On Sun, Apr 08, 2001 at 01:04:25PM -0400, Perry E. Metzger wrote:
> Is this a win???? Does it have to be that slow?

Will, I understand that the original speedup was because of a bug... so part
of it is inavoidable. I can't judge whether a better implementation might
regain some of that speed.

(I'm familiar with this sort of thing; in 1.4someting NetBSD/Amiga
was faster at process creation, because caches were not synchronized
correctly. This was detected by running some regression test.  Lots
of complaints afterwards.)

Regards,
	Ignatios