Moved stuff from mapblock{h,cpp} to mapblock_mesh.{h,cpp} and content_mapblock.{h,cpp}

src/mapblock_mesh.cpp

@@ -0,0 +1,789 @@
+/*
+Minetest-c55
+Copyright (C) 2010-2011 celeron55, Perttu Ahola <celeron55@gmail.com>
+
+This program is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 2 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License along
+with this program; if not, write to the Free Software Foundation, Inc.,
+51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+*/
+
+#include "mapblock_mesh.h"
+#include "light.h"
+#include "mapblock.h"
+#include "map.h"
+#include "main.h" // For g_settings and g_texturesource
+#include "content_mapblock.h"
+
+void MeshMakeData::fill(u32 daynight_ratio, MapBlock *block)
+{
+	m_daynight_ratio = daynight_ratio;
+	m_blockpos = block->getPos();
+
+	v3s16 blockpos_nodes = m_blockpos*MAP_BLOCKSIZE;
+	
+	/*
+		There is no harm not copying the TempMods of the neighbors
+		because they are already copied to this block
+	*/
+	m_temp_mods.clear();
+	block->copyTempMods(m_temp_mods);
+	
+	/*
+		Copy data
+	*/
+
+	// Allocate this block + neighbors
+	m_vmanip.clear();
+	m_vmanip.addArea(VoxelArea(blockpos_nodes-v3s16(1,1,1)*MAP_BLOCKSIZE,
+			blockpos_nodes+v3s16(1,1,1)*MAP_BLOCKSIZE*2-v3s16(1,1,1)));
+
+	{
+		//TimeTaker timer("copy central block data");
+		// 0ms
+
+		// Copy our data
+		block->copyTo(m_vmanip);
+	}
+	{
+		//TimeTaker timer("copy neighbor block data");
+		// 0ms
+
+		/*
+			Copy neighbors. This is lightning fast.
+			Copying only the borders would be *very* slow.
+		*/
+		
+		// Get map
+		NodeContainer *parentcontainer = block->getParent();
+		// This will only work if the parent is the map
+		assert(parentcontainer->nodeContainerId() == NODECONTAINER_ID_MAP);
+		// OK, we have the map!
+		Map *map = (Map*)parentcontainer;
+
+		for(u16 i=0; i<6; i++)
+		{
+			const v3s16 &dir = g_6dirs[i];
+			v3s16 bp = m_blockpos + dir;
+			MapBlock *b = map->getBlockNoCreateNoEx(bp);
+			if(b)
+				b->copyTo(m_vmanip);
+		}
+	}
+}
+
+/*
+	vertex_dirs: v3s16[4]
+*/
+void getNodeVertexDirs(v3s16 dir, v3s16 *vertex_dirs)
+{
+	/*
+		If looked from outside the node towards the face, the corners are:
+		0: bottom-right
+		1: bottom-left
+		2: top-left
+		3: top-right
+	*/
+	if(dir == v3s16(0,0,1))
+	{
+		// If looking towards z+, this is the face that is behind
+		// the center point, facing towards z+.
