Factorization of polynomials over some representations of function fields has been fixed.
When a non-simple extension is constructed it is now checked that the resulting extension will be separable and if not an error occurs.
Memory handling has been fixed when constructing a quotient of an order by an existing ideal.
The parametrization routine for degree 9 Del Pezzo surfaces over the rationals ParametrizeDegree9DelPezzo has been speeded up by use of additional basis normalisations and an alternative algebra trivialisation algorithm. It also now returns an isomorphism with inverse map data included. ParametrizeDelPezzoDeg9 is now simply a synonym for this main intrinsic.
A bug in qCoverDescent reported by N. Freitas has been fixed.
A problem when coercing between function fields (affecting PrimaryDecomposition) has been fixed. Reported by D. Kohel.
A crash in SyzygyModule has been fixed. Reported by M. Reid.
A bug in TorsionBound for elliptic curves over a number field has been fixed.
A crash that could arise when computing automorphisms of planes has been fixed. Reported by J. Key.
Magma-level printing of planes now uses the correct constructor name.
The version of nauty included within Magma has been upgraded to 2.5r2.
A crash when computing the socle of a permutation group has been fixed.
A crash computing AbelianQuotient of an fp-group has been fixed. Bug reported by Derek Holt.
A crash, and some issues with the algorithm, when computing Centre of a pc-group have been fixed. Crash reported by Stephen Humphries.
A bug in IsIrreducible applied to a group character constructed by Induction has been fixed. Bug reported by Derek Holt.
A bug where the PolycyclicGroup constructor failed to detect an uncollected word appearing as the RHS of a relation has been fixed. Such a relation will now give a runtime error. Reported by E. O'Brien.
A small collection of problems that can arise when computing the socle of a primitive permutation group having degree in the range 106 to 107 has been fixed.
A crash due to the incorrect construction of the order of the matrix group GOPlus(2,4) has been fixed. Bug reported by U. Thiel.
A bug with computing the Hodge structure of a tensor product was fixed. Reported by D. Roberts.
The variable parameters to SmallRoots have been changed to allow more specific types than just RngElt.
A bad slowness in exact division of multivariate polynomials has been fixed. Reported by M. Grassl.
The new intrinsic EdgeFacetIncidenceMatrix can be used to calculate the edge-facet incidence matrix of a polyhedron.
The new intrinsic NormalEdgeCones can be used to recover the (outer) normal cones of the edges of a polyhedron. When working with polytopes, this is significantly faster than the equivalent sequence of calls to NormalCone.
A bug in QuadraticClassGroupTwoPart for integers with large even 2-valuation and an odd number prime factors 3 mod 4 was fixed. Reported by J. Stankewicz.
There are new intrinsics for Segre embedding - SegreEmbedding and SegreProduct - and for blowing up points or subschemes: BlowUp. These have already been documented in the online version of the Handbook.
The RemoveLinearRelations intrinsic now copies across attributes from the original scheme to the scheme that is returned.
The dense back end of the sparse Smith algorithm has been improved. Problem reported by E. O'Brien.
Creating a real field with 52 bits of precision no longer produces a copy of the parent of timings.
A problem with indicating the wrong source line on bad sequence indexing has been fixed. Reported by S. Humphries.