---
title: "Magma V2.21-5 Patch Notes"
canonical: "https://docs.magma-maths.org/patchnotes/2.21-5.html"
source: "https://docs.magma-maths.org/patchnotes/2.21-5.md"
---

<a id="magma-v2-21-5-patch-notes"></a>
# Magma V2.21-5 Patch Notes

## Algebraic Number Fields

- An iterator has been provided to iterate through elements of quotients of orders of relative number fields by their ideals.  Reported by M. Kirschmer.
- It is now possible to construct `Random` elements of quotients of orders of relative number fields by their ideals.
- Compatibility has been improved for orders of number fields. This allows the comparison of orders having the same coefficient ring using `eq`. Reported by M. Kirschmer.

## Commutative Algebra

- A problem in comparison of module elements has been fixed.  Reported by C. Quitte.

## Dirichlet Characters

- A problem with assigning names to a Dirichlet group generators was fixed.

## Elliptic Curves

- The [`MinimalQuadraticTwist`](../ArithmeticGeometry/EllipticCurves/creation.md#function-minimalquadratictwist-crvell) intrinsic when applied to a rational whose numerator has a prime-power divisor $p^e$ with $6|e$ and $p>3$ congruent to 3 mod 4 could return a curve with a different $j$-invariant. Reported by A. Sutherland.

## Groups

- A bug in `Subgroups` when using `IndexEqual` set to an impossible index has been fixed.  Reported by A. Sutherland.
- A crash using the transversal map with groups of type GrpFPCox has been fixed.  Reported by B. Howlett.
- If explicitly setting all of the conjugacy classes of a group, they are no longer re-ordered if they are already sorted.  Previously, elements with the same order and length could be re-ordered in this case.  Reported by S. Goldstein.
- A crash when computing preimages with a [`BlocksAction`](../FiniteGroups/PermutationGroups/actions.md#function-blocksaction-grpperm-any) map has been fixed. Bug reported by Juergen Klueners.

## Lattices

- Using the `Max` parameter with `ShortestVectors` could lead to incorrect caching of partial information, and give erroneous results for later enumerations. Reported by C. Doberstein.

## Series Rings

- The `IsSquare` intrinsic for series over the rationals whose constant coefficient (after removing the valuation) was not 1 could give wrong answers due to `Log`/`Exp` not working. The same problem can also occur with `IsPower` more generally. Reported by M. Stoll.
- `Roots`, to a given precision, of polynomials over series rings over inexact rings have been disallowed.
