[m-rev.] diff: alternative fix for bug 264 test case

[Mark Brown]

On Mon, Nov 23, 2015 at 9:52 AM, Julien Fischer <jfischer at opturion.com> wrote:
>> The last test case I am sure is freely distributable,
>> so it is attached.
>
>
> There's a copy of this in benchmarks/progs/training_cars.

Running this case with mmc -e gets < 0.5% slowdown. This only really
exercises inst_match_final, but the effect on performance should be
similar for inst_match_initial and inst_merge.

A worse case, but less realistic, would be a large bound inst with
lots of higher-order info that matches the default.

Mark