11 files changed, 774 insertions, 109 deletions

diff --git a/thirdparty/README.md b/thirdparty/README.md
index bc820634bb..732c08fdea 100644
--- a/thirdparty/README.md
+++ b/thirdparty/README.md
@@ -346,7 +346,7 @@ Collection of single-file libraries used in Godot components.
   * License: Public Domain
  - `clipper.{cpp,hpp}`
   * Upstream: https://sourceforge.net/projects/polyclipping
-  * Version: 6.4.2
+  * Version: 6.4.2 + Godot changes (added optional exceptions handling)
   * License: BSL-1.0
 - `fastlz.{c,h}`
   * Upstream: https://github.com/ariya/FastLZ
diff --git a/thirdparty/assimp/code/res/resource.h b/thirdparty/assimp/code/res/resource.h
deleted file mode 100644
index 37d39284fe..0000000000
--- a/thirdparty/assimp/code/res/resource.h
+++ /dev/null
@@ -1,14 +0,0 @@
-//{{NO_DEPENDENCIES}}
-// Microsoft Visual C++ generated include file.
-// Used by assimp.rc
-
-// Nächste Standardwerte für neue Objekte
-// 
-#ifdef APSTUDIO_INVOKED
-#ifndef APSTUDIO_READONLY_SYMBOLS
-#define _APS_NEXT_RESOURCE_VALUE        101
-#define _APS_NEXT_COMMAND_VALUE         40001
-#define _APS_NEXT_CONTROL_VALUE         1001
-#define _APS_NEXT_SYMED_VALUE           101
-#endif
-#endif
diff --git a/thirdparty/bullet/BulletCollision/CollisionDispatch/btCollisionWorld.cpp b/thirdparty/bullet/BulletCollision/CollisionDispatch/btCollisionWorld.cpp
index 782e9efaf1..b30ce03164 100644
--- a/thirdparty/bullet/BulletCollision/CollisionDispatch/btCollisionWorld.cpp
+++ b/thirdparty/bullet/BulletCollision/CollisionDispatch/btCollisionWorld.cpp
@@ -19,9 +19,10 @@ subject to the following restrictions:
 #include "BulletCollision/CollisionShapes/btCollisionShape.h"
 #include "BulletCollision/CollisionShapes/btConvexShape.h"
 #include "BulletCollision/NarrowPhaseCollision/btGjkEpaPenetrationDepthSolver.h"
-#include "BulletCollision/CollisionShapes/btSphereShape.h"                 //for raycasting
-#include "BulletCollision/CollisionShapes/btBvhTriangleMeshShape.h"        //for raycasting
-#include "BulletCollision/CollisionShapes/btScaledBvhTriangleMeshShape.h"  //for raycasting
+#include "BulletCollision/CollisionShapes/btSphereShape.h" //for raycasting
+#include "BulletCollision/CollisionShapes/btBvhTriangleMeshShape.h" //for raycasting
+#include "BulletCollision/CollisionShapes/btScaledBvhTriangleMeshShape.h" //for raycasting
+#include "BulletCollision/CollisionShapes/btHeightfieldTerrainShape.h" //for raycasting
 #include "BulletCollision/NarrowPhaseCollision/btRaycastCallback.h"
 #include "BulletCollision/CollisionShapes/btCompoundShape.h"
 #include "BulletCollision/NarrowPhaseCollision/btSubSimplexConvexCast.h"
@@ -413,6 +414,18 @@ void btCollisionWorld::rayTestSingleInternal(const btTransform& rayFromTrans, co
 				rcb.m_hitFraction = resultCallback.m_closestHitFraction;
 				triangleMesh->performRaycast(&rcb, rayFromLocalScaled, rayToLocalScaled);
 			}
+			else if (collisionShape->getShapeType()==TERRAIN_SHAPE_PROXYTYPE)
+			{
+				///optimized version for btHeightfieldTerrainShape
+				btHeightfieldTerrainShape* heightField = (btHeightfieldTerrainShape*)collisionShape;
+				btTransform worldTocollisionObject = colObjWorldTransform.inverse();
+				btVector3 rayFromLocal = worldTocollisionObject * rayFromTrans.getOrigin();
+				btVector3 rayToLocal = worldTocollisionObject * rayToTrans.getOrigin();
+
+				BridgeTriangleRaycastCallback rcb(rayFromLocal,rayToLocal,&resultCallback,collisionObjectWrap->getCollisionObject(),heightField,colObjWorldTransform);
+				rcb.m_hitFraction = resultCallback.m_closestHitFraction;
+				heightField->performRaycast(&rcb, rayFromLocal, rayToLocal);
+			}
 			else
 			{
 				//generic (slower) case
diff --git a/thirdparty/bullet/BulletCollision/CollisionShapes/btHeightfieldTerrainShape.cpp b/thirdparty/bullet/BulletCollision/CollisionShapes/btHeightfieldTerrainShape.cpp
index c85ce2498e..4adf27e6bb 100644
--- a/thirdparty/bullet/BulletCollision/CollisionShapes/btHeightfieldTerrainShape.cpp
+++ b/thirdparty/bullet/BulletCollision/CollisionShapes/btHeightfieldTerrainShape.cpp
@@ -73,6 +73,10 @@ void btHeightfieldTerrainShape::initialize(
 	m_useZigzagSubdivision = false;
 	m_upAxis = upAxis;
 	m_localScaling.setValue(btScalar(1.), btScalar(1.), btScalar(1.));
+	m_vboundsGrid = NULL;
+	m_vboundsChunkSize = 0;
+	m_vboundsGridWidth = 0;
+	m_vboundsGridLength = 0;
 
 	// determine min/max axis-aligned bounding box (aabb) values
 	switch (m_upAxis)
@@ -108,6 +112,7 @@ void btHeightfieldTerrainShape::initialize(
 
 btHeightfieldTerrainShape::~btHeightfieldTerrainShape()
 {
+	clearAccelerator();
 }
 
 void btHeightfieldTerrainShape::getAabb(const btTransform& t, btVector3& aabbMin, btVector3& aabbMax) const
@@ -323,6 +328,8 @@ void btHeightfieldTerrainShape::processAllTriangles(btTriangleCallback* callback
 		}
 	}
 
