No OneTemporary
Actions

Authored By

Unknown

Size

177 KB

Referenced Files

None

Subscribers

None

View Options

	diff --git a/util/sdl/bitmap.cpp b/util/sdl/bitmap.cpp
	index 5d073f04..9eff2f47 100644
	--- a/util/sdl/bitmap.cpp
	+++ b/util/sdl/bitmap.cpp
	@@ -1,1896 +1,1925 @@
	#include "../bitmap.h"
	#include "../lit_bitmap.h"
	#include "../trans-bitmap.h"
	#include "../funcs.h"
	#include "../../util/debug.h"
	#include "../system.h"
	#include "sprig/sprig.h"
	+#include "stretch/SDL_stretch.h"
	#include <SDL.h>
	#include "image/SDL_image.h"
	#include <math.h>
	#include "exceptions/exception.h"
	#include <string>
	#include <sstream>

	class BitmapException: public Exception::Base {
	public:
	BitmapException(const std::string & file, int line, const std::string & reason):
	Base(file, line),
	reason(reason){
	}

	BitmapException(const BitmapException & copy):
	Base(copy),
	reason(copy.reason){
	}

	virtual void throwSelf() const {
	throw *this;
	}

	virtual ~BitmapException() throw () {
	}

	protected:
	virtual const std::string getReason() const {
	return reason;
	}

	std::string reason;
	};

	static const int WINDOWED = 0;
	static const int FULLSCREEN = 1;
	/* bits per pixel */
	static int SCREEN_DEPTH = 16;
	static SDL_Surface * screen;

	static SDL_PixelFormat format565;

	typedef unsigned int (*blender)(unsigned int color1, unsigned int color2, unsigned int alpha);

	/* taken from allegro 4.2: src/colblend.c, _blender_trans16 */
	/* this function performs a psuedo-SIMD operation on the pixel
	* components in RGB 5-6-5 format. To get this to work for some
	* other format probably all that needs to happen is to change
	* the 0x7E0F81F constant to something else. 5-5-5:
	* binary: 0011 1110 000 0111 1100 0001 1111
	* hex: 0x174076037
	*/
	static inline unsigned int transBlender(unsigned int x, unsigned int y, unsigned int n){
	unsigned long result;

	if (n)
	n = (n + 1) / 8;

	/* hex: 0x7E0F81F
	* binary: 0111 1110 0000 1111 1000 0001 1111
	*/
	x = ((x & 0xFFFF) \| (x << 16)) & 0x7E0F81F;
	y = ((y & 0xFFFF) \| (y << 16)) & 0x7E0F81F;

	result = ((x - y) * n / 32 + y) & 0x7E0F81F;

	return ((result & 0xFFFF) \| (result >> 16));
	}

	static inline unsigned int multiplyBlender(unsigned int x, unsigned int y, unsigned int n){
	Uint8 redX = 0;
	Uint8 greenX = 0;
	Uint8 blueX = 0;
	SDL_GetRGB(x, &format565, &redX, &greenX, &blueX);
	Uint8 redY = 0;
	Uint8 greenY = 0;
	Uint8 blueY = 0;
	SDL_GetRGB(y, &format565, &redY, &greenY, &blueY);

	int r = redX * redY / 256;
	int g = greenX * greenY / 256;
	int b = blueX * blueY / 256;
	return transBlender(Bitmap::makeColor(r, g, b), y, n);
	}

	static inline unsigned int addBlender(unsigned int x, unsigned int y, unsigned int n){
	Uint8 redX = 0;
	Uint8 greenX = 0;
	Uint8 blueX = 0;
	SDL_GetRGB(x, &format565, &redX, &greenX, &blueX);
	Uint8 redY = 0;
	Uint8 greenY = 0;
	Uint8 blueY = 0;
	SDL_GetRGB(y, &format565, &redY, &greenY, &blueY);

	int r = redY + redX * n / 256;
	int g = greenY + greenX * n / 256;
	int b = blueY + blueX * n / 256;

	r = Util::min(r, 255);
	g = Util::min(g, 255);
	b = Util::min(b, 255);

	return Bitmap::makeColor(r, g, b);
	}

	static inline int iabs(int x){
	return x < 0 ? -x : x;
	}

	static inline unsigned int differenceBlender(unsigned int x, unsigned int y, unsigned int n){
	Uint8 redX = 0;
	Uint8 greenX = 0;
	Uint8 blueX = 0;
	SDL_GetRGB(x, &format565, &redX, &greenX, &blueX);
	Uint8 redY = 0;
	Uint8 greenY = 0;
	Uint8 blueY = 0;
	SDL_GetRGB(y, &format565, &redY, &greenY, &blueY);

