This float->half conversion matches how "F32TO16" works on Intel GPU's. Input cannot be negative, Inf or Nan.

Cheap TRS extract for rigid + per-axis-scale matrices (no skew/projective handling); the shared decomposition for all bind-pose math so results stay b

Returns true if and only if the test failed.

Helpers to retrieve ImRect (we don't need to store BuildWorkRect as every access tend to change it, hence the code asymmetry)

Uncompress the center of mass position

Get pointer

Calculate the support vector for this convex shape.

This is provided for consistency (but we don't actually use this)

Case-insensitive compare of a view against a NUL-terminated literal.

Backend data stored in io.BackendRendererUserData to allow support for multiple Dear ImGui contexts It is STRONGLY preferred that you use docking bran

Backend data stored in io.BackendRendererUserData to allow support for multiple Dear ImGui contexts It is STRONGLY preferred that you use docking bran

Replaces every Token in Source (single pass, inserted text not rescanned); splices the module-stack chunks into the function-body tokens in ParticleSi

Mark this class as embedded, this means the type can be used in a compound or constructed on the stack. The Release function will never destruct the o

* DISCLAIMER: There are known *seed-dependent* multicollisions here due to * multiplication by zero, affecting hashes of lengths 17 to 240. * * How

@brief returns an iterator to the first element @sa https://json.nlohmann.me/api/basic_json/begin/

@brief returns a const iterator to the first element @sa https://json.nlohmann.me/api/basic_json/cbegin/

boost::hash_combine

small_block = num_blk_pixels < 31

Returns 1 if key is found, 0 otherwise.

Important: pPixels is either 32-bit/texel or 64-bit/texel.

Performs a linear interpolation between two vectors. (@p v1 toward @p v2) @param[out] dest The result vector. [(x, y, x)] @param[in] v1 The starting

init