+	// TODO If m_vboundsGrid is available, use it to determine if we really need to process this area
+
 	for (int j = startJ; j < endJ; j++)
 	{
 		for (int x = startX; x < endX; x++)
@@ -373,3 +380,416 @@ const btVector3& btHeightfieldTerrainShape::getLocalScaling() const
 {
 	return m_localScaling;
 }
+
+
+
+struct GridRaycastState
+{
+	int x; // Next quad coords
+	int z;
+	int prev_x; // Previous quad coords
+	int prev_z;
+	btScalar param; // Exit param for previous quad
+	btScalar prevParam; // Enter param for previous quad
+	btScalar maxDistanceFlat;
+	btScalar maxDistance3d;
+};
+
+
+// TODO Does it really need to take 3D vectors?
+/// Iterates through a virtual 2D grid of unit-sized square cells,
+/// and executes an action on each cell intersecting the given segment, ordered from begin to end.
+/// Initially inspired by http://www.cse.yorku.ca/~amana/research/grid.pdf
+template <typename Action_T>
+void gridRaycast(Action_T &quadAction, const btVector3 &beginPos, const btVector3 &endPos)
+{
+	GridRaycastState rs;
+	rs.maxDistance3d = beginPos.distance(endPos);
+	if (rs.maxDistance3d < 0.0001)
+		// Consider the ray is too small to hit anything
+		return;
+
+	btScalar rayDirectionFlatX = endPos[0] - beginPos[0];
+	btScalar rayDirectionFlatZ = endPos[2] - beginPos[2];
+	rs.maxDistanceFlat = btSqrt(rayDirectionFlatX * rayDirectionFlatX + rayDirectionFlatZ * rayDirectionFlatZ);
+
+	if(rs.maxDistanceFlat < 0.0001)
+	{
+		// Consider the ray vertical
+		rayDirectionFlatX = 0;
+		rayDirectionFlatZ = 0;
+	}
+	else
+	{
+		rayDirectionFlatX /= rs.maxDistanceFlat;
+		rayDirectionFlatZ /= rs.maxDistanceFlat;
+	}
+
+	const int xiStep = rayDirectionFlatX > 0 ? 1 : rayDirectionFlatX < 0 ? -1 : 0;
+	const int ziStep = rayDirectionFlatZ > 0 ? 1 : rayDirectionFlatZ < 0 ? -1 : 0;
+
+	const float infinite = 9999999;
+	const btScalar paramDeltaX = xiStep != 0 ? 1.f / btFabs(rayDirectionFlatX) : infinite;
+	const btScalar paramDeltaZ = ziStep != 0 ? 1.f / btFabs(rayDirectionFlatZ) : infinite;
+
+	// pos = param * dir
+	btScalar paramCrossX; // At which value of `param` we will cross a x-axis lane?
+	btScalar paramCrossZ; // At which value of `param` we will cross a z-axis lane?
+
+	// paramCrossX and paramCrossZ are initialized as being the first cross
+	// X initialization
+	if (xiStep != 0)
+	{
+		if (xiStep == 1)
+			paramCrossX = (ceil(beginPos[0]) - beginPos[0]) * paramDeltaX;
+		else
+			paramCrossX = (beginPos[0] - floor(beginPos[0])) * paramDeltaX;
+	}
+	else
+		paramCrossX = infinite; // Will never cross on X
+
+	// Z initialization
+	if (ziStep != 0)
+	{
+		if (ziStep == 1)
+			paramCrossZ = (ceil(beginPos[2]) - beginPos[2]) * paramDeltaZ;
+		else
+			paramCrossZ = (beginPos[2] - floor(beginPos[2])) * paramDeltaZ;
+	}
+	else
+		paramCrossZ = infinite; // Will never cross on Z
+
+	rs.x = static_cast<int>(floor(beginPos[0]));
+	rs.z = static_cast<int>(floor(beginPos[2]));
+
+	// Workaround cases where the ray starts at an integer position
+	if (paramCrossX == 0.0)
+	{
+		paramCrossX += paramDeltaX;
+		// If going backwards, we should ignore the position we would get by the above flooring,
+		// because the ray is not heading in that direction
+		if (xiStep == -1)
+			rs.x -= 1;
+	}
+
+	if (paramCrossZ == 0.0)
+	{
+		paramCrossZ += paramDeltaZ;
+		if (ziStep == -1)
+			rs.z -= 1;
+	}
+
+	rs.prev_x = rs.x;
+	rs.prev_z = rs.z;
+	rs.param = 0;
+
+	while (true)
+	{
+		rs.prev_x = rs.x;
+		rs.prev_z = rs.z;
+		rs.prevParam = rs.param;
+
+		if (paramCrossX < paramCrossZ)
+		{
+			// X lane
+			rs.x += xiStep;
+			// Assign before advancing the param,
+			// to be in sync with the initialization step
+			rs.param = paramCrossX;
+			paramCrossX += paramDeltaX;
+		}
+		else
+		{
+			// Z lane
+			rs.z += ziStep;
+			rs.param = paramCrossZ;
+			paramCrossZ += paramDeltaZ;
+		}
+
+		if (rs.param > rs.maxDistanceFlat)
+		{
+			rs.param = rs.maxDistanceFlat;
+			quadAction(rs);
+			break;
+		}
+		else
+			quadAction(rs);
+	}
+}
+
+
+struct ProcessTrianglesAction
+{
+	const btHeightfieldTerrainShape *shape;
+	bool flipQuadEdges;
+	bool useDiamondSubdivision;
+	int width;
+	int length;
+	btTriangleCallback* callback;
+
+	void exec(int x, int z) const
+	{
+		if(x < 0 || z < 0 || x >= width || z >= length)
+			return;
+
+		btVector3 vertices[3];
+
+		// Check quad
+		if (flipQuadEdges || (useDiamondSubdivision && (((z + x) & 1) > 0)))
+		{
+			// First triangle
+			shape->getVertex(x, z, vertices[0]);
+			shape->getVertex(x + 1, z, vertices[1]);
+			shape->getVertex(x + 1, z + 1, vertices[2]);
+			callback->processTriangle(vertices, x, z);
+
+			// Second triangle
+			shape->getVertex(x, z, vertices[0]);
+			shape->getVertex(x + 1, z + 1, vertices[1]);
+			shape->getVertex(x, z + 1, vertices[2]);
+			callback->processTriangle(vertices, x, z);
+		}
+		else
+		{
+			// First triangle
+			shape->getVertex(x, z, vertices[0]);
+			shape->getVertex(x, z + 1, vertices[1]);
+			shape->getVertex(x + 1, z, vertices[2]);
+			callback->processTriangle(vertices, x, z);
+
+			// Second triangle
+			shape->getVertex(x + 1, z, vertices[0]);
+			shape->getVertex(x, z + 1, vertices[1]);
+			shape->getVertex(x + 1, z + 1, vertices[2]);
+			callback->processTriangle(vertices, x, z);
+		}
+	}
+
+	void operator ()(const GridRaycastState &bs) const
+	{
+		exec(bs.prev_x, bs.prev_z);
+	}
+};
+
+
+struct ProcessVBoundsAction
+{
+	const btHeightfieldTerrainShape::Range *vbounds;
+	int width;
+	int length;
+	int chunkSize;
+
+	btVector3 rayBegin;
+	btVector3 rayEnd;
+	btVector3 rayDir;
+
+	ProcessTrianglesAction processTriangles;
+
+	void operator ()(const GridRaycastState &rs) const
+	{
+		int x = rs.prev_x;
+		int z = rs.prev_z;
+
+		if(x < 0 || z < 0 || x >= width || z >= length)
+			return;
+
+		const btHeightfieldTerrainShape::Range chunk = vbounds[x + z * width];
+
+		btVector3 enterPos;
+		btVector3 exitPos;
+
+		if (rs.maxDistanceFlat > 0.0001)
+		{
+			btScalar flatTo3d = chunkSize * rs.maxDistance3d / rs.maxDistanceFlat;
+			btScalar enterParam3d = rs.prevParam * flatTo3d;
+			btScalar exitParam3d = rs.param * flatTo3d;
+			enterPos = rayBegin + rayDir * enterParam3d;
+			exitPos = rayBegin + rayDir * exitParam3d;
