>> Has anyone implemented any new advances into trace(1) such that you
>> can tell it to follow children as they are created?

(trace(1) doesn't exist on NetBSD, lest anyone get confused.)

> ktrace(1) does this just fine.  I know der Mouse doesn't find that
> ktrace does everything he'd like it to do, but it works well enough
> for me.

And it works well enough for me, too, in many cases.  But it doesn't,
and in a few cases couldn't be made to, do some things I sometimes
want.  (Most of the things I want could potentially be added - the
first example that comes to mind is printing the address for the
connect() and bind() calls - but anything that calls for run-time
judgement is paradigmatically impossible.  An example might be an
"untrusted hostile program" environment provider, which might (say)
refuse to write to any but a very limited set of files...or forbid
network connections except to a few hosts...)

					der Mouse

			    mouse@collatz.mcrcim.mcgill.edu