NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/sys/amiga/programmer/472

Internet Message Format | 1996-08-05 | 5.8 KB

Path: cs.uwa.edu.au!jasonb
From: [e-mail] (Jason S Birch)
Newsgroups: comp.sys.amiga.programmer
Subject: Re: PPC compilers
Date: 8 Jan 96 10:51:43 GMT
Organization: The University of Western Australia
Message-ID: <[e-mail]>
References: <[e-mail]> <4b77tq$[e-mail]> <MQAQx*[e-mail]> <4bqhnf$[e-mail]> <[e-mail]> <4c9i2l$[e-mail]> <4cf0ep$233@ra.i  <4ck47h$[e-mail]> <[e-mail]> <4com6v$[e-mail]>
NNTP-Posting-Host: decadence.cs.uwa.oz.au
X-Newsreader: NN version 6.5.0 #3 (NOV)

[e-mail] (Stephan Schaem) writes:

>Jason S Birch ([e-mail]) wrote:
>: [e-mail] (Stephan Schaem) writes:
>: >Lars Duening ([e-mail]) wrote:
>: >: If a have to access hardware registers (never had to so far), I will
>: >: not do this by direct pointers, but instead define a structure
>: >: representing them. This is then the only place where I have to worry
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>: >: about .c, .w and .l, everywhere else I can write
      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>: >: 'chipregs->blt0con = *b++;' and let the compiler figure the rest out.
      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>: >: HLLs win again here: the compiler will warn if *b is of the wrong
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>: >: type, and if I defined the structure wrong (your .w/.l example), I
      ^^^^
>: >: just have to correct the structure, not the n statements where it is
>: >: used. And the compiler would even take care of bitshifts where
>: >: necessary.

>: > You let the compiler tell you AFTER you write your code about the type?

>: That's not what he said. He said the only place he has to worry about .c, .w
>: and .l is where he's defining his structures (although even then, it would be
>: extremely rare) and after that he can just use his new types, relying on the
>: compiler to give a warning if he's made a mistake.

> I dont think he can declare like this:   chipregs.bltcon0  con0;
> So he will need so 'worry' about it when he declare con0, he will
> read ushort when peeking the structure (Like the asm coder) or
> maybe he typedefed  ushort  chipreg; and will use this instead.
> *b could be a pointer to float for all he cares, dont you think its
> wise to know what data your are manipulating so you can declare your
> variable correctly?!?!

Again: *If* he's trying to create a data type with certain
implementation-dependent size-requirements (such as in the bltcon0
case, assuming no headers to give us one) *then*, at that point in
time, he has to know the size of the type in order to define it
correctly. He stated that earlier - check what I've underlined.
However, once done, he can then safely create variables of that type
(the abstract one he's defined, which happens to be the right size) and
safely assign appropriate values to them, without having to worry if
the eventual assignments are implemented as a .c, .w, or .l. He can
safely rely on the compiler to produce the correct code, and warn him
if he's trying to assign an inappropriate value. *b must be a pointer
to the correct type that he's defined, and as he said, the compiler
will warn him if *b is the wrong type. That's all.

>: He needs to know how he's defined it, yes - but not whether it turns into a .c,
>: .w, or .l. It could be written as a UCHAR, a LONG, or even typedef'd to a
>: BILBY. Read it again.

> He needs to know how he declared it too... meaning he has to know when
> he write the code what is the type on bltcon0.. a uword or a BILBY (typdef
> for uword)... And if he as a few brain cell left he will remember that
> when he use the variable. EX when I declare float, long in a function
> I know wich is wich otherwise I might store a float temp result
> in a long and bang!

The question is not about whether you need to know if a variable is a
float or a long, since they are "abstract" types (and can be different
sizes on different machines). The question is whether you need to know
how many bits are in each, and what each bit means. In assembler, you
do need to know the former, and occasionally the latter.

>: > Yea, I agree... move.w (ax)+,(a1)+ just mean : "Copy 2 bytes from a0 to
>: > a1, and increment both address by 2"
>: > In C the equivalent to move.? (a0)+,(a1)+ is
>: > A++ = B++;
>: > And that beying defined somewhere:
>: > ushort    *A,*B;

>: That's only the C equivalent on certain platforms. C itself says nothing about
>: the actual size of its various types, other than a relative ordering.

> ANSI define very well the minimum size of its type.

It does *not* define that ushort is the same as two bytes, which is
what you are claiming above as being equivalent. It only guarantees
that char <= short <= long.

> Yes, thats very true... in asm you only have fixed type. But typedef is
> not really a part of the compiler but preprocessor, and could be easely
> integrated to an assembler.(But its not, so its hell when you modify
> structure)

>#typedef .w .chipreg

> move.chipreg    (GLOBAL_x1,a4),(CHIPREG_bltcon0,a1)

typedef really is part of the compiler, not the preprocessor.
Unfortunately, given the notice C takes of typedefs you've made, it
may as well be part of the preprocessor. :-(
(ie. typedef xxx foo;
     typedef xxx bar;

     foo x;
     bar y;

I would hope x = y to at least give a warning, even if they are
actually the same type.)

> Stephan

--
Jason S Birch                        ,-_|\ email: [e-mail]
Department of Computer Science      /     \ Tel (work): [phone]
The University of Western Australia *_.-._/ Fax (work): [phone]
Nedlands  W. Australia  6907             v  Tel (home): [phone]