Problems with GC, trees and array concatenation

[Regan Heath]

Oskar Linde Wrote:
> I have no idea why the Individual->real substitution above would make 
> any difference, if not by pure chance.

In D1.001 a new type aware GC was implemented:

http://www.digitalmars.com/d/changelog.html#new1_001

Is it possible that this is why substituting 'Individual' for 'real' makes a difference.  When you use 'real' your memory consists of mainly 'real's which are non-pointer types, greatly reducing the risk of spurious pointer collisions.

Regan Heath