• Kyle G. (22/22) Oct 02 2007 Hi,
• BCS (3/32) Oct 02 2007 its a pitty AA.remove is of type void. If it wasn't then this would work...
• Kyle G. (6/43) Oct 02 2007 I suppose you could define a remove alternative:
• Steven Schveighoffer (8/11) Oct 04 2007 The question I would ask you is: why?

"Kyle G." <kyle.james.gibson gmail.com> writes:

Hi,

I want to declare an AA such that it's starting length is 0, yet, it is 
not null.

For instance:

1 string[int] aa;
2 assert (aa.length == 0);
3 assert (aa is null);
4 aa[100] = "";
5 aa.remove(100);
6 assert (aa.length == 0);
7 assert (aa !is null);

The above runs without error, as it should. Notice that assertions 3 and 
7 are opposite.

However, is there anything shorter/more efficient that I can do 
(preferably on line 1) which would generate an AA which satisfies the 
assertions 6 and 7?

This appears to function in the same way:

string[int] aa = [100:""];
aa.remove(100);
assert (aa.length == 0);
assert (aa !is null);

Thank you.

BCS <ao pathlink.com> writes:

Reply to Kyle G.,

 Hi,
 
 I want to declare an AA such that it's starting length is 0, yet, it
 is not null.
 
 For instance:
 
 1 string[int] aa;
 2 assert (aa.length == 0);
 3 assert (aa is null);
 4 aa[100] = "";
 5 aa.remove(100);
 6 assert (aa.length == 0);
 7 assert (aa !is null);
 The above runs without error, as it should. Notice that assertions 3
 and 7 are opposite.
 
 However, is there anything shorter/more efficient that I can do
 (preferably on line 1) which would generate an AA which satisfies the
 assertions 6 and 7?
 
 This appears to function in the same way:
 
 string[int] aa = [100:""];
 aa.remove(100);
 assert (aa.length == 0);
 assert (aa !is null);
 Thank you.
 

its a pitty AA.remove is of type void. If it wasn't then this would work:

([int.init:char[].init].remove(int.init))

"Kyle G." <kyle.james.gibson gmail.com> writes:

BCS wrote:
 Reply to Kyle G.,
 
 Hi,

 I want to declare an AA such that it's starting length is 0, yet, it
 is not null.

 For instance:

 1 string[int] aa;
 2 assert (aa.length == 0);
 3 assert (aa is null);
 4 aa[100] = "";
 5 aa.remove(100);
 6 assert (aa.length == 0);
 7 assert (aa !is null);
 The above runs without error, as it should. Notice that assertions 3
 and 7 are opposite.

 However, is there anything shorter/more efficient that I can do
 (preferably on line 1) which would generate an AA which satisfies the
 assertions 6 and 7?

 This appears to function in the same way:

 string[int] aa = [100:""];
 aa.remove(100);
 assert (aa.length == 0);
 assert (aa !is null);
 Thank you.

 
 its a pitty AA.remove is of type void. If it wasn't then this would work:
 
 ([int.init:char[].init].remove(int.init))
 
 

I suppose you could define a remove alternative:

V[K] bremove(V,K) (ref V[K] ar, K k) {
  ar.remove(k);
  return ar;
}

"Steven Schveighoffer" <schveiguy yahoo.com> writes:

"Kyle G." wrote
 Hi,

 I want to declare an AA such that it's starting length is 0, yet, it is 
 not null.

The question I would ask you is: why?

I understand that it is weird, and I agree it should be consistent, but 
using the length property is consistent.  Why not just ignore the comparison 
to null and only use the length check?

IMO, arrays should not be comparable to null, since they are internally 
structures.

-Steve