On Monday, 23 February 2015 at 21:33:35 UTC, Steven Schveighoffer
wrote:
It means that it can do that, and does do that in dcollections.
std.container does not have all the code to do it,
dcollections.RBTree has some extra calls that would probably
help. However, the allocation is done in one spot so the code
can easily use an allocator.
Have anybody done any benchmarks on which allocators are suitable
for which containers? I'm especially interested in optimizing
RBTree because I want a fast Dikjstra implementation in D:
http://rosettacode.org/wiki/Dijkstra%27s_algorithm#D
