Magma V2.18-5 Patch Notes

Algebraic Geometry¶

• A trivial problem (resulting in runtime errors) in equality testing for resolution graphs has been fixed. Reported by C. Gorzel.

Associative Arrays¶

• The operator # now works for associative arrays (returning the number of keys).

Coding Theory¶

• A bug in ParityCheckMatrix has been fixed; this bug could cause the parent of the returned matrix to have the wrong number of rows when the code was defined over $Z_4$. Reported by M. Grassl.

Commutative Algebra¶

• Some crashes in PrimaryDecomposition when weights are present have been fixed. Reported by S. Jambor.
• A crash in Variety over real fields has been fixed.

Elliptic Curves¶

• A bug in TatePairing and ReducedTatePairing has been fixed. This was producing wrong results for the pairing of P and Q when P=Q is of order two. Reported by B. Allombert.
• Saving a workspace when the elliptic curve database is around no longer causes the database to return read errors thereafter.
• The elliptic curve database has been updated to conductor 229,999.
• A bug with ManinConstant of 27a4 and 80b4 was fixed. Reported via the Sage bug tracking system.
• MinimalModel was accidentally disabled for non-totally real fields; it is now restored to normal. Reported by E. Gonzalez-Jimenez.

Function Fields¶

• A fix has been made to the bounds used in GaloisSubgroup and GaloisQuotient in characteristic p and to bounds used for reducible polynomials.
• Improvements have been made to MaximalOrder computations in Artin--Schreier extensions. The use of a special algorithm for these extensions has been limited to function fields whose constant field is perfect.

Groups¶

• A crash when computing the AbelianQuotient of a permutation group has been fixed.
• A bug in the IsTransitive filter for the Subgroups command, which, for larger groups, caused the whole group to be missed, has been fixed. Reported by G. Royle and by D. Holt.
• The ClassicalMaximals routine has been fixed in the cases of $Sp(12, 3)$, $O_+(10,q)$ and $O_-(10,q)$, where there were errors in the subgroups returned. Fixed by D. Holt, who noticed the errors.
• The LowIndexNormalSubgroups routine has been given an improved method of computing maximal submodules of an elementary abelian section, following comments by M. Conder.
• A bug in LowIndexSubgroups for permutation groups has been fixed. Reported by A-S. Elsenhans.
• A crash in matrix groups that showed up as a crash in Subgroups has been fixed.
• A new permutation group intrinsic pCoreQuotient has been added at the request of D. Holt.
• A crash when computing the automorphism group of a matrix with trivial automorphism group has been fixed. Reportd by M. Grassl.

Hyperelliptic Curves¶

• A bug in TwoCoverDescent involving excessively high levels of precision in p-adic fields has been fixed. This was preventing the algorithm from terminating in some cases. Reported by E. Gonzalez-Jimenez.

Integers¶

• A crash when calling Binomial or NumberOfPermutations on large integers has been fixed.

Lie Theory¶

• The intrinsic IsSimple for Lie algebras has been fixed to return false for Lie algebras of dimension 1. Reported by R. Zeier.
• A crash in the package for Lie algebras generated by extremal elements has been fixed.

Number Fields¶

• Null sequences are now handled in PseudoMatrix given a sequence and a matrix.

Quadratic Fields¶

• A crash when the map from the class group is reconstructed and applied has been fixed. Reported by M. Kirschmer.

Quaternion Algebras¶

• Trivial errors in implementation of the mass formula have been corrected. (This caused bugs in enumerating left or right ideal classes.) Reported by D. Smertnig.

Rationals¶

• An obscure bug in rational subtraction has been fixed. Reported by E. Thomé.

System¶

• A bug has been fixed that could cause crashes when using eval with a database around.