+		vertex_dirs[0] = v3s16(-1,-1, 1);
+		vertex_dirs[1] = v3s16( 1,-1, 1);
+		vertex_dirs[2] = v3s16( 1, 1, 1);
+		vertex_dirs[3] = v3s16(-1, 1, 1);
+	}
+	else if(dir == v3s16(0,0,-1))
+	{
+		// faces towards Z-
+		vertex_dirs[0] = v3s16( 1,-1,-1);
+		vertex_dirs[1] = v3s16(-1,-1,-1);
+		vertex_dirs[2] = v3s16(-1, 1,-1);
+		vertex_dirs[3] = v3s16( 1, 1,-1);
+	}
+	else if(dir == v3s16(1,0,0))
+	{
+		// faces towards X+
+		vertex_dirs[0] = v3s16( 1,-1, 1);
+		vertex_dirs[1] = v3s16( 1,-1,-1);
+		vertex_dirs[2] = v3s16( 1, 1,-1);
+		vertex_dirs[3] = v3s16( 1, 1, 1);
+	}
+	else if(dir == v3s16(-1,0,0))
+	{
+		// faces towards X-
+		vertex_dirs[0] = v3s16(-1,-1,-1);
+		vertex_dirs[1] = v3s16(-1,-1, 1);
+		vertex_dirs[2] = v3s16(-1, 1, 1);
+		vertex_dirs[3] = v3s16(-1, 1,-1);
+	}
+	else if(dir == v3s16(0,1,0))
+	{
+		// faces towards Y+ (assume Z- as "down" in texture)
+		vertex_dirs[0] = v3s16( 1, 1,-1);
+		vertex_dirs[1] = v3s16(-1, 1,-1);
+		vertex_dirs[2] = v3s16(-1, 1, 1);
+		vertex_dirs[3] = v3s16( 1, 1, 1);
+	}
+	else if(dir == v3s16(0,-1,0))
+	{
+		// faces towards Y- (assume Z+ as "down" in texture)
+		vertex_dirs[0] = v3s16( 1,-1, 1);
+		vertex_dirs[1] = v3s16(-1,-1, 1);
+		vertex_dirs[2] = v3s16(-1,-1,-1);
+		vertex_dirs[3] = v3s16( 1,-1,-1);
+	}
+}
+
+inline video::SColor lightColor(u8 alpha, u8 light)
+{
+	return video::SColor(alpha,light,light,light);
+}
+
+struct FastFace
+{
+	TileSpec tile;
+	video::S3DVertex vertices[4]; // Precalculated vertices
+};
+
+void makeFastFace(TileSpec tile, u8 li0, u8 li1, u8 li2, u8 li3, v3f p,
+		v3s16 dir, v3f scale, v3f posRelative_f,
+		core::array<FastFace> &dest)
+{
+	FastFace face;
+	
+	// Position is at the center of the cube.
+	v3f pos = p * BS;
+	posRelative_f *= BS;
+
+	v3f vertex_pos[4];
+	v3s16 vertex_dirs[4];
+	getNodeVertexDirs(dir, vertex_dirs);
+	for(u16 i=0; i<4; i++)
+	{
+		vertex_pos[i] = v3f(
+				BS/2*vertex_dirs[i].X,
+				BS/2*vertex_dirs[i].Y,
+				BS/2*vertex_dirs[i].Z
+		);
+	}
+
+	for(u16 i=0; i<4; i++)
+	{
+		vertex_pos[i].X *= scale.X;
+		vertex_pos[i].Y *= scale.Y;
+		vertex_pos[i].Z *= scale.Z;
+		vertex_pos[i] += pos + posRelative_f;
+	}
+
+	f32 abs_scale = 1.;
+	if     (scale.X < 0.999 || scale.X > 1.001) abs_scale = scale.X;
+	else if(scale.Y < 0.999 || scale.Y > 1.001) abs_scale = scale.Y;
+	else if(scale.Z < 0.999 || scale.Z > 1.001) abs_scale = scale.Z;
+
+	v3f zerovector = v3f(0,0,0);
+	
+	u8 alpha = tile.alpha;
+	/*u8 alpha = 255;
+	if(tile.id == TILE_WATER)
+		alpha = WATER_ALPHA;*/
+
+	float x0 = tile.texture.pos.X;
+	float y0 = tile.texture.pos.Y;
+	float w = tile.texture.size.X;
+	float h = tile.texture.size.Y;
+
+	/*video::SColor c = lightColor(alpha, li);
+
+	face.vertices[0] = video::S3DVertex(vertex_pos[0], v3f(0,1,0), c,
+			core::vector2d<f32>(x0+w*abs_scale, y0+h));
