A bug in testing membership of these groups has been fixed. Bug reported by Kasper Andersen.

ConstaCyclicCode now checks that the shift factor is non-zero. Reported by M. Grassl.

A bug in meet for additive codes has been fixed; this bug could cause too large an intersection to be returned. Reported by D. Smith.

Applying Aut to a code that is not defined over a finite field now produces an error message instead of crashing.

The application of LeeWeightEnumerator to a code defined over a ring other than ℤ₄ now produces an error message instead of crashing.

The weight computation of codes over the finite ring (not field) with two elements was either incorrect or crashing, depending on platform. This bug has been fixed. Reported by A. Munemasa.

Fixed a minor leak in cyclic code creation.

Fixed some leaks in the quadratic residue code creation.

A problem with domains in maps for TwoDescent has been fixed. Reported by N. Bruin.

Some package-level issues with HeegnerPoint have been fixed; it now works again.

A minor leak in the rank computation has been fixed.

A leak in IsomorphismData has been fixed.

A bug in Rewrite which produced incorrect results when the subgroup had a trivial generator has been fixed. Bug reported by Eamonn O'Brien.

The calculation of the genus of a function field in non-simple representation has been fixed. Reported by M. Grassl.

Some problems which arose when computing Galois groups of function fields in characteristic p have been fixed.

The intrinsic HypergeometricSeries2F1 has been extended to allow rational and integral values of its complex argument.

Fixed an occasional crash when putting RWS groups into sequences. Reported by S. Humphries.

The description of Round for infinities no longer makes mention of the Gaussian integers.

The implementation of ShiftRight for negative integers has been modified so that its results now always match the documented behaviour. Reported by I. Morel.

Improved the bounds checking of Binomial so that large parameters which lead to (relatively) small results are computed now.

Three omissions in RestrictionMatrix were fixed. Bug reported and solution provided by R. Zeier.

Some assertion failures that could arise when using universal enveloping algebras have been fixed. Reported by A. Maurischat.

Factorization of inseparable polynomials over series rings has been fixed. Reported by R. Butenuth.

Loss of precision in some operations in General Local Fields has been reduced, specifically in application of automorphisms, although they are no longer created as hom<>s as a result.

Precision issues in FixedField have been better handled.

Improved matrix coercion of sequences of sequences, the case when the inner sequences were intended for coercion into the base ring is handled better. Fixes a bug reported by E. O'Brien.

The intrinsic Minors no longer takes a Signed parameter, nor are the signs adjusted any more; a new intrinsic Cofactors has been added for situations where the signs are wanted. The cofactors (and minors) are returned in such an order that coercing the sequence of cofactors into the matrix ring will produce the adjugate.

The universe of sequences returned by RowSequence is now correctly set when the matrix has no rows (or columns).

Installed a fix in the function RecogniseSU4 to improve performance. Further minor problems have been fixed in this version. Bugs reported by E. O'Brien, fixed by P. Brooksbank.

The intrinsic IsConjugate for subgroups has been corrected in the case in which the two subgroups had different orders. This had caused a runtime error when two variables were assigned. Bug reported by K. Andersen.

A crash that occurred when IsDisjoint was called on multisets has been fixed.

The operators diff, diff:=, sdiff, and sdiff:= now work for multiset arguments. This lack was pointed out by K. Andersen.

A crash which occurred when attempting to call MinimumCut having the same vertex as source and sink has been fixed (it now raises an error instead.) Reported by M. Grassl.

A missing case in OptimizedRepresentation for orders that could cause a runtime error in the code for solving relative norm equations has been fixed. Reported by E. Thome.

The quo constructor has been modified to always perform a degree reduction. Inconsistency reported by Kasper Andersen.

The description of Round has been corrected to indicate that it rounds away from zero, not toward infinity.

A bug in the printing of complex numbers as polynomial coefficients has been fixed. This could result in the sign of the imaginary component being printed incorrectly in the case of the constant term. Reported by D. Yasaki.

The Handbook description of Round has been corrected to indicate that it rounds away from zero, not toward infinity.

A bug in printing real numbers in printf has been fixed.

A bug in WriteGModuleOver that would cause an internal error runtime has been fixed. Reported by J. Schmidt.

A typo in Minimize for representations that prevented the All flag from being be passed on has been fixed. Reported by J. Schmidt.

Some memory leaks in the scheme machinery have been fixed. Reported by A-S. Elsenhans.

A problem in TangentCone for non-hypersurfaces has been fixed. The function now employs a local Groebner basis computation.

Changes have been made to Is(Non)Singular and SingularSubscheme to address the problem of possibly erroneous results for subschemes of weighted projective space and other ambients not covered by standard affine patches. For a weighted projective scheme X, calls to these intrinsics will now give a runtime error if it is determined that the current singularity-checking methods made fail on a non-empty subset of X. For schemes in newer, more general toric ambients, a runtime error is now automatically generated (we will try to update the methods used to deal with this problem in later releases). Problem reported by Tony Varilly.

A bug has been fixed that could cause a crash when coercing symmetric functions over certain rings (in particular, rational function fields and number fields) into polynomial rings.

A bug has been fixed that could cause the result to lie in an underlying ring of the polynomial ring argument under some circumstances.

Several bugs relating to the conversion of symmetric functions to and from polynomial ideals were fixed. These bugs could cause the result to crash or erroneously succeed when the result did not lie in the ideal.

A minor leak that could occasionally arise when hashing certain types of objects has been removed.