develooper Front page | perl.perl5.porters | Postings from March 2003

Re: Improving smoke coverage with graco latin squares

Thread Previous
From:
Nicholas Clark
Date:
March 12, 2003 13:43
Subject:
Re: Improving smoke coverage with graco latin squares
Message ID:
20030312205245.GB276@Bagpuss.unfortu.net
On Thu, Feb 27, 2003 at 01:42:49PM +0100, Rafael Garcia-Suarez wrote:

> The only combinations that aren't tested are ab, ac, bc and the like
> (combinations of symbols in the same group.) This doesn't matter
> if we choose carefully options (example :
>     a -> <nothing>
>     b -> -Duse64bitint
>     c -> -Duse64bitall )
> 
> With a 4x4 square, one could test 16 options in 16 builds.
> With a 5x5 square, one could test 20 options in 25 builds.
> Smaller build times, more options. Could be a great complement to the
> usual smoking matrix.

I think that this is a good idea. Although I think that I'd be happier
combining it with Abe's smart smoke idea - ie if the patchlevel advances
you start smoking the square. But if at the end of a "run" it is unchanged
then you start filling in the gaps to complete the brute force trial of
every option permutation against every other.

This might be rather simplistic - maybe it's better to do a complementary
graco latin square that has ab ac bc in combination, so that by the second
run everything has been tried with everything else in some way.

(And then fill in the gaps)

As usual I'm trying to optimise before the simple version is even working.

Nicholas Clark

Thread Previous


nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About