www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - Type deduction on templated constructor.

reply "francesco cattoglio" <francesco.cattoglio gmail.com> writes:
So, I have this code (also on http://dpaste.dzfl.pl/3f767b17e83c)
This Vector(T) struct is taken from gfm.math.vector.

struct Vector(T) {
	T x, y, z;
	this(X : T, Y : T, Z : T)(X x_, Y y_, Z z_)
	{
		x = x_; y = y_; z = z_;	
	}
}

void main()
{
	Vector!ubyte test = Vector!ubyte(1, 1, 1);	
}

It fails to compile because "template 
f508.Vector!ubyte.Vector.__ctor cannot deduce function from 
argument types !()(int, int, int)".
Note that if one just defines a constructor as this(T x_, T y_, T 
z_) everything works.

My question is: should this code compile? I understand that the 
literal "1" is "int" therefore it can screw type deduction, but I 
wonder if the compiler should be smart enough to deduce it 
correctly.
Jul 24 2014
parent reply "bearophile" <bearophileHUGS lycos.com> writes:
francesco cattoglio:


 should this code compile? I understand that the literal "1" is 
 "int" therefore it can screw type deduction, but I wonder if 
 the compiler should be smart enough to deduce it correctly.


To keep both the compiler and programmers sane, D templates don't 
perform implicit type conversions. This sometimes is not handy, 
but on the whole saves from a large number of troubles.

So you can write (D V.2.066):

Vector!ubyte(ubyte(1), ubyte(1), ubyte(1));

Or you can create a little helper function that makes that code 
more DRY.

Bye,
bearophile
Jul 24 2014
parent reply "francesco cattoglio" <francesco.cattoglio gmail.com> writes:
On Thursday, 24 July 2014 at 09:38:14 UTC, bearophile wrote:

 francesco cattoglio:


 should this code compile? I understand that the literal "1" is 
 "int" therefore it can screw type deduction, but I wonder if 
 the compiler should be smart enough to deduce it correctly.


 To keep both the compiler and programmers sane, D templates 
 don't perform implicit type conversions. This sometimes is not 
 handy, but on the whole saves from a large number of troubles.

 So you can write (D V.2.066):

 Vector!ubyte(ubyte(1), ubyte(1), ubyte(1));

 Or you can create a little helper function that makes that code 
 more DRY.

 Bye,
 bearophile


I expected such an answer and I do understand the decisions 
behind it. Yet, you gave me a really GOOD news! Having to write 
cast(ubyte) 1 was way too much verbose for my liking, while the 
new ubyte(1) is reasonable enough.
Jul 24 2014
next sibling parent Philippe Sigaud via Digitalmars-d-learn writes:
 I expected such an answer and I do understand the decisions behind it. Yet,
 you gave me a really GOOD news! Having to write cast(ubyte) 1 was way too
 much verbose for my liking, while the new ubyte(1) is reasonable enough.


Why not use `1u`?
Jul 30 2014
prev sibling parent Philippe Sigaud via Digitalmars-d-learn writes:
On Wed, Jul 30, 2014 at 11:46 AM, Philippe Sigaud
<philippe.sigaud gmail.com> wrote:

 I expected such an answer and I do understand the decisions behind it. Yet,
 you gave me a really GOOD news! Having to write cast(ubyte) 1 was way too
 much verbose for my liking, while the new ubyte(1) is reasonable enough.


 Why not use `1u`?


Ow! Ignore that. 1u is an uint, not an ubyte.
Jul 30 2014