--- title: "Magma V2.28-26 Patch Notes" canonical: "https://docs.magma-maths.org/patchnotes/2.28-26.html" source: "https://docs.magma-maths.org/patchnotes/2.28-26.md" --- # Magma V2.28-26 Patch Notes ## Algebraic Number Fields - A crash has been fixed in the computation of `UnitGroup`s. - A crash in `ext<>` for number fields when given a sequence of integers has been fixed. Reported by Lewis Combes. ## Elliptic Curves - The functions `WeilPairing(P, Q, n)` and [`TatePairing(P, Q, n)`](../ArithmeticGeometry/EllipticCurvesOverFiniteFields/pairings.md#function-tatepairing-ptell-ptell-rngintelt) now first attempt to coerce Q into the parent of P if these parents are not identical. Issue reported by Sachi Hashimoto. ## Hyperelliptic Curves - The function [`HeightConstant`](../ArithmeticGeometry/HyperellipticCurves/jacobians-number-fields.md#function-heightconstant-jachyp) now sometimes gives a better bound for the difference between naive and canonical height on Jacobians of genus 2 curves, for instance in Example 6.2 of the recent paper of Garcia-Fritz and Pasten. Improved code supplied by Steffen Mueller and Michael Stoll. ## Integers - Fixed a bug with [`ShiftRight`](../BasicRings/Integers/operation-element.md#function-shiftright-rngintelt-rngintelt) that would return an answer one too low if a negative number was shifted and the result should have been exact. ## Language - A crash in assignment (in a certain situation involving multiple RHS values) has been fixed. Reported by Antoine Joux. ## Lattices - A crash in [`ClosestVector`](../LatticesQuadraticForms/Lattices/enumeration.md#function-closestvector-lat-modtuprngelt) involving a lattice with very large entries in its basis matrix has been fixed. Reported by Markus Kirschmer. - A crash in [`HKZ`](../LatticesQuadraticForms/Lattices/reduction.md#function-lat-hkz) with the `Unique` parameter and a transformation matrix matrix has been fixed. Reported by Hao Guo. ## Matrices - A bad slowdown when computing the Smith Normal Form with transformation matrices for a sparse matrix defined over a ring which is not the integer ring has been fixed. ## Matrix Groups - An incorrect result for `CompositionTreeCompositionFactors` applied to the Lyons group has been fixed. Reported by David Craven. - A new optional parameter `Sign` has been added to [`DoubleCoverSymmetricGroup`](../FiniteGroups/AlmostSimpleGroups/double-covers-alt-sym.md#function-doublecoversymmetricgroup-rngintelt) which is the opposite of the existing `Type`. Added by Derek Holt after discussion with David Craven. ## Polynomial Rings - A crash in `Discriminant` (arising from [`IsEasySnAn`](../GlobalFields/GaloisGroupsAndAutomorphisms/galois.md#function-iseasysnan-rngupolelt)) for very large integer polynomials has been fixed. The algorithm has been sped up at the same time. Reported by Nick Katz.