An equality test in the unit tests failed on MacM1, so I relaxed its tolerance.

Elliptic alpha function.

Rogers-Ramanujan functions.

Jacobi theta function with characteristics.

Allows a negative nome.

Some conformal mappings.

Nome in function of the parameter.

The logarithms of the Jacobi theta functions were not in the principal branch.

The Dedekind eta function is now vectorized.

Changed the expression of the

`kleinj`

function in order that its factors avoid a possible float overflow.Major changes in the implementation of the Jacobi theta functions, following the new Fortran implementation by Mikael Fremling.

More unit tests.

The

`halfPeriods`

function did not work for a pair of real (`numeric`

) numbers. This has been fixed with the help of`as.complex`

.

Lemniscate elliptic functions.

Dixon elliptic functions.

Some values of the Jacobi theta functions were wrong as of version 2.1.0.

Added some unit tests.

New function

`halfPeriods`

, computing the half-periods from the elliptic invariants.New function

`ellipticInvariants`

, computing the elliptic invariants from the half-periods.

The case when the elliptic invariant

`g2`

is zero is now handled.The method computing the half-periods ratio when the elliptic invariants are given led a wrong sign sometimes.

- Minor fix in the C++ code.

Weierstrass sigma function.

Weierstrass zeta function.

Costa surface.

Vectorization.

Better accuracy.

First release.