Dear Mrs. and Mr. Forum,
yesterday Akihiro Munemasa sent two bug reports to 'gap-trouble'.
The first was about an error message that occurs in some cases
in the computation of (rational) conjugacy classes.
For example, 'ConjugacyClasses( GU(2,5) )' will produce this.
The second was about the groups 'GU(2,q)' and 'SU(2,q)'. Here one
gets no error message. But the groups one gets with these commands
are not the required unitary groups.
Two other bugs in GAP-3.3 should be mentioned which meanwhile
were found in Aachen.
The first is in the new implementation of the LLL algorithm;
in some cases the computed basis is not LLL-reduced, or it contains
The other bug is in one of the programs that are used to compute the
canonical representative of conjugacy classes in Ag groups; in some
cases the question whether two elements are conjugate is answered ``no''
although in fact the elements are conjugate.
All these bugs will be fixed in the next version of GAP.