Magma V2.13-15 Patch Notes

Elliptic Curves¶

• The identity point on an elliptic curve over a p-adic field behaves somewhat better now. In particular, it is possible to coerce the coordinates of the identity point back into the curve.
• IsogenyFromKernelFactored applied to a group scheme now works (formerly triggered a runtime error).
• Fixed a bug in IsogenyFromKernel that would cause the resulting isogeny to be incorrect when the kernel contained the full two-torsion and the base field was not the rationals or a small finite field.

Hyperelliptic curves¶

• The map phi returned by TwoSelmerGroup was completely wrong in some cases, and has been corrected. (This problem was introduced in V2.13.) Reported by Patrick Corn.

Invariant Theory¶

• A bug in Module, when the secondaries were set via the attribute, has been fixed. Reported by D. Wehlau.

Lattices¶

• Problems lattice vector enumeration has been fixed (when the bound was close to the machine int size). Problem reported by Ronald van Luijk.
• A problem with the construction of sub-lattices of lattices with real valued gram matrices has been fixed. Without this, Magma would encounter internal errors.

Models of genus one curves¶

• A mistake in the HyperellipticCurve constructor has been corrected. Prior to this, for a model $f(X)$ of degree $2$ without cross-terms, HyperellipticCurve returned $y^2 = -f(x)$ instead of $y^2 = f(x)$. Reported by Bjorn Poonen.

Modules over Dedekind domains¶

• A bug in the handling and printing of parents of pseudo-matrices has been fixed. Without this fix, a sequence of pseudo matrices could cause internal errors. Bug reported by U. Staemmler.

Number Fields¶

• Several minor problems in the class group computation of number fields have been fixes. Together, they would trigger internal errors. Reported by N. Bruin.
• A bug in the unit group computation that, especially in fields with unit rank $\ge 5$ and small regulator could result in infinite loops, has been fixed. Bug reported by M. Kirschmer.

Polynomial Rings¶

• A bug in factorization of polynomials over function fields over quotient rings has been fixed. Reported by N. Bruin.

System¶

• Some bugs relating to interactions between integer factorization sequences (RngIntEltFact) and normal sequences (SeqEnum) have been fixed.