Table of syndromes and covering radius

Ticket #19913 proposes a generic, Guava-free implementation to compute the covering radius of a code.

We know that the covering radius is the biggest weight possible for an error vector in the table of syndromes used in the syndrome decoder (#19623).

In order to save some computation time, we should have the following mechanism:

if the user computed his code's covering radius and then uses a Syndrome decoder, the loop which fills the table will stop when the weight of the errors grows bigger than the covering radius.
if the user already computed the table of syndromes for this code (at maximal decoding radius, of course) and tries to compute the covering radius, look for the biggest weight of errors in the table of syndromes and return this weight instead of running the exhaustive search for the covering radius.

Depending on which ticket makes its way to Sage first, I'll update the remaining one with this mechanism.

Comments (1)