+
+			// We did enter the flat projection of the AABB,
+			// but we have to check if we intersect it on the vertical axis
+			if (enterPos[1] > chunk.max && exitPos[1] > chunk.max)
+				return;
+			if (enterPos[1] < chunk.min && exitPos[1] < chunk.min)
+				return;
+		}
+		else
+		{
+			// Consider the ray vertical
+			// (though we shouldn't reach this often because there is an early check up-front)
+			enterPos = rayBegin;
+			exitPos = rayEnd;
+		}
+
+		gridRaycast(processTriangles, enterPos, exitPos);
+		// Note: it could be possible to have more than one grid at different levels,
+		// to do this there would be a branch using a pointer to another ProcessVBoundsAction
+	}
+};
+
+
+// TODO How do I interrupt the ray when there is a hit? `callback` does not return any result
+/// Performs a raycast using a hierarchical Bresenham algorithm.
+/// Does not allocate any memory by itself.
+void btHeightfieldTerrainShape::performRaycast(btTriangleCallback* callback, const btVector3& raySource, const btVector3& rayTarget) const
+{
+	// Transform to cell-local
+	btVector3 beginPos = raySource / m_localScaling;
+	btVector3 endPos = rayTarget / m_localScaling;
+	beginPos += m_localOrigin;
+	endPos += m_localOrigin;
+
+	ProcessTrianglesAction processTriangles;
+	processTriangles.shape = this;
+	processTriangles.flipQuadEdges = m_flipQuadEdges;
+	processTriangles.useDiamondSubdivision = m_useDiamondSubdivision;
+	processTriangles.callback = callback;
+	processTriangles.width = m_heightStickWidth - 1;
+	processTriangles.length = m_heightStickLength - 1;
+
+	// TODO Transform vectors to account for m_upAxis
+	int iBeginX = static_cast<int>(floor(beginPos[0]));
+	int iBeginZ = static_cast<int>(floor(beginPos[2]));
+	int iEndX = static_cast<int>(floor(endPos[0]));
+	int iEndZ = static_cast<int>(floor(endPos[2]));
+
+	if (iBeginX == iEndX && iBeginZ == iEndZ)
+	{
+		// The ray will never cross quads within the plane,
+		// so directly process triangles within one quad
+		// (typically, vertical rays should end up here)
+		processTriangles.exec(iBeginX, iEndZ);
+		return;
+	}
+
+	if (m_vboundsGrid == NULL)
+	{
+		// Process all quads intersecting the flat projection of the ray
+		gridRaycast(processTriangles, beginPos, endPos);
+	}
+	else
+	{
+		btVector3 rayDiff = endPos - beginPos;
+		btScalar flatDistance2 = rayDiff[0] * rayDiff[0] + rayDiff[2] * rayDiff[2];
+		if (flatDistance2 < m_vboundsChunkSize * m_vboundsChunkSize)
+		{
+			// Don't use chunks, the ray is too short in the plane
+			gridRaycast(processTriangles, beginPos, endPos);
+		}
+
+		ProcessVBoundsAction processVBounds;
+		processVBounds.width = m_vboundsGridWidth;
+		processVBounds.length = m_vboundsGridLength;
+		processVBounds.vbounds = m_vboundsGrid;
+		processVBounds.rayBegin = beginPos;
+		processVBounds.rayEnd = endPos;
+		processVBounds.rayDir = rayDiff.normalized();
+		processVBounds.processTriangles = processTriangles;
+		processVBounds.chunkSize = m_vboundsChunkSize;
+		// The ray is long, run raycast on a higher-level grid
+		gridRaycast(processVBounds, beginPos / m_vboundsChunkSize, endPos / m_vboundsChunkSize);
+	}
+}
+
+
+/// Builds a grid data structure storing the min and max heights of the terrain in chunks.
+/// if chunkSize is zero, that accelerator is removed.
+/// If you modify the heights, you need to rebuild this accelerator.
+void btHeightfieldTerrainShape::buildAccelerator(int chunkSize)
+{
+	if (chunkSize <= 0)
+	{
+		clearAccelerator();
+		return;
+	}
+
+	m_vboundsChunkSize = chunkSize;
+	int nChunksX = m_heightStickWidth / chunkSize;
+	int nChunksZ = m_heightStickLength / chunkSize;
+
+	if (m_heightStickWidth % chunkSize > 0)
+		++nChunksX; // In case terrain size isn't dividable by chunk size
+	if (m_heightStickLength % chunkSize > 0)
+		++nChunksZ;
+
+	if(m_vboundsGridWidth != nChunksX || m_vboundsGridLength != nChunksZ)
+	{
+		clearAccelerator();
+		m_vboundsGridWidth = nChunksX;
+		m_vboundsGridLength = nChunksZ;
+	}
+
+	if (nChunksX == 0 || nChunksZ == 0)
+		return;
+
+	// TODO What is the recommended way to allocate this?
+	// This data structure is only reallocated if the required size changed
+	if (m_vboundsGrid == NULL)
+		m_vboundsGrid = new Range[nChunksX * nChunksZ];
+
+	// Compute min and max height for all chunks
+	for (int cz = 0; cz < nChunksZ; ++cz)
+	{
+		int z0 = cz * chunkSize;
+
+		for (int cx = 0; cx < nChunksX; ++cx)
+		{
+			int x0 = cx * chunkSize;
+
+			Range r;
+
+			r.min = getRawHeightFieldValue(x0, z0);
+			r.max = r.min;
+
+			// Compute min and max height for this chunk.
+			// We have to include one extra cell to account for neighbors.
+			// Here is why:
+			// Say we have a flat terrain, and a plateau that fits a chunk perfectly.
+			//
+			//   Left        Right
+			// 0---0---0---1---1---1
+			// |   |   |   |   |   |
+			// 0---0---0---1---1---1
+			// |   |   |   |   |   |
+			// 0---0---0---1---1---1
+			//           x
+			//
+			// If the AABB for the Left chunk did not share vertices with the Right,
+			// then we would fail collision tests at x due to a gap.
+			//
+			for (int z = z0; z < z0 + chunkSize + 1; ++z)
+			{
+				if (z >= m_heightStickLength)
+					continue;
+
+				for (int x = x0; x < x0 + chunkSize + 1; ++x)
+				{
+					if (x >= m_heightStickWidth)
+						continue;
+
+					btScalar height = getRawHeightFieldValue(x, z);
+
+					if (height < r.min)
+						r.min = height;
+					else if (height > r.max)
+						r.max = height;
+				}
+			}
+
+			m_vboundsGrid[cx + cz * nChunksX] = r;
+		}
+	}
+}
+
+
+void btHeightfieldTerrainShape::clearAccelerator()
+{
+	if (m_vboundsGrid)
+	{
+		// TODO What is the recommended way to deallocate this?
+		delete[] m_vboundsGrid;
+		m_vboundsGrid = 0;
+	}
+}
+
+
diff --git a/thirdparty/bullet/BulletCollision/CollisionShapes/btHeightfieldTerrainShape.h b/thirdparty/bullet/BulletCollision/CollisionShapes/btHeightfieldTerrainShape.h
index 8a50a57e31..e23b548cb2 100644
--- a/thirdparty/bullet/BulletCollision/CollisionShapes/btHeightfieldTerrainShape.h
+++ b/thirdparty/bullet/BulletCollision/CollisionShapes/btHeightfieldTerrainShape.h
@@ -18,6 +18,7 @@ subject to the following restrictions:
 
