68k details

[Megol]

Originally Posted by meynaf

And you pretend it is PORTABLE C ? This kind of hack is everything but portable !

As it is indeed portable C I do claim that.

And the registers ? How do you save/reload them in portable C ?
Registers are completely 100% dependent of cpu architecture !

At least try to have a basic level of knowledge before trying to discuss something. I already told you BTW.

[meynaf]

Originally Posted by Megol

As it is indeed portable C I do claim that.

By definition hacks are not portable. Else they wouldn't be hacks.

Originally Posted by Megol

At least try to have a basic level of knowledge before trying to discuss something. I already told you BTW.

No you did not tell anything and your sentence is insulting (as usual ?).

[kamelito]

Originally Posted by Bruce Abbott

Soon there may be another choice. I am working on an interactive disassembler that - most of the time - does get it 100% right, and only needs a few mouse clicks to change the type when it is wrong.

Great news, please make it happens

[Thorham]

That you can get all kinds of things to work with hacks is well known. Now the question is whether or not you can do an OS in portable C without any hacks and do it properly? I'm betting on no.

[plasmab]

realistically you get a C compiler to do the heavy lifting..

e.g. gcc

{
register int sp asm ("sp");
printf("%x", sp);
}

Now thats going work on any CPU which the gnu assembler calls the stack pointer "sp".... which i bet is a probably a lot of them. They often have an alias for it..

Now its absolutely not standard C89/C99 etc. but its portable on the same compiler family.

So split hairs about whether its real C etc. fine i dont care about that argument. i accept its not standard. And sure you get crapper/slower code etc. Compilers all do this different so you end up with hacks on a per compiler basis. But they all do this and you have hacks if you try and make portable assembler that works on all assemblers..

[Thorham]

So it can be done but it's not very useful. Who made a point out of this anyway? There doesn't seem to be much of a point

[plasmab]

Originally Posted by Thorham

So it can be done but it's not very useful. Who made a point out of this anyway? There doesn't seem to be much of a point

Not I. I'm very bored now.

There is pretty much ways to do all the assembly marshalling you need for an OS in C but they are compiler family specific. Which is absolutely fine by me.

BTW: I compiled the above code on x86, ARM and 68K then i got bored.

[Thorham]

Originally Posted by plasmab

There is pretty much ways to do all the assembly marshalling you need for an OS in C but they are compiler family specific. Which is absolutely fine by me.

I'm willing to bet that they don't do this in Linux and BSD. Best way seems to do it by hand.

[alpine9000]

I wrote a multi tasking OS in C, but yeah the context switches had to be asm. And a few other little things like atomic operations for locks etc.

[plasmab]

Originally Posted by Thorham

I'm willing to bet that they don't do this in Linux and BSD. Best way seems to do it by hand.

NetBSD is a bit purist this way... maybe there. but yes. its a bit stupid to try 100% C and its a bit stupid to do 100% ASM. Happy medium.

But its wrong to say you cant do this stuff at all because the compiler writers do help you out a lot.

Originally Posted by alpine9000

I wrote a multi tasking OS in C, but yeah the context switches had to be asm. And a few other little things like atomic operations for locks etc.

Yes i cannot see how you do SMP locks in pure C (but happy for someone to point how to do it).

Anyways.. im so bored im going to watch Forrest Gump.

[meynaf]

Originally Posted by plasmab

{
register int sp asm ("sp");
printf("%x", sp);
}

Not at all my definition of "portable C". There is even the string "asm" in it

[plasmab]

Originally Posted by meynaf

Not at all my definition of "portable C". There is even the string "asm" in it

You are entitled to that opinion. I knew would be your only way of defending the lack of knowledge you demonstrated here.

It’s portable on all GCC compliers. VBCC uses slightly different syntax. All you’re pointing out is that C989/C99 doesn’t have a defined standard for it. It’s possible that future standards will.

Like I say different assemblers use different syntax notation. There is no assembly standard so your argument is laughable

[meynaf]

Originally Posted by plasmab

You are entitled to that opinion. I knew would be your only way of defending the lack of knowledge you demonstrated here.

I'm not more entitled with my opinion than you with yours. And again you return against me your own lack of knowledge, because you are completely unable to counter any of my arguments.

Originally Posted by plasmab

It’s portable on all GCC compliers. VBCC uses slightly different syntax. All you’re pointing out is that C989/C99 doesn’t have a defined standard for it. It’s possible that future standards will.

Using disguised asm is doable on just about every compiler. It's still asm and not "portable C".
Now of course several folks here pretend an OS can be fully written in portable C and even has been, but where is the source code showing this ?

Originally Posted by plasmab

Like I say different assemblers use different syntax notation. There is no assembly standard so your argument is laughable

The asm syntax has absolutely nothing to do with my arguments here !
It's your above sentence that's laughable.


But anyway what was your point originally ? It was that the type of CPU does not matter. But as "portable" things require at least some amount of inefficient hack leading to code bloat, it's the proof that the CPU type actually *does* matter. Not wanting to be proven wrong, you attempted to redirect the discussion to something else. Nice try.

