faster splitter

[Andrei Alexandrescu via Digitalmars-d]

On 05/27/2016 06:19 PM, qznc wrote:
> manual find: 118 ±24
> qznc find:   109 ±13  <--- using the sentinel trick
> Chris find:  142 ±27
>
> It is normal that the numbers of the other tests change, since those are
> relative to the fastest one in each run. When qznc find 'wins' more
> often, the others get more slowdown.

What compiler?

> This means another special case for mutable ranges could be worthwhile.

Also mind the throwing possibilities.


Andrei