 #include "btConcaveShape.h"
 
+
 ///btHeightfieldTerrainShape simulates a 2D heightfield terrain
 /**
   The caller is responsible for maintaining the heightfield array; this
@@ -71,6 +72,12 @@ subject to the following restrictions:
 ATTRIBUTE_ALIGNED16(class)
 btHeightfieldTerrainShape : public btConcaveShape
 {
+public:
+	struct Range {
+		btScalar min;
+		btScalar max;
+	};
+
 protected:
 	btVector3 m_localAabbMin;
 	btVector3 m_localAabbMax;
@@ -100,9 +107,14 @@ protected:
 
 	btVector3 m_localScaling;
 
+	// Accelerator
+	Range *m_vboundsGrid;
+	int m_vboundsGridWidth;
+	int m_vboundsGridLength;
+	int m_vboundsChunkSize;
+
 	virtual btScalar getRawHeightFieldValue(int x, int y) const;
 	void quantizeWithClamp(int* out, const btVector3& point, int isMax) const;
-	void getVertex(int x, int y, btVector3& vertex) const;
 
 	/// protected initialization
 	/**
@@ -154,6 +166,13 @@ public:
 	virtual void setLocalScaling(const btVector3& scaling);
 
 	virtual const btVector3& getLocalScaling() const;
+	
+	void		getVertex(int x,int y,btVector3& vertex) const;
+
+	void performRaycast (btTriangleCallback* callback, const btVector3& raySource, const btVector3& rayTarget) const;
+
+	void buildAccelerator(int chunkSize=16);
+	void clearAccelerator();
 
 	//debugging
 	virtual const char* getName() const { return "HEIGHTFIELD"; }
diff --git a/thirdparty/cvtt/LICENSE.txt b/thirdparty/cvtt/LICENSE.txt
index 6964f6e43f..d79aec44e3 100644
--- a/thirdparty/cvtt/LICENSE.txt
+++ b/thirdparty/cvtt/LICENSE.txt
@@ -1,45 +1,45 @@
-Convection Texture Tools Stand-Alone Kernels

-

-Copyright (c) 2018 Eric Lasota

-

-Permission is hereby granted, free of charge, to any person obtaining

-a copy of this software and associated documentation files (the

-"Software"), to deal in the Software without restriction, including

-without limitation the rights to use, copy, modify, merge, publish,

-distribute, sublicense, and/or sell copies of the Software, and to

-permit persons to whom the Software is furnished to do so, subject

-to the following conditions:

-

-The above copyright notice and this permission notice shall be included

-in all copies or substantial portions of the Software.

-

-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS

-OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF

-MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.

-IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY

-CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,

-TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE

-SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

-

-**************************************************************************

-

-Based on DirectX Texture Library

-

-Copyright (c) 2018 Microsoft Corp

-

-Permission is hereby granted, free of charge, to any person obtaining a copy of this 

-software and associated documentation files (the "Software"), to deal in the Software 

-without restriction, including without limitation the rights to use, copy, modify, 

-merge, publish, distribute, sublicense, and/or sell copies of the Software, and to 

-permit persons to whom the Software is furnished to do so, subject to the following 

-conditions: 

-

-The above copyright notice and this permission notice shall be included in all copies 

-or substantial portions of the Software.  

-

-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, 

-INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A 

-PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT 

-HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF 

-CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE 

+Convection Texture Tools Stand-Alone Kernels
+
+Copyright (c) 2018 Eric Lasota
+
+Permission is hereby granted, free of charge, to any person obtaining
+a copy of this software and associated documentation files (the
+"Software"), to deal in the Software without restriction, including
+without limitation the rights to use, copy, modify, merge, publish,
+distribute, sublicense, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so, subject
+to the following conditions:
+
+The above copyright notice and this permission notice shall be included
+in all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
+CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+**************************************************************************
+
+Based on DirectX Texture Library
+
+Copyright (c) 2018 Microsoft Corp
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of this 
+software and associated documentation files (the "Software"), to deal in the Software 
+without restriction, including without limitation the rights to use, copy, modify, 
+merge, publish, distribute, sublicense, and/or sell copies of the Software, and to 
+permit persons to whom the Software is furnished to do so, subject to the following 
+conditions: 
+
+The above copyright notice and this permission notice shall be included in all copies 
+or substantial portions of the Software.  
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, 
+INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A 
+PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT 
+HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF 
+CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE 
 OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
\ No newline at end of file
diff --git a/thirdparty/glad/glad.c b/thirdparty/glad/glad.c
index e7c64a7e17..9704c1079f 100644
--- a/thirdparty/glad/glad.c
+++ b/thirdparty/glad/glad.c
@@ -1,6 +1,6 @@
 /*
 
-    OpenGL loader generated by glad 0.1.29 on Mon Mar  4 12:47:22 2019.
+    OpenGL loader generated by glad 0.1.29 on Wed May  1 23:16:34 2019.
 
     Language/Generator: C/C++
     Specification: gl
@@ -9,6 +9,8 @@
     Extensions:
         GL_ARB_debug_output,
         GL_ARB_framebuffer_object,
+        GL_EXT_framebuffer_blit,
+        GL_EXT_framebuffer_multisample,
         GL_EXT_framebuffer_object
     Loader: True
     Local files: False
@@ -16,9 +18,9 @@
     Reproducible: False
 
     Commandline:
-        --profile="compatibility" --api="gl=3.3" --generator="c" --spec="gl" --extensions="GL_ARB_debug_output,GL_ARB_framebuffer_object,GL_EXT_framebuffer_object"
+        --profile="compatibility" --api="gl=3.3" --generator="c" --spec="gl" --extensions="GL_ARB_debug_output,GL_ARB_framebuffer_object,GL_EXT_framebuffer_blit,GL_EXT_framebuffer_multisample,GL_EXT_framebuffer_object"
     Online:
-        https://glad.dav1d.de/#profile=compatibility&language=c&specification=gl&loader=on&api=gl%3D3.3&extensions=GL_ARB_debug_output&extensions=GL_ARB_framebuffer_object&extensions=GL_EXT_framebuffer_object
+        https://glad.dav1d.de/#profile=compatibility&language=c&specification=gl&loader=on&api=gl%3D3.3&extensions=GL_ARB_debug_output&extensions=GL_ARB_framebuffer_object&extensions=GL_EXT_framebuffer_blit&extensions=GL_EXT_framebuffer_multisample&extensions=GL_EXT_framebuffer_object
 */
 
