• =?ISO-8859-1?Q?Anders_F_Bj=F6rklund?= (14/35) Feb 08 2005 1)

=?ISO-8859-1?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:

Given this test code snippet:

 void test1(char[] s) { }
 void test1(wchar[] s) { }
 
 void test2(byte[] s) { }
 void test2(int[] s) { }
 
 void main()
 {
   char[] s = "foo";
   test1(s);
 
   static int[] i = [ 0, 1, 2 ];
   test2(i);
 
   test1(cast(char[]) "bar");
 //test2(cast(byte[]) [0, 1, 2]);
 }


1)
Why does "i" require a static initializer ?
(or making as the variable static, as above)

 literals.d:8: variable literals.main.i is not a static and
               cannot have static initializer

It seems to work just fine for the string ?

2)
Why could not the eg. int[] literals work as strings do ?
(in that you have to insert a cast to resolve overloads)

 literals.d:16: function literals.test1 overloads void(char[]s) and
                void(wchar[]s) both match argument list for test1

I understand ambiguity is the biggest stop for those ?


Would this be helped if those literal arrays had
the proposed .readonly keyword attached to them ?

Or will those literals and the hash literals not
come until D 2.0 ("second coming"), no matter what ?

--anders