Re: 16-fyi-bad-argc-interface.patch

[akim]

On Sat, Oct 13, 2001 at 03:50:39PM +0100, Gary V. Vaughan wrote:
> On Sat, Oct 13, 2001 at 10:57:11AM +0200, Akim Demaille wrote:
> 
> I fine with this, except that I had wanted to save on the overhead of a
> function call when no checks were necessary... this deep in the inner
> loop it must make a difference.

I guessed that was the point indeed, but I really prefer losing some 
performance and gain maintainability, which led me to think of...

> Why don't we have a `M4_BAD_ARGC' macro, that only calls the func if at
> least one of the range limits is non-negative?

this :)  We seem to agree, so let's go for that.  I was not sure the
gain would be significant, but if you do think it might be.