How to flatten N-dimensional array?

[Pavel Shkadzko]

On Sunday, 24 May 2020 at 11:21:00 UTC, 9il wrote:
> On Saturday, 23 May 2020 at 18:15:32 UTC, Pavel Shkadzko wrote:
>> [...]
>
> If the common nd-array isn't jugged (a parallelotop), you can 
> use fuse function.
>
> ----------
> /+dub.sdl:
> dependency "mir-algorithm" version="~>3.8.12"
> +/
> import std.stdio: writeln;
> import mir.ndslice;
>
> void main() {
>     auto arr =
>     [[[0, 1, 2, 3, 4],
>       [5, 6, 7, 8, 9]],
>      [[10, 11, 12, 13, 14],
>       [15, 16, 17, 18, 19]],
>      [[20, 21, 22, 23, 24],
>       [25, 26, 27, 28, 29]]];
>     auto flatten = arr.fuse.field;
>
>     static assert(is(typeof(flatten) == int[]));
>    	assert(flatten == 30.iota);
> }
> ----------
>
> It performs exactly one allocation.

Yep, I know about Mir fuse. I was more wondering about the 
absence of flatten in Phobos. But on dlang forum some people 
claimed that when it comes to anything multidimensional, use Mir.