Re: Test all macros for cache (`configure -C') safety

[Noah Misch]

On Fri, Feb 04, 2005 at 10:54:52PM -0800, Paul Eggert wrote:
> Also, isn't it a bit simpler just to invoke "diff" directly on the
> input files when comparing them?  I guess this would result in a
> diff-of-a-diff output on failure, but that's OK.

I do not recognize the advantage.  Perhaps I'm missing something obvious.  It is
a minor point, though.

> (One minor question: why is AT_CMP's m4_fatal useful?  I left it in
> though.)

It was a debugging check.  I suppose I omitted the second argument more than
once and grew tired of not discovering that until `cat ' hung waiting for
terminal input.