[ic] Different mv_substring_match for diff. fields (resolved + patch)

[Davor Ocelic]

On Thu, 3 Jul 2008 14:18:33 -0400
Mike Heins <mike at perusion.com> wrote:

> > So the proposed solution is:
> > 
> > Instead of turning off mv_coordinate, let's just ensure that the
> > number of search specs matches the number of search fields. (Taking
> > user's setting of mv_coordinate=1 as a sign that he wants it to
> > happen).
> 
> Can't happen. We could add an "mv_force_coordinate" or something,
> but we can't just change this behavior willy nilly.

Sure.

But as I figure it, having mv_coordinate=1 and a non-matching
number of  search specs and search fields is not a legitimate
combination anyway, so this change would not break any valid search.

I suppose it'd break only for people who had an incorrect search
in the first place.

I had mv_force_coordinate in mind myself too (which would also have a
'fc' shorthand available) but do we need it?

I mean, this isn't a "glitch" in the code that turned out to have
unexpected side-benefits and that people started relying. This is just
some unuseful code path that was never mentioned or advertised by
anyone.

-doc