2.5.7

• Fixed a segmentation fault occurring after using invalid types in tuples or buffers as function parameters (#169)

2.5.6

• Fixed sign, uaddCarry, usubBorrow, umulExtended, imulExtended not accepting integer inputs

• Changed bitfieldExtract, bitfieldInsert, bitfieldReverse to require ctypes types to determine the int type to use

• Fixed bitCount, findLSB, findMSB not using the optimal int type

• Fixed equal

• Fixed integer overflows causing a crash (now raising a warning instead)

• Deprecated frexp warning

• Added tests from glm

• Added support for aarch64 (#165, #166) and s390x linux wheels, as well as arm64 macOS wheels.

• Also added musllinux builds

Thanks to @odidev for helping out with the aarch64 wheels!

2.5.5

Fixed #163

2.5.4

• Fixed two memory leaks (#159)

2.5.3

• Made iterators iterable (#156)

2.5.2

• Fixed tuple type checking (#153)
• Added official support for Python 3.10

2.5.1

Fixed mvec instantiation (#152)

2.5.0

• Enhancements related to #139
  • rotate, translate and scale no longer require an input matrix
  • Added support for matrix initialization from quaternions (mat3, mat4)
  • Added support for iterable arguments for min and max
  • Added support for homogenous multiplication (#140)
• Improved performance comparison (#144)
• Allowing for C contiguity in matrix initialization from buffers (#138)
• Fixed a bug where NxM buffers would be interpreted as MxN matrices
• Fixed a few other bugs (2dd31f3)

2.4.0

• Added bitwise operators to int vec types
  • Added invert operator
  • Added bitwise-and operator
  • Added bitwise-or operator
  • Added bitwise-xor operator
  • Added lshift operator
  • Added rshift operator
• Improved array sorting algorithm
• Updated doc

2.3.1

• Fixed CodeQL alerts
• Fixed Python 3.10 incompatibility