D~~? 11:26:49 yes i have a sister 11:26:54 hi dax :) 11:26:55 sh's 12 11:26:58 got a problem 11:27:03 fluxbox wont compile 11:27:05 dax :/ 11:27:06 a really hot 26-year old one 11:27:11 OHHH 11:27:13 huntrckr: hmmm... what error 11:27:17 dax : address? 11:27:30 lynx: she's married already 11:27:39 dax : *shrug* 11:27:41 Windowmenu.cc: In constructor `Windowmenu::Windowmenu(FluxboxWindow*)': 11:27:42 Windowmenu.cc:111: `WindowmenuRemember' undeclared (first use this function) 11:27:42 Windowmenu.cc:111: (Each undeclared identifier is reported only once for each 11:27:42 function it appears in.) 11:27:42 Windowmenu.cc: In member function `void Windowmenu::Remembermenu::update()': 11:27:42 Windowmenu.cc:419: `WindowmenuRememberWorkspace' undeclared (first use this 11:27:44 function) 11:27:45 Windowmenu.cc:420: `WindowmenuRememberDimensions' undeclared (first use this 11:27:47 function) 11:27:49 dax's pfonenumber 092214062 11:27:49 Windowmenu.cc:421: `WindowmenuRememberPosition' undeclared (first use this 11:27:51 function) 11:27:53 Windowmenu.cc:422: `WindowmenuRememberShaded' undeclared (first use this 11:27:55 function) 11:27:56 oops 11:27:57 Windowmenu.cc:423: `WindowmenuRememberTab' undeclared (first use this function) 11:28:00 ewwwwwww 11:28:01 Windowmenu.cc:424: `WindowmenuRememberDeco' undeclared (first use this 11:28:03 function) 11:28:05 Windowmenu.cc:425: `WindowmenuRememberSticky' undeclared (first use this 11:28:07 function) 11:28:09 Windowmenu.cc:426: `WindowmenuRememberJump' undeclared (first use this 11:28:11 function) 11:28:12 rsync and retry :) 11:28:13 Windowmenu.cc:427: `WindowmenuRememberSave' undeclared (first use this 11:28:15 function) 11:28:17 Liesbeth : what is the country numba of freaky deaky belgium? 11:28:26 32 11:28:41 +32 92214962 ? 11:29:05 whos number is that 11:29:06 ? 11:29:09 03292214062 11:29:13 ... 11:29:15 oh 11:29:19 ... 11:29:27 shall i callx0r? 11:29:30 mocht dat jantje? 11:29:37 03292303137 and ask for "Sexy Simon" 11:29:43 sorrie, ik ga't nie meer doen 11:29:43 nah don't call 11:29:45 HEYYYYYYYYYYYY 11:29:56 don't you dare! 11:30:01 hrm.. ok 11:30:09 maybe i will call eks then... 11:30:14 * lynx wunna phonex0r 11:30:18 phreak calls... hmm 11:30:34 that's my number 11:30:47 cool 11:30:51 call me anytime 11:31:05 my brother even gets phonecalls from Iran 11:31:07 is it really yours or is it sexy simon? 11:31:14 mine and simon's 11:31:17 sexy simon is her bro 11:31:22 ah 11:31:29 he's the best 11:31:34 yes, sure 11:31:36 i'm still better 11:31:45 ubersexy liesbeth 11:31:48 hrm 11:31:52 --- quit: do ("Quit") 11:32:00 is that the time where i have to ask for pics? 11:32:03 lol 11:32:08 hmm 11:32:09 maybe 11:32:13 --- join: do (~green@194.85.84.244) joined #osdev 11:32:14 i've some 11:32:19 taken in paris 11:32:25 paris! 11:32:28 i have some , too 11:32:29 oui 11:32:30 taken in hrm 11:32:35 vienna and party.. 11:32:35 hrm 11:32:56 i've got some from florence and laiatico 11:33:01 all sorts of 11:33:01 Liesbeth : where have you been in paris? 11:33:03 i have a pic too... taken from my newest movie 11:33:10 huntrckr : url? 11:33:14 all over the place, montmarte 11:33:31 la/le? tourre eiffel 11:33:38 centre pompidou 11:33:41 la defense 11:33:43 la toure eifeel 11:33:44 just watch "Sum of all fears"... the hunky guy that play the main part... /me 11:33:51 tour 11:33:53 :p 11:34:01 and of course, disneyland paris 11:34:04 huntrckr : dunno that 11:34:05 Liesbeth AHHH 11:34:12 Liesbeth : i have been there, too... in 1996 11:34:17 or 1994? 11:34:21 just after it opened 11:34:23 i went 2 times 11:34:29 i as scared of captain EO :P 11:34:30 1 time i got sick 11:34:33 reeeally sick 11:34:39 lol 11:34:42 * lynx too 11:34:49 i at in a hotelroom the whole stay 11:34:52 *sat 11:34:54 which hotel? 11:34:58 i was in "santa fe" 11:35:01 lo 11:35:02 l 11:35:07 can't remember 11:35:17 what color did it have? 11:35:21 the one michael jackson staid in 11:35:23 pink! 11:35:29 YEAH! 11:35:34 i think that was santa fe :P 11:35:40 oh 11:35:42 hrm 11:35:51 michael jackson = captain eo 11:36:06 hehehe 11:36:15 well 11:36:25 that was his role in the star wars thingy 11:36:34 the dutchmen were funny: whats jusd'orange in french? 