+	face.vertices[1] = video::S3DVertex(vertex_pos[1], v3f(0,1,0), c,
+			core::vector2d<f32>(x0, y0+h));
+	face.vertices[2] = video::S3DVertex(vertex_pos[2], v3f(0,1,0), c,
+			core::vector2d<f32>(x0, y0));
+	face.vertices[3] = video::S3DVertex(vertex_pos[3], v3f(0,1,0), c,
+			core::vector2d<f32>(x0+w*abs_scale, y0));*/
+
+	face.vertices[0] = video::S3DVertex(vertex_pos[0], v3f(0,1,0),
+			lightColor(alpha, li0),
+			core::vector2d<f32>(x0+w*abs_scale, y0+h));
+	face.vertices[1] = video::S3DVertex(vertex_pos[1], v3f(0,1,0),
+			lightColor(alpha, li1),
+			core::vector2d<f32>(x0, y0+h));
+	face.vertices[2] = video::S3DVertex(vertex_pos[2], v3f(0,1,0),
+			lightColor(alpha, li2),
+			core::vector2d<f32>(x0, y0));
+	face.vertices[3] = video::S3DVertex(vertex_pos[3], v3f(0,1,0),
+			lightColor(alpha, li3),
+			core::vector2d<f32>(x0+w*abs_scale, y0));
+
+	face.tile = tile;
+	//DEBUG
+	//f->tile = TILE_STONE;
+	
+	dest.push_back(face);
+}
+	
+/*
+	Gets node tile from any place relative to block.
+	Returns TILE_NODE if doesn't exist or should not be drawn.
+*/
+TileSpec getNodeTile(MapNode mn, v3s16 p, v3s16 face_dir,
+		NodeModMap &temp_mods)
+{
+	TileSpec spec;
+	spec = mn.getTile(face_dir);
+	
+	/*
+		Check temporary modifications on this node
+	*/
+	/*core::map<v3s16, NodeMod>::Node *n;
+	n = m_temp_mods.find(p);
+	// If modified
+	if(n != NULL)
+	{
+		struct NodeMod mod = n->getValue();*/
+	NodeMod mod;
+	if(temp_mods.get(p, &mod))
+	{
+		if(mod.type == NODEMOD_CHANGECONTENT)
+		{
+			MapNode mn2(mod.param);
+			spec = mn2.getTile(face_dir);
+		}
+		if(mod.type == NODEMOD_CRACK)
+		{
+			/*
+				Get texture id, translate it to name, append stuff to
+				name, get texture id
+			*/
+
+			// Get original texture name
+			u32 orig_id = spec.texture.id;
+			std::string orig_name = g_texturesource->getTextureName(orig_id);
+
+			// Create new texture name
+			std::ostringstream os;
+			os<<orig_name<<"^[crack"<<mod.param;
+
+			// Get new texture
+			u32 new_id = g_texturesource->getTextureId(os.str());
+			
+			/*dstream<<"MapBlock::getNodeTile(): Switching from "
+					<<orig_name<<" to "<<os.str()<<" ("
+					<<orig_id<<" to "<<new_id<<")"<<std::endl;*/
+			
+			spec.texture = g_texturesource->getTexture(new_id);
+		}
+	}
+	
+	return spec;
+}
+
+u8 getNodeContent(v3s16 p, MapNode mn, NodeModMap &temp_mods)
+{
+	/*
+		Check temporary modifications on this node
+	*/
+	/*core::map<v3s16, NodeMod>::Node *n;
+	n = m_temp_mods.find(p);
+	// If modified
+	if(n != NULL)
+	{
+		struct NodeMod mod = n->getValue();*/
+	NodeMod mod;
+	if(temp_mods.get(p, &mod))
+	{
+		if(mod.type == NODEMOD_CHANGECONTENT)
+		{
+			// Overrides content
+			return mod.param;
+		}
+		if(mod.type == NODEMOD_CRACK)
+		{
+			/*
+				Content doesn't change.
