Re: [boost] Regression tests: how long is reasonable?

From: Beman Dawes (bdawes_at_[hidden])
Date: 2002-08-12 20:38:16

At 05:13 PM 8/12/2002, Gennadiy Rozental wrote:

 >On the other hand Boost.Test users are not required to use bjam but still 

 >could be interecting in timing.

That's true. In production programs, I always call a timer directly and 
display the results on job termination.  I would say about 5%-10% of all 
problems first show up as unexpected timing deviations.

Developers hearing that often say that isn't their experience, but then 
when you ask them, they don't routinely supply a run timer, so they have no 
idea if timing deviations are happening unless the deviation is large in 
absolute terms.

--Beman