11:36:47 hrm 11:36:50 lol 11:36:53 i hated that thing the starwars thingy 11:36:57 ye 11:37:14 did you have to wait in the queue for like 1.2 hours, too? 11:37:19 * lynx hated that 11:37:22 that irritating french droid babbling the whole time 11:37:23 * lynx HATES mickey 11:37:27 LOL 11:37:34 * lynx thinks he remembers that 11:37:36 blablabla AUUUSECOUUUUUURSSSSSSSSSSSSSSSSSSSSSSS 11:37:43 lmao 11:37:46 that was the droid 11:38:09 it was noisy 11:38:17 gtg 11:38:44 ok 11:38:47 bye 11:39:06 see ya! 11:39:22 --- quit: Liesbeth () 11:40:02 hrm 11:40:05 anyone else up? 11:40:19 * lynx is prolly releasing something for mekka&symposium 11:40:20 no 11:40:51 ok 11:40:57 dax : gimme your phone# 11:41:06 no lol 11:41:11 why? 11:41:28 i used to phone eks regularily :) 11:42:08 brb 11:45:14 03292303137 11:45:27 but using the phone atm 11:46:32 thanks, i will save it on my palm 11:46:53 --- join: file (proxy@mctn1-2922.nb.aliant.net) joined #osdev 11:48:06 Hey file 11:51:17 --- quit: lodda (Remote closed the connection) 11:56:17 --- join: ChillySpy (ChillySpy@ppp150.ppp4.cleveland.nccw.net) joined #osdev 11:56:46 --- part: Javanx left #osdev 11:56:53 --- join: Javanx (~javanx@213.45.18.12) joined #osdev 11:59:33 --- quit: file (Killed (NickServ (Ghost: JoshNet!proxy@mctn1-1278.nb.aliant.net))) 12:00:01 --- join: file (proxy@mctn1-1278.nb.aliant.net) joined #osdev 12:00:42 --- quit: tirloni ("going home..") 12:03:41 --- nick: oink -> ziga 12:06:49 --- quit: ChillySpy () 12:11:55 --- join: ChillySpy (ChillySpy@ppp150.ppp4.cleveland.nccw.net) joined #osdev 12:14:10 --- join: DRF (Daniel@host213-121-68-149.surfport24.v21.co.uk) joined #osdev 12:22:45 --- nick: HeavyJoost -> HeavyYoast 12:24:15 --- quit: huntrckr (Remote closed the connection) 12:30:44 --- join: huntrckr (~huntrckr@myr53-01-p144.gt.saix.net) joined #osdev 12:30:55 dax: you here? 12:31:17 <-- needs some beginners help with fluxbox 12:31:25 ? :) 12:31:28 * ziga uses fluxbox 12:31:29 er 12:31:30 I used 12:33:28 --- quit: huntrckr (Remote closed the connection) 12:34:53 --- join: huntrckr (~huntrckr@myr53-01-p144.gt.saix.net) joined #osdev 12:35:20 hmmm... nevermind 12:35:35 * huntrckr is heading for a bed 12:35:44 CYa all tomorrow night :) 12:35:51 --- nick: HeavyYoast -> HeavyJoost 12:35:51 --- nick: huntrckr -> huntrckr_zzzz 12:36:07 --- join: indigo1 (~apple@64.3.7.219) joined #osdev 12:36:42 --- part: indigo1 left #osdev 12:37:04 --- join: indigo1 (~apple@64.3.7.219) joined #osdev 12:38:14 moo? 12:38:16 anyone here? 12:38:23 this irc client sucks... 12:38:33 Which client are you using? 12:38:33 --- part: indigo1 left #osdev 12:38:40 --- join: indigo1 (~apple@64.3.7.219) joined #osdev 12:38:52 IRCstep 12:38:55 hum 12:38:58 Never heard of it 12:38:59 i can't see who's here 12:39:08 neither have I, but it's all i could find 12:39:13 anyway, i have a bit of a problem 12:39:26 i need to stuff stuff into the keyboard buffer in DOS 12:39:36 Ok 12:39:57 is there a dos service to do that, or something i can do to the keyboard controller? 12:40:39 There are BIOS interrupts. They will read daa from the keyboard but apart from that I don't think you will have much luck. 12:41:06 i need to put data into the buffer, not take it out 12:41:14 as if someone were typing... 12:41:18 You will prob have to do it the harder way, more similar to Keyboard drivers in pmode 12:41:31 hrmph 12:41:33 is eks here? 12:41:54 Afraid not 12:42:09 humph 12:42:26 What do you want to put data into the keyboard buffer for? 12:42:40 kicks and giggles 12:42:48 Fair enough 12:42:53 indigo1 :) 12:42:56 Evening 12:43:00 howdy 12:43:08 hm 12:43:22 i'm at work, and there's a program that does it in windows, but not in dos 12:43:35 there was supposed to be a dos one too, but the guy who was writing it dissapeared 12:43:59 I don't think that there is a shortcut. You will have to make a specilised DOS keyboard driver I think 12:44:21 how would i do that? 12:44:40 make a TSR and grab int 21h? 12:45:16 Errm, I wouldn't really know. You have to play around with the keyboard controller for realmode -> pmode bootloaders so it must be possible. But you probably will have to refer to specific memory locations etc 12:45:31 hm 12:46:14 Playing with the int21h table will allow you to see what the person writes and play around in that sense but not take control of the keyboard. 12:46:43 Unless you wanted to disable it and replace it with a connection to your keyboard somehow. 12:47:50 --- quit: ChillySpy () 12:48:23 Basicly if I was really motivated I could do it on my OS. Which is similar to DOS. But it's not on my prioritory list currently. (FAT12 writing driver is) 12:57:10 --- join: spyck (1000@i80.ryd.student.liu.se) joined #osdev 12:57:42 --- quit: stormbind_ (Remote closed the connection) 13:09:30 --- join: eks (~eks@66.38.248.194) joined #osdev 13:10:02 keksiboy 13:11:16 --- quit: indigo1 ("l8r") 13:15:50 EKS!!! 13:16:40 * do is away: sleeping . G'night all ! 