[GAP Forum] FGA: bugfix and git repository

[sievers at math2.nat.tu-bs.de]

Dear GAP Forum,

yesterday saw the release of version 1.3.0 of the FGA package,
a package for computing with free groups that is usually loaded
automatically.

It is available at the usual place at
http://www.icm.tu-bs.de/ag_algebra/software/FGA/FGA-1.3.0.tar.gz

This new version fixes a bug reported by Ignat Soroko, that in a few
cases let

   RepresentiveAction(f,g,h)

(where f, g and h are subgroups of free groups)
return an element as a conjugating element when in fact none existed.

That in turn also caused

   IsConjugate(f,g,h)

to falsely return true in these cases.

(Computations with single elements as second and third argument are
 not affected by this bug.)

In detail, the bug would be triggered whenever g and h have the same rank
and after cyclic reduction g and h have the same number of what Sims calls
"important cosets", and there is an x in f such that g^x is a proper
subgroup of h.


Additionally, I'd like to announce that there is now a public git
repository for the FGA package at
https://github.com/chsievers/fga


All the best,
Christian Sievers