+				
+				face_contents works just like it should, because
+				there should not be faces between differently cracked
+				nodes.
+
+				If a semi-transparent node is cracked in front an
+				another one, it really doesn't matter whether there
+				is a cracked face drawn in between or not.
+			*/
+		}
+	}
+
+	return mn.d;
+}
+
+v3s16 dirs8[8] = {
+	v3s16(0,0,0),
+	v3s16(0,0,1),
+	v3s16(0,1,0),
+	v3s16(0,1,1),
+	v3s16(1,0,0),
+	v3s16(1,1,0),
+	v3s16(1,0,1),
+	v3s16(1,1,1),
+};
+
+// Calculate lighting at the XYZ- corner of p
+u8 getSmoothLight(v3s16 p, VoxelManipulator &vmanip, u32 daynight_ratio)
+{
+	u16 ambient_occlusion = 0;
+	u16 light = 0;
+	u16 light_count = 0;
+	for(u32 i=0; i<8; i++)
+	{
+		MapNode n = vmanip.getNodeNoEx(p - dirs8[i]);
+		if(content_features(n.d).param_type == CPT_LIGHT)
+		{
+			light += decode_light(n.getLightBlend(daynight_ratio));
+			light_count++;
+		}
+		else
+		{
+			if(n.d != CONTENT_IGNORE)
+				ambient_occlusion++;
+		}
+	}
+
+	if(light_count == 0)
+		return 255;
+	
+	light /= light_count;
+
+	if(ambient_occlusion > 4)
+	{
+		ambient_occlusion -= 4;
+		light = (float)light / ((float)ambient_occlusion * 0.5 + 1.0);
+	}
+
+	return light;
+}
+
+// Calculate lighting at the given corner of p
+u8 getSmoothLight(v3s16 p, v3s16 corner,
+		VoxelManipulator &vmanip, u32 daynight_ratio)
+{
+	if(corner.X == 1) p.X += 1;
+	else              assert(corner.X == -1);
+	if(corner.Y == 1) p.Y += 1;
+	else              assert(corner.Y == -1);
+	if(corner.Z == 1) p.Z += 1;
+	else              assert(corner.Z == -1);
+	
+	return getSmoothLight(p, vmanip, daynight_ratio);
+}
+
+void getTileInfo(
+		// Input:
+		v3s16 blockpos_nodes,
+		v3s16 p,
+		v3s16 face_dir,
+		u32 daynight_ratio,
+		VoxelManipulator &vmanip,
+		NodeModMap &temp_mods,
+		bool smooth_lighting,
+		// Output:
+		bool &makes_face,
+		v3s16 &p_corrected,
+		v3s16 &face_dir_corrected,
+		u8 *lights,
+		TileSpec &tile
+	)
+{
+	MapNode n0 = vmanip.getNodeNoEx(blockpos_nodes + p);
+	MapNode n1 = vmanip.getNodeNoEx(blockpos_nodes + p + face_dir);
+	TileSpec tile0 = getNodeTile(n0, p, face_dir, temp_mods);
+	TileSpec tile1 = getNodeTile(n1, p + face_dir, -face_dir, temp_mods);
+	
+	// This is hackish
+	u8 content0 = getNodeContent(p, n0, temp_mods);
+	u8 content1 = getNodeContent(p + face_dir, n1, temp_mods);
+	u8 mf = face_contents(content0, content1);
+
+	if(mf == 0)
+	{
+		makes_face = false;
+		return;
+	}
+
+	makes_face = true;
+	
+	if(mf == 1)
+	{
+		tile = tile0;
+		p_corrected = p;
+		face_dir_corrected = face_dir;
+	}
+	else
+	{
+		tile = tile1;
+		p_corrected = p + face_dir;