	// int r = iabs(redY - redX);
	// int g = iabs(greenY - greenX);
	// int b = iabs(blueY - blueX);
	int r = redY - redX;
	int g = greenY - greenX;
	int b = blueY - blueX;
	if (r < 0){
	r = 0;
	}
	if (g < 0){
	g = 0;
	}
	if (b < 0){
	b = 0;
	}
	return transBlender(Bitmap::makeColor(r, g, b), y, n);
	}

	static inline unsigned int burnBlender(unsigned int x, unsigned int y, unsigned int n){
	Uint8 redX = 0;
	Uint8 greenX = 0;
	Uint8 blueX = 0;
	SDL_GetRGB(x, &format565, &redX, &greenX, &blueX);
	Uint8 redY = 0;
	Uint8 greenY = 0;
	Uint8 blueY = 0;
	SDL_GetRGB(y, &format565, &redY, &greenY, &blueY);

	int r = redX - redY;
	int g = greenX - greenY;
	int b = blueX - blueY;
	if (r < 0){
	r = 0;
	}
	if (g < 0){
	g = 0;
	}
	if (b < 0){
	g = 0;
	}
	return transBlender(Bitmap::makeColor(r, g, b), y, n);
	}

	static inline unsigned int noBlender(unsigned int a, unsigned int b, unsigned int c){
	return a;
	}

	struct BlendingData{
	BlendingData():
	red(0), green(0), blue(0), alpha(0), currentBlender(noBlender){}

	int red, green, blue, alpha;
	blender currentBlender;
	};

	static BlendingData globalBlend;
	static int drawingMode = Bitmap::MODE_SOLID;

	static int drawingAlpha(){
	switch (::drawingMode){
	case Bitmap::MODE_SOLID : return 255;
	case Bitmap::MODE_TRANS : return globalBlend.alpha;
	default : return 255;
	}
	}

	static void paintown_applyTrans16(SDL_Surface * dst, const int color);
	static void paintown_replace16(SDL_Surface * dst, const int original, const int replace);
	static void paintown_draw_sprite_ex16(SDL_Surface * dst, SDL_Surface * src, int dx, int dy, int mode, int flip, Bitmap::Filter * filter);
	static void paintown_draw_sprite_filter_ex16(SDL_Surface * dst, SDL_Surface * src, int x, int y, Bitmap::Filter * filter);
	static void paintown_light16(SDL_Surface * dst, const int x, const int y, int width, int height, const int start_y, const int focus_alpha, const int edge_alpha, const int focus_color, const int edge_color);

	int Bitmap::MaskColor(){
	static int mask = makeColor(255, 0, 255);
	return mask;
	}

	static SDL_Surface * optimizedSurface(SDL_Surface * in){
	/* SDL_DisplayFormat will return 0 if a graphics context is not set,
	* like if a test is running instead of the real game.
	*/
	// SDL_Surface * out = SDL_DisplayFormat(in);
	SDL_Surface * out = SDL_ConvertSurface(in, &format565, SDL_SWSURFACE);
	if (out == NULL){
	// out = SDL_CreateRGBSurface(SDL_SWSURFACE, in->w, in->h, in->format->BitsPerPixel, 0, 0, 0, 0);
	out = SDL_CreateRGBSurface(SDL_SWSURFACE, in->w, in->h, SCREEN_DEPTH, format565.Rmask, format565.Gmask, format565.Bmask, format565.Amask);
	if (out == NULL){
	std::ostringstream out;
	out << "Could not create RGB surface of size " << in->w << ", " << in->h << ". Memory usage: " << System::memoryUsage();
	throw BitmapException(__FILE__, __LINE__, out.str());
	}
	SDL_Rect source;
	SDL_Rect destination;
	source.w = in->w;
	source.h = in->h;
	source.x = 0;
	source.y = 0;

	destination.w = in->w;
	destination.h = in->h;
	destination.x = 0;
	destination.y = 0;

	SDL_BlitSurface(in, &source, out, &destination);
	}
	return out;
	}

	Bitmap * Bitmap::Screen = NULL;
	static Bitmap * Scaler = NULL;
	static Bitmap * Buffer = NULL;

	void BitmapData::setSurface(SDL_Surface * surface){
	this->surface = surface;
	clip_left = 0;
	clip_top = 0;
	if (surface){
	clip_right = surface->w;
	clip_bottom = surface->h;
	} else {
	clip_right = 0;
	clip_bottom = 0;
	}
	}

	Bitmap::Bitmap():
	own(NULL),
	mustResize(false),
	bit8MaskColor(0){
	/* TODO */
	}

	Bitmap::Bitmap(const char * data, int length):
	own(NULL),
	mustResize(false),
	bit8MaskColor(0){
	SDL_RWops * ops = SDL_RWFromConstMem(data, length);
	SDL_Surface * loaded = IMG_Load_RW(ops, 1);
	if (loaded){
	getData().setSurface(optimizedSurface(loaded));
	SDL_FreeSurface(loaded);
	} else {
	std::ostringstream out;
	out << "Could not load surface from memory " << (void*) data << " length " << length;
	throw BitmapException(__FILE__, __LINE__, out.str());
	}

	own = new int;
	*own = 1;
	}

	Bitmap::Bitmap(SDL_Surface * who, bool deep_copy):
	own(NULL),
	mustResize(false),
	bit8MaskColor(0){
	if (deep_copy){
	SDL_Surface * surface = SDL_CreateRGBSurface(SDL_SWSURFACE, who->w, who->h, SCREEN_DEPTH, format565.Rmask, format565.Gmask, format565.Bmask, format565.Amask);
	SDL_Rect source;
	SDL_Rect destination;
	source.w = surface->w;
	source.h = surface->h;
	source.x = 0;
	source.y = 0;

	destination.w = surface->w;
	destination.h = surface->h;
	destination.x = 0;
	destination.y = 0;

