Testing the zero ideal for being square is now handled.
MinimalPolynomial has been fixed for non integral elements when it differs to the characteristic polynomial. Reported by F. Voloch.
Testing the zero ideal for being square is now handled.
Computing GCDs of elements of quotients of an order of a number field by an ideal has been fixed.
A crash in pFundamentalUnits has been fixed.
A specialised implementation for Trace of elements of cyclotomic fields has been provided which will use less memory in large degree fields. Reported by J. Voight.
A number of bugs have been fixed in ResolveSingularSurface for blow-up desingularization and also in the IntersectionMatrix computation of intersection numbers for the blow-up divisors. The result of the bugs was that irreducible blow-up components were sometimes erroneously duplicated and that, because of both related and independent problems, intersection numbers were incorrect or would cause runtime errors. Both problems have now been fixed.
The difference between a divisor and a place of a curve has been fixed when the base fields of the curves of the divisor and the place are different.
A problem with non-monic models has been fixed. Reported by S. Hashimoto.
A crash in SemidirectProduct has been fixed.
A crash in ClassicalMaximals (with type "U", dimension 16, field of characteristic 3, and normaliser or general flag set) has been fixed.
An improved recognition of the cyclotomic case for Jacobi motives has been applied, and an alternative method to reconstruct Euler factors from their p-adic roots is used in large degree. Noted by F. Rodriguez Villegas.
A modification to the factorization routine has been applied to remove large powers more efficiently. Reported by M. Grassl.
The BKZ intrinsic now gives an error when stability problems are detected, rather than simply returning the partially reduced lattice.
The multiplication algorithm for matrices defined over GF(3) has been speed up in the AVX executable. The CUDA algorithm over the same type of matrices has also had a fix in large dimension.
The Irreducible parameter on ChevalleyGroup for F4 in characteristic 3 now correctly returns the 25-dimensional representation, not a 26-dimensional one.
A problem with Basis for half-integral weight forms has been fixed, involving auxiliary calculations in trivial spaces. Reported by M. Somos.
A problem with NewformDecomposition in square level has been patched. Reported by A. Sutherland.
A numerical round-off problem with InnerTwists has been fixed. Reported by A. Sutherland.
Reduction to compute flat closures in projective space now works properly over number fields. Bug reported by J. Rouse.
A crash in AttachSpec has been fixed. Reported by E. O'Brien/J. Maglione.