• William Bashir (4/4) Sep 08 2010 Why not include most of DCollections in Phobos? std.container will surel...
• bearophile (4/7) Sep 08 2010 Those data structures will be added. But Andrei wants to organize them i...
• Steven Schveighoffer (10/14) Sep 08 2010 FWIW, dcollections' RBTree is going to be in phobos std.container. I

William Bashir <williambashir mailinator.com> writes:

Why not include most of DCollections in Phobos? std.container will surely
benefit from having a rbtree, list and other data structures and implementing
new versions for just for Phobos seems like NIH.

Thanks

bearophile <bearophileHUGS lycos.com> writes:

William Bashir:

 Why not include most of DCollections in Phobos? std.container will surely
 benefit from having a rbtree, list and other data structures and implementing
 new versions for just for Phobos seems like NIH.

Those data structures will be added. But Andrei wants to organize them in a
different way.

Bye,
bearophile

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

On Wed, 08 Sep 2010 13:52:11 -0400, William Bashir  
<williambashir mailinator.com> wrote:

 Why not include most of DCollections in Phobos? std.container will surely
 benefit from having a rbtree, list and other data structures and  
 implementing
 new versions for just for Phobos seems like NIH.

FWIW, dcollections' RBTree is going to be in phobos std.container.  I  
submitted a std.container-ified to Andrei, it lacks documentation and full  
unit tests, but I copied almost exactly my RBNode that implements the  
algorithm.

As far as the other things, they are pretty rote (hash tables, arrays,  
linked lists), so NIH doesn't really apply.  The RBTree took a long time  
to get right (started with pascal-like pseudocode from my algorithms book).

-Steve