[with patch, positive review] improve sbox linear and differences matrices computation

[ylchapuy]

In particular, use walsh transform for linear_approximation_matrix.

[malb]

Hi there, it is embarrassing how bad my naive original code was. Here's a comparison (for the release tour)

**Old***

sage: S = mq.SR(1,4,4,8).sbox()
sage: %time _ = S.difference_distribution_matrix()
CPU times: user 82.14 s, sys: 0.01 s, total: 82.15 s
Wall time: 82.15 s

sage: %time _ = S.linear_approximation_matrix()
CPU times: user 145.10 s, sys: 0.02 s, total: 145.12 s
Wall time: 145.12 s

***New***

sage: S = mq.SR(1,4,4,8).sbox()
sage: %time _ = S.difference_distribution_matrix()
CPU times: user 0.32 s, sys: 0.00 s, total: 0.32 s
Wall time: 0.32 s
sage: %time _ = S.linear_approximation_matrix()
CPU times: user 1.10 s, sys: 0.00 s, total: 1.10 s
Wall time: 1.10 s

The code looks good, doctests pass.

The only issue: the sage -coverage script will pick up _walsh_transform and complain that it isn't documented and doctested.

Of course it is impossible to doctest this inner function directly, but the keyword # indirect doctest will do the trick.

This is a positive review except for the missing documentation.

[ylchapuy]

Both patches should be applied.

I added an indirect doctest. I give myself a positive review, feel free to correct me if you disagree.

[malb]

All good, definitively a positive review.

[mvngu]

Just to let people know, this has been merged in sage-4.1.1-alpha0. I can't close this ticket because I don't have the privilege to do so. Sorry, folks :-(