The bound used internally to compute class groups of cyclotomic extensions has been increased. The old bound was too small for certain examples.
A bug in the computation of RayResidueRing (and hence RayClassGroup has been fixed. The bug resulted in too small groups to be returned. Reported by C. Weir.
A MinimumLeeWeight attribute has been added for codes.
A bug that was causing runtime errors in IsIsomorphic for coherent sheaves has been fixed. Reported by C. Quitte.
A problem in computing a free resolution of a graded module with negative weights has been fixed. Reported by C. Quitte.
Fixed a bug in CremonaReference that stopped it identifying non-minimal curves. Reported by E. G. Jiménez.
Calling HeightPairingMatrix on a null sequence no longer crashes. Additionally, a memory leak when specifying alternative precisions was fixed.
EllipticCurve now works for a hyperelliptic curve and a non-integral Weierstrass point. Reported by M. Stoll.
IntegralPoints occasionally missed points due to a trivial error, which has been fixed.
The Cremona database has been updated to include curves up to conductor 140000.
An infinite recursion bug in the case of ideals with non-trivial gradings has been fixed. Reported by E. Rains.
A runtime error when computing the character table of a matrix group over the ring of integers modulo m has been fixed. Bug reported by I.M. Isaacs.
A crash due to memory management problems with left and right cosets of a permutation group has been fixed.
The orders of minus type, and all conformal, orthogonal groups are now stored with the group on creation. Omission reported by E. O'Brien.
A bug in PCGroup when applied to a pc-group has been fixed. Bug reported and fixed by D. Howden.
P. Brooksbank's package for constructive recognition of SU3, SU4 and Sp4 in even characteristic has been updated.
A crash when applying Evaluate with 2nd argument a group automorphism group has been fixed. Bug reported by D. Howden.
A crash in AutomorphismGroup for lattices has been fixed. Reported by G. Bellot.
A crash in AutomorphismGroup for lattices has been fixed. Reported by G. Collinet.
Fixed a bug where ChevalleyBasis and related functions could fail due to an unlucky choice of random elements.
Fixed a problem where MatrixLieAlgebra (to make a Lie algebra from a normal matrix algebra) would return a matrix Lie algebra with too few elements. Bug reported by A. Previtali.
Changed the output of CartanName so that the string returned does no longer contain any trailing whitespace.
Fixed a problem in DirectSum for root data that resulted in crashes in the algorithms for representation decompositions and in ReductiveType for Lie algebras. Bug reported independently by W. de Graaf and R. Zeier.
Fixed a crash in SymmetricPower for representation decompositions of non-semisimple root data. Bug reported by W. de Graaf.
Fixed a number of memory leaks in the algorithms for calculating with representation decompositions.
Fixed a number of bugs in the construction of representations for non- semisimple groups of Lie type (in particular those returned by the AdjointRepresentation and StandardRepresentation intrinsics) where the map returned would not be a homomorphism.
A bug in HighestWeightModule and HighestWeightRepresentation, occurring in certain cases due to incompatibility of root orders, has been fixed. Bug reported and fix provided by W. de Graaf.
The strategy in SplittingField for exact polynomials over local rings has been changed to avoid unneccessary precision loss. Reported by T. Dokchitser.
SplittingField for exact polynomials over local rings now supports a parameter to not compute a optimized presentation of the result to avoid precision problems.
HenselLift of a root of a polynomial over a p-adic field to a default precision no longer returns roots to the wrong precision. Reported by T. Matev.
Fix the function Ngens for matrix algebras of degree 0. Bug reported by A. Previtali.
Calling MaximalOrder with the Disriminant parameter and a non-monic defining polynomial exposed an internal error. Reported by N. Bruin.
The calculator exhibited a problem in the sub<|> constructor for number fields: in fields with too many small degree subfields it occaisonly failed to find a primitive element.
In some strange circumstances, the regulator computed using Regulator of some order was wrong due to badly chosen internal precision.
A crash involving IsSymmetric has been fixed. Reported by N. Bruin.
A bug in polynomial squaring over non-commutative rings has been fixed. Reported by A. Previtali.
Added the intrinsic IsIntegrallyClosed. Requested by G. Hegedüs.
A crash has been fixed when dividing power series defined over a ring without a division algorithm.
IsPrincipal is now allowed for ideals over rings of type RngQuad and RngUPol.
Fixed IsIsomorphic when twosided and onesided ideals where mixed.
Printing of symmetric functions as coefficients of polynomials or series has been improved to appropriately bracket the terms.
Printing expressions in the debugger now works when reference parameters to procedures are involved. Reported by D. Howden.
An oversight in the deletion of tuples which could cause internal errors in rare situations has been fixed.