+		face_dir_corrected = -face_dir;
+	}
+	
+	if(smooth_lighting == false)
+	{
+		lights[0] = lights[1] = lights[2] = lights[3] =
+				decode_light(getFaceLight(daynight_ratio, n0, n1, face_dir));
+	}
+	else
+	{
+		v3s16 vertex_dirs[4];
+		getNodeVertexDirs(face_dir_corrected, vertex_dirs);
+		for(u16 i=0; i<4; i++)
+		{
+			lights[i] = getSmoothLight(blockpos_nodes + p_corrected,
+					vertex_dirs[i], vmanip, daynight_ratio);
+		}
+	}
+	
+	return;
+}
+
+/*
+	startpos:
+	translate_dir: unit vector with only one of x, y or z
+	face_dir: unit vector with only one of x, y or z
+*/
+void updateFastFaceRow(
+		u32 daynight_ratio,
+		v3f posRelative_f,
+		v3s16 startpos,
+		u16 length,
+		v3s16 translate_dir,
+		v3f translate_dir_f,
+		v3s16 face_dir,
+		v3f face_dir_f,
+		core::array<FastFace> &dest,
+		NodeModMap &temp_mods,
+		VoxelManipulator &vmanip,
+		v3s16 blockpos_nodes,
+		bool smooth_lighting)
+{
+	v3s16 p = startpos;
+	
+	u16 continuous_tiles_count = 0;
+	
+	bool makes_face;
+	v3s16 p_corrected;
+	v3s16 face_dir_corrected;
+	u8 lights[4];
+	TileSpec tile;
+	getTileInfo(blockpos_nodes, p, face_dir, daynight_ratio,
+			vmanip, temp_mods, smooth_lighting,
+			makes_face, p_corrected, face_dir_corrected, lights, tile);
+
+	for(u16 j=0; j<length; j++)
+	{
+		// If tiling can be done, this is set to false in the next step
+		bool next_is_different = true;
+		
+		v3s16 p_next;
+		
+		bool next_makes_face = false;
+		v3s16 next_p_corrected;
+		v3s16 next_face_dir_corrected;
+		u8 next_lights[4] = {0,0,0,0};
+		TileSpec next_tile;
+		
+		// If at last position, there is nothing to compare to and
+		// the face must be drawn anyway
+		if(j != length - 1)
+		{
+			p_next = p + translate_dir;
+			
+			getTileInfo(blockpos_nodes, p_next, face_dir, daynight_ratio,
+					vmanip, temp_mods, smooth_lighting,
+					next_makes_face, next_p_corrected,
+					next_face_dir_corrected, next_lights,
+					next_tile);
+			
+			if(next_makes_face == makes_face
+					&& next_p_corrected == p_corrected
+					&& next_face_dir_corrected == face_dir_corrected
+					&& next_lights[0] == lights[0]
+					&& next_lights[1] == lights[1]
+					&& next_lights[2] == lights[2]
+					&& next_lights[3] == lights[3]
+					&& next_tile == tile)
+			{
+				next_is_different = false;
+			}
+		}
+
+		continuous_tiles_count++;
+		
+		// This is set to true if the texture doesn't allow more tiling
+		bool end_of_texture = false;
+		/*
+			If there is no texture, it can be tiled infinitely.
+			If tiled==0, it means the texture can be tiled infinitely.
+			Otherwise check tiled agains continuous_tiles_count.
