Core: Integrate math constants in structs directly #107723
Conversation
| static const AABB ZERO; | ||
| static const AABB ONE; | ||
| static const AABB INF; | ||
| static const AABB NaN; |
There was a problem hiding this comment.
Note: In order to make these constants constexpr, they had to first be declared in the structs as const. Attempting to define them as constexpr at this stage raises an incomplete-type error
| #undef MIN | ||
| #undef MAX |
There was a problem hiding this comment.
These #undefs were necessary to get compilation working on Apple devices, because somewhere along the line there's an include which defines MIN/MAX macros that didn't get removed by typedefs.h.
| @@ -78,19 +90,19 @@ struct [[nodiscard]] Vector2i { | |||
| } | |||
|
|
|||
| Vector2i min(const Vector2i &p_vector2i) const { | |||
| return Vector2i(MIN(x, p_vector2i.x), MIN(y, p_vector2i.y)); | |||
| return Vector2i(::MIN(x, p_vector2i.x), ::MIN(y, p_vector2i.y)); | |||
There was a problem hiding this comment.
The global namespace needed to be explicitly called, as this would otherwise conflict with the new constants
While this and the #undefs were effective solutions, they're still workarounds for the bigger issue of having these pseudo-macros on the global namespace. Once those functions migrate to Math, no further workarounds will be necessary.
Extends our defined math constants to the mathmatical structs directly. This has the benefit of reducing redundancy across the codebase with a singular variable, as well as making developer intent much clearer on whether an arbitrary struct value is a magic number or referencing one of our constants. For usability, this implements the constants
ZERO/ONEfor all structs,INF/NANfor float structs, andMAX/MINfor integral structs; no additional bindings were made.