What's the preferred way in D to implement single-allocation variable-sized arrays such as

/** Single-Allocation Array. */
struct ArrayN
{
    ubyte length;  // <= maxLength
    size room;     // allocated length
    ubyte[0] data; // `room` number of bytes follows
}

where insertion/deletion typically is done via

    ArrayN* pushBack(ArrayN*, ubyte element);
    ArrayN* popBack(ArrayN*);

which, when needed, will reallocate a new larger/smaller `ArrayN`

?

Further, what's the official name for this structure?

On Wednesday, 18 May 2016 at 21:28:56 UTC, Nordlöw wrote:
> What's the preferred way in D to implement single-allocation variable-sized arrays such as
>
> /** Single-Allocation Array. */
> struct ArrayN
> {
>     ubyte length;  // <= maxLength
>     size room;     // allocated length
>     ubyte[0] data; // `room` number of bytes follows
> }
>
> where insertion/deletion typically is done via
>
>     ArrayN* pushBack(ArrayN*, ubyte element);
>     ArrayN* popBack(ArrayN*);
>
> which, when needed, will reallocate a new larger/smaller `ArrayN`
>
> ?
>
> Further, what's the official name for this structure?

In C it's called a variable-length struct or object. I don't think D implements them, but this could probably work:

struct Foo {
    size_t len;

    ubyte[] data() @property {
        auto thisptr = cast(ubyte*)(&this);
        return thisptr[Foo.sizeof..(Foo.sizeof+len)];
    }
}

On Wednesday, 18 May 2016 at 23:45:07 UTC, Alex Parrill wrote:
> In C it's called a variable-length struct or object. I don't think D implements them, but this could probably work:

Thanks!