Anybody tested DOS/65 so far ? Interested in feedback... - CP/M

This is a discussion on Anybody tested DOS/65 so far ? Interested in feedback... - CP/M ; Hi everybody, Richard finished his work so far to create a working copy of his DOS/65 (his CP/M-alike 6502 based OS), I've posted several times news about it. Ruud as far as I could remember was also very interested, but ...

+ Reply to Thread
Results 1 to 11 of 11

Thread: Anybody tested DOS/65 so far ? Interested in feedback...

  1. Anybody tested DOS/65 so far ? Interested in feedback...

    Hi everybody,

    Richard finished his work so far to create a working copy of his DOS/65
    (his CP/M-alike 6502 based OS), I've posted several times news about it.
    Ruud as far as I could remember was also very interested, but I've heard
    nothing about it then. So I am interested in getting feedback about his
    implementation and the idea.

    Regards
    Peter

    --
    ** Interested in DOS/65 ? See http://www.z80.eu/dos65.html **

  2. Re: Anybody tested DOS/65 so far ? Interested in feedback...



    "Peter Dassow" wrote in message
    news:481ff830$0$6542$9b4e6d93@newsspool3.arcor-online.net...
    > Hi everybody,
    >
    > Richard finished his work so far to create a working copy of his DOS/65
    > (his CP/M-alike 6502 based OS), I've posted several times news about it.
    > Ruud as far as I could remember was also very interested, but I've heard
    > nothing about it then. So I am interested in getting feedback about his
    > implementation and the idea.


    I like it but it needs PIP for that CP/M feel!

    Tom Lake


  3. Re: Anybody tested DOS/65 so far ? Interested in feedback...

    Tom Lake wrote:
    >> Richard finished his work so far to create a working copy of his
    >> DOS/65 (his CP/M-alike 6502 based OS), I've posted several times news
    >> about it. So I am interested in getting feedback
    >> about his implementation and the idea.

    >
    > I like it but it needs PIP for that CP/M feel!
    >

    Tom, there is a COPY command which is in fact much more sophisticated
    than PIP and I am sure you will like it.

    Regards
    Peter

  4. Re: Anybody tested DOS/65 so far ? Interested in feedback...

    On May 9, 1:36*pm, Peter Dassow wrote:
    > Tom Lake wrote:
    > >> Richard finished his work so far to create a working copy of his
    > >> DOS/65 (his CP/M-alike 6502 based OS), I've posted several times news
    > >> about it. So I am interested in getting feedback
    > >> about his implementation and the idea.

    >
    > > I like it but it needs PIP for that CP/M feel!

    >
    > Tom, there is a COPY command which is in fact much more sophisticated
    > than PIP and I am sure you will like it.
    >
    > Regards
    > * Peter


    Is it like PPIP ... that is, accepting both:

    PPIP a4:hforthm4.com a0:hforthm.com
    and
    PPIP a0:hforthm.com=a4:hforthm4.com

    That's sophisticated.


  5. Re: Anybody tested DOS/65 so far ? Interested in feedback...

    Peter Dassow wrote:
    [...]
    > I didn't expect such an answer because I asked for another thing, but
    > .... anyway ... I've started the compiler and compiled a (very) small
    > program, but I didn't developed a whole program suite with it yet ;-)


    What's the code quality like?

    I ask because the ACK has a 6502 backend, targeted for the BBC Micro,
    but I haven't made any attempt to get it into working order because my
    initial experiments showed the code was laughably bad --- mostly due to
    the 6502 being entirely unsuited for running C-like languages on.

    But if there's any interest it ought to be easy enough to get it up and
    running again.

    --
    ┌─── dg*cowlark.com ───── http://www.cowlark.com ─────
    │ "I have always wished for my computer to be as easy to use as my
    │ telephone; my wish has come true because I can no longer figure out
    │ how to use my telephone." --- Bjarne Stroustrup

  6. Re: Anybody tested DOS/65 so far ? Interested in feedback...

    --{ David Given a plop ceci: }--

    >
    > I ask because the ACK has a 6502 backend, targeted for the BBC Micro,


    Did you mean "Amsterdam Compiler Kit" ? And if the answer is "Yes",
    did you have an url for more informations ?


    --
    > apparemment Gopher ne fonctionne pas avec Outlook Express


  7. Re: Anybody tested DOS/65 so far ? Interested in feedback...

    On May 10, 5:39 am, "Thierry B." wrote:
    > --{ David Given a plop ceci: }--
    > > I ask because the ACK has a 6502 backend, targeted for the BBC Micro,


    > Did you mean "Amsterdam Compiler Kit" ? And if the answer is "Yes",
    > did you have an url for more informations ?


    Seems like it, because BBC is the 6502 target that the ACK docs
    mention.

    http://tack.sourceforge.net/

  8. Re: Anybody tested DOS/65 so far ? Interested in feedback...

    BruceMcF wrote:
    [...]
    > Seems like it, because BBC is the 6502 target that the ACK docs
    > mention.
    >
    > http://tack.sourceforge.net/


    Yup, that's it.

    I haven't touched it for a while because it will, if I let it, soak up
    every spare moment I've got, and because it turns out the register
    allocator is sufficiently naff that it's not really worth using on RISC
    machines, but it does work, and produces decent enough binaries on the
    various supported architectures.

    If you want 6502 support, you'll need to get the old 5 release rather
    than the new 6 release I'm doing.

