On 29/09/2010 22:22, Steven Schveighoffer wrote:
On Wed, 29 Sep 2010 16:14:27 -0400, BLS <[email protected]> wrote:
On 29/09/2010 21:19, Jonathan M Davis wrote:
On Wednesday, September 29, 2010 11:58:01 BLS wrote:
Hope is was not a pipe dream...
I've no idea who has published it and where to find this snippet but in
case that I am not completely wrong there should exist something like
Implements!() for structures..
???
Bjoern
And what would the implement? I thought that structs didn't work with
interfaces. And if they did, you'd use is() with : to test that. So,
what are
you really looking for here? What would Implements!() do if it exists?
- Jonathan M Davis
Hi Jonathan,
Just because structures don't work with interfaces I am still hoping
for Implements!().
What do I expect from Implements! A contract which ensures (at compile
time) that a structure implements f.i. all functions required by a
random access range.
well, as said, I'm not even sure that Implements! exist at all.
bjoern
Well, there are the isRandomAccessRange and friends, I think that's the
standard mode of operation for testing if a struct implements a
particular compile-time interface.
The thing is, there's no syntax to *define* a compile-time interface
except this way :)
So Implements!(myStruct, ???) Where ??? is the compile time interface,
what do you pass in there?
-Steve
Implements!(*I*MyInterface) // where IMyInterface is a plain vanilla
Interface
This was at least my hope 'cause IMHO this could make just f.i. the
range stuff more reliable.
Unfortunately we are not able to use classes to implement the basic
data-structures for collections/graphs. if we could use class Node,
class Tree : IxRange, class Map : Tree ...
Bjoern
ps have send you an email