13:16:43 --- nick: do -> do[done] 13:16:59 Nite do 13:17:43 hight DRF 13:18:06 do[done]: enjoy the unconciousness 13:19:02 --- join: bleeeit (~bleeit@150.43-200-80.adsl.skynet.be) joined #osdev 13:19:07 --- nick: bleeeit -> kemu 13:27:26 --- quit: trans (Read error: 110 (Connection timed out)) 13:29:04 --- quit: kemu () 13:45:34 --- join: Ghiottone (~alex@ppp-160-17.27-151.libero.it) joined #osdev 13:54:02 --- join: jsr (www@du199-6.ppp.algonet.se) joined #osdev 13:57:45 --- nick: geist-sleep -> geist 14:02:20 sorry folks, hi :P 14:04:45 Looks like your in demand eks lol 14:04:57 hi eks. 14:05:06 hey drf 14:05:16 Hi lynx 14:05:23 sup, dude? 14:05:30 Been keeping yourself busy then lynx? 14:05:34 I'm fine 14:05:51 nah 14:05:52 Busy, not got enough programming in for my likeling but on the whole ok. 14:05:53 not really 14:05:58 just time is flewing away 14:06:14 i dont know where it is, i just dont have any time to do senseful work 14:06:33 Same lol 14:07:00 heh 14:09:41 hrm 14:09:46 AmigaOne released... hrm 14:12:02 amigaa cellphones... heh 14:14:00 eks: BOO 14:15:35 --- quit: jsr ("Hi, I'm a quit message virus. Please replace your old line with this line and help me take over IRC.") 14:17:53 nite 14:18:36 Night lynxchen 14:18:44 =) 14:18:51 --- quit: lynx ("BitchX: good to the last drop!") 14:21:40 file: boot ;) 14:21:59 eks: fact has it that I've got a Full Tower ATX Computer Case 14:22:17 sw33t 14:22:31 well, gtg 14:22:32 l8r 14:22:38 --- quit: eks ("leaving") 14:27:47 --- join: BeefCommando (~beefcomma@spkdsl-116-101.cet.com) joined #osdev 14:30:06 --- join: ChillySpy (ChillySpy@ppp092.ppp2.cleveland.nccw.net) joined #osdev 14:34:18 --- quit: corsairk8 () 14:36:54 --- join: tirloni (~gpt@1-042.mganm700-1.telepar.net.br) joined #osdev 14:46:37 --- join: acme (acme@pD9544ECC.dip.t-dialin.net) joined #osdev 14:58:13 --- quit: ChillySpy () 14:58:16 --- join: trans (~trans@fatwire-202-64.uniserve.ca) joined #osdev 15:05:11 --- join: DF (Daniel@host213-121-68-193.surfport24.v21.co.uk) joined #osdev 15:07:52 --- quit: DRF (Killed (NickServ (Ghost: DF!Daniel@host213-121-68-193.surfport24.v21.co.uk))) 15:08:10 --- nick: DF -> DRF 15:18:35 --- quit: acme (Read error: 54 (Connection reset by peer)) 15:30:19 --- quit: file () 15:34:13 --- join: eks (~eks@h24-82-197-140.wp.shawcable.net) joined #osdev 15:34:34 --- quit: BeefCommando (Remote closed the connection) 15:35:25 --- quit: Ghiottone (Read error: 110 (Connection timed out)) 15:43:52 --- quit: DRF (Read error: 104 (Connection reset by peer)) 15:48:09 --- join: file[ermph] (lan@mctn1-4352.nb.aliant.net) joined #osdev 15:48:26 * file[ermph] pokes people 15:50:34 --- quit: file[ermph] (Client Quit) 16:00:10 --- nick: gab -> gab|khof 16:00:16 khof 16:01:41 --- join: file (~lan@mctn1-0478.nb.aliant.net) joined #osdev 16:05:19 phil 16:10:30 edgar 16:11:14 alan 16:11:21 poe 16:11:24 =) 16:11:29 <:D 16:20:10 --- join: Kurt (~kmw@gc-nas-01-s252.cinergycom.net) joined #osdev 16:32:01 --- quit: Javanx (Read error: 113 (No route to host)) 16:32:50 --- join: Javanx (~javanx@213.45.18.12) joined #osdev 16:33:12 --- quit: Kurt ("Connection reset by queer") 16:37:13 --- join: n0vice (n0vice@pcp02102499pcs.towson01.md.comcast.net) joined #osdev 16:37:16 hey 16:46:12 sup 16:47:42 nothing you 16:49:15 do you know about os developement? 16:49:23 sure 16:51:02 what should i read to start out 16:51:12 like i know a little c, but no assembly 16:51:26 what assembly tutorials should i read? 16:52:09 get real familiar with C, get real familiar with with compilers output (the asm, how stacks work, etc) 16:52:23 then get super familiar with how your computer works 16:52:36 ie, how it boots, what the BIOS is, how to switch into protected mode 16:52:43 and those are technical details you'll need to know 16:53:05 after that, you should become very familiar with how oses really work (schedulers, timers, interrupts, etc) 16:53:24 oh ok 16:53:35 where is a good asm tutorial..? 16:53:37 and lots of classic computer science stuff like mutual exclusion 16:53:43 I dunno, I never used any 16:54:01 I dont know much about tutorials and that kind of stuff, others might be better help 16:54:28 ok 16:54:33 thank you 16:54:37 no prob 16:54:51 it's a long journey, dont expect anything any time soon, but it's very educational 16:57:49 yea 16:57:55 obviously I didn't tell you to much specifically, but I dont have a lot of specifics to tell, it's been a while for me 16:58:23 I learned my os dev by just reading every book I saw and digging through code and diving in 16:59:24 and also by the time I eventually got into osdev stuff, my core level of know-how was there so that I didn't have too many technical hurdles to get over. In other words, learning the specifics of asm or bios or whatever was the least of my problems 16:59:33 it was mostly design problems I dealt with 16:59:37 that's the best way to learn 17:00:07 it's sort of liking being an accomplished mechanic before you start to learn how to work on cars 17:00:42 cool 17:04:47 --- join: sleep- (~ivan@adsl-68-21-169-91.dsl.chcgil.ameritech.net) joined #osdev 17:05:15 re 17:05:35 hey 17:06:05 this is great! 