Fix Variant conversion to float instead of double

[Demindiro]

real_t is defined as a float (or a double depending on a flag it seems?) in core/math/math_defs.h and Variant::construct was thus casting to a float.

Instead of using a double, maybe it is better to fix whatever is causing the flag not to be set?

godot/core/math/math_defs.h

Lines 109 to 113 in 2a4f0bc

	#ifdef REAL_T_IS_DOUBLE
	typedef double real_t;
	#else
	typedef float real_t;
	#endif

Closes #44303

[aaronfranke]

Good catch. Variant::REAL is a double, not real_t (which is used for vectors/etc). I also checked on master, the problem doesn't exist there, so this is a 3.x specific problem. Note: I haven't actually checked that this solves the linked issue.

Instead of using a double, maybe it is better to fix whatever is causing the flag not to be set?

See godotengine/godot-proposals#892

[qarmin]

Checked and it fixed my issue.

[akien-mga]

Thanks!