Why does config.guess come up with such wacky stuff?

[Mike Hopkirk(hops)]

On Wed Dec 01,1999 (04:58:17PM -0800), jbuck@synopsys.com(Joe Buck) wrote:
> Jeff writes:
> > For building our releases we override config.guess so that we can exercise
> > more control over what systems the release will work on [...]
> > But I consider this behavior the exception rather than the rule.
(presumably bugs in configure scripts and overriding config.guess values)

> > 
> > I agree 100% with Robert that config.guess is supposed to cater to the random
> > person that picks up GCC (or other tools) off the net and wants to get it
> > built with a minimum of hassle.

But not necessarily the most optimal build
 
> This means that in any case where config.guess can't be sure it has the
> right answer, it should report a safe answer.  This will mean that it
> might assume an older processor in some cases and not use the latest
> instructions, but at least it will work.

Doesn't it also mean that apps should use  config.guess results as 
provided and build only the most generic version with those values 
unless specifically told to do different.

iff it is told to do different its build should then start finding out exactly 
what it needs to know to optimise better/more ??
 ... i.e put the smarts in the app builds that need or want to care  ...

( thats not to say there shouldn't be something common ( or a 
    mode of autoconf/config.guess) that can dig out this info but it shouldn't
    be the base case )...




-- 
- hops

Everything disclaimed (including disclaimer)
------<hops@sco.com>--------------------------------------
Mike Hopkirk (hops)  |       Whenever possible steal code.
SCO Inc              |          Tom Duff. (ex) Bell Labs