Could it be a valid option to have a shorthand template syntax if the template body has only one statement? Something like:

template Tuple(A...) alias Tuple = A;

Another suggestion would be to have the ?: syntax supported for static statements, as we're currently forced to use static ifs. It really adds a lot of code to templates for even simple conditions.

Jeremie Pelletier wrote:
> Could it be a valid option to have a shorthand template syntax if the template body has only one statement? Something like:
> 
> template Tuple(A...) alias Tuple = A;

Aside from that not being how alias works,

template Tuple(A...) alias A Tuple;

template Tuple(A...) { alias A Tuple; }

You're saving two characters; hardly seems worth it, to me.  That said, I can't think of any technical reason why you couldn't change the grammar to allow for a single declaration.

> Another suggestion would be to have the ?: syntax supported for static statements, as we're currently forced to use static ifs. It really adds a lot of code to templates for even simple conditions.

I don't think using ?: for statements is a good idea; it's only ever suitable for very small expressions.

The problem there is that there would be no way to distinguish between a compile-time ?: and a run-time ?: for either the compiler or the programmer.  I have enough grief with CTFE and kinda-static foreach, I don't want to see another construct like that added.

Jeremie Pelletier wrote:
> Another suggestion would be to have the ?: syntax supported for static statements, as we're currently forced to use static ifs. It really adds a lot of code to templates for even simple conditions.


?: already works at compile time, at least in D2. I use it all the time.

  template cat(string a, string b, bool addSpace)
  {
      enum string cat = a ~ (addSpace ? " " : "") ~ b;
  }

  static assert (cat!("foo", "bar", false) == "foobar");
  static assert (cat!("foo", "bar", true) == "foo bar");

Is that what you meant?

-Lars