• "H. S. Teoh" <hsteoh quickfur.ath.cx> Mar 06 2012
• =?UTF-8?B?QWxpIMOHZWhyZWxp?= <acehreli yahoo.com> Mar 06 2012
• simendsjo <simendsjo gmail.com> Mar 06 2012
• Andrej Mitrovic <andrej.mitrovich gmail.com> Mar 06 2012

"H. S. Teoh" <hsteoh quickfur.ath.cx> writes:

Code:

	// test.d
	class A {}
	class B : A {}
	class C : A {}
	void main() {
		A[] objs = [ new B, new C ];	// line 6
	}

Compiler error:

	test.d(6): Error: cannot implicitly convert expression (new B) of type test.A
to test.C
	test.d(6): Error: cannot implicitly convert expression ([(__error),new C]) of
type C[] to A[]

I thought D was supposed to automatically infer the most derived common
base class to assign to an array literal of different types? At least,
that's the impression I got from TDPL.

What am I doing wrong?


T

-- 
MAS = Mana Ada Sistem?

=?UTF-8?B?QWxpIMOHZWhyZWxp?= <acehreli yahoo.com> writes:

On 03/06/2012 11:27 AM, H. S. Teoh wrote:
 Code:

 	// test.d
 	class A {}
 	class B : A {}
 	class C : A {}
 	void main() {
 		A[] objs = [ new B, new C ];	// line 6
 	}

 Compiler error:

 	test.d(6): Error: cannot implicitly convert expression (new B) of type test.A
to test.C
 	test.d(6): Error: cannot implicitly convert expression ([(__error),new C]) of
type C[] to A[]

 I thought D was supposed to automatically infer the most derived common
 base class to assign to an array literal of different types? At least,
 that's the impression I got from TDPL.

 What am I doing wrong?


 T


That's probably this bug:

   http://d.puremagic.com/issues/show_bug.cgi?id=5498

Luckily it has a pull request.

Ali

simendsjo <simendsjo gmail.com> writes:

On Tue, 06 Mar 2012 20:27:56 +0100, H. S. Teoh <hsteoh quickfur.ath.cx>  
wrote:

 Code:

 	// test.d
 	class A {}
 	class B : A {}
 	class C : A {}
 	void main() {
 		A[] objs = [ new B, new C ];	// line 6
 	}

 Compiler error:

 	test.d(6): Error: cannot implicitly convert expression (new B) of type  
 test.A to test.C
 	test.d(6): Error: cannot implicitly convert expression ([(__error),new  
 C]) of type C[] to A[]

 I thought D was supposed to automatically infer the most derived common
 base class to assign to an array literal of different types? At least,
 that's the impression I got from TDPL.

 What am I doing wrong?


 T


I have no idea (and haven't read TDPL), but adding a new A somewhere in  
the array makes it work. Compiler bug..?

Andrej Mitrovic <andrej.mitrovich gmail.com> writes:

You can use a cast as a workaround:
A[] objs = [cast(A)new B, new C ];

This bug has been around for a while, the first time I've seen it
mentioned was in the second code snippet in this blog post:
http://klickverbot.at/blog/2010/11/announcing-d-support-in-swig/