[plasmab]

You say I am unable to counter your argument? Ok you said you can’t make portable C that access registers. I posted a damn code example and you put your fingers in your ears and say “la la la.. I can’t hear!”

It’s actually now defined in the C11 spec.

No matter what anyone says you will argue the opposite.

And my point about the assembly was you have more hacks make an OS work on multiple assemblers on the same platform than multiple C compilers on different platforms. But that went to miles over your head. Really your level isn’t even worth engaging with anymore.

[meynaf]

Originally Posted by plasmab

You say I am unable to counter your argument? Ok you said you can’t make portable C that access registers. I posted a damn code example and you put your fingers in your ears and say “la la la.. I can’t hear!”

It’s actually now defined in the C11 spec.

Now try to access FPU, XMM, YMM registers in "portable" way. How can the access to something not existing on all machines be "portable" ?
You're doing cherry-picking here, just finding a small, useless example where it remotely appears to work. Pityful.

Originally Posted by plasmab

No matter what anyone says you will argue the opposite.

Try saying intelligent things.

Originally Posted by plasmab

And my point about the assembly was you have more hacks make an OS work on multiple assemblers on the same platform than multiple C compilers on different platforms.

Which you didn't say explicitly and is just wrong anyway.

Originally Posted by plasmab

But that went to miles over your head. Really your level isn’t even worth engaging with anymore.

Look up. Not down. Really your level is too low and i don't want to be bring down there.


I have disassembled megabytes of code, often compiled, and ported several games from one computer to another without even having the source code.
(Not only 68k, i did some 6502 too and would be doing x86 if it weren't so bad.)
Of course copy protection had to be removed in the process.
I have written whole utilities like picture viewer in 100% asm.
I am using my own disassembler, my own debugger.

And now it will be *my* lack of knowledge ?
Especially from people who apparently do mainly - if not only - hardware ?
This is very poor and very stupid.

I really wish i could find a suitable reply to such aggressive nonsense other than clicking "report post". Sigh.
(As it is obvious writing normal argumentation doesn't work.)

[plasmab]

For the sake of finishing this I will concede that I know nothing about computers. My bachelors degrees is electronics and my PhD in physics..

My issue is that you make sweeping statements that can be proven wrong very easily. Even by someone as stupid as me. You also have lied a couple of times here. Like when you claimed AmigaOS had the best interrupt latency.

You are getting upset because people are calling out your sweeping generalisation and lies. Maybe/probably you have some facts further down the line that are worth talking about but we never get to them because you are so imprecise and sweeping with what you say that people get stuck on those.

Finally when you fail with your arguments and get frustrated you are resorting to chest beating / “d$ck swinging” about your achievements.

Seriously, talk about this stuff objectively and with facts and you will get better engagement and the respect you probably deserve deep down. I tried to tell you this ~100 posts ago.

[meynaf]

Originally Posted by plasmab

My issue is that you make sweeping statements that can be proven wrong very easily. Even by someone as stupid as me.

So far nothing has been "proven wrong". So what are you waiting ? Just do it.
It's not me doing "sweeping statements". You're reversing the charge of the proof here. You can prove something exists by actually finding one and showing it. You can't objectively prove it doesn't exist.

Originally Posted by plasmab

You also have lied a couple of times here. Like when you claimed AmigaOS had the best interrupt latency.

And to the strawman fallacy you go again ! (And also to abusive generalization, by giving one example and saying there were others.)
I've never claimed AmigaOS had the best interrupt latency.
Just that it had the best context switching speed of the systems i tested some years in the past - and i don't pretend having tested them all, so far not.

Originally Posted by plasmab

You are getting upset because people are calling out your sweeping generalisation and lies. Maybe/probably you have some facts further down the line that are worth talking about but we never get to them because you are so imprecise and sweeping with what you say that people get stuck on those.

No I am getting upset because you keep on insulting me instead of just speaking about the subject.
The worse i concede is this : you assert things without proof and i dismiss them without proof.

Originally Posted by plasmab

Finally when you fail with your arguments and get frustrated you are resorting to chest beating / “d$ck swinging” about your achievements.

This was just to show how much your insults were ridiculous.

Originally Posted by plasmab

Seriously, talk about this stuff objectively and with facts and you will get better engagement and the respect you probably deserve deep down. I tried to tell you this ~100 posts ago.

Facts ? What facts ? Again, how can I prove something can not be done ???
This is impossible. The burden of the proof isn't on my side.

[meynaf]

So far i have seen :
- personal attacks
- strawman fallacy
- cherry picking
- reversing the burden of the proof
- abusive generalization

I wonder what will come next ?

EDIT:
Forgot this one :
- red herring

[plasmab]

Ok I’m done. It’s pointless talking to someone with their fingers in their ears.

[Megol]

Originally Posted by meynaf

So far i have seen :
- personal attacks
- strawman fallacy
- cherry picking
- reversing the burden of the proof
- abusive generalization

I wonder what will come next ?

EDIT:
Forgot this one :
- red herring

Yes I wonder what you will bring next. Pot kettle etc.