17:06:11 all my worst nightmares are coming true... 17:06:17 what? 17:06:31 the CEO and software engineer 17:06:34 of my company 17:06:49 are at the trade show - our demo server was trashed by the shippers 17:07:06 the motherboard got ripped out of its mounting and cracked the VGA in half :( 17:07:16 hes trying to get it up and running without a console :P 17:07:20 ? what the hell did they do 17:07:24 i dunno 17:07:28 somehow it got really trashed 17:07:56 damn 17:08:11 oh well - thats a typical situation in a startup i guess 17:08:33 start up company? 17:08:39 yeah 17:08:47 were just starting up ;) 17:08:52 what do you do? 17:09:17 we make various products for linux clustering 17:09:24 our first release is a PXE server appliance 17:09:42 basically, its a network boot server used to manage large numbers of linux server nodes 17:10:01 oh cool 17:10:08 its fun 17:10:16 are you a programmer for the company? 17:10:24 i decided never to work on projects that arent related to high performance computing 17:10:29 yeah 17:10:33 i started the company actually 17:10:37 what language 17:10:37 but im just a programmer 17:10:38 oh cool 17:10:42 C++ 17:10:51 sweet 17:10:53 and lots of XML 17:11:11 its getting to the stress part - the part where we are nearing the 1.0 release 17:11:18 thats always a fun ride :) 17:11:49 i guess im going to be up all night getting the software demo working 17:11:51 *sigh* 17:12:08 that sucks 17:12:31 well - i think i must like this - i always end up in this exact position 17:12:41 there must be a sub-conscious part of me that makes it happen :) 17:13:16 lol 17:14:17 anyway - its not as bad as it could be - i just hope the demo unit still runs 17:19:43 oh thank god - they just called - the box is pinable! 17:19:49 s/pinable/pingable/ 17:25:09 moo 17:25:11 splended 17:25:16 eks: moo! 17:26:21 anyone ever heard of The High Level Assembler 17:26:41 yes 17:26:43 it sucks 17:26:48 lol ok 17:26:54 why 17:27:02 it's something like the worst of asm and C all in one 17:27:19 and it's very ugly 17:27:28 it's actually just a set of macros for TASM i think 17:27:41 so instead of "push eax" you have "push( eax )" 17:28:04 it's really lame 17:28:08 so just don't do it 17:28:45 indigo was molested by the High Level Assembler when he was a kid 17:29:06 geist: shh...it will hurt me if you tell 17:29:32 * indigo works on tauga a bit 17:29:55 indigo: show me on the doll 17:30:05 heh 17:30:43 i wondered what those dolls were for... 17:32:46 humph 17:32:49 ahhhh i need a good asm tutorial 17:32:57 * indigo wishes tauga was further along 17:33:20 n0vice: i started to write one... 17:33:26 maybe i'll work on that 17:33:46 maybe you could teach me and post the convo 17:33:47 lol 17:33:50 just kidding 17:35:10 i can only find operating system specific asm guides, but am i correct in saying that it wont help since i want to eventually program an operating system? 17:36:37 http://sempiternity.org/articles/learning_ia32_asm/ 17:37:09 there's a start of one 17:37:23 you need to be able to read asm far more than write it 17:37:42 reading it is pretty easy. I'd suggest teaching yourself by dissassembling C code and looking at the intel asm manual 17:37:44 novice: do you know C? 17:37:50 trace what the compiler is doing 17:38:38 that's basically how I learned asm 17:38:45 yes i do know c 17:38:55 --- quit: tirloni ("ircII EPIC4-1.0.1 -- Are we there yet?") 17:38:56 since then I've gotten pretty proficient at 3 or 4 different architectures. the basics are pretty much the same 17:39:02 novice: compile programs with 'gcc -S' and read the assembly output 17:39:09 thats a good way to learn 17:39:15 yep, that's how I did it 17:39:37 that sounds too hard lol 17:39:46 if that's too hard, then forget ever writing an os 17:39:50 hehe 17:39:50 lol 17:40:08 get on it! right now 17:40:12 ok ok 17:40:16 I want a report on what you learned in 30 minutes 17:40:41 compile with -O0 for now, it'll generate lots of dumb code but very straightforward 17:40:59 yeah, thats a good idrea 17:41:01 s/rea/ea/ 17:41:04 then build a list of specific questions and I betcha someone will answer 17:41:18 like how the stack works and how args are passed 17:41:36 write a function that passes args to other functions, see how the args are passed, etc 17:41:43 see how the result is gotten back 17:41:56 lol 17:41:59 write simple functions like int foo(int a) { return a + 1; } 17:42:01 see what it does 17:42:02 how bout i just do a hello world program 17:42:05 no 17:42:10 do much much simpler 17:42:21 that'll just have a function call to an external library function 17:42:22 eeh I oh 17:43:33 thats a lot of shit for i hello world program 17:44:32 what what options do i use with gcc 17:45:46 novice: to convert C to assembly use gcc -S 17:46:26 create a file 17:46:28 blah.c 17:46:30 put 17:46:35 i dont think that is all asm 17:46:37 int foo() { return 0; } 17:46:41 compile that with gcc -S 17:46:48 .file "hello.c" 17:47:02 dont mess with hello world yet 17:47:06 you wont get anything useful out of it 17:47:06 ok 17:48:00 gcc -S foo.c -o - 17:48:10 will spit it