Skip to content

Latest commit

 

History

History
159 lines (120 loc) · 6.69 KB

CHANGELOG.md

File metadata and controls

159 lines (120 loc) · 6.69 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Next

Added

  • Package extensions for LinearAlgebra and StaticArraysCore to add methods to LinearAlgebra.dot, LinearAlgebra.normalize, StaticArraysCore.similar_type.
  • Package extension for Quaternions.jl to allow for interoperability of the Quaternion type with AbstractCliffordNumber.

Removed

  • [BREAKING] CliffordNumbers.normalize is no longer exported to avoid a name conflict with LinearAlgebra.normalize.
  • [BREAKING] Support for Julia 1.8 due to the use of package extensions.

0.1.10 - 2024-11-06

Added

Fixed

  • isinf, isnan, isreal, isinteger, iseven, and isodd did not have working methods for AbstractCliffordNumber arguments.

0.1.9 - 2024-10-28

Added

  • A package extension for interoperability with Unitful.jl (only supported on julia 1.9 and up). It provides support for wedge products for Quantity, including Quantity{<:Real} and Quantity{<:AbstractCliffordNumber} (geometric products were already supported).
  • nonzero_grades(::Complex) is defined and returns 0:0, like nonzero_grades(::Real).
  • print(::IO, ::AbstractCliffordNumber) shows a prettier (but not parseable) representation.

Changed

  • Indexing and equality checking of BitIndices is slightly more efficient.

0.1.8 - 2024-09-23

Added

  • Implementations of flipsign and copysign for BitIndex.
  • +(::BitIndex) returns the input.
  • Faster multiplication methods for Rational coefficients.

Changed

  • Geometric products of BitIndex objects are faster.
  • scalar_product is implemented with a generated function independent of CliffordNumbers.mul.
  • normalize returns x if abs2(x) is zero.
  • Indexing of Clifford numbers with Rational scalar types is significantly faster due to the use of flipsign instead of multiplication.
  • KVector indexing is now faster.

Fixed

  • scalar_product had no method to handle mismatched scalar types.
  • abs now returns a Real result: it is equal to sqrt(abs(abs2(x))).
  • Indexing KVector does not throw an error if the first entry is 1//0.

0.1.7 - 2024-09-11

Added

  • Base.literal_pow definitions for KVector{1} that produce KVector results.

Changed

  • Natural exponential are only optimized for KVector types whose BitIndices uniformly square to the same value.

Fixed

  • OddCliffordNumber constructors now fail to operate on scalars unless there is exactly one odd element in the algebra.
  • Broken bullets in AbstractCliffordNumber docstring.
  • StackOverflowError in KVector{0} exponentiation.
  • Incorrect natural exponentials in algebras with non-uniform sign signatures.
  • isequal(x::AbstractCliffordNumber, y::AbstractCliffordNumber) gives the correct result, no longer equal to x == y for signed zeros and NaNs.

0.1.6 - 2024-06-27

Added

  • Definition of signature(::BitIndex{Q}).
  • Unexported aliases CliffordNumbers.CliffordScalar{Q,T} === KVector{0,Q,T,1}, CliffordNumbers.CliffordVector{Q,T,L} === KVector{1,Q,T,L}, and CliffordNumbers.CliffordBivector{Q,T,L} === KVector{2,Q,T,L}.
  • @basis_vars macro that defines basis 1-blade variables for an algebra.

Changed

  • Split up src/math.jl to separate files in src/math/.

0.1.5 - 2024-06-10

Fixed

  • Missing exports for regressive and ∨ (#14 - thanks @ajahraus!)
  • Actually changed the implementations of scalar_product, abs2, and abs.

0.1.4 - 2024-06-06

Added

  • Linear extension of left and right complements to all Clifford numbers.
  • Regressive product (∨) definition.

Changed

  • scalar_product, abs2, and abs are much faster: they just extract the scalar portion of the result of CliffordNumbers.mul(x, y, CliffordNumbers.GradeFilter{:*}())

Fixed

  • Incorrect products for elements of algebras with negative-squaring elements.
  • Incorrect complements for elements of algebras with non-positive-definite signatures.
  • Incorrect normalization of division of scalars by Clifford numbers.

0.1.3 - 2024-06-04

Added

  • Left complement (left_complement()) and right complement (right_complement()) of BitIndex objects.

Fixed

  • Incorrect definition of abs2 for non-positive-definite and degenerate metrics.

0.1.2 - 2024-05-31

Added

  • Base.float and Base.big definitions for AbstractCliffordNumber types and instances.
  • Base.literal_pow definitions for AbstractCliffordNumber instances raised to constant powers, allowing the powers of KVector, EvenCliffordNumber, or OddCliffordNumber to be inferred as either EvenCliffordNumber or OddCliffordNumber depending on the exponent.

Changed

  • Geometric products involving pseudoscalars (KVector{K,Q} where K === dimension(Q)) now promote to smaller types if possible.
  • CliffordNumbers.mul iterates through the indices of the smaller argument, which drastically reduces the performance discrepancy when the multiplication arguments are reversed if SIMD vectorization is utilized.

0.1.1 - 2024-05-28

Changed

  • @inline annotations have been provided for all products.

0.1.0 - 2024-05-17

Initial release of CliffordNumbers.jl