* Try running tests on all three operating systems
* Try adding an MSVC build
* Make tests work on Windows
* Try reconfiguring Travis for this
* Maybe it's because we're in git bash
* Try explicitly doing something else for travis
* Remove a part I think is unnecessary
* Run the test EXEs as they compile
* Add coverage features and add it, laboriously, to everything
* Fix easy tests
* Add tests for != operators
* Clean up test framework a little
* Add documentation of coverage macros
* Fix tests for mat4 to quaternion
* Slightly improve formatting of coverage output
* Trailing whitespace must die
* Add array subscript operators for all types
* Taking the parameter for the operator[] as a reference. This should allow it to be inlined
* I guess you can't do that.
* Update version and readme
* Convert everything to new inline scheme
* Add both extern and inline for some SSE vs. non-SSE functions
Some functions had nice, compact SSE implementations but bulky non-SSE
implementations, so this commit inlines just the implementations that make
sense. Also updated documentation.
* Convert HINLINE and HEXTERN to HMM_INLINE and HMM_EXTERN
* Fixed bug when using handmademath in C mode
* SSEd vec4 operations
* Fixed hmm_vec4 for non-sse builds. Added SSE option for HMM_Vec4 to load in one instruction
* Whoops. We were loading in the wrong order
* SSEd more things
* SSEd more functions
* Minor fixups
* SSE'd hmm_vec4 initialization
* Removed zeroing
* Vector normalization should be zero'd
* Removed old comments
* SSEd HMM_MultiplyMat4 and HMM_Transpose. And added HMM_LinearCombineSSE
* Maybe Travis doesn't support SSE?
* Fix compile process so the SSE option is consistently defined
* Fix link error
* Documentation
* Added function prototype for operator ==
* Added != operator for hmm_vec2, hmm_vec3, hmm_vec4
* Add C versions of equality checks
Also made the C++ tests actually run...😳
* Update documentation
* Handle zero-vector normalization.
When normalizing a vectors, we have to check whether vector length is
not zero, to avoid dividing by zero when normalizing zero-vectors.
* Test for normalization of zero vectors
* at one place HANDMADE_NO_SSE instead of HANDMADE_MATH_NO_SSE was used
* introduce HANDMADE_MATH__USE_SSE for the SSE #ifdefs throughout code
- use #ifdef HANDMADE_MATH_NO_SSE at only one place
* only use SSE (#define HANDMADE_MATH__USE_SSE) if the targetplatform
actually supports it
=> users don't have to #define HANDMADE_MATH_NO_SSE on ARM etc
* at one place HMM_SqrtF instead of HMM_SquareRootF was used
* Switch to custom unit testing that is compatible with C
* Remove Google Test framework
* Attempt to fix missing math functions
* Link against C math library
* Try forcing c99 again
* Include -lm at compile time
* Move -lm argument to the end
* Update README and gitignore
