Are there any default dmd optimizations

[Nick Sabalausky]

On Wed, 27 Feb 2013 11:42:53 +0100
Jacob Carlborg <doob at me.com> wrote:

> On 2013-02-27 00:37, Andrei Alexandrescu wrote:
> 
> > Agreed, but it does happen often that a language feature is later
> > superseded by a generalization thereof.
> 
> In this case it would be two features:
> 
> 1. Allow to run arbitrary code at top level
> 2. Allow to pass a delegate to a function after the parameter list
> 
> void unittest (void delegate () dg)
> 
> unittest {
>      assert(true);
> }
> 
> Would be lowered to:
> 
> unittest({
>      assert(true);
> });
> 
> Then we also can easily support named unit tests:
> 
> void unittest (string name, void delegate () dg)
> 
> unittest("foo") {
>      assert(true);
> }
> 
> Would be lowered to:
> 
> unittest("foo", {
>      assert(true);
> });
> 
> I think it would be nice if D could get better at declarative
> programming.
> 

I like that, but "run arbitrary code at top level" may be a bit of a
problem because it conflicts with allowing forward references.

Ie, for example:

    void foo() { bar(); }
    void bar() { i = 3; }
    int i;

vs:

    void main() {
        void foo() { bar(); }
        void bar() { i = 3; }
        int i;
    }


The first one works, but the second doesn't. And my understanding is
that the second one not working is a deliberate thing related to
not being in a declaration-only context.