[GAP Forum] Table inconsistency

[Mathieu Dutour]

I think there is a problem in the tables of the irreducible
finite groups of GLn(Z) in GAP.

If you do 
gap> DisplayImfInvariants(23,3,0);
#I Z-class 23.3.1:  Size = 2^19*3^6*5^3*7*11*23
#I   isomorphism type = C2 x Co2
#I   elementary divisors = 1^23
#I   orbit size = 4600, minimal norm = 3
#I Z-class 23.3.2:  Size = 2^19*3^6*5^3*7*11*23
#I   isomorphism type = C2 x Co2
#I   elementary divisors = 1^22*4
#I   orbit size = 93150, minimal norm = 4
#I Z-class 23.3.3:  Size = 2^19*3^6*5^3*7*11*23
#I   isomorphism type = C2 x Co2
#I   elementary divisors = 1*4^22
#I   orbit size = 4600, minimal norm = 12
then the representations of C2 x Co2 up to Z-equivalence correspond to
the lattices O23, Lambda_{23} and Lamnda_{23}^* respectively, which is
all ok.

But if one does the same operation
DisplayImfInvariants(22,x); with x varying between 1 and 12, then one gets
only one Z-class for all of them! This is highly unlikely since some of the
lattices are not self-dual. For example, I can produce at least 3
inequivalent representations of the group given by
DisplayImfInvariants(22,7) when only one is listed.

  Mathieu