> Would anyone be opposed to replaing gawk with mawk?  mawk is smaller,
> faster, and cleaner.  It has fewer features, but most of the "features"
> it lacks are in perl anyway.  :)
> 
> $ ls -l lang/*awk*
> -rw-rw-rw-  1 seebs      520032 Jun 23  1995 lang/gawk-2.15.6.tgz
> -rw-rw-rw-  1 seebs      821008 Jan 12 13:40 lang/gawk-3.0.0-doc.tgz
> -rw-rw-rw-  1 seebs      881969 Jan 12 13:42 lang/gawk-3.0.0.tgz
> -rw-rw-rw-  1 seebs      193169 Feb  9 14:33 lang/mawk1.2.2.tgz
> 
> I think this would help NetBSD achieve the goal of being small, fast,
> and desireable.

NetBSD once used mawk.  

I made the decision to change to gawk because tests showed that mawk
was not significantly faster for any of my awk scripts, and even more
importantly, it would not pass gawk's posix regression scripts.  Also,
at the time it looked like mawk development had been discontinued, 
while gawk was still being enhanced.

Since they are both gpl'd I am willing to re-evaluate the decision.
However, I must be convinced of mawk's future and that it is posix
compliant.

	--jtc