to stdout 17:48:11 ahh that looks a little bit more readable lol 17:48:32 are % used in asm? 17:48:47 that is asm 17:48:55 it's at&t syntax 17:49:11 and what about ".file "blah.c" ".text" ".align 2" 17:49:13 the bummer is it wont match exactly with the syntax in the intel manuals unfortunately 17:49:17 forget that 17:49:28 where does the ctual asm code start 17:49:41 asm instructions are indented 17:49:48 xxxx: is a label 17:50:02 most of it is indented 17:50:05 stuff like .globl, and .text are asm directives that you dont need to see 17:50:08 then most of it is asm 17:50:25 but you said .file .text and .align wernt and they are indented 17:50:29 sigh 17:50:31 dude get a clue 17:50:42 if it's .xxx it's a directive probably 17:50:55 ok ok sorry 17:51:00 especially if it has some shit after it that the cpu obviously isn't going to parse directly 17:51:19 asm opcodes *never* start with . on intel 17:51:36 ok 17:51:40 like I said, if it starts at the first column and ends with a ':', it's a label 17:51:44 what about amd 17:51:46 the target of a branch 17:51:55 amd is the same thing 17:51:59 ok i got it 17:52:09 i dont see how this is going to help me lol... 17:52:13 the x86 instruction set is the same except for a few extensions (like mmx, sse, 3dnow, etc) 17:52:20 then you probably arne't ready for it 17:52:20 sorry 17:52:27 go find a tutorial 17:52:29 novice: the .file and .line stuff allows you to match up the line in the C program with the assembly code 17:52:35 ..how do i get ready for it 17:52:46 novice: read each line of C and find that line in the asm output 17:53:02 novice: some times there are many lines of asm for one C line 17:53:09 DWARF 17:53:12 oh ok 17:53:17 the c was 1 line...let me change that 17:53:30 novice: but the .line makes it real easy to see what part of the C code generated what asm lines 17:53:45 wli: as in the object format? 17:54:38 n0vice: see if you can find some asm tuturials as well 17:54:44 one was stated a while back 17:54:51 to get the basics (registers, etc) 17:55:33 since it's all processed by the cpu, which is inheirently a dumb piece of silicon, asm code tends to be a buttload of very very simple things all together 17:55:45 you'll be suprised sometimes how many line sof asm it takes to do something very simple 17:56:01 ok...but does it matter if i read one that was made for a certain OS? 17:56:06 will taht effect me in the long run? 17:56:08 no 17:56:12 ok 17:56:12 it's all the same thing 17:56:21 but which should i read, linux, or windows , or dos 17:56:30 i use all 3... 17:56:32 linux 17:56:37 it's easiest for now 17:56:50 but again, the code you'll be looking at is exactly the same 17:57:03 the cpu is the same, it just matters what libraries it links into 17:57:12 ok 17:57:14 except dos, but forget dos 17:57:22 cool thanks, im going to go read now 17:57:24 forget you ever heard of dos 17:57:27 * n0vice is reading 17:57:30 lol ok i will 17:57:30 the tutorial? 17:57:32 yea 17:57:34 good 17:57:42 and get a copy of the intel x86 manuals 17:57:48 you can get them from developer.intel.com 17:58:11 they have everything you'd ever want to know about the x86 architecture 17:58:31 unfortunately the asm syntax they have in the documents do not match the one produced by gcc 17:58:39 o well 17:58:47 so that might confuse you until you get the hang of thigns 17:59:00 whats the best asm compiler for linux? 17:59:07 gas 17:59:09 fasm? 17:59:11 oh ok 17:59:31 gas == gcc 17:59:36 sort of 17:59:56 ie, if you have an asm file (blahblah.S) and you give it to gcc, gcc will start gas for you 18:00:07 as a background, gcc is basically a frontend for a set of programs 18:00:18 cc1, gas, ld, etc 18:00:22 oh so gas comes with gcc/unix 18:00:26 yes 18:00:40 whats the command? 18:00:48 normally when you compile a .c file, gcc starts up cc1 and another stage I forget, and then passes the result to gas 18:01:00 when you run gcc with -S, it basically just stops at the stage before it gives it to gas 18:01:15 so what you see is what gas gets passed when you compile a .c file 18:01:29 gcc can figure out what to call by the file itself 18:01:38 so you can pass .S files to gcc and it'll call gas on it 18:01:42 hey, od the intel manuals have the nasm systax? because i just found a converter, nasm to gas 18:01:49 yes 18:01:56 nasm == intel syntax 18:01:56 oh cool 18:01:59 gas == at&t syntax 18:02:12 which is a bummer 18:02:27 but there is a converter.. 18:02:48 most asm programmers prefer intel syntax, but gcc produces at&t, which you'll look at a lot if you debug code, so I figure you may as well get used to at&t syntax 18:03:25 ok 18:03:39 and gcc compile both syntaxes 18:03:43 can* 18:03:46 not that I know of 18:03:47 not and 18:03:51 oh 18:04:02 gcc == gas and gas is at&t 18:06:07 --- join: ChillySpy (ChillySpy@ppp031.ppp1.cleveland.nccw.net) joined #osdev 18:06:27 --- join: malenfant (~malenfant@ati2362cy13e4.bc.hsia.telus.net) joined #osdev 18:08:26 anyway, have fun. I'm too busy to help any more today 18:08:30 --- part: geist left #osdev 18:16:27 --- quit: indigo (Read error: 110 (Connection timed out)) 18:23:04 --- nick: minddog[test] -> minddog 18:38:53 --- quit: Zenton (Remote closed the connection) 18:50:38 one nightmare after another ... 