error in ConjugacyClasses computation

[ThomasBreuer]

Gabriel Navarro has reported the following problem.
It happens in GAP 4.10.2 as well as in the current master branch.

gap> g:= Group( (1,2,3)(7,8)(12,13), (1,3)(2,4,5)(6,7,9,12,11,8,10,13), 
>               (1,4,3)(6,7)(8,11)(9,13)(10,12) );;
gap> max:= MaximalSubgroupClassReps( g );;
gap> ConjugacyClasses( max[1] );
Error, should not happen nonslv stabstack at /.../gapmaster/gap/lib/fitfree.gi:662 called from
OrbitsRepsAndStabsVectorsMultistage( o.stabradgens, 
 List( o.stabradgens, function ( x )
        return bas * npcgsact( x ) * basinv;
    end ), pcisom, o.stabrsubsz, solvtriv, o.stabfacgens, 
 List( o.stabfacgens, function ( x )
        return bas * npcgsact( x ) * basinv;
    end ), o.stabfacimgs, hom, o.subsz, OnRight, el 
 ) at /.../gapmaster/gap/lib/clashom.gi:2148 called from
LiftClassesEATrivRep( G, mpcgs, i, fants, hom, pcisom, solvtriv 
 ) at /.../gapmaster/gap/lib/clashom.gi:2478 called from
ConjugacyClassesViaRadical( G 
 ) at /.../gapmaster/gap/lib/clashom.gi:3087 called from
<function "ConjugacyClasses perm group">( <arguments> )
 called from read-eval loop at *stdin*:4
you can 'quit;' to quit to outer loop, or
you can 'return;' to continue

[fingolfin]

That error message was inserted in a code change by @hulpke on 2013-02-22.

The error also happens in GAP 4.8 and 4.9. It works in GAP 4.6; in GAP 4.7, there is a different error:

gap> g:= Group( (1,2,3)(7,8)(12,13), (1,3)(2,4,5)(6,7,9,12,11,8,10,13),
>               (1,4,3)(6,7)(8,11)(9,13)(10,12) );;
gap> max:= MaximalSubgroupClassReps( g );;
gap> ConjugacyClasses( max[1] );
Error, matrix <mat> must be invertible called from
ProjectiveOrder( mat ) called from
Order( imgs[x] ) called from
func( elm ) called from
Filtered( [ 1 .. Length( imgs ) ], function ( x )
      return Order( imgs[x] ) > 1;
  end ) called from
LiftClassesEATrivRep( G, mpcgs, i, fants, hom, pcisom, solvtriv ) called from
...  at line 4 of *stdin*
you can 'quit;' to quit to outer loop, or
you can 'return;' to continue
brk>

[hulpke]

The problem is that the fitting free data structure for the subgroup cannot just re-use the pc isomorphism for the radical of the whole group. Fixed.