Converting a ulong to a byte array and constructing a ulong from it
Paul Backus
snarwin at gmail.com
Thu Oct 24 13:50:54 UTC 2019
On Thursday, 24 October 2019 at 13:33:30 UTC, 9898287 wrote:
> What's the function for converting a ulong to a native-endian
> byte array?
> For example,
>
> auto bytes = 0x1234567890123456u64.to_ne_bytes();
> // should yield
> // [0x12, 0x34, 0x56, 0x78, 0x90, 0x12, 0x34, 0x56] in
> big-endian and
> // [0x56, 0x34, 0x12, 0x90, 0x78, 0x56, 0x34, 0x12] in
> little-endian systems
>
> Also, what's the function for constructing a ulong from a
> native-endian byte array?
> For example,
>
> auto value = from_be_bytes!ulong([0x12, 0x34, 0x56, 0x78, 0x90,
> 0x12, 0x34, 0x56]);
> // value is 0x1234567890123456
>
> Rust equivalent:
> https://doc.rust-lang.org/std/primitive.u64.html#method.to_ne_bytes
Use a cast:
ulong n = 0x1122334455667788;
ubyte[] bytes = (cast(ubyte*) &n)[0 .. n.sizeof];
ulong m = *cast(ulong*) bytes.ptr;
assert(m == n);
Note that if you're writing code that cares about the native byte
order, there's a good chance you're making a mistake. Rob Pike
has written a blog post titled "The byte order fallacy"
discussing this in more detail:
https://commandcenter.blogspot.com/2012/04/byte-order-fallacy.html
If you need to convert between native byte order and big/little
endian byte arrays (e.g., for serialization/deserialization),
there are functions in `std.bitmanip` that you can use:
https://dlang.org/phobos/std_bitmanip.html
More information about the Digitalmars-d-learn
mailing list