On Fri, Nov 16, 2007 at 04:37:08AM -0800, Paul Goyette wrote:
> 
> Yeah, 4 or 5 minutes is in the right ball-park.  I think that much of 
> the time is actually spent updating the count-down;  it might go a bit 
> faster if we didn't print out each and every 1MB step.  At the least, 
> watching the countdown makes it _feel_ like it takes longer!

One possibility is to output 100 before the first write, 99 after it,
then 95 90 (etc) as % of the way through the dump.
That is less than one line of text, but conveys (almost all of) the
useful information.

	David

-- 
David Laight: david@l8s.co.uk