digitalmars.D.learn - Ambiguous template parameter names
- jmh530 (16/16) May 02 2018 In the function below, there is a template parameter and a normal
In the function below, there is a template parameter and a normal parameter both with the same name. However, the function returns the normal parameter. The template parameter is effectively ignored. I was surprised by this behavior. Is this a bug or intentional? I did not see it documented anywhere. ``` int foo(int val)(int val) { return val; } void main() { assert(foo!1(2) == 2); } ```
May 02 2018
On Wednesday, 2 May 2018 at 20:32:43 UTC, jmh530 wrote:In the function below, there is a template parameter and a normal parameter both with the same name. However, the function returns the normal parameter. The template parameter is effectively ignored. I was surprised by this behavior. Is this a bug or intentional? I did not see it documented anywhere. ``` int foo(int val)(int val) { return val; } void main() { assert(foo!1(2) == 2); } ```i think this is a bug and that it should be reported. The main problem is that you cant even use the val (the template parameter one) in a static if.
May 02 2018
On Wednesday, 2 May 2018 at 20:32:43 UTC, jmh530 wrote:In the function below, there is a template parameter and a normal parameter both with the same name. However, the function returns the normal parameter. The template parameter is effectively ignored. I was surprised by this behavior. Is this a bug or intentional? I did not see it documented anywhere. ``` int foo(int val)(int val) { return val; } void main() { assert(foo!1(2) == 2); } ```It's not a big per se. It's a consequence of the declaration expanding to the real template function form (I can't type it all out as I'm on my phone), thus the inner `val` from the function shadows the one from the template.
May 02 2018
On Thursday, 3 May 2018 at 00:52:58 UTC, Meta wrote:[snip] It's not a big per se. It's a consequence of the declaration expanding to the real template function form (I can't type it all out as I'm on my phone), thus the inner `val` from the function shadows the one from the template.That makes sense. Before creating the example, I would have assumed that when you instantiate it as foo!1, then the `val=1` would flow through to the inner foo function. template foo(int val) { int foo(int val) { return val; } }
May 02 2018
On Thursday, 3 May 2018 at 02:48:10 UTC, jmh530 wrote:On Thursday, 3 May 2018 at 00:52:58 UTC, Meta wrote:If you want that, you might be able to do `int val = val` on the inner function, though I'm not sure that'll work.[snip] It's not a big per se. It's a consequence of the declaration expanding to the real template function form (I can't type it all out as I'm on my phone), thus the inner `val` from the function shadows the one from the template.That makes sense. Before creating the example, I would have assumed that when you instantiate it as foo!1, then the `val=1` would flow through to the inner foo function. template foo(int val) { int foo(int val) { return val; } }
May 02 2018
On Thursday, 3 May 2018 at 02:51:18 UTC, Meta wrote:If you want that, you might be able to do `int val = val` on the inner function, though I'm not sure that'll work.It does not work and will do nothing.
May 03 2018
On Thursday, 3 May 2018 at 13:30:03 UTC, bauss wrote:On Thursday, 3 May 2018 at 02:51:18 UTC, Meta wrote:Below compiles, but my guess is that the ASM is saying that the foo function only has a run-time argument. template foo(int val) { int foo(int val = val) { return val; } } void main() { assert(foo!1 == 1); assert(foo!1(2) == 2); }If you want that, you might be able to do `int val = val` on the inner function, though I'm not sure that'll work.It does not work and will do nothing.
May 04 2018