Is this valid:

  ulong length = 1;
  ubyte[8] test;

  test[] = cast(ubyte[])length;
or
  test = cast(ubyte[])length;
or
  test[] = cast(ubyte[])length[0..8];
or
  ?

if not, why not?


what would it do exactly...

- convert length from a ulong to a ubyte and assign to test[0], test[1], test[2], ..etc..
- point ubyte's data pointer at length (or a copy of length).

I'm hoping for the latter because as ulong is 64 bit and ubyte is 8 bit and ubyte[8] is 64 bits worth of data it could do it.

Without this possibility you're left going:
  memcpy(&test[0],length,length.sizeof);

which, while not bad, is not true D style.

Thoughts?

Regan.

-- 
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/

"Regan Heath" <regan@netwin.co.nz> wrote in message news:opr9tddfna5a2sq9@digitalmars.com...
> Is this valid:
>
>    ulong length = 1;
>    ubyte[8] test;
>
>    test[] = cast(ubyte[])length;
> or
>    test = cast(ubyte[])length;
> or
>    test[] = cast(ubyte[])length[0..8];
> or
>    ?
>
> if not, why not?
>

They're just not compatible types. You can't cast a struct to an int either. You can get around it using a pointer:

test[] = (cast(ubyte*)&length)[0 .. length.sizeof];

On Fri, 18 Jun 2004 21:49:52 -0400, Vathix <vathixSpamFix@dprogramming.com> wrote:
> "Regan Heath" <regan@netwin.co.nz> wrote in message
> news:opr9tddfna5a2sq9@digitalmars.com...
>> Is this valid:
>>
>>    ulong length = 1;
>>    ubyte[8] test;
>>
>>    test[] = cast(ubyte[])length;
>> or
>>    test = cast(ubyte[])length;
>> or
>>    test[] = cast(ubyte[])length[0..8];
>> or
>>    ?
>>
>> if not, why not?
>>
>
> They're just not compatible types. You can't cast a struct to an int either.

Do you mean "an int to a struct"? Why not? All you're saying is pretend this is one of these, all it has to do is point at the address given and interpret it as the new type.


So I don't see why this:

  test[] = cast(ubyte[])length;

should not be allowed, it's explicit, and it makes sense that it should do exactly this:

  test[] = (cast(ubyte*)&length)[0 .. length.sizeof];


Given that you know the size of each type you could cast any array of any basic type to any other array of another basic type, or even a user defined type.

Casting from a small type to a large type creates the case where there isn't enough data to make up a complete last item, in this case you can either pad or throw an "Error: lengths don't match for array copy" error.

I think this is very useful.

> You can get around it using a pointer:
>
> test[] = (cast(ubyte*)&length)[0 .. length.sizeof];

Thanks. This will help for now. But pointers really aren't the D thing.

Regan.

-- 
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/

Regan Heath wrote:

> On Fri, 18 Jun 2004 21:49:52 -0400, Vathix <vathixSpamFix@dprogramming.com> wrote:
>>
>> They're just not compatible types. You can't cast a struct to an int either.
> 
> Do you mean "an int to a struct"? Why not? All you're saying is pretend this is one of these, all it has to do is point at the address given and interpret it as the new type.

Arrays in D are just a struct with a length field and a pointer to the data. So what you need to do is change the pointer so that it points to the address where the ulong is stored.


-- 
Julio César Carrascal Urquijo
http://jcesar.f2o.org/

On Sat, 19 Jun 2004 12:31:57 -0500, Julio César Carrascal Urquijo <adnoctum@phreaker.net> wrote:

> Regan Heath wrote:
>
>> On Fri, 18 Jun 2004 21:49:52 -0400, Vathix <vathixSpamFix@dprogramming.com> wrote:
>  >>
>>> They're just not compatible types. You can't cast a struct to an int either.
>>
>> Do you mean "an int to a struct"? Why not? All you're saying is pretend this is one of these, all it has to do is point at the address given and interpret it as the new type.
>
> Arrays in D are just a struct with a length field and a pointer to the data. So what you need to do is change the pointer so that it points to the address where the ulong is stored.

Yes. And calculate the length. Is there an official way to do this? Or are you suggesting an ugly hack?

Regan.

-- 
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/