 #include <stdio.h>
@@ -992,11 +994,15 @@ PFNGLWINDOWPOS3SPROC glad_glWindowPos3s = NULL;
 PFNGLWINDOWPOS3SVPROC glad_glWindowPos3sv = NULL;
 int GLAD_GL_ARB_debug_output = 0;
 int GLAD_GL_ARB_framebuffer_object = 0;
+int GLAD_GL_EXT_framebuffer_blit = 0;
+int GLAD_GL_EXT_framebuffer_multisample = 0;
 int GLAD_GL_EXT_framebuffer_object = 0;
 PFNGLDEBUGMESSAGECONTROLARBPROC glad_glDebugMessageControlARB = NULL;
 PFNGLDEBUGMESSAGEINSERTARBPROC glad_glDebugMessageInsertARB = NULL;
 PFNGLDEBUGMESSAGECALLBACKARBPROC glad_glDebugMessageCallbackARB = NULL;
 PFNGLGETDEBUGMESSAGELOGARBPROC glad_glGetDebugMessageLogARB = NULL;
+PFNGLBLITFRAMEBUFFEREXTPROC glad_glBlitFramebufferEXT = NULL;
+PFNGLRENDERBUFFERSTORAGEMULTISAMPLEEXTPROC glad_glRenderbufferStorageMultisampleEXT = NULL;
 PFNGLISRENDERBUFFEREXTPROC glad_glIsRenderbufferEXT = NULL;
 PFNGLBINDRENDERBUFFEREXTPROC glad_glBindRenderbufferEXT = NULL;
 PFNGLDELETERENDERBUFFERSEXTPROC glad_glDeleteRenderbuffersEXT = NULL;
@@ -1807,6 +1813,14 @@ static void load_GL_ARB_framebuffer_object(GLADloadproc load) {
 	glad_glRenderbufferStorageMultisample = (PFNGLRENDERBUFFERSTORAGEMULTISAMPLEPROC)load("glRenderbufferStorageMultisample");
 	glad_glFramebufferTextureLayer = (PFNGLFRAMEBUFFERTEXTURELAYERPROC)load("glFramebufferTextureLayer");
 }
+static void load_GL_EXT_framebuffer_blit(GLADloadproc load) {
+	if(!GLAD_GL_EXT_framebuffer_blit) return;
+	glad_glBlitFramebufferEXT = (PFNGLBLITFRAMEBUFFEREXTPROC)load("glBlitFramebufferEXT");
+}
+static void load_GL_EXT_framebuffer_multisample(GLADloadproc load) {
+	if(!GLAD_GL_EXT_framebuffer_multisample) return;
+	glad_glRenderbufferStorageMultisampleEXT = (PFNGLRENDERBUFFERSTORAGEMULTISAMPLEEXTPROC)load("glRenderbufferStorageMultisampleEXT");
+}
 static void load_GL_EXT_framebuffer_object(GLADloadproc load) {
 	if(!GLAD_GL_EXT_framebuffer_object) return;
 	glad_glIsRenderbufferEXT = (PFNGLISRENDERBUFFEREXTPROC)load("glIsRenderbufferEXT");
@@ -1831,6 +1845,8 @@ static int find_extensionsGL(void) {
 	if (!get_exts()) return 0;
 	GLAD_GL_ARB_debug_output = has_ext("GL_ARB_debug_output");
 	GLAD_GL_ARB_framebuffer_object = has_ext("GL_ARB_framebuffer_object");
+	GLAD_GL_EXT_framebuffer_blit = has_ext("GL_EXT_framebuffer_blit");
+	GLAD_GL_EXT_framebuffer_multisample = has_ext("GL_EXT_framebuffer_multisample");
 	GLAD_GL_EXT_framebuffer_object = has_ext("GL_EXT_framebuffer_object");
 	free_exts();
 	return 1;
@@ -1912,6 +1928,8 @@ int gladLoadGLLoader(GLADloadproc load) {
 	if (!find_extensionsGL()) return 0;
 	load_GL_ARB_debug_output(load);
 	load_GL_ARB_framebuffer_object(load);
+	load_GL_EXT_framebuffer_blit(load);
+	load_GL_EXT_framebuffer_multisample(load);
 	load_GL_EXT_framebuffer_object(load);
 	return GLVersion.major != 0 || GLVersion.minor != 0;
 }
diff --git a/thirdparty/glad/glad/glad.h b/thirdparty/glad/glad/glad.h
index 6345de6f7a..b398faf627 100644
--- a/thirdparty/glad/glad/glad.h
+++ b/thirdparty/glad/glad/glad.h
@@ -1,6 +1,6 @@
 /*
 
-    OpenGL loader generated by glad 0.1.29 on Mon Mar  4 12:47:22 2019.
+    OpenGL loader generated by glad 0.1.29 on Wed May  1 23:16:34 2019.
 
     Language/Generator: C/C++
     Specification: gl
@@ -9,6 +9,8 @@
     Extensions:
         GL_ARB_debug_output,
         GL_ARB_framebuffer_object,
+        GL_EXT_framebuffer_blit,
+        GL_EXT_framebuffer_multisample,
         GL_EXT_framebuffer_object
     Loader: True
     Local files: False
@@ -16,9 +18,9 @@
     Reproducible: False
 
     Commandline:
-        --profile="compatibility" --api="gl=3.3" --generator="c" --spec="gl" --extensions="GL_ARB_debug_output,GL_ARB_framebuffer_object,GL_EXT_framebuffer_object"
+        --profile="compatibility" --api="gl=3.3" --generator="c" --spec="gl" --extensions="GL_ARB_debug_output,GL_ARB_framebuffer_object,GL_EXT_framebuffer_blit,GL_EXT_framebuffer_multisample,GL_EXT_framebuffer_object"
     Online:
-        https://glad.dav1d.de/#profile=compatibility&language=c&specification=gl&loader=on&api=gl%3D3.3&extensions=GL_ARB_debug_output&extensions=GL_ARB_framebuffer_object&extensions=GL_EXT_framebuffer_object
+        https://glad.dav1d.de/#profile=compatibility&language=c&specification=gl&loader=on&api=gl%3D3.3&extensions=GL_ARB_debug_output&extensions=GL_ARB_framebuffer_object&extensions=GL_EXT_framebuffer_blit&extensions=GL_EXT_framebuffer_multisample&extensions=GL_EXT_framebuffer_object
 */
 
