On Friday, 19 June 2015 at 10:13:42 UTC, Ilya Yaroshenko wrote:
>
> You can slice fixed size arrays:
>
>
> auto myFun()
> {
>      float[4096] data;
>      auto tensor = data[].sliced(256, 16);
>      ///use tensor
> }
>

After playing around with some stuff more, I keep finding the syntax for the static and dynamic arrays confusing (more general to D than your code). It just seems weird that you have to treat them differently when applying a function to them.

Also, whenever you have an expanlation point followed by a type that could be an array, you need to make sure to put parentheses around it. For instance,

to!(real[])(x);
instead of
to!real[](x);

or

auto fp = &function!(real[]);

instead of
auto fp = &function!real[];

I wish the documentation made it a bit more clear that you have to do this.

On Saturday, 20 June 2015 at 09:17:22 UTC, Ilya Yaroshenko wrote:
> autodata is hard to understand without HTML documentation.
> Automated documentation based on https://github.com/kiith-sa/harbored-mod
> can be found at
> http://ddocs.org/autodata/~master/index.html,  and it is empty. You may want to read http://dlang.org/ddoc.html
>
> Regards,
> Ilya

Thanks for the info. I have now documented the lib (and its dependencies) and pushed the updates, but the ddocs site seems to be down for me, so I can't see if the docs have made their way to the site.
In any case, you can take a look if you like, once the site is working again.