This compiles and runs but I think it is a bug.
---
import std.stdio;
mixin template TestMixin(T) {
    int testMixin;
}
struct Test {
    int testMixin;
    mixin TestMixin!Test;
}
void main() {
    Test t;
    t.testMixin = 10;
    writefln("%s", t);
}
---
$ ./hack
Test(10, 0)

I can also change the struct to be:
---
struct Test {
   string testMixin;
   mixin TestMixin!Test;
}
// and the assignment in main() to
t.testMixin = "some string";
---
$ ./hack
Test("some string", 0)

so whether this is a bug or not, I think it is bad. Any thoughts?

G.

The symbols that are introduced by mixin template never override/conflict
with the formally defined symbols.
It's a designed behavior.

Kenji Hara


2013/9/23 growler <growlercab@gmail.com>

> This compiles and runs but I think it is a bug.
> ---
> import std.stdio;
> mixin template TestMixin(T) {
>     int testMixin;
> }
> struct Test {
>     int testMixin;
>     mixin TestMixin!Test;
> }
> void main() {
>     Test t;
>     t.testMixin = 10;
>     writefln("%s", t);
> }
> ---
> $ ./hack
> Test(10, 0)
>
> I can also change the struct to be:
> ---
> struct Test {
>    string testMixin;
>    mixin TestMixin!Test;
> }
> // and the assignment in main() to
> t.testMixin = "some string";
> ---
> $ ./hack
> Test("some string", 0)
>
> so whether this is a bug or not, I think it is bad. Any thoughts?
>
> G.
>

On Monday, 23 September 2013 at 12:08:37 UTC, Kenji Hara wrote:
> The symbols that are introduced by mixin template never override/conflict
> with the formally defined symbols.
> It's a designed behavior.
>
> Kenji Hara
>
>

Thanks Kenji for the information. I guess I'd better go back and read the docs again...and scurry off back to D.learn :D

G.