	SDL_BlitSurface(who, &source, surface, &destination);
	getData().setSurface(surface);
	own = new int;
	*own = 1;
	} else {
	getData().setSurface(who);
	}
	}

	Bitmap::Bitmap(int w, int h):
	mustResize(false),
	bit8MaskColor(0){
	SDL_Surface * surface = SDL_CreateRGBSurface(SDL_SWSURFACE, w, h, SCREEN_DEPTH, format565.Rmask, format565.Gmask, format565.Bmask, format565.Amask);
	if (surface == NULL){
	std::ostringstream out;
	out << "Could not create surface with dimensions " << w << ", " << h;
	throw BitmapException(__FILE__, __LINE__, out.str());
	}
	getData().setSurface(surface);
	own = new int;
	*own = 1;
	}

	Bitmap::Bitmap( const char * load_file ):
	own(NULL),
	mustResize(false),
	bit8MaskColor(0){
	internalLoadFile(load_file);
	}

	Bitmap::Bitmap( const std::string & load_file ):
	own(NULL),
	mustResize(false){
	internalLoadFile(load_file.c_str());
	}

	Bitmap::Bitmap( const char * load_file, int sx, int sy ):
	own(NULL),
	mustResize(false),
	bit8MaskColor(0){
	Bitmap temp(load_file);
	SDL_Surface * surface = SDL_CreateRGBSurface(SDL_SWSURFACE, sx, sy, SCREEN_DEPTH, format565.Rmask, format565.Gmask, format565.Bmask, format565.Amask);
	getData().setSurface(surface);
	own = new int;
	*own = 1;

	temp.Stretch(*this);
	}

	/* unused */
	Bitmap::Bitmap( const char * load_file, int sx, int sy, double accuracy ):
	own(NULL),
	mustResize(false),
	bit8MaskColor(0){
	throw BitmapException(__FILE__, __LINE__, "Unimplemented constructor");
	}

	Bitmap::Bitmap( const Bitmap & copy, bool deep_copy):
	own(NULL),
	mustResize(false),
	bit8MaskColor(copy.bit8MaskColor){
	if (deep_copy){
	SDL_Surface * who = copy.getData().getSurface();
	SDL_Surface * surface = SDL_CreateRGBSurface(SDL_SWSURFACE, who->w, who->h, SCREEN_DEPTH, format565.Rmask, format565.Gmask, format565.Bmask, format565.Amask);
	SDL_Rect source;
	SDL_Rect destination;
	source.w = surface->w;
	source.h = surface->h;
	source.x = 0;
	source.y = 0;

	destination.w = surface->w;
	destination.h = surface->h;
	destination.x = 0;
	destination.y = 0;

	SDL_BlitSurface(who, &source, surface, &destination);
	getData().setSurface(surface);
	own = new int;
	*own = 1;
	} else {
	getData().setSurface(copy.getData().getSurface());
	own = copy.own;
	if (own){
	*own += 1;
	}
	}
	}

	Bitmap::Bitmap( const Bitmap & copy, int sx, int sy ):
	own(NULL),
	mustResize(false),
	bit8MaskColor(copy.bit8MaskColor){
	/* TODO */
	}

	Bitmap::Bitmap( const Bitmap & copy, int sx, int sy, double accuracy ):
	own(NULL),
	mustResize(false),
	bit8MaskColor(copy.bit8MaskColor){
	/* TODO */
	}

	static inline Uint8* computeOffset(SDL_Surface * surface, int x, int y){
	int bpp = surface->format->BytesPerPixel;
	return ((Uint8)surface->pixels) + y surface->pitch + x * bpp;
	}

	Bitmap::Bitmap( const Bitmap & copy, int x, int y, int width, int height ):
	own(NULL),
	mustResize(false),
	bit8MaskColor(copy.bit8MaskColor){
	path = copy.getPath();
	SDL_Surface * his = copy.getData().getSurface();
	if (x < 0)
	x = 0;
	if (y < 0)
	y = 0;
	if (width + x > his->w )
	width = his->w - x;
	if (height + y > his->h)
	height = his->h - y;

