A memory problem in the complex field has been fixed. Previously certain operations (computation of zeros and -1st powers of complex numbers) would cause the memory to fill up.
A crash in creation functions for group databases has been fixed. Reported by M. Kirschmer.
The HasCM function now uniformly returns the actual (negative) discriminant, rather than its absolute value. Reported by Wolfgang Kroworsch.
Some problems with the use of the modular equation databases by SEA for point counting on elliptic curves have been fixed. It should be noted that the standard Magma exported database has modular data for primes up to 400. There is, however, a (rather large) optional database containing extra modular data for most primes below 1000. A warning message has also been added to SEA if the algorithm tries to use auxilliary primes above 400 and no modular data is found in the library files. This informs the user that the extended optional database is available.
A bug when computing NormalClosure in PC-groups has been fixed.
A bug when remaking PC-groups from a compact presentation (integer sequence) has been fixed. The bug showed up by causing crashes in Centre, Descendants and possibly other routines too. Reported by H. Nover.
Bugs which caused crashes when constructing subgroups of finitely-presented groups have been fixed. Reported by J. Ellenberg and D. Holt.
A bug when the pQuotient function was used on a group with trivial quotient, and the quotient map is also requested, has been fixed. Reported by D. Holt.
Fixed sign issues in odd characteristic A-infinity computations.
A slowdown when a large number of identifiers are assigned has been fixed. Reported by Y. Edel.
A missing error check for the function Fix has been put in. Reported by T. Dokchitser.
Several problems in Newton polygons defined by rational points have been fixed. The previous version, when used on polygons where the numerator or denominator exceed 214 could cause memory corruption.
A bug the factorisation of polynomials over number fields has been fixed. The bug would cause the memory corruption and trigger (eventually) internal errors, eg. during the inclusion test (IsSubfield) for number fields. Reported by N. Bruin.
A bug in the computation of over structures for quadratic fields has been corrected, the wrong over structure computed previously whould trigger internal errors. Reported by D. Taylor.
A bug the the handling of non-simple fields for subfield lattices has been fixed. Previously, instead of runtime errors, internal errors could be triggered. Reported by M. Grassl.
A bug in the creation of the underlying algerba of a number field (especially a cyclotomic field) has been fixed, without this fix internal errors would be triggered. Reported by M. Grassl.
A bug the computation of ring class groups of quadratic fields has been fixed, certain examples could trigger runtime errors.