Transitive groups issue with ConjugacyClassesSubgroups

[CAPiedade]

While testing the ConjugacyClassesSubgroups function on some groups of the TransitiveGroups library, I have found some occasional strange behaviour in some groups.

gap> g := TransitiveGroup(12,288);
[(L(6):2)^2]2=L(6):2wr2
gap> ConjugacyClassesSubgroups(g);;
Error, the collection <elm> must be contained in Source(<map>) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/mapping.gi:126 called from
Image( lhom, gps[x] ) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/oprtglat.gi:214 called from
func( C[i] ) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/coll.gi:663 called from
List( c, function ( x )
      return Image( lhom, gps[x] );
  end ) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/oprtglat.gi:214 called from
ClusterConjugacyPermgroups( G, ShallowCopy( dom ) 
 ) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/oprtglat.gi:480 called from
SubgroupsOrbitsAndNormalizers( i[1], i[2], false 
 ) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/grppccom.gi:1145 called from
...  at *stdin*:2
type 'quit;' to quit to outer loop

The same problem happened again later also with TransitiveGroup(10,43).

gap> g := TransitiveGroup(10,43);
[S(5)^2]2
gap> ConjugacyClassesSubgroups(g);
Error, the collection <elm> must be contained in Source(<map>) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/mapping.gi:126 called from
Image( lhom, gps[x] ) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/oprtglat.gi:214 called from
func( C[i] ) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/coll.gi:663 called from
List( c, function ( x )
      return Image( lhom, gps[x] );
  end ) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/oprtglat.gi:214 called from
ClusterConjugacyPermgroups( G, ShallowCopy( dom ) 
 ) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/oprtglat.gi:480 called from
SubgroupsOrbitsAndNormalizers( i[1], i[2], false 
 ) at /opt/homebrew/Cellar/gap/4.13.0/libexec/lib/grppccom.gi:1145 called from
...  at *stdin*:2
type 'quit;' to quit to outer loop

These two groups are the same, the (A5xA5):D8.
I ran the ConjugacyClassesSubgroups function for the list of SmallGroups of size {1 .. 959}/{256,384,512,640,768,896} and this problem did not show up elsewhere

[fingolfin]

Thank you for taking the time to report this issue, much appreciated!

This works in GAP 4.12.2.

According to git bisect this regression was introduced by f64e305 in PR #5514 by @hulpke

[hulpke]

Was already fixed in branch hulpke/work, now put into PR.