	SDL_Surface * sub = SDL_CreateRGBSurfaceFrom(computeOffset(his, x, y), width, height, SCREEN_DEPTH, his->pitch, format565.Rmask, format565.Gmask, format565.Bmask, format565.Amask);
	getData().setSurface(sub);

	own = new int;
	*own = 1;
	}

	void Bitmap::internalLoadFile(const char * path){
	this->path = path;
	SDL_Surface * loaded = IMG_Load(path);
	if (loaded){
	getData().setSurface(optimizedSurface(loaded));
	SDL_FreeSurface(loaded);
	} else {
	std::ostringstream out;
	out << "Could not load file '" << path << "'";
	throw BitmapException(__FILE__, __LINE__, out.str());
	}
	own = new int;
	*own = 1;
	}

	int Bitmap::getWidth() const {
	if (getData().getSurface() != NULL){
	return getData().getSurface()->w;
	}

	return 0;
	}

	int Bitmap::getHeight() const {
	if (getData().getSurface() != NULL){
	return getData().getSurface()->h;
	}

	return 0;
	}

	int Bitmap::getRed(int c){
	Uint8 red = 0;
	Uint8 green = 0;
	Uint8 blue = 0;
	SDL_GetRGB(c, &format565, &red, &green, &blue);
	return red;
	}

	int Bitmap::getBlue(int c){
	Uint8 red = 0;
	Uint8 green = 0;
	Uint8 blue = 0;
	SDL_GetRGB(c, &format565, &red, &green, &blue);
	return blue;
	}

	int Bitmap::getGreen(int c){
	Uint8 red = 0;
	Uint8 green = 0;
	Uint8 blue = 0;
	SDL_GetRGB(c, &format565, &red, &green, &blue);
	return green;
	}

	int Bitmap::makeColor(int red, int blue, int green){
	return SDL_MapRGB(&format565, red, blue, green);
	}

	void Bitmap::initializeExtraStuff(){
	/* this is as good a place as any to initialize our format */
	format565.palette = 0;
	format565.BitsPerPixel = 16;
	format565.BytesPerPixel = 2;
	format565.Rloss = 3;
	format565.Gloss = 2;
	format565.Bloss = 3;
	format565.Aloss = 0;
	format565.Rshift = 11;
	format565.Gshift = 5;
	format565.Bshift = 0;
	format565.Ashift = 0;
	format565.Rmask = 63488;
	format565.Gmask = 2016;
	format565.Bmask = 31;
	format565.Amask = 0;
	#ifndef PS3
	format565.colorkey = 0;
	format565.alpha = 255;
	#endif
	}

	int Bitmap::setGraphicsMode(int mode, int width, int height){
	initializeExtraStuff();

	switch (mode){
	case WINDOWED : {
	// screen = SDL_SetVideoMode(width, height, SCREEN_DEPTH, SDL_HWSURFACE \| SDL_DOUBLEBUF \| SDL_RESIZABLE);
	screen = SDL_SetVideoMode(width, height, SCREEN_DEPTH, SDL_SWSURFACE \| SDL_RESIZABLE);
	SDL_ShowCursor(0);
	// screen = SDL_SetVideoMode(width, height, SCREEN_DEPTH, SDL_SWSURFACE \| SDL_DOUBLEBUF);
	if (!screen){
	return 1;
	}
	break;
	}
	case FULLSCREEN : {
	screen = SDL_SetVideoMode(width, height, SCREEN_DEPTH, SDL_HWSURFACE \| SDL_DOUBLEBUF \| SDL_FULLSCREEN);
	SDL_ShowCursor(0);
	// screen = SDL_SetVideoMode(width, height, SCREEN_DEPTH, SDL_SWSURFACE \| SDL_DOUBLEBUF);
	if (!screen){
	return 1;
	}
	break;
	}
	}

	Global::debug(1) << "SDL Screen format palette: " << screen->format->palette <<
	" bits per pixel: " << (int) screen->format->BitsPerPixel <<
	" bytes per pixel: " << (int) screen->format->BytesPerPixel <<
	" rloss: " << (int) screen->format->Rloss <<
	" gloss: " << (int) screen->format->Gloss <<
	" bloss: " << (int) screen->format->Bloss <<
	" aloss: " << (int) screen->format->Aloss <<
	" rshift: " << (int) screen->format->Rshift <<
	" gshift: " << (int) screen->format->Gshift <<
	" bshift: " << (int) screen->format->Bshift <<
	" ashift: " << (int) screen->format->Ashift <<
	" rmask: " << (int) screen->format->Rmask <<
	" gmask: " << (int) screen->format->Gmask <<
	" bmask: " << (int) screen->format->Bmask <<
	" amask: " << (int) screen->format->Amask <<
	#ifndef PS3
	" colorkey: " << (int) screen->format->colorkey <<
	" alpha: " << (int) screen->format->alpha << std::endl;
	#else
	std::endl;
	#endif


	if (SCALE_X == 0){
	SCALE_X = width;
	}

	SCALE_X = width;

	if (SCALE_Y == 0){
	SCALE_Y = height;
	}

	SCALE_Y = height;