 
@@ -3633,6 +3635,13 @@ GLAPI PFNGLSECONDARYCOLORP3UIVPROC glad_glSecondaryColorP3uiv;
 #define GL_DEBUG_SEVERITY_HIGH_ARB 0x9146
 #define GL_DEBUG_SEVERITY_MEDIUM_ARB 0x9147
 #define GL_DEBUG_SEVERITY_LOW_ARB 0x9148
+#define GL_READ_FRAMEBUFFER_EXT 0x8CA8
+#define GL_DRAW_FRAMEBUFFER_EXT 0x8CA9
+#define GL_DRAW_FRAMEBUFFER_BINDING_EXT 0x8CA6
+#define GL_READ_FRAMEBUFFER_BINDING_EXT 0x8CAA
+#define GL_RENDERBUFFER_SAMPLES_EXT 0x8CAB
+#define GL_FRAMEBUFFER_INCOMPLETE_MULTISAMPLE_EXT 0x8D56
+#define GL_MAX_SAMPLES_EXT 0x8D57
 #define GL_INVALID_FRAMEBUFFER_OPERATION_EXT 0x0506
 #define GL_MAX_RENDERBUFFER_SIZE_EXT 0x84E8
 #define GL_FRAMEBUFFER_BINDING_EXT 0x8CA6
@@ -3704,6 +3713,20 @@ GLAPI PFNGLGETDEBUGMESSAGELOGARBPROC glad_glGetDebugMessageLogARB;
 #define GL_ARB_framebuffer_object 1
 GLAPI int GLAD_GL_ARB_framebuffer_object;
 #endif
+#ifndef GL_EXT_framebuffer_blit
+#define GL_EXT_framebuffer_blit 1
+GLAPI int GLAD_GL_EXT_framebuffer_blit;
+typedef void (APIENTRYP PFNGLBLITFRAMEBUFFEREXTPROC)(GLint srcX0, GLint srcY0, GLint srcX1, GLint srcY1, GLint dstX0, GLint dstY0, GLint dstX1, GLint dstY1, GLbitfield mask, GLenum filter);
+GLAPI PFNGLBLITFRAMEBUFFEREXTPROC glad_glBlitFramebufferEXT;
+#define glBlitFramebufferEXT glad_glBlitFramebufferEXT
+#endif
+#ifndef GL_EXT_framebuffer_multisample
+#define GL_EXT_framebuffer_multisample 1
+GLAPI int GLAD_GL_EXT_framebuffer_multisample;
+typedef void (APIENTRYP PFNGLRENDERBUFFERSTORAGEMULTISAMPLEEXTPROC)(GLenum target, GLsizei samples, GLenum internalformat, GLsizei width, GLsizei height);
+GLAPI PFNGLRENDERBUFFERSTORAGEMULTISAMPLEEXTPROC glad_glRenderbufferStorageMultisampleEXT;
+#define glRenderbufferStorageMultisampleEXT glad_glRenderbufferStorageMultisampleEXT
+#endif
 #ifndef GL_EXT_framebuffer_object
 #define GL_EXT_framebuffer_object 1
 GLAPI int GLAD_GL_EXT_framebuffer_object;
diff --git a/thirdparty/misc/clipper-exceptions.patch b/thirdparty/misc/clipper-exceptions.patch
new file mode 100644
index 0000000000..537afd59b3
--- /dev/null
+++ b/thirdparty/misc/clipper-exceptions.patch
@@ -0,0 +1,154 @@
+diff --git a/thirdparty/misc/clipper.cpp b/thirdparty/misc/clipper.cpp
+index 8c3a59c4ca..c67045d113 100644
+--- a/thirdparty/misc/clipper.cpp
++++ b/thirdparty/misc/clipper.cpp
+@@ -48,6 +48,38 @@
+ #include <ostream>
+ #include <functional>
+ 
++//Explicitly disables exceptions handling for target platform
++//#define CLIPPER_NOEXCEPTION
++
++#define CLIPPER_THROW(exception) std::abort()
++#define CLIPPER_TRY if(true)
++#define CLIPPER_CATCH(exception) if(false)
++
++#if defined(__cpp_exceptions) || defined(__EXCEPTIONS) || defined(_CPPUNWIND)
++  #ifndef CLIPPER_NOEXCEPTION
++    #undef CLIPPER_THROW
++    #define CLIPPER_THROW(exception) throw exception
++    #undef CLIPPER_TRY
++    #define CLIPPER_TRY try
++    #undef CLIPPER_CATCH
++    #define CLIPPER_CATCH(exception) catch(exception)
++  #endif
++#endif
++
++//Optionally allows to override exception macros
++#if defined(CLIPPER_THROW_USER)
++	#undef CLIPPER_THROW
++	#define CLIPPER_THROW CLIPPER_THROW_USER
++#endif
++#if defined(CLIPPER_TRY_USER)
++	#undef CLIPPER_TRY
++	#define CLIPPER_TRY CLIPPER_TRY_USER
++#endif
++#if defined(CLIPPER_CATCH_USER)
++	#undef CLIPPER_CATCH
++	#define CLIPPER_CATCH CLIPPER_CATCH_USER
++#endif
++
+ namespace ClipperLib {
+ 
+ static double const pi = 3.141592653589793238;
+@@ -898,7 +930,7 @@ void RangeTest(const IntPoint& Pt, bool& useFullRange)
+   if (useFullRange)
+   {
+     if (Pt.X > hiRange || Pt.Y > hiRange || -Pt.X > hiRange || -Pt.Y > hiRange) 
+-      throw clipperException("Coordinate outside allowed range");
++      CLIPPER_THROW(clipperException("Coordinate outside allowed range"));
+   }
+   else if (Pt.X > loRange|| Pt.Y > loRange || -Pt.X > loRange || -Pt.Y > loRange) 
+   {
+@@ -1046,10 +1078,10 @@ bool ClipperBase::AddPath(const Path &pg, PolyType PolyTyp, bool Closed)
+ {
+ #ifdef use_lines
+   if (!Closed && PolyTyp == ptClip)
+-    throw clipperException("AddPath: Open paths must be subject.");
++    CLIPPER_THROW(clipperException("AddPath: Open paths must be subject."));
+ #else
+   if (!Closed)
+-    throw clipperException("AddPath: Open paths have been disabled.");
++    CLIPPER_THROW(clipperException("AddPath: Open paths have been disabled."));
+ #endif
+ 
+   int highI = (int)pg.size() -1;
+@@ -1062,7 +1094,7 @@ bool ClipperBase::AddPath(const Path &pg, PolyType PolyTyp, bool Closed)
+ 
+   bool IsFlat = true;
+   //1. Basic (first) edge initialization ...
+-  try
++  CLIPPER_TRY
+   {
+     edges[1].Curr = pg[1];
+     RangeTest(pg[0], m_UseFullRange);
+@@ -1075,10 +1107,10 @@ bool ClipperBase::AddPath(const Path &pg, PolyType PolyTyp, bool Closed)
+       InitEdge(&edges[i], &edges[i+1], &edges[i-1], pg[i]);
+     }
+   }
+-  catch(...)
++  CLIPPER_CATCH(...)
+   {
+     delete [] edges;
+-    throw; //range test fails
++    CLIPPER_THROW(); //range test fails
+   }
+   TEdge *eStart = &edges[0];
+ 
+@@ -1442,7 +1474,7 @@ void ClipperBase::SwapPositionsInAEL(TEdge *Edge1, TEdge *Edge2)
+ void ClipperBase::UpdateEdgeIntoAEL(TEdge *&e)
+ {
+   if (!e->NextInLML) 
+-    throw clipperException("UpdateEdgeIntoAEL: invalid call");
++    CLIPPER_THROW(clipperException("UpdateEdgeIntoAEL: invalid call"));
+ 
+   e->NextInLML->OutIdx = e->OutIdx;
+   TEdge* AelPrev = e->PrevInAEL;
+@@ -1510,7 +1542,7 @@ bool Clipper::Execute(ClipType clipType, Paths &solution,
+ {
+   if( m_ExecuteLocked ) return false;
+   if (m_HasOpenPaths)
+-    throw clipperException("Error: PolyTree struct is needed for open path clipping.");
++    CLIPPER_THROW(clipperException("Error: PolyTree struct is needed for open path clipping."));
+   m_ExecuteLocked = true;
+   solution.resize(0);
+   m_SubjFillType = subjFillType;
+@@ -1560,7 +1592,7 @@ void Clipper::FixHoleLinkage(OutRec &outrec)
+ bool Clipper::ExecuteInternal()
+ {
+   bool succeeded = true;
+-  try {
++  CLIPPER_TRY {
+     Reset();
+     m_Maxima = MaximaList();
+     m_SortedEdges = 0;
+@@ -1583,7 +1615,7 @@ bool Clipper::ExecuteInternal()
+       InsertLocalMinimaIntoAEL(botY);
+     }
+   }
+-  catch(...) 
++  CLIPPER_CATCH(...) 
+   {
+     succeeded = false;
+   }
+@@ -2827,18 +2859,18 @@ void Clipper::ProcessHorizontal(TEdge *horzEdge)
+ bool Clipper::ProcessIntersections(const cInt topY)
+ {
+   if( !m_ActiveEdges ) return true;
+-  try {
++  CLIPPER_TRY {
+     BuildIntersectList(topY);
+     size_t IlSize = m_IntersectList.size();
+     if (IlSize == 0) return true;
+     if (IlSize == 1 || FixupIntersectionOrder()) ProcessIntersectList();
+     else return false;
+   }
+-  catch(...) 
++  CLIPPER_CATCH(...) 
+   {
+     m_SortedEdges = 0;
+     DisposeIntersectNodes();
+-    throw clipperException("ProcessIntersections error");
++    CLIPPER_THROW(clipperException("ProcessIntersections error"));
+   }
+   m_SortedEdges = 0;
+   return true;
+@@ -3002,7 +3034,7 @@ void Clipper::DoMaxima(TEdge *e)
+     DeleteFromAEL(eMaxPair);
+   } 
+ #endif
+-  else throw clipperException("DoMaxima error");
++  else CLIPPER_THROW(clipperException("DoMaxima error"));
+ }
+ //------------------------------------------------------------------------------
+ 
diff --git a/thirdparty/misc/clipper.cpp b/thirdparty/misc/clipper.cpp
index d3143fe5ab..c67045d113 100644
--- a/thirdparty/misc/clipper.cpp
+++ b/thirdparty/misc/clipper.cpp
@@ -48,6 +48,38 @@
 #include <ostream>
 #include <functional>
 
