struct Test{
	int x;
	this(int v){
		x = v;
		writeln(x.to!string ~ " Created");
	}
	~this(){
		writeln(x.to!string ~ " Destroyed");
	}
	void opOpAssign(string op, Type)(ref Type s){
		x = s.x;
	}
}
void showme(Type)(ref Type t){
	writeln(t.x);
}
void main(){
	auto t = Test(1);
	auto t2 = Test(2);
	showme(t);
	showme(t2);
	t = t2;
}

Prints:
1 Created
2 Created
1
2
1 Destroyed
2 Destroyed
2 Destroyed

this line:
t = t2
Causes the dtor to be called.

Why?
I expected nothing to happen because "ref" its a simple pointer, right?
Or I am missing something here?

On 8/1/16 12:01 PM, Patric wrote:

> I expected nothing to happen because "ref" its a simple pointer, right?
> Or I am missing something here?
>

You want opAssign, not opOpAssign. opOpAssign is for things like +=.

Your code had me worried for a while :)

-Steve

On Monday, 1 August 2016 at 16:05:51 UTC, Steven Schveighoffer wrote:
> On 8/1/16 12:01 PM, Patric wrote:
>
>> I expected nothing to happen because "ref" its a simple pointer, right?
>> Or I am missing something here?
>>
>
> You want opAssign, not opOpAssign. opOpAssign is for things like +=.
>
> Your code had me worried for a while :)
>
> -Steve

Same thing:
(Remembered now of DPaste)
(and I understand your concern :P)
https://dpaste.dzfl.pl/af512b5f6288

On Monday, 1 August 2016 at 16:05:51 UTC, Steven Schveighoffer wrote:
> On 8/1/16 12:01 PM, Patric wrote:
>
>> I expected nothing to happen because "ref" its a simple pointer, right?
>> Or I am missing something here?
>>
>
> You want opAssign, not opOpAssign. opOpAssign is for things like +=.
>
> Your code had me worried for a while :)
>
> -Steve

Sorry, silly me, forgot to remove the "string op" now its ok :)

On Monday, 1 August 2016 at 16:14:31 UTC, Patric wrote:
> On Monday, 1 August 2016 at 16:05:51 UTC, Steven Schveighoffer wrote:
>> On 8/1/16 12:01 PM, Patric wrote:
>>
>>> I expected nothing to happen because "ref" its a simple pointer, right?
>>> Or I am missing something here?
>>>
>>
>> You want opAssign, not opOpAssign. opOpAssign is for things like +=.
>>
>> Your code had me worried for a while :)
>>
>> -Steve
>
> Same thing:
> (Remembered now of DPaste)
> (and I understand your concern :P)
> https://dpaste.dzfl.pl/af512b5f6288

You've implemented opAssign incorrectly. See:

https://dlang.org/spec/operatoroverloading.html#assignment

On Monday, 1 August 2016 at 16:05:51 UTC, Steven Schveighoffer wrote:
> On 8/1/16 12:01 PM, Patric wrote:
>
>> I expected nothing to happen because "ref" its a simple pointer, right?
>> Or I am missing something here?
>>
>
> You want opAssign, not opOpAssign. opOpAssign is for things like +=.
>
> Your code had me worried for a while :)
>
> -Steve

But still.
If it was the case of "+=" wasn´t wrong to call the dtor since is a ref var?

On Monday, 1 August 2016 at 16:17:02 UTC, Patric wrote:
> On Monday, 1 August 2016 at 16:05:51 UTC, Steven Schveighoffer wrote:
>> On 8/1/16 12:01 PM, Patric wrote:
>>
>>> I expected nothing to happen because "ref" its a simple pointer, right?
>>> Or I am missing something here?
>>>
>>
>> You want opAssign, not opOpAssign. opOpAssign is for things like +=.
>>
>> Your code had me worried for a while :)
>>
>> -Steve
>
> Sorry, silly me, forgot to remove the "string op" now its ok :)

Well, you beat me to it  :)

On Monday, 1 August 2016 at 16:18:32 UTC, Patric wrote:

> But still.
> If it was the case of "+=" wasn´t wrong to call the dtor since is a ref var?

No. It was working as expected. You never implemented opAssign, so default assignment was being used. There was no ref variable.

On Monday, 1 August 2016 at 16:21:16 UTC, Mike Parker wrote:
> On Monday, 1 August 2016 at 16:18:32 UTC, Patric wrote:
>
>> But still.
>> If it was the case of "+=" wasn´t wrong to call the dtor since is a ref var?
>
> No. It was working as expected. You never implemented opAssign, so default assignment was being used. There was no ref variable.

Iep, I realized this by now. Thanks anyway :)