18:51:28 --- part: ChillySpy left #osdev 18:56:28 --- quit: trans (Read error: 110 (Connection timed out)) 18:56:39 --- join: indigo (indigo@bgp01105107bgs.wbrmfd01.mi.comcast.net) joined #osdev 19:14:00 hum 19:19:52 moo 19:23:52 --- quit: n0vice () 19:24:02 --- quit: Javanx ("I don't feel a thing, and i stopped remembering. The days are just like moments tourned to hours") 19:31:06 gn 19:31:55 gn indigo 19:33:31 --- quit: Teratogen ("bailing out near line 1") 19:50:46 --- join: Lathiat (~VrS@005.pth0201.pth.iprimus.net.au) joined #osdev 20:23:37 --- join: trans (~trans@fatwire-202-64.uniserve.ca) joined #osdev 20:27:20 --- quit: Lathiat ("Leaving") 20:35:29 --- join: revanthn (~samsweetk@202.62.85.2) joined #osdev 20:36:52 hi guys 20:39:26 howdy 20:41:34 heya sleep- 20:43:17 how's your OS going 20:44:25 not going at all 20:44:32 im working 100% on work stuff :( 20:44:36 busy right 20:44:39 yeah 20:44:51 big trade show tomorrow - gotta get the demo patched up 20:45:17 oh 20:45:20 we already had one disaster - the demo hardware was nearly destroyed in transit L:0 20:45:34 but we lucked out - its still functional 20:46:03 what are you giving a demo of 20:46:10 our network boot server 20:46:17 its a PXE server appliance 20:46:32 we're demoing a pre-release version of it tomorrow 20:46:54 cool 20:47:11 yeah, *hopefully* it will be cool :0 20:47:46 what are your target people 20:48:05 administrators of large linux clusters 21:06:44 --- join: geist (~geist@tkgeisel.com) joined #osdev 21:09:44 --- join: beefcommando (~beefcomma@spkdsl-116-101.cet.com) joined #osdev 21:11:36 hello my fellow os devers 21:12:09 geist:any new improvements to newOS 21:12:20 newOS seems mainly for SMP systems 21:16:06 remember: "one processor is not enough" 21:16:42 not mainly for smp 21:16:46 it just supports smp 21:16:54 though I do all my testing on smp boxes 21:19:02 http://apollo.backplane.com/FreeBSDSmp/ 21:24:00 --- quit: trans (Read error: 110 (Connection timed out)) 21:26:39 --- quit: huntrckr_zzzz (Read error: 110 (Connection timed out)) 21:27:02 yah 21:27:41 I feel sorry for those guys 21:27:50 retrofitting a kernel to do smp later is no fun at all 21:27:53 i want a good cluster filesystem 21:28:21 I want a Honda 21:30:51 * geist like 21:30:51 http://www.hondamotorcycles.com/motorcycles/cruiser_standard/model.asp?ModelName=Shadow+VLX+Deluxe&ModelYear=2003&ModelId=VT600CD3 21:31:40 fs requirements: (1) uses distributed storage available in pc-based clusters (2) transaction protection (3) fail-over (4) fast 21:31:46 oh, and free 21:32:17 sleep-: guess you gotta roll your own :) 21:32:24 hmm 21:32:31 isnt there *anything*? 21:32:40 that seems like such a needed thing 21:32:45 not that I know of 21:34:02 ok, ill make the damn thing 21:34:18 ill have to read about distributed stuff a bit first 21:37:05 the sum of all fears is a pretty good movie 21:39:23 --- quit: do[done] ("Quit") 21:39:27 --- join: do (~green@194.85.84.244) joined #osdev 21:40:44 * do is away: bbl 21:40:48 --- nick: do -> do[done] 21:44:54 --- nick: geist -> geist-gaming 21:46:54 * revanthn wonders what everyone is doing 21:55:29 --- quit: eks ("Zzzzz") 21:55:57 anyone intrested in numbers http://www.math.uga.edu/~ntheory/number_theory.html 22:05:58 --- join: trans (~trans@fatwire-202-64.uniserve.ca) joined #osdev 22:08:51 --- join: Lathiat (~VrS@082.b.002.pth.iprimus.net.au) joined #osdev 22:09:13 i followed the thing at http://osdev.neopages.net/tutorials/basickernel.php and downloaded the grub-compatible source but cant get grub to boot it 22:09:25 i did root (fd0) and kernel /kernel.bin --type=multiboot but it says invalid binary format 22:09:43 am i missing something? 22:10:20 do one thing download the GRUB image and boot with it 22:11:19 ? 22:11:35 i dont understand what you mean 22:12:00 i have made a grub boot floppy.. i booted from that set the root to fd0 where my kernel is, it recognizes and tab completes my kernel.bin file 22:12:36 http://osdev.neopages.net/downloads/disk_images/grubimage.zip 22:13:08 oh i didnt see a link to that anywhere 22:13:37 --- quit: revanthn (Read error: 104 (Connection reset by peer)) 22:14:23 --- join: revanthn (~samsweetk@202.62.85.2) joined #osdev 22:14:29 with that image 22:14:31 i get file nto found 22:14:42 or is the kernel not in the image? 