+//Explicitly disables exceptions handling for target platform
+//#define CLIPPER_NOEXCEPTION
+
+#define CLIPPER_THROW(exception) std::abort()
+#define CLIPPER_TRY if(true)
+#define CLIPPER_CATCH(exception) if(false)
+
+#if defined(__cpp_exceptions) || defined(__EXCEPTIONS) || defined(_CPPUNWIND)
+  #ifndef CLIPPER_NOEXCEPTION
+    #undef CLIPPER_THROW
+    #define CLIPPER_THROW(exception) throw exception
+    #undef CLIPPER_TRY
+    #define CLIPPER_TRY try
+    #undef CLIPPER_CATCH
+    #define CLIPPER_CATCH(exception) catch(exception)
+  #endif
+#endif
+
+//Optionally allows to override exception macros
+#if defined(CLIPPER_THROW_USER)
+	#undef CLIPPER_THROW
+	#define CLIPPER_THROW CLIPPER_THROW_USER
+#endif
+#if defined(CLIPPER_TRY_USER)
+	#undef CLIPPER_TRY
+	#define CLIPPER_TRY CLIPPER_TRY_USER
+#endif
+#if defined(CLIPPER_CATCH_USER)
+	#undef CLIPPER_CATCH
+	#define CLIPPER_CATCH CLIPPER_CATCH_USER
+#endif
+
 namespace ClipperLib {
 
 static double const pi = 3.141592653589793238;
@@ -898,7 +930,7 @@ void RangeTest(const IntPoint& Pt, bool& useFullRange)
   if (useFullRange)
   {
     if (Pt.X > hiRange || Pt.Y > hiRange || -Pt.X > hiRange || -Pt.Y > hiRange) 
-      throw clipperException("Coordinate outside allowed range");
+      CLIPPER_THROW(clipperException("Coordinate outside allowed range"));
   }
   else if (Pt.X > loRange|| Pt.Y > loRange || -Pt.X > loRange || -Pt.Y > loRange) 
   {
@@ -1046,10 +1078,10 @@ bool ClipperBase::AddPath(const Path &pg, PolyType PolyTyp, bool Closed)
 {
 #ifdef use_lines
   if (!Closed && PolyTyp == ptClip)
-    throw clipperException("AddPath: Open paths must be subject.");
+    CLIPPER_THROW(clipperException("AddPath: Open paths must be subject."));
 #else
   if (!Closed)
-    throw clipperException("AddPath: Open paths have been disabled.");
+    CLIPPER_THROW(clipperException("AddPath: Open paths have been disabled."));
 #endif
 
   int highI = (int)pg.size() -1;
@@ -1062,7 +1094,7 @@ bool ClipperBase::AddPath(const Path &pg, PolyType PolyTyp, bool Closed)
 
   bool IsFlat = true;
   //1. Basic (first) edge initialization ...
-  try
+  CLIPPER_TRY
   {
     edges[1].Curr = pg[1];
     RangeTest(pg[0], m_UseFullRange);
@@ -1075,10 +1107,10 @@ bool ClipperBase::AddPath(const Path &pg, PolyType PolyTyp, bool Closed)
       InitEdge(&edges[i], &edges[i+1], &edges[i-1], pg[i]);
     }
   }
-  catch(...)
+  CLIPPER_CATCH(...)
   {
     delete [] edges;
-    throw; //range test fails
+    CLIPPER_THROW(); //range test fails
   }
   TEdge *eStart = &edges[0];
 
@@ -1442,7 +1474,7 @@ void ClipperBase::SwapPositionsInAEL(TEdge *Edge1, TEdge *Edge2)
 void ClipperBase::UpdateEdgeIntoAEL(TEdge *&e)
 {
   if (!e->NextInLML) 
-    throw clipperException("UpdateEdgeIntoAEL: invalid call");
+    CLIPPER_THROW(clipperException("UpdateEdgeIntoAEL: invalid call"));
 
   e->NextInLML->OutIdx = e->OutIdx;
   TEdge* AelPrev = e->PrevInAEL;
@@ -1510,7 +1542,7 @@ bool Clipper::Execute(ClipType clipType, Paths &solution,
 {
   if( m_ExecuteLocked ) return false;
   if (m_HasOpenPaths)
-    throw clipperException("Error: PolyTree struct is needed for open path clipping.");
+    CLIPPER_THROW(clipperException("Error: PolyTree struct is needed for open path clipping."));
   m_ExecuteLocked = true;
   solution.resize(0);
   m_SubjFillType = subjFillType;
@@ -1560,7 +1592,7 @@ void Clipper::FixHoleLinkage(OutRec &outrec)
 bool Clipper::ExecuteInternal()
 {
   bool succeeded = true;
-  try {
+  CLIPPER_TRY {
     Reset();
     m_Maxima = MaximaList();
     m_SortedEdges = 0;
@@ -1583,7 +1615,7 @@ bool Clipper::ExecuteInternal()
       InsertLocalMinimaIntoAEL(botY);
     }
   }
-  catch(...) 
+  CLIPPER_CATCH(...) 
   {
     succeeded = false;
   }
@@ -2827,18 +2859,18 @@ void Clipper::ProcessHorizontal(TEdge *horzEdge)
 bool Clipper::ProcessIntersections(const cInt topY)
 {
   if( !m_ActiveEdges ) return true;
-  try {
+  CLIPPER_TRY {
     BuildIntersectList(topY);
     size_t IlSize = m_IntersectList.size();
     if (IlSize == 0) return true;
     if (IlSize == 1 || FixupIntersectionOrder()) ProcessIntersectList();
     else return false;
   }
-  catch(...) 
+  CLIPPER_CATCH(...) 
   {
     m_SortedEdges = 0;
     DisposeIntersectNodes();
-    throw clipperException("ProcessIntersections error");
+    CLIPPER_THROW(clipperException("ProcessIntersections error"));
   }
   m_SortedEdges = 0;
   return true;
@@ -3002,7 +3034,7 @@ void Clipper::DoMaxima(TEdge *e)
     DeleteFromAEL(eMaxPair);
   } 
 #endif
-  else throw clipperException("DoMaxima error");
+  else CLIPPER_THROW(clipperException("DoMaxima error"));
 }
 //------------------------------------------------------------------------------
 
