[m-rev.] for review: string switches using tries

[Peter Wang]

On Mon, 02 Mar 2015 19:56:49 +1100 (EST), "Zoltan Somogyi" <zoltan.somogyi at runbox.com> wrote:
> 
> Any idea whether my guess is in the right ballpark?

The results may mean nothing.  No predicates which contain a string
switch are called very frequently overall.  The top two of the
predicates are only called 1482 and 586 times in a ~3 second run,
which is nowhere on the profile.

Peter