22:15:09 do some reading and figure it out i g2g no 22:15:10 now 22:15:11 sorry 22:15:15 --- part: revanthn left #osdev 22:16:22 --- quit: file (Read error: 113 (No route to host)) 22:22:32 --- join: mur (nobody@66.234.170.97) joined #osdev 22:22:36 HA! 22:22:37 HA! 22:22:40 found new client 22:22:45 sup 22:22:48 Rrroobert 22:23:00 neato, this irc gateway works 22:23:05 nuffin much 22:24:27 you? 22:24:45 irc gateway? 22:24:56 http://schematic.org/irc.php 22:25:00 --- quit: beefcommando (Remote closed the connection) 22:25:19 yes i dont use client, this is web page (no java applet IRC client) 22:25:52 trying multi chans 22:25:54 --- part: mur left #osdev 22:26:31 --- join: mur (nobody@66.234.170.97) joined #osdev 22:26:42 no multchan 22:29:47 ROBERT, RICO, SLEEP-, TRANS, wake up already! 22:29:56 --- join: stormbind (~stormbind@p50835F6C.dip.t-dialin.net) joined #osdev 22:29:56 --- quit: mur (Remote closed the connection) 22:30:39 howdy 22:30:46 * sleep- is miserable - deadline approaching :( 22:31:26 for? 22:31:27 --- join: mur (nobody@66.234.170.97) joined #osdev 22:31:29 bah ! :P 22:31:33 lol 22:31:36 not working well? 22:31:45 better than nothing :P 22:31:51 cgi irc rocks 22:32:02 there are a few public versions installed in places 22:32:05 google for cgi:irc 22:32:07 trans, robert, rico, sleep- wake up already get up 22:32:16 mur: sleep- is awake 22:34:13 sleep- is sleeping in irc 22:36:37 hehe 22:36:37 --- quit: mur (Remote closed the connection) 22:36:40 or so it seems 22:37:07 i need to finish this gui demo 22:37:17 before i kill myself out of frustration 22:40:47 bbl 22:42:57 Stick the letters on with sticky tape. If you want to change the layout you can easily pull them off. 22:43:35 wrt the /. article on how to mutate keycaps but leds use too much power 22:44:47 top story on /. right now is pretty funny 22:45:56 Ask Slashdot: Server Side Virus Scanning Options???? 22:46:42 umm - no, i meant the doom for nokia phones 22:47:25 --- join: redblue (star@ppp085.216-96-207.sherb.mt.videotron.ca) joined #osdev 22:47:31 * sleep- wonders if his http cache is fucked 22:48:39 ogm 22:48:41 i dont beleive that 22:48:48 i just got spammed with a netbios message 22:49:41 --- join: revanthn (revanthn@202.9.183.175) joined #osdev 22:50:01 hehe 22:50:14 heya sleep- 22:50:33 hiya 22:51:51 lath: you should turn all that crap off 22:52:11 what crap 22:52:27 he said he got spammed with a netbios message 22:52:42 * sleep- thinks any open port is a bad idea 22:52:53 besides maybe shs 22:52:55 ssh 22:53:13 yes openports are not good 22:53:38 i say only one port must be open with all communications going through it with Different ID's or something 22:53:54 dhcpd is *so* annoying - changing the prefix in the makefiles in such a pain 22:54:04 no configure 22:54:05 :( 22:54:28 sleep-:what are networks servers anyway 22:54:29 i actually have to sed all the makefiles 22:54:34 rev: ? 22:55:00 i didnt understand that question 22:55:00 you asked me something about network servers right 22:55:01 or is it someone else 22:55:09 i dunno 22:55:23 i make networked servers as my job :) 22:55:41 are you good with thin clients also 22:55:45 well 22:55:47 im trying to be 22:55:56 since my PXE server can deliver thin client code 22:56:07 its a good thing for me to be aware of 22:56:12 clarify things for me will ya 22:56:15 sure 22:56:16 if i can 22:56:30 where do you store the OS in a thin client,a ROM? 22:57:05 depends 22:57:08 on the system 22:57:16 some things boot off a static ROM 22:57:23 --- join: roy (~lover_boy@wl-parikesit-12.teras.net.id) joined #osdev 22:57:23 what is usual way it is done 22:57:26 but i bet more and more will be network delivered at boot time 22:57:37 i dont know what the "usual" method is 22:57:38 what if they want to make changes to the core OS or something 22:57:54 i think network delivery will be the most common approach in the near future 22:58:02 hi it's me suhendra, is quistis there? 22:58:11 because that centralizes the management and eases upgrading, etc. 22:58:22 --- nick: roy -> hendra 22:58:37 who's suhendra? 22:58:39 at least, that seems like a better method to me 22:58:59 is there anyone know about compiling kernel with Linux GCC 22:58:59 could be that since I make a network boot server, my view of the world is biased though :) 22:59:11 hendra: yes 23:00:09 --- join: eirikn (eirik@odin.eirikn.net) joined #osdev 23:00:11 hendra:this is #osdev almost all of us here know that 23:00:46 bbl 23:01:07 --- quit: hendra (Client Quit) 23:02:24 i would hope we all at least know how to compile our own kernels with gcc 23:03:47 rev: were you going somewhere with that thin client question? 