new std.variant (was Re: The Right Approach to Exceptions)

[Juan Manuel Cabo]

On Thursday, 23 February 2012 at 01:36:32 UTC, Juan Manuel Cabo 
wrote:
> If all I want is binary search on a big appender, then it
> is O(k * n * log(n)), and that k right there doesn't
> bother me.

(Where binary search is of course O(log(n))
and accessing individual elements with the proposed
Appender is O(N / (4080/T.sizeof)), so k == 4080/T.sizeof)

--jm