+		*/
+		if(tile.texture.atlas != NULL && tile.texture.tiled != 0)
+		{
+			if(tile.texture.tiled <= continuous_tiles_count)
+				end_of_texture = true;
+		}
+		
+		// Do this to disable tiling textures
+		//end_of_texture = true; //DEBUG
+		
+		if(next_is_different || end_of_texture)
+		{
+			/*
+				Create a face if there should be one
+			*/
+			if(makes_face)
+			{
+				// Floating point conversion of the position vector
+				v3f pf(p_corrected.X, p_corrected.Y, p_corrected.Z);
+				// Center point of face (kind of)
+				v3f sp = pf - ((f32)continuous_tiles_count / 2. - 0.5) * translate_dir_f;
+				v3f scale(1,1,1);
+
+				if(translate_dir.X != 0)
+				{
+					scale.X = continuous_tiles_count;
+				}
+				if(translate_dir.Y != 0)
+				{
+					scale.Y = continuous_tiles_count;
+				}
+				if(translate_dir.Z != 0)
+				{
+					scale.Z = continuous_tiles_count;
+				}
+				
+				makeFastFace(tile, lights[0], lights[1], lights[2], lights[3],
+						sp, face_dir_corrected, scale,
+						posRelative_f, dest);
+			}
+
+			continuous_tiles_count = 0;
+			
+			makes_face = next_makes_face;
+			p_corrected = next_p_corrected;
+			face_dir_corrected = next_face_dir_corrected;
+			lights[0] = next_lights[0];
+			lights[1] = next_lights[1];
+			lights[2] = next_lights[2];
+			lights[3] = next_lights[3];
+			tile = next_tile;
+		}
+		
+		p = p_next;
+	}
+}
+
+scene::SMesh* makeMapBlockMesh(MeshMakeData *data)
+{
+	// 4-21ms for MAP_BLOCKSIZE=16
+	// 24-155ms for MAP_BLOCKSIZE=32
+	//TimeTaker timer1("makeMapBlockMesh()");
+
+	core::array<FastFace> fastfaces_new;
+
+	v3s16 blockpos_nodes = data->m_blockpos*MAP_BLOCKSIZE;
+	
+	// floating point conversion
+	v3f posRelative_f(blockpos_nodes.X, blockpos_nodes.Y, blockpos_nodes.Z);
+	
+	/*
+		Some settings
+	*/
+	//bool new_style_water = g_settings.getBool("new_style_water");
+	//bool new_style_leaves = g_settings.getBool("new_style_leaves");
+	bool smooth_lighting = g_settings.getBool("smooth_lighting");
+	
+	/*
+		We are including the faces of the trailing edges of the block.
+		This means that when something changes, the caller must
+		also update the meshes of the blocks at the leading edges.
+
+		NOTE: This is the slowest part of this method.
+	*/
+	
+	{
+		// 4-23ms for MAP_BLOCKSIZE=16
+		//TimeTaker timer2("updateMesh() collect");
+
+		/*
+			Go through every y,z and get top(y+) faces in rows of x+
+		*/
+		for(s16 y=0; y<MAP_BLOCKSIZE; y++){
+			for(s16 z=0; z<MAP_BLOCKSIZE; z++){
+				updateFastFaceRow(data->m_daynight_ratio, posRelative_f,
+						v3s16(0,y,z), MAP_BLOCKSIZE,
+						v3s16(1,0,0), //dir
+						v3f  (1,0,0),
+						v3s16(0,1,0), //face dir
+						v3f  (0,1,0),
+						fastfaces_new,
+						data->m_temp_mods,
+						data->m_vmanip,
+						blockpos_nodes,
+						smooth_lighting);
+			}
+		}
+		/*
+			Go through every x,y and get right(x+) faces in rows of z+
+		*/
+		for(s16 x=0; x<MAP_BLOCKSIZE; x++){
+			for(s16 y=0; y<MAP_BLOCKSIZE; y++){
+				updateFastFaceRow(data->m_daynight_ratio, posRelative_f,
+						v3s16(x,y,0), MAP_BLOCKSIZE,
+						v3s16(0,0,1),
+						v3f  (0,0,1),
+						v3s16(1,0,0),
+						v3f  (1,0,0),
+						fastfaces_new,
