www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Casting byte[] to int/long

reply Henrik <zodiachus gmail.com> writes:
Hello!

If I have a dynamic array of bytes, say

byte[] a; a = [1,2,3,4];

Is it possible to somehow cast a number of these bytes into an int or a long,
like so;

int b = cast(int)a[0..2];
long c = cast(long)a;

or something? The goal is to convert 2 * 8 bit into 1 * 16 bit or 4 * 8 bit
into 1 * 64 bit. 1) Is it possible? 2) Is it possible without confusing the
GC? 3) How? ;)
Nov 29 2006
parent reply Bill Baxter <wbaxter gmail.com> writes:
Henrik wrote:

 Hello!
 
 If I have a dynamic array of bytes, say
 
 byte[] a; a = [1,2,3,4];
 
 Is it possible to somehow cast a number of these bytes into an int or a long,
 like so;
 
 int b = cast(int)a[0..2];
 long c = cast(long)a;
 
 or something? The goal is to convert 2 * 8 bit into 1 * 16 bit or 4 * 8 bit
 into 1 * 64 bit. 1) Is it possible? 2) Is it possible without confusing the
 GC? 3) How? ;)



You need something like:
   int b = *cast(int*)&a[0];
   long c = *cast(long*)a.ptr;

--bb
Nov 29 2006
parent reply Gregor Richards <Richards codu.org> writes:
Bill Baxter wrote:

 Henrik wrote:
 

 Hello!

 If I have a dynamic array of bytes, say

 byte[] a; a = [1,2,3,4];

 Is it possible to somehow cast a number of these bytes into an int or 
 a long,
 like so;

 int b = cast(int)a[0..2];
 long c = cast(long)a;

 or something? The goal is to convert 2 * 8 bit into 1 * 16 bit or 4 * 
 8 bit
 into 1 * 64 bit. 1) Is it possible? 2) Is it possible without 
 confusing the
 GC? 3) How? ;)



 
 You need something like:
   int b = *cast(int*)&a[0];
   long c = *cast(long*)a.ptr;
 
 --bb



You probably know this, but be very careful while doing this, endianness 
makes all the difference:

int main()
{
     long a = 10;
     int *b = cast(int *) &a;
     writefln("%d", *b);
}

(Untested) That code will print 10 on a little-endian system, but will 
print 0 on a big-endian system.

  - Gregor Richards
Nov 29 2006
parent Henrik <zodiachus gmail.com> writes:
Gregor Richards wrote:

(Untested) That code will print 10 on a little-endian system, but will




Shit, you are right. I would have forgotten about it :|
std.system to the rescue!
Nov 29 2006