| |
| |
| |
|
Page: 1 2 3 4 5 6 7
Comments:
<0> yaaawn and streeeeeetch
<0> ****, middle east is totally blowing up
<0> this will spread
<1> indeed
<0> geist, do you have an electronic copy of the ddj article you wrote?
<0> searching for it has been futile
<1> no
<1> it's on teh web, we looked it up the other day
<1> december 2001 iirc
<0> oh, ddj has opened up
<0> yay, found it on ddj - april 2002
<1> it's not that great
<0> nice overview
<1> it's okay. frankly I expected them to have a bunch of constructive criticism when I wrote it, but when I sent it to them they pretty much printed it verbatim
<1> the tough part was keeping it under the size limit
<0> right
<0> i'm thinking i can solve the object-recursive spinlock problem by using the object id (or address) + refcount as the lock flag rather than 1
<0> this requires coffee
<2> Would anyone here know how to get the current OS from gmake?
<3> OS = $(shell uname -s)
<2> that's what I was missing, sh
<2> ty
<4> Why do it in gmake?
<4> If you have configure, export the the variable there.
<2> We don't use configure to build the OS
<2> and since it's in a repo, it's easier to do that than to revert the changes to a config file each time
<4> nod, cool
<2> wouldn't happen to know about conditional statements in gmake would you?
<4> .
<2> I get /bin/sh: -c: line 0: syntax error near unexpected token `"Linux","FreeBSD"'
<2> from: ifeq ("$(OS)","FreeBSD")
<2> and that's pretty much identical to the manual's example
<0> Calm down with http://www.youtube.com/watch?v=Mo0XUdPHdKA&search=royksopp
<0> or even better: http://www.youtube.com/watch?v=04nuJnNDf8I&search=royksopp
<0> Poor Leno...
<5> Cody`: Try without the quotes?
<2> I have, but I can again
<2> same error
<5> Beyond that, maybe $(OS) isn't set correctly?
<5> Oh, guess it is. Hrm...
<2> meh I'll just make a bsd specific rule
<5> Ah. I think it's expecting /bin/sh to be bash.
<2> I'm testing it on bash
<2> I'm on linux atm
<6> hi
<0> Cody`, do as any sane person would do and make your own build system.
<6> lol
<0> meta
<0> and the mystery unravels
<0> iret
<1> yawn
<1> pretty quiet here tonight
<7> Whee.
<7> Now I get to experience the joy of bounce buffering.
<1> excellent
<1> i wish I could go on about all the fun work stuff
<7> But if you did the thought police would come and take you away?
<1> pretty much
<1> KHAN!!!!!
<7> From hell's heart, I stab at thee.
<8> morning
<1> morning ronny
<8> im off - next exam is in 1.5 hours
<8> automates and formal languages :/
<9> i don't suppose nyef's awake atm?
<5> No, I'm not. ^_-
<1> absolutely not
<9> lol
<5> What's up?
<9> have u looked at http://www.forthos.org/ nyef? i just gave 'forthos' a googling, and it seems fairly complete... ive just read the main page so far so i don't know though :)
<5> No, that's a new one by me...
<9> hmmm if it actually has protected mode multitasking i'd consider using it... especially because i have an old 386 in my room and i wanna make it do something cool :D
<9> not so sure about it's OO extensions though
<5> I have about a third of this feature set implemented, another third planned, and the last third is uninteresting to me.
<9> i'd rather something like CLOS - generic words sounds a lot better than the hackish C++ style OO it seems to use
<9> yeah if i use it, i'd cut it back a lot and re-implement some
<5> The problem with generics is that you really need type information.
<9> yeah - but you don't really *need* to guess the type of something that isn't an object
<9> hmmm
<9> wait
<9> yeah that wouldn't really work
<9> because you'd have to know if something on the stack is an int or a float or a pointer to an object before you go looking for cl*** info
<5> Is this 0x418327 an integer or a pointer to a string?
<9> but still - i'd prefer something that looks like a normal forth word... forthos.org's OO seems to have lotsa confusing looking grammar
<5> I'm completely unconvinced that OO is a good thing for Forth.
<9> even if only the first argument is tested for type
<9> if it was fairly transparent it *could* be
<9> but yeah it sorta screws with forth's simplicity
<1> yeah, it's like the brain**** JIT. I mean why mess with something that's already perfect?
<9> haha
<5> ... A vi-like block editor, and an emacs-like input line editor? WTF?
<9> haha
<7> Object... oriented... forth?
<9> deep... fried... mars bar?
<7> Ham... dog?
<9> pota... to?
<10> pizza... milkshake?
<5> Heh. And ForthOS is incapable of doing its own installation. Neat.
<9> i'd like to see Unix re-implemented in forth, and compare it's crappiness to C Unices
<5> Heh.
<5> I was actually thinking of grafting a hypervisor onto my Forth system and using it to try and run linux usermode processes.
<5> Just to see how far I could get with it.
<9> if you made an entirely OO forth - where everything on the stack are pointers to objects... i can't imagine it would be any worse than other strictly OO languages
<9> it would still be easier to implement than other such languages too
<5> I'm wondering if I should redo my Forth system from scratch. Get a hosted Forth system and use it to cross-compile a kernel and whatnot.
<9> as in write the kernel in forth?
<5> Well, my existing Forth system metacompiles itself.
<5> Except for the bootsector, that is. It just gets copied from build to build.
<9> i had a play with it this morning btw
<5> Oh?
<1> 4796259934
<9> i can't really say much about it - i don't know enough forth to do anything useful :) but i defined a word and stuff - it worked at least
<5> Ah, yeah.
<1> 1238471623
<5> There's a block editor in it that you can get to by running 'edit2'.
<5> That's at least halfway impressive.
<9> cool
<5> Another thing I was thinking I might try is an all-ASM OS of some sort.
<9> mm that could be interesting
<9> but be sure to have some structure to it... don't go and do a menuetos
<9> *shudder*
<9> i cannot believe there is a web browser for it that renders apple.com :|
<1> yeah totally
<5> That's actually kindof funny.
<1> hacking in asm doesn't excuse you from writing good code
<1> if anything it's that much more important
<9> yeah
<5> Believe me, I know.
<9> also these days it's not so much CPU speed that's the bottleneck anyway
<1> exactly
<9> and it's better to use a clever algorithm and do it in C than use a ****ty one that's easier to implement, and optimize it to hell in asm
<5> So many project ideas, so little time...
<9> same
<9> plus im too lazy :)
<9> SchemeOS is still tempting
<9> mostly i just cbf implementing or porting a scheme for it though
<5> "cbf"?
<9> plus i'm worried about scheme's memory usage... i guess it's just it's high level nature, but it [or some implementations] apparently eats up as much RAM as java - whereas D and other GC languages don't use as much
<9> cant be ****ed
<5> Ah.
<5> The more I think about it, the more fun an all-asm os seems.
<9> i'd of thought the opposite
<1> yeah that's pretty much my feeling
<9> it's so... static and... crap...
<5> Okay, an all-asm -kernel-, then.
Return to
#osdev
or
Go to some related
logs:
#bash
open application/x-httpd-php
#perl
#css
perl iselement
setfont+fedora+dataflex
who owns livemercial
#sql
#mysql
gentoo xstart
|
|