email-based regression search server

[Joe Buck]

I wrote:
> > If instead you have
> > 
> > P P F F P P P P F F F F
> > 
> > then you know that you will find an F that is immediately preceded by
> > a P, but you don't know which one you'll get.
> > 
> > Right?

On Wed, Feb 05, 2003 at 03:38:17PM -0800, Janis Johnson wrote:
> It will converge, but the result won't make any sense, in that the patch
> that is found won't have anything to do with the test.  I ran into that
> with some of the regression hunts I did, and ended up identifying patches
> for different target support being responsible for regressions on i686.

Then perhaps the answer is for people to try the binary search first,
and then if they get a result that makes no sense, to use the linear
search operation that Phil suggested, ideally with endpoints supplied
so that the linear search is over a narrow range.