A bug in converting sparse cyclotomic elements with denominators into complex fields has been fixed.
Coercing elements of cyclotomic fields into a complex field has been made more efficient. Reported by M. Grassl.
A problem with LLL precision in EllipticCurveSearch has been remedied via the usage of try/catch. Reported by A. Brumer.
Groebner basis computation over number fields (when the coefficients lie in the rational field) has been sped up further. Problem example supplied by C. Bonnafe.
The ConjugacyClasses function for type GrpPermCox has been made faster. Problem reported by U. Thiel.
A crash in SemidirectProduct has been fixed. Bug reported by Maarten Derickx.
A bug in the map returned by AbsoluteRepresentation has been fixed. Reported by E. O'Brien.
The ClassGroup computation for quadratic forms with discriminant 1 mod 4 has been corrected. Reported by E. Mulder.
Real and Complex fields may now be included in sets.
A bug in root finding for a non-monic binomial polynomial has been fixed. Reported by N. Bruin and M. Stoll.
Universes and extended types of sequences of sequences of factorization sequences now correctly retain the factorization sequence information.
Equality comparison of sets no longer checks every element when both sets are arithmetic progressions.
The main sparse Gaussian elimination algorithm has been greatly improved for very large sparse matrices, in reduction of memory usage and time taken. Even greater reduction of memory usage and time taken has been achieved for very large sparse matrices over GF(2). New parameter MakeWider has been added to Rank for sparse matrices, to specify that the elimination algorithm should transpose the input if need be to ensure that the number of columns is greater than or equal than the number of rows (this help a lot with certain types of input). Sparse matrix inputs originally causing memory blowup reported by S. Goldstein.