bullet: Sync with current upstream master branch

This updates our local copy to commit 5ec8339b6fc491e3f09a34a4516e82787f053fcc.
We need a recent master commit for some new features that we use in Godot
(see #25543 and #28909).

To avoid warnings generated by Bullet headers included in our own module,
we include those headers with -isystem on GCC and Clang.

Fixes #29503.

1 files changed, 7 insertions, 6 deletions

diff --git a/thirdparty/bullet/BulletDynamics/ConstraintSolver/btGeneric6DofSpring2Constraint.cpp b/thirdparty/bullet/BulletDynamics/ConstraintSolver/btGeneric6DofSpring2Constraint.cpp
index 49c8d9bbf7..9a3b39e6f8 100644
--- a/thirdparty/bullet/BulletDynamics/ConstraintSolver/btGeneric6DofSpring2Constraint.cpp
+++ b/thirdparty/bullet/BulletDynamics/ConstraintSolver/btGeneric6DofSpring2Constraint.cpp
@@ -32,7 +32,7 @@ Cons:
 
 /*
 2007-09-09
-btGeneric6DofConstraint Refactored by Francisco Le?n
+btGeneric6DofConstraint Refactored by Francisco Leon
 email: projectileman@yahoo.com
 http://gimpact.sf.net
 */
@@ -40,6 +40,7 @@ http://gimpact.sf.net
 #include "btGeneric6DofSpring2Constraint.h"
 #include "BulletDynamics/Dynamics/btRigidBody.h"
 #include "LinearMath/btTransformUtil.h"
+#include <cmath>
 #include <new>
 
 btGeneric6DofSpring2Constraint::btGeneric6DofSpring2Constraint(btRigidBody& rbA, btRigidBody& rbB, const btTransform& frameInA, const btTransform& frameInB, RotateOrder rotOrder)
@@ -310,9 +311,9 @@ void btGeneric6DofSpring2Constraint::calculateAngleInfo()
 		case RO_XYZ:
 		{
 			//Is this the "line of nodes" calculation choosing planes YZ (B coordinate system) and xy (A coordinate system)? (http://en.wikipedia.org/wiki/Euler_angles)
-			//The two planes are non-homologous, so this is a Tait–Bryan angle formalism and not a proper Euler
+			//The two planes are non-homologous, so this is a Tait-Bryan angle formalism and not a proper Euler
 			//Extrinsic rotations are equal to the reversed order intrinsic rotations so the above xyz extrinsic rotations (axes are fixed) are the same as the zy'x" intrinsic rotations (axes are refreshed after each rotation)
-			//that is why xy and YZ planes are chosen (this will describe a zy'x" intrinsic rotation) (see the figure on the left at http://en.wikipedia.org/wiki/Euler_angles under Tait–Bryan angles)
+			//that is why xy and YZ planes are chosen (this will describe a zy'x" intrinsic rotation) (see the figure on the left at http://en.wikipedia.org/wiki/Euler_angles under Tait-Bryan angles)
 			// x' = Nperp = N.cross(axis2)
 			// y' = N = axis2.cross(axis0)
 			// z' = z
@@ -845,7 +846,7 @@ int btGeneric6DofSpring2Constraint::get_limit_motor_info2(
 		if (m_rbA.getInvMass() == 0) m = mB; else
 		if (m_rbB.getInvMass() == 0) m = mA; else
 			m = mA*mB / (mA + mB);
-		btScalar angularfreq = sqrt(ks / m);
+		btScalar angularfreq = btSqrt(ks / m);
 
 		//limit stiffness (the spring should not be sampled faster that the quarter of its angular frequency)
 		if (limot->m_springStiffnessLimited && 0.25 < angularfreq * dt)
@@ -865,7 +866,7 @@ int btGeneric6DofSpring2Constraint::get_limit_motor_info2(
 		// vel + f / m * (rotational ? -1 : 1)
 		// so in theory this should be set here for m_constraintError
 		// (with m_constraintError we set a desired velocity for the affected body(es))
-		// however in practice any value is fine as long as it is greater then the "proper" velocity,
+		// however in practice any value is fine as long as it is greater than the "proper" velocity,
 		// because the m_lowerLimit and the m_upperLimit will determinate the strength of the final pulling force
 		// so it is much simpler (and more robust) just to simply use inf (with the proper sign)
 		// (Even with our best intent the "new" velocity is only an estimation. If we underestimate
@@ -1085,7 +1086,7 @@ void btGeneric6DofSpring2Constraint::setServoTarget(int index, btScalar targetOr
 		btScalar target = targetOrg + SIMD_PI;
 		if (1)
 		{
-			btScalar m = target - SIMD_2_PI * floor(target / SIMD_2_PI);
+			btScalar m = target - SIMD_2_PI * std::floor(target / SIMD_2_PI);
 			// handle boundary cases resulted from floating-point cut off:
 			{
 				if (m >= SIMD_2_PI)