Re: [Gnu-arch-users] more on the merge-fest

[Andrew Suffield]

On Wed, Nov 26, 2003 at 01:34:32PM +0100, Thomas Zander wrote:
> On Wednesday 26 November 2003 12:21, Andrew Suffield wrote:
> > > > If cause and effect are separated by 500 lines of code, then no
> > > > amount of testing in 20-line chunks is going to demonstrate the
> > > > problem.
> > >
> > > Uhh, yes.  Because those intervening 500 lines of code must,
> > > themselves, also be unit-tested.
> >
> > So, now you know that somewhere in that 500 line chunk lies a bug. Are
> > you seriously suggesting that you can often spot it by inspection?
> 
> Read up on your testing methods; there are very good ways to test each if 
> and each switch in those 500 lines and let a certain test fail.
> Its all about writing more tests and knowing the 'border' cases.

This is a clone of the "more granularity" appeal, and it's still
wrong. You are operating under the assumption that the tests form a
complete, correct description of the behaviour of the program (if this
were true, then you could simply deploy the test suite and skip
writing the program).

-- 
  .''`.  ** Debian GNU/Linux ** | Andrew Suffield
 : :' :  http://www.debian.org/ |
 `. `'                          |
   `-             -><-          |

signature.asc
Description: Digital signature