Patch Notes
Magma V2.17-12 Patch Notes
Patch release for Magma V2.17-12. Updated areas: Algebraic Function Fields, Algebraic Number Fields, Finitely Presented Algebras, Global Arithmetic Fields, Graphs and 6 more.
Algebraic Function Fields¶
- Deletion of a prime ideal after an element has been expanded at it has been fixed.
Evaluateat a place of an element with product representation has been sped up.- A fix has been made to multivariate factorization and primary decomposition over non-simple function fields.
- Some defining polynomials of infinite orders now have smaller coefficients.
- The computation of the
Discriminantof a polynomial over an order of a function field is faster now using the subresultant algorithm. Computation of theDiscriminantof an order in relative representation is also faster as a result.
Algebraic Number Fields¶
- An error instead of a crash now occurs when using
quo<FldNum : | >with a reducible polynomial on the right hand side. - An improvement has been made to principal ideal testing when the norm of the ideal is large. Reported by F. Bouyer.
Finitely Presented Algebras¶
- Some problems in finitely presented algebras have been fixed. Reported by A. Kasprzyk.
Global Arithmetic Fields¶
- Powering of elements of coefficient rings of extensions has been optimized by using the coefficient ring for the powering instead of the extension which is the parent. Reported by M. Grassl.
- Some
GaloisGroupcomputations have been fixed. GaloisGroupof a non-simple extension now gives an error instead of a crash. Reported by D. Grünewald.- A internal fix has been made to powering of elements.
- The computation of primes in the discriminant of a Kummer or Artin--Schreier extension in the
MaximalOrdercomputation has been improved. - The
Alparameter toMaximalOrderhas been improved. SelectingAl := "Round2"orAl := "Round4"will avoid the use of the special Kummer and Artin-Schreier extension algorithms.
Graphs¶
AssignWeightsnow produces an error message instead of crashing when weights are provided for non-existent edge indices.
Groups¶
- A bug where the subgroups of a pc-group were not stored with the group when computed has been fixed.
- A crash in code for finding the socle of a primitive permutation group has been fixed.
- Easy inputs for
IsCompletelyReducibleandIsNilpotentByFiniteare handled more sensibly. - Parameters for
LieTypeandRecogniseSU4are altered to improve performance for certain types of input. - A crash when computing the conjugate of a character by an element of a supergroup ahs been fixed.
Modules¶
- An incorrect result for
IsIrreduciblefor some modules over the rational field has been fixed. Reported by M. Kuenzer.
Polytopes and Polyhedra¶
IsSmoothwas erroneously returning true for non-Fano polytopes.
Schemes¶
- A memory fix has been made to maps between schemes. Reported by D. Kohel.
Sequences¶
- A bug has been fixed in
ParallelSort; this bug could cause incorrect results when exactly one of the sequences contained booleans. Reported by T. Feulner.
System¶
- A possible crash when calling a variadic intrinsic has been fixed. This could occur if the first variable argument was an expression whose value was not stored elsewhere. Report by A. Kasprzyk.