This benchmark is odd currently because it uses inconsistent cases
between benchmark iterations, and each iteration actually does a bit of
testing.

This separates the two benchmark cases into two separate benchmarks and
removes the testing done on each iteration. The unit tests above
suffice.

The benchmark being more succinct will make it easier to gauge the
benefits of any future MIME header reading changes.

Change-Id: I2399fab28067f1aeec3d9b16951d39d787f8b39c
Reviewed-on: https://go-review.googlesource.com/134225Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>