@@ -4329,10 +4361,10 @@ double DistanceFromLineSqrd(
   const IntPoint& pt, const IntPoint& ln1, const IntPoint& ln2)
 {
   //The equation of a line in general form (Ax + By + C = 0)
-  //given 2 points (x¹,y¹) & (x²,y²) is ...
-  //(y¹ - y²)x + (x² - x¹)y + (y² - y¹)x¹ - (x² - x¹)y¹ = 0
-  //A = (y¹ - y²); B = (x² - x¹); C = (y² - y¹)x¹ - (x² - x¹)y¹
-  //perpendicular distance of point (x³,y³) = (Ax³ + By³ + C)/Sqrt(A² + B²)
+  //given 2 points (x¹,y¹) & (x²,y²) is ...
+  //(y¹ - y²)x + (x² - x¹)y + (y² - y¹)x¹ - (x² - x¹)y¹ = 0
+  //A = (y¹ - y²); B = (x² - x¹); C = (y² - y¹)x¹ - (x² - x¹)y¹
+  //perpendicular distance of point (x³,y³) = (Ax³ + By³ + C)/Sqrt(A² + B²)
   //see http://en.wikipedia.org/wiki/Perpendicular_distance
   double A = double(ln1.Y - ln2.Y);
   double B = double(ln2.X - ln1.X);
diff --git a/thirdparty/xatlas/xatlas.cpp b/thirdparty/xatlas/xatlas.cpp
index eb0824a517..2cc2905eee 100644
--- a/thirdparty/xatlas/xatlas.cpp
+++ b/thirdparty/xatlas/xatlas.cpp
@@ -4388,7 +4388,7 @@ private:
 class Solver
 {
 public:
-	// Solve the symmetric system: At·A·x = At·b
+	// Solve the symmetric system: At·A·x = At·b
 	static bool LeastSquaresSolver(const sparse::Matrix &A, const FullVector &b, FullVector &x, float epsilon = 1e-5f)
 	{
 		xaDebugAssert(A.width() == x.dimension());
@@ -4477,22 +4477,22 @@ private:
 	* Gradient method.
 	*
 	* Solving sparse linear systems:
-	* (1)		A·x = b
+	* (1)		A·x = b
 	*
 	* The conjugate gradient algorithm solves (1) only in the case that A is
 	* symmetric and positive definite. It is based on the idea of minimizing the
 	* function
 	*
-	* (2)		f(x) = 1/2·x·A·x - b·x
+	* (2)		f(x) = 1/2·x·A·x - b·x
 	*
 	* This function is minimized when its gradient
 	*
-	* (3)		df = A·x - b
+	* (3)		df = A·x - b
 	*
 	* is zero, which is equivalent to (1). The minimization is carried out by
 	* generating a succession of search directions p.k and improved minimizers x.k.
-	* At each stage a quantity alfa.k is found that minimizes f(x.k + alfa.k·p.k),
-	* and x.k+1 is set equal to the new point x.k + alfa.k·p.k. The p.k and x.k are
+	* At each stage a quantity alfa.k is found that minimizes f(x.k + alfa.k·p.k),
+	* and x.k+1 is set equal to the new point x.k + alfa.k·p.k. The p.k and x.k are
 	* built up in such a way that x.k+1 is also the minimizer of f over the whole
 	* vector space of directions already taken, {p.1, p.2, . . . , p.k}. After N
 	* iterations you arrive at the minimizer over the entire vector space, i.e., the
@@ -4520,7 +4520,7 @@ private:
 		float delta_new;
 		float alpha;
 		float beta;
-		// r = b - A·x;
+		// r = b - A·x;
 		sparse::copy(b, r);
 		sparse::sgemv(-1, A, x, 1, r);
 		// p = r;
@@ -4529,24 +4529,24 @@ private:
 		delta_0 = delta_new;
 		while (i < i_max && delta_new > epsilon * epsilon * delta_0) {
 			i++;
-			// q = A·p
+			// q = A·p
 			mult(A, p, q);
-			// alpha = delta_new / p·q
+			// alpha = delta_new / p·q
 			alpha = delta_new / sparse::dot( p, q );
-			// x = alfa·p + x
+			// x = alfa·p + x
 			sparse::saxpy(alpha, p, x);
 			if ((i & 31) == 0) { // recompute r after 32 steps
-				// r = b - A·x
+				// r = b - A·x
 				sparse::copy(b, r);
 				sparse::sgemv(-1, A, x, 1, r);
 			} else {
-				// r = r - alpha·q
+				// r = r - alpha·q
 				sparse::saxpy(-alpha, q, r);
 			}
 			delta_old = delta_new;
 			delta_new = sparse::dot( r, r );
 			beta = delta_new / delta_old;
-			// p = beta·p + r
+			// p = beta·p + r
 			sparse::scal(beta, p);
 			sparse::saxpy(1, r, p);
 		}
@@ -4572,35 +4572,35 @@ private:
 		float delta_new;
 		float alpha;
 		float beta;
-		// r = b - A·x
+		// r = b - A·x
 		sparse::copy(b, r);
 		sparse::sgemv(-1, A, x, 1, r);
-		// p = M^-1 · r
+		// p = M^-1 · r
 		preconditioner.apply(r, p);
 		delta_new = sparse::dot(r, p);
 		delta_0 = delta_new;
 		while (i < i_max && delta_new > epsilon * epsilon * delta_0) {
 			i++;
-			// q = A·p
+			// q = A·p
 			mult(A, p, q);
-			// alpha = delta_new / p·q
+			// alpha = delta_new / p·q
 			alpha = delta_new / sparse::dot(p, q);
-			// x = alfa·p + x
+			// x = alfa·p + x
 			sparse::saxpy(alpha, p, x);
 			if ((i & 31) == 0) { // recompute r after 32 steps
-				// r = b - A·x
+				// r = b - A·x
 				sparse::copy(b, r);
 				sparse::sgemv(-1, A, x, 1, r);
 			} else {
-				// r = r - alfa·q
+				// r = r - alfa·q
 				sparse::saxpy(-alpha, q, r);
 			}
-			// s = M^-1 · r
+			// s = M^-1 · r
 			preconditioner.apply(r, s);
 			delta_old = delta_new;
 			delta_new = sparse::dot( r, s );
 			beta = delta_new / delta_old;
-			// p = s + beta·p
+			// p = s + beta·p
 			sparse::scal(beta, p);
 			sparse::saxpy(1, s, p);
 		}