    --
    ┌─── dg*cowlark.com ───── http://www.cowlark.com ─────
    │ "I have always wished for my computer to be as easy to use as my
    │ telephone; my wish has come true because I can no longer figure out
    │ how to use my telephone." --- Bjarne Stroustrup

  9. Re: Anybody tested DOS/65 so far ? Interested in feedback...

    On May 9, 8:26*pm, David Given wrote:
    > initial experiments showed the code was laughably bad --- mostly due to
    > the 6502 being entirely unsuited for running C-like languages on.
    >


    You should have told me this before I wrote all that C code on the
    Apple IIe and C64. As a matter of fact some of that unsuitable code,
    even some which I got paid to write, is quite available.

    Here take a look here:

    http://www.commodorefree.com/magazine/issue18.pdf

    I wonder though what kind of languages you might consider are suited
    to the 6502. Or did you mean "the 6502 being entirely insulted by
    running on C-like languages"

    Yes I too have found that some folks who like to write in BASIC
    augmented by machine language seem to recoil at the nicely structured
    and organized C programming environment on 6502 machines. That is
    really too bad and their loss.

    Cheers,

    Bill

  10. Re: Anybody tested DOS/65 so far ? Interested in feedback...

    Bill Buckels wrote:
    [...]
    > You should have told me this before I wrote all that C code on the
    > Apple IIe and C64. As a matter of fact some of that unsuitable code,
    > even some which I got paid to write, is quite available.


    I never said it was impossible. I simply said that the 6502 was
    extremely unsuited for it (with a certain amount of hyperbole).

    It's much the same problem as with the Z80, but more so --- the
    designers never thought that people would want to do stack-relative
    loads and stores as frequently as stack-based languages like C do. The
    fact that the stack's limited to 256 bytes is also a bit of a problem!

    If anyone's interested, the ACK works round all of this by using an
    external stack. This is largely because of architectural restrictions
    that enforce a particular stack frame format that assumes sizeof(sp) ==
    sizeof(void*) --- any custom compiler will be able to do better.
    Unfortunately, this makes certain infrequently used operations like
    'return from subroutine' insanely complicated:

    http://tack.cvs.sourceforge.net/tack....2&view=markup

    It doesn't help that it's using this external stack *in addition to* the
    CPU stack, rather than instead of. This is the longjmp code:

    http://tack.cvs.sourceforge.net/tack....2&view=markup

    This is not how I would have done it. Needless to say, I would not
    recommend the use of this code by anyone.

    > http://www.commodorefree.com/magazine/issue18.pdf


    ....contains no technical information at all, I'm afraid (or at least
    that I could find wading through the PDF). What sort of code did Aztec C
    produce? How did it deal with the stack issue?

    (Thinking about it, it ought to be possible to do a real stack-relative
    load with something like:

    tsx
    lda 0x0200+offset,X
    ldy 0x0201+offset,X

    ....leaving the result in AY; this is only 7 bytes, which is comparable
    to the Z80's 6. This technique may even allow you to perform 2op and 3op
    operations sourcing directly from the stack, rather than having to load
    the values and copying them into zero page. Hmm. Of course, it still
    doesn't get round the very limited stack size; and both processors
    suffer hugely from the fact that doing this sort of thing takes loads of
    instructions, when they've only got 64kB of addressable memory. A
    frickin' *ARM* can do it in 4 bytes. 2, if you're talking about Thumb.)

    (Thinking about it some more, the 6502 would probably be far more
    suitable for a 8-bit int, 16-bit long, 16-bit pointer compiler
    architecture than the Z80; I tried that briefly on the Z80 but
    eventually gave up because I couldn't push or pop bytes to the stack.
    That's not a problem with the 6502. That ought to vastly improve int
    performance at the expense of being hideously non-de-facto-standard.)

    --
    ┌─── dg*cowlark.com ───── http://www.cowlark.com ─────
    │ "I have always wished for my computer to be as easy to use as my
    │ telephone; my wish has come true because I can no longer figure out
    │ how to use my telephone." --- Bjarne Stroustrup

  11. Re: Anybody tested DOS/65 so far ? Interested in feedback...

    On May 13, 9:14 pm, David Given wrote:
    > If anyone's interested, the ACK works round all of this by using an
    > external stack.


    Yes, that's how you do it, with post-indexed zero-page addressing,
    LDA (z),Y
    and relatives. Its the 65816 that is designed for a straightforward C
    implementation, with a 64K stack and
    LDA 15,S
    and
    LDA (15,S),Y
    address modes. The 6502 was designed long before suitability for
    compiled C code was a consideration ... it was popular because its
    small transistor count and early production process led to high yields
    which for a period made it a very attractive price/performance
    proposition. However, that low transistor count comes at a price.

    > (Thinking about it, it ought to be possible to do a real stack-relative
    > load with something like:


    > tsx
    > lda 0x0200+offset,X
    > ldy 0x0201+offset,X


    > ...leaving the result in AY; this is only 7 bytes, which is comparable
    > to the Z80's 6.


    More normally results will be on the zero page, and the stack relative
    load would be integrated with an operation ...

    TSX
    CLC
    LDA W
    ADC $100+offset,X
    STA W
    LDA W+1
    ADC $101+offset,X
    STA W+1

    .... and of course doing everything a byte at a time is another reason
    why the 65816 is the more natural C-oriented processor in the
    family ... when in 16-bit accumulator mode, a current integer can ride
    in the accumulator for much longer before it needs to come to rest in
    a zero-page location:

    ....
    CLC
    ADC offset1,S
    SEC
    SBC offset2,S
    ....

    .... but there are almost certainly not enough C64+SuperCPU and
    AppleIIgs systems out there to justify a DOS/65816!!!

+ Reply to Thread