+						data->m_temp_mods,
+						data->m_vmanip,
+						blockpos_nodes,
+						smooth_lighting);
+			}
+		}
+		/*
+			Go through every y,z and get back(z+) faces in rows of x+
+		*/
+		for(s16 z=0; z<MAP_BLOCKSIZE; z++){
+			for(s16 y=0; y<MAP_BLOCKSIZE; y++){
+				updateFastFaceRow(data->m_daynight_ratio, posRelative_f,
+						v3s16(0,y,z), MAP_BLOCKSIZE,
+						v3s16(1,0,0),
+						v3f  (1,0,0),
+						v3s16(0,0,1),
+						v3f  (0,0,1),
+						fastfaces_new,
+						data->m_temp_mods,
+						data->m_vmanip,
+						blockpos_nodes,
+						smooth_lighting);
+			}
+		}
+	}
+
+	// End of slow part
+
+	/*
+		Convert FastFaces to SMesh
+	*/
+
+	MeshCollector collector;
+
+	if(fastfaces_new.size() > 0)
+	{
+		// avg 0ms (100ms spikes when loading textures the first time)
+		//TimeTaker timer2("updateMesh() mesh building");
+
+		video::SMaterial material;
+		material.setFlag(video::EMF_LIGHTING, false);
+		material.setFlag(video::EMF_BILINEAR_FILTER, false);
+		material.setFlag(video::EMF_FOG_ENABLE, true);
+		//material.setFlag(video::EMF_ANTI_ALIASING, video::EAAM_OFF);
+		//material.setFlag(video::EMF_ANTI_ALIASING, video::EAAM_SIMPLE);
+
+		for(u32 i=0; i<fastfaces_new.size(); i++)
+		{
+			FastFace &f = fastfaces_new[i];
+
+			const u16 indices[] = {0,1,2,2,3,0};
+			const u16 indices_alternate[] = {0,1,3,2,3,1};
+			
+			video::ITexture *texture = f.tile.texture.atlas;
+			if(texture == NULL)
+				continue;
+
+			material.setTexture(0, texture);
+			
+			f.tile.applyMaterialOptions(material);
+
+			const u16 *indices_p = indices;
+			
+			/*
+				Revert triangles for nicer looking gradient if vertices
+				1 and 3 have same color or 0 and 2 have different color.
+			*/
+			if(f.vertices[0].Color != f.vertices[2].Color
+					|| f.vertices[1].Color == f.vertices[3].Color)
+				indices_p = indices_alternate;
+			
+			collector.append(material, f.vertices, 4, indices_p, 6);
+		}
+	}
+
+	/*
+		Add special graphics:
+		- torches
+		- flowing water
+		- fences
+		- whatever
+	*/
+
+	mapblock_mesh_generate_special(data, collector);
+	
+	/*
+		Add stuff from collector to mesh
+	*/
+	
+	scene::SMesh *mesh_new = NULL;
+	mesh_new = new scene::SMesh();
+	
+	collector.fillMesh(mesh_new);
+
+	/*
+		Do some stuff to the mesh
+	*/
+
+	mesh_new->recalculateBoundingBox();
+
+	/*
+		Delete new mesh if it is empty
+	*/
+
+	if(mesh_new->getMeshBufferCount() == 0)
+	{
+		mesh_new->drop();
+		mesh_new = NULL;
+	}
+
+	if(mesh_new)
+	{
+#if 0
+		// Usually 1-700 faces and 1-7 materials
+		std::cout<<"Updated MapBlock has "<<fastfaces_new.size()<<" faces "
+				<<"and uses "<<mesh_new->getMeshBufferCount()
+				<<" materials (meshbuffers)"<<std::endl;
+#endif
+
+		// Use VBO for mesh (this just would set this for ever buffer)
+		// This will lead to infinite memory usage because or irrlicht.
+		//mesh_new->setHardwareMappingHint(scene::EHM_STATIC);
+
+		/*
+			NOTE: If that is enabled, some kind of a queue to the main
+			thread should be made which would call irrlicht to delete
+			the hardware buffer and then delete the mesh
+		*/
+	}
+
+	return mesh_new;
+	
+	//std::cout<<"added "<<fastfaces.getSize()<<" faces."<<std::endl;
+}
+