new error message in 2.066, type bool (const)

[Paul D Anderson via Digitalmars-d-learn]

On Wednesday, 20 August 2014 at 20:46:20 UTC, Paul D Anderson 
wrote:
> Re-compiling existing code with version 2.066 generates a lot 
> of errors complaining about implicit conversion to const. 
> Typical is this call (inside a struct with properties 1 & 2):
>
> 	z.sign = x.sign ^ y.sign;
>
> Error: None of the overloads of 'sign' are callable using 
> argument types bool (const), candidates are:
>
> 1)	@property
> 	@safe
> 	bool sign() const
> 	{
> 		return signed;
> 	}
>
> 2)	@property
> 	@safe
> 	bool sign(in bool value)
> 	{
> 		signed = value;
> 		return signed;
> 	}
>
> What changed? It ran okay with early beta versions, but not 
> with the release.
>
> Paul

The problem (not obvious from the above) is that a templated type 
is bringing its qualifiers into the template.

Here is a small example:

T add(T)(in T x, in T y)
{
	T z;
	z = x + y;
	return z;
}

void main()
{
	const double a = 1.0;
	const double b = 2.0;
	double c;
	c = add(a,b);
}

Error: Cannot modify immutable expression z

---

Since a and b are const double the deduced template type is also 
const double. Then when z is declared as type T it is also const.

This compiled and ran in beta 5 but not in beta 6.

If I modify the above so that declaration and assignment are 
performed at the same time:

	T z = x + y;

This compiles and assigns the (const double) value 3.0 to z. 
After the function call c contains the double value 3.0, but it 
is no longer const.

The error can be avoided by declaring z to be of type Unqual!T in 
most cases.

I don't know if this is expected behavior that just wasn't 
enforced before, or if this is something new. Either way I don't 
like it. And I'm a little surprised I'm the only one affected by 
this. I'll keep digging.

Paul