	/* does this need to be here? I think configuration will set SCALE_ */
	SCALE_X = 640;
	SCALE_Y = 480;

	if ( Screen != NULL ){
	delete Screen;
	Screen = NULL;
	}

	if ( Scaler != NULL ){
	delete Scaler;
	Scaler = NULL;
	}

	if ( Buffer != NULL ){
	delete Buffer;
	Buffer = NULL;
	}

	if (width != 0 && height != 0){
	Screen = new Bitmap(screen);
	Scaler = new Bitmap(width, height);
	/*
	if ( width != 0 && height != 0 && (width != SCALE_X \|\| height != SCALE_Y) ){
	Scaler = new Bitmap(width, height);
	Buffer = new Bitmap(SCALE_X, SCALE_Y);
	}
	*/
	}

	for (std::vector<Bitmap*>::iterator it = needResize.begin(); it != needResize.end(); it++){
	Bitmap * who = *it;
	who->resize(width, height);
	}

	return 0;
	}

	void Bitmap::shutdown(){
	delete Screen;
	Screen = NULL;
	delete Scaler;
	Scaler = NULL;
	delete Buffer;
	Buffer = NULL;
	}

	void Bitmap::addBlender( int r, int g, int b, int a ){
	globalBlend.red = r;
	globalBlend.green = g;
	globalBlend.blue = b;
	globalBlend.alpha = a;
	globalBlend.currentBlender = ::addBlender;
	}

	void Bitmap::multiplyBlender( int r, int g, int b, int a ){
	globalBlend.red = r;
	globalBlend.green = g;
	globalBlend.blue = b;
	globalBlend.alpha = a;
	globalBlend.currentBlender = ::multiplyBlender;
	}

	void Bitmap::differenceBlender( int r, int g, int b, int a ){
	globalBlend.red = r;
	globalBlend.green = g;
	globalBlend.blue = b;
	globalBlend.alpha = a;
	globalBlend.currentBlender = ::differenceBlender;
	}

	void Bitmap::burnBlender(int r, int g, int b, int a){
	globalBlend.red = r;
	globalBlend.green = g;
	globalBlend.blue = b;
	globalBlend.alpha = a;
	globalBlend.currentBlender = ::burnBlender;
	}

	Bitmap & Bitmap::operator=(const Bitmap & copy){
	releaseInternalBitmap();
	path = copy.getPath();
	getData().setSurface(copy.getData().getSurface());
	// own = false;
	own = copy.own;
	if (own)
	*own += 1;
	return *this;
	}

	int Bitmap::setGfxModeText(){
	/* TODO */
	return 0;
	}

	int Bitmap::setGfxModeFullscreen(int x, int y){
	return setGraphicsMode(FULLSCREEN, x, y);
	}

	int Bitmap::setGfxModeWindowed( int x, int y ){
	return setGraphicsMode(WINDOWED, x, y);
	}

	void Bitmap::drawingMode(int type){
	::drawingMode = type;
	}

	void Bitmap::transBlender( int r, int g, int b, int a ){
	globalBlend.red = r;
	globalBlend.green = g;
	globalBlend.blue = b;
	globalBlend.alpha = a;
	globalBlend.currentBlender = ::transBlender;
	}

	void Bitmap::setClipRect( int x1, int y1, int x2, int y2 ) const {
	SDL_Rect area;
	area.x = x1;
	area.y = y1;
	area.w = x2 - x1;
	area.h = y2 - y1;
	SDL_SetClipRect(getData().getSurface(), &area);
	SDL_GetClipRect(getData().getSurface(), &area);
	getData().setClip(area.x, area.y, area.x + area.w, area.y + area.h);
	}

	void Bitmap::getClipRect(int & x1, int & y1, int & x2, int & y2) const {
	x1 = getData().clip_left;
	y1 = getData().clip_top;
	x2 = getData().clip_right;
	y2 = getData().clip_bottom;
	}

	void Bitmap::destroyPrivateData(){
	SDL_FreeSurface(getData().getSurface());
	}

	static void doPutPixel(SDL_Surface * surface, int x, int y, int pixel, bool translucent){

	if (SDL_MUSTLOCK(surface)){
	SDL_LockSurface(surface);
	}

	int bpp = surface->format->BytesPerPixel;
	/* Here p is the address to the pixel we want to set */
	Uint8 *p = computeOffset(surface, x, y);

	switch(bpp) {
	case 1:
	*p = pixel;
	break;

	case 2:
	if (translucent){
	(Uint16 )p = globalBlend.currentBlender(pixel, (Uint16)p, globalBlend.alpha);
	} else {
	(Uint16 )p = pixel;
	}
	break;
	case 3:
	if(SDL_BYTEORDER == SDL_BIG_ENDIAN) {
	p[0] = (pixel >> 16) & 0xff;
	p[1] = (pixel >> 8) & 0xff;
	p[2] = pixel & 0xff;
	} else {
	p[0] = pixel & 0xff;
	p[1] = (pixel >> 8) & 0xff;
	p[2] = (pixel >> 16) & 0xff;
	}
	break;
	case 4:
	(Uint32 )p = pixel;
	break;
	}

	if (SDL_MUSTLOCK(surface)){
	SDL_UnlockSurface(surface);
	}

	}

	void Bitmap::putPixel(int x, int y, int pixel) const {
	/* clip it */
	if (getData().isClipped(x, y)){
	return;
	}

	SDL_Surface * surface = getData().getSurface();
	doPutPixel(surface, x, y, pixel, false);
	}

	void Bitmap::putPixelNormal(int x, int y, int col) const {
	putPixel(x, y, col);
	}

	void TranslucentBitmap::putPixelNormal(int x, int y, int color) const {
	if (getData().isClipped(x, y)){
	return;
	}