23:03:56 btw, you should check out PXES 23:04:01 it seems interesting 23:06:12 --- join: revanth (revanthn@202.9.183.237) joined #osdev 23:06:31 http://www.electricnews.net/Files/1032513230-canesta.jpg 23:06:33 http://www.canesta.com/products.htm 23:06:41 who thinks that is real? 23:06:50 anyone heard anything about it? 23:07:23 re rev 23:07:28 hey 23:07:30 air:is that real 23:07:35 whoa 23:07:52 can any1 reconnommend a good os dev tutoriall.. ive googled and found heaps of cool stuff but cant find anything that goes step by step showing me what say bootloader asm does.. its great cut and pasting 500 lines of asm that does cool stuff - but if uc ant understand it is pointless 23:07:57 weird 23:08:05 air: if thats real, its pretty cool 23:08:15 its damn cool 23:08:16 Lathiat:you need patience 23:08:16 lath: read the linux source 23:08:23 yeah it's rocks 23:08:36 like i realise its not the easiest thing in the world.... 23:08:39 air: where did you find that? 23:08:50 lath: there are some ok books - a lot of them are out of print though 23:08:54 but i want something that will help me understand how it works, not show me the source adn tell me what it does 23:08:58 sleep-: /. 23:09:04 and online tutorial would be nice... 23:09:11 lath: find a xinu book - those were clear and concise 23:09:18 sleep-: in the comments for the mutating keyboard 23:09:24 lath: but are you more interested in PC internals, or OS theory? 23:09:50 air: thats a cool thing - not sure how it would feel to type on nothing? :) 23:09:52 more pc internals 23:10:02 lath: there are quite a few books that go over that stuff 23:10:02 im not trying to create the next linux 23:10:13 sleep-:guide him to the OS-FAQ it's good for newbies 23:10:50 rev: you do it, i dunno where that is :) 23:11:04 i have a basic udnerstanding of assembly 23:11:07 and i can program C 23:11:09 (and C++) 23:11:25 sleep-: u would be typing on yer desk 23:11:31 or any flat surface 23:11:35 do a search for OS-FAQ by Dark Fiber it's good 23:11:50 air:how the hell will it recognize what we typed 23:11:51 Lathiat: qzx.com/library 23:12:01 * Lathiat looks 23:12:16 revanth: Canesta's electronic perception technology will instantly resolve the user's finger movements into ordinary serial keystroke data that is easily utilized by the wireless or mobile device. 23:12:22 http://www.mega-tokyo.com/os/os-faq.html <-- that one? 23:12:26 lath: dont listen to them - you should read Trans Pmode32 DPMI/VCPI DOS extender code first 23:12:44 yeah 23:13:04 air:but what about those who can't type 23:13:06 man, /me feels old all of a sudden 23:13:12 there fingers move one at a time 23:13:21 why sleep- 23:13:24 hunt and peck? 23:13:47 im probably the only one here who used pmode32 23:14:15 sleep-: geezer :) 23:14:30 hehe :) 23:14:34 pmode32 was cool 23:14:49 for its time anyway 23:15:04 used pmode32 for what 23:15:14 it was a DOS extender 23:15:24 DPMI == DOS Protected Mode Interface 23:15:34 VCPI == Virtual Control Programmers Interface 23:15:44 they were two ways to get extended memory over 1MB in DOS 23:15:54 i know that 23:15:56 ;0 23:16:04 most ppl have forgotten about that shit 23:16:15 revanth: an IR light source illuminates the user's fingers to the perception chipset can track em 23:16:22 Tran was a cool coder who wrote pmode32 which interfaced with both 23:16:24 all in asm 23:16:42 i was all impressed because it did paging and stuff 23:17:05 but won't IR a bit harmfull for our body on long term exposure 23:17:12 er *isn't 23:17:13 that was from the PC demo scene 23:17:35 revanth: has yer remote control killed u yet? 23:17:56 remote control is not faced towards me 23:18:01 so it's less radiation 23:18:08 besides it's only around 33Hz 23:18:27 but inorder to get good perception of fingers with IR we need higher frequencies 23:18:48 i am pretty sure ir is not dangerous 23:19:55 anyways even if it a concept or it is for real,it rocks 23:20:03 there is so much radiation from all the devices we use, whats a little IR gonna do? 23:20:08 yah 23:20:19 but i dont want to have to be near a flat surface 23:20:21 radiations differ 23:20:25 i want to be able to type sitting anywhere 23:20:41 the usual light that comes out of the tube is also radiation but isn't harmful for humans 23:21:17 * Lathiat is away (bbl) 23:22:09 i guess you can adjust the size of the keyboard by just moving the PDA farther away 23:29:55 --- quit: revanthn (Connection timed out) 23:30:18 --- nick: revanth -> revanthn 23:32:25 --- quit: malenfant ("sleep...") 23:33:02 i wonder if u have to keep yer fingers raised 23:33:34 air:is that a concept or is it working 23:33:49 it say it has many pratical difficulties 23:34:07 i dont know 23:34:11 besides PDA's or for handheld use and you can't find a flat surface everywhere 23:35:52 * sleep- wonders how they know when your finger hits the surface 23:36:29 good question sleep- 23:36:47 --- join: lodda (~lothar@p508FD92E.dip.t-dialin.net) joined #osdev 23:36:50 hoi 23:37:27 hi lodda 23:37:44 hello revanthn 23:41:40 g2g guys ttyl 23:41:41 --- quit: revanthn () 23:46:04 --- quit: sleep- ("Leaving") 23:49:44 --- quit: Lathiat ("Leaving") 23:59:59 --- log: ended osdev/02.11.05