	SDL_Surface * surface = getData().getSurface();
	doPutPixel(surface, x, y, color, true);
	}

	bool Bitmap::getError(){
	/* TODO */
	return false;
	}

	void Bitmap::rectangle( int x1, int y1, int x2, int y2, int color ) const {
	SPG_Rect(getData().getSurface(), x1, y1, x2, y2, color);
	}

	void TranslucentBitmap::rectangle( int x1, int y1, int x2, int y2, int color ) const {
	int alpha = globalBlend.alpha;
	SPG_RectBlend(getData().getSurface(), x1, y1, x2, y2, color, alpha);
	}

	void Bitmap::rectangleFill( int x1, int y1, int x2, int y2, int color ) const {
	SPG_RectFilled(getData().getSurface(), x1, y1, x2, y2, color);
	}

	void TranslucentBitmap::rectangleFill(int x1, int y1, int x2, int y2, int color) const {
	int alpha = globalBlend.alpha;
	SPG_RectFilledBlend(getData().getSurface(), x1, y1, x2, y2, color, alpha);
	}

	void Bitmap::circleFill(int x, int y, int radius, int color) const {
	switch (::drawingMode){
	case MODE_SOLID : {
	SPG_CircleFilled(getData().getSurface(), x, y, radius, color);
	break;
	}
	case MODE_TRANS : {
	int alpha = globalBlend.alpha;
	SPG_CircleFilledBlend(getData().getSurface(), x, y, radius, color, alpha);
	break;
	}
	}
	}

	void TranslucentBitmap::circleFill(int x, int y, int radius, int color) const {
	int alpha = globalBlend.alpha;
	SPG_CircleFilledBlend(getData().getSurface(), x, y, radius, color, alpha);
	}

	void Bitmap::circle(int x, int y, int radius, int color) const {
	// Uint8 red, green, blue;
	// SDL_GetRGB(color, getData().getSurface()->format, &red, &green, &blue);
	// int alpha = 255;
	switch (::drawingMode){
	case MODE_SOLID : {
	SPG_Circle(getData().getSurface(), x, y, radius, color);
	break;
	}
	case MODE_TRANS : {
	int alpha = globalBlend.alpha;
	SPG_CircleBlend(getData().getSurface(), x, y, radius, color, alpha);
	break;
	}
	}

	// circleRGBA(getData().getSurface(), x, y, radius, red, green, blue, alpha);
	}

	void Bitmap::line( const int x1, const int y1, const int x2, const int y2, const int color ) const {
	switch (::drawingMode){
	case MODE_SOLID : {
	SPG_Line(getData().getSurface(), x1, y1, x2, y2, color);
	break;
	}
	case MODE_TRANS : {
	int alpha = globalBlend.alpha;
	SPG_LineBlend(getData().getSurface(), x1, y1, x2, y2, color, alpha);
	break;
	}
	}
	}

	void Bitmap::draw(const int x, const int y, const Bitmap & where) const {
	if (getData().getSurface() != NULL){
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_NORMAL, Bitmap::SPRITE_NO_FLIP, NULL);
	/*
	SDL_SetColorKey(getData().getSurface(), SDL_SRCCOLORKEY, makeColor(255, 0, 255));
	Blit(x, y, where);
	*/
	}
	}

	void Bitmap::drawHFlip(const int x, const int y, const Bitmap & where) const {
	paintown_draw_sprite_ex16( where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_NORMAL, Bitmap::SPRITE_H_FLIP, NULL);
	}

	void Bitmap::drawHFlip(const int x, const int y, Filter * filter, const Bitmap & where) const {
	paintown_draw_sprite_ex16( where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_NORMAL, Bitmap::SPRITE_H_FLIP, filter);
	}

	void Bitmap::drawVFlip( const int x, const int y, const Bitmap & where ) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_NORMAL, Bitmap::SPRITE_V_FLIP, NULL);
	}

	void Bitmap::drawVFlip( const int x, const int y, Filter * filter, const Bitmap & where ) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_NORMAL, Bitmap::SPRITE_V_FLIP, filter);
	}

	void Bitmap::drawHVFlip( const int x, const int y, const Bitmap & where ) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_NORMAL, Bitmap::SPRITE_V_FLIP \| Bitmap::SPRITE_H_FLIP, NULL);
	}

	void Bitmap::drawHVFlip( const int x, const int y, Filter * filter, const Bitmap & where ) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_NORMAL, Bitmap::SPRITE_V_FLIP \| Bitmap::SPRITE_H_FLIP, filter);
	}

	void TranslucentBitmap::draw(const int x, const int y, const Bitmap & where) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_TRANS, Bitmap::SPRITE_NO_FLIP, NULL);
	}

	void TranslucentBitmap::draw( const int x, const int y, Filter * filter, const Bitmap & where ) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_TRANS, Bitmap::SPRITE_NO_FLIP, filter);
	}

	void TranslucentBitmap::drawHFlip( const int x, const int y, const Bitmap & where ) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_TRANS, Bitmap::SPRITE_H_FLIP, NULL);
	}

	void TranslucentBitmap::drawHFlip( const int x, const int y, Filter * filter, const Bitmap & where ) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_TRANS, Bitmap::SPRITE_H_FLIP, filter);
	}

	void TranslucentBitmap::drawVFlip( const int x, const int y, const Bitmap & where ) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_TRANS, Bitmap::SPRITE_V_FLIP, NULL);
	}

	void TranslucentBitmap::drawVFlip( const int x, const int y, Filter * filter, const Bitmap & where ) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_TRANS, Bitmap::SPRITE_V_FLIP, filter);
	}

	void TranslucentBitmap::drawHVFlip( const int x, const int y, const Bitmap & where ) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_TRANS, SPRITE_V_FLIP \| SPRITE_H_FLIP, NULL);
	}

	void TranslucentBitmap::drawHVFlip( const int x, const int y, Filter * filter,const Bitmap & where ) const {
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_TRANS, SPRITE_V_FLIP \| SPRITE_H_FLIP, filter);
	}

	void Bitmap::drawStretched( const int x, const int y, const int new_width, const int new_height, const Bitmap & who ) const {

	if (getData().getSurface() != NULL){
	SDL_SetColorKey(getData().getSurface(), SDL_SRCCOLORKEY, makeColor(255, 0, 255));

	SDL_Surface * src = getData().getSurface();
	SDL_Surface * dst = who.getData().getSurface();

	int myWidth = src->w;
	int myHeight = src->h;
	int hisWidth = new_width;
	int hisHeight = new_height;
	int useX = x;
	int useY = y;
	int myX = 0;
	int myY = 0;

	float xscale = (float) hisWidth / (float) myWidth;
	float yscale = (float) hisHeight / (float) myHeight;

	/* sprig wont do the clipping right, if you start drawing from a negative offset
	* sprig will do nothing.
	*/
	if (useX < 0){
	useX = 0;
	myX = -x / xscale;
	}

	if (useY < 0){
	useY = 0;
	myY = -y / yscale;
	}

	SPG_TransformX(src, dst, 0, xscale, yscale, myX, myY, useX, useY, SPG_TCOLORKEY);
	}
	}

	void Bitmap::Blit( const Bitmap & where ) const {
	Blit(0, 0, where);
	}

	void Bitmap::Blit( const int x, const int y, const Bitmap & where ) const {
	Blit(0, 0, x, y, where);
	}

	void Bitmap::Blit( const int mx, const int my, const int wx, const int wy, const Bitmap & where ) const {
	Blit(mx, my, getWidth(), getHeight(), wx, wy, where);
	}

	static void doBlit(SDL_Surface * mine, const int mx, const int my, const int width, const int height, const int wx, const int wy, const Bitmap & where ){
	SDL_Rect source;
	SDL_Rect destination;
	source.w = width;
	source.h = height;
	source.x = mx;
	source.y = my;

	destination.w = width;
	destination.h = height;
	destination.x = wx;
	destination.y = wy;

	SDL_BlitSurface(mine, &source, where.getData().getSurface(), &destination);
	}

	void Bitmap::Blit( const int mx, const int my, const int width, const int height, const int wx, const int wy, const Bitmap & where ) const {
	SDL_SetColorKey(getData().getSurface(), 0, MaskColor());
	doBlit(getData().getSurface(), mx, my, width, height, wx, wy, where);

	/* FIXME: this is a hack, maybe put a call here for the other bitmap to update stuff
	* like where->Blitted()
	*/
	if (&where == Screen){
	SDL_Flip(Screen->getData().getSurface());
	}
	}

	void Bitmap::BlitMasked( const int mx, const int my, const int width, const int height, const int wx, const int wy, const Bitmap & where ) const {
	SDL_SetColorKey(getData().getSurface(), SDL_SRCCOLORKEY, MaskColor());

	doBlit(getData().getSurface(),mx, my, width, height, wx, wy, where);

	/* FIXME: this is a hack, maybe put a call here for the other bitmap to update stuff
	* like where->Blitted()
	*/
	if (&where == Screen){
	SDL_Flip(Screen->getData().getSurface());
	}
	}

	void Bitmap::BlitToScreen(const int upper_left_x, const int upper_left_y) const {
	if (getWidth() != Screen->getWidth() \|\| getHeight() != Screen->getHeight()){
	/*
	this->Blit( upper_left_x, upper_left_y, *Buffer );
	Buffer->Stretch(*Scaler);
	Scaler->Blit(0, 0, 0, 0, *Screen);
	*/

	this->Stretch(*Scaler, 0, 0, getWidth(), getHeight(), upper_left_x, upper_left_y, Scaler->getWidth(), Scaler->getHeight());
	Scaler->Blit(0, 0, 0, 0, *Screen);
	} else {
	this->Blit( upper_left_x, upper_left_y, *Screen );
	}

	/*
	if ( Scaler == NULL ){
	this->Blit( upper_left_x, upper_left_y, *Screen );
	} else {
	this->Blit( upper_left_x, upper_left_y, *Buffer );
	Buffer->Stretch(*Scaler);
	Scaler->Blit(0, 0, 0, 0, *Screen);
	}
	*/

	// SDL_Flip(Screen->getData().getSurface());
	// SDL_UpdateRect(Screen->getData().getSurface(), 0, 0, Screen->getWidth(), Screen->getHeight());
	}

	void Bitmap::BlitAreaToScreen(const int upper_left_x, const int upper_left_y) const {
	if ( Scaler != NULL ){
	double mult_x = (double) Scaler->getWidth() / (double) SCALE_X;
	double mult_y = (double) Scaler->getHeight() / (double) SCALE_Y;

	int x = (int)(upper_left_x * mult_x);
	int y = (int)(upper_left_y * mult_y);
	int w = (int)(this->getWidth() * mult_x);
	int h = (int)(this->getHeight() * mult_y);

	// printf("ux %d uy %d uw %d uh %d. x %d y %d w %d h %d\n", upper_left_x, upper_left_y, getWidth(), getHeight(), x, y, w, h );

	this->Stretch( *Scaler, 0, 0, this->getWidth(), this->getHeight(), x, y, w, h );

	Bitmap tmp(*Scaler, x, y, w, h );
	tmp.Blit( x, y, *Screen );

	// Scaler->Blit( x, y, w, h, *Screen );
	} else {
	this->Blit( upper_left_x, upper_left_y, *Screen );
	}
	}

	void Bitmap::Stretch( const Bitmap & where ) const {
	if (getWidth() == where.getWidth() && getHeight() == where.getHeight()){
	Blit(where);
	} else {
	Stretch(where, 0, 0, getWidth(), getHeight(), 0, 0, where.getWidth(), where.getHeight());
	}
	}

	void Bitmap::Stretch( const Bitmap & where, const int sourceX, const int sourceY, const int sourceWidth, const int sourceHeight, const int destX, const int destY, const int destWidth, const int destHeight ) const {

	+ Bitmap subSource(*this, sourceX, sourceY, sourceWidth, sourceHeight);
	+ Bitmap subDestination(where, destX, destY, destWidth, destHeight);
	+
	+ SDL_Surface * src = subSource.getData().getSurface();
	+ SDL_Surface * dst = subDestination.getData().getSurface();
	+
	+ if (SDL_MUSTLOCK(src)){
	+ SDL_LockSurface(src);
	+ }
	+
	+ if (SDL_MUSTLOCK(dst)){
	+ SDL_LockSurface(dst);
	+ }
	+
	+ SDL_StretchSurfaceRect(src, NULL, dst, NULL);
	+
	+ if (SDL_MUSTLOCK(src)){
	+ SDL_UnlockSurface(src);
	+ }
	+
	+ if (SDL_MUSTLOCK(dst)){
	+ SDL_UnlockSurface(dst);
	+ }
	+
	+ /*
	SDL_Surface * src = getData().getSurface();
	SDL_Surface * dst = where.getData().getSurface();
	+ */

	+ /*
	float xscale = (float) destWidth / (float) sourceWidth;
	float yscale = (float) destHeight / (float) sourceHeight;
	SDL_SetColorKey(src, 0, MaskColor());
	SPG_TransformX(src, dst, 0, xscale, yscale, sourceX, sourceY, destX, destY, SPG_NONE);
	+ */

	/*
	SDL_Rect source;
	SDL_Rect destination;
	source.x = sourceX;
	source.y = sourceY;
	source.w = sourceWidth;
	source.h = sourceHeight;

	destination.x = destX;
	destination.y = destY;
	destination.w = destWidth;
	destination.h = destHeight;

	// SDL_StretchSurfaceRect(getData().getSurface(), &source, where.getData().getSurface(), &destination);

	SDL_Surface * src = getData().getSurface();
	SDL_Surface * dst = where.getData().getSurface();

	if (SDL_MUSTLOCK(src)){
	SDL_LockSurface(src);
	}

	if (SDL_MUSTLOCK(dst)){
	SDL_LockSurface(dst);
	}

	SDL_StretchSurfaceRect(src, &source, dst, &destination);

	if (SDL_MUSTLOCK(src)){
	SDL_UnlockSurface(src);
	}

	if (SDL_MUSTLOCK(dst)){
	SDL_UnlockSurface(dst);
	}
	*/
	}

	void Bitmap::save( const std::string & str ) const {
	/* TODO */
	}

	void Bitmap::triangle( int x1, int y1, int x2, int y2, int x3, int y3, int color ) const {

	switch (::drawingMode){
	case MODE_SOLID : {
	SPG_TrigonFilled(getData().getSurface(), x1, y1, x2, y2, x3, y3, color);
	break;
	}
	case MODE_TRANS : {
	int alpha = globalBlend.alpha;
	SPG_TrigonFilledBlend(getData().getSurface(), x1, y1, x2, y2, x3, y3, color, alpha);
	break;
	}
	}
	}

	void Bitmap::ellipse( int x, int y, int rx, int ry, int color ) const {
	switch (::drawingMode){
	case MODE_SOLID : {
	SPG_Ellipse(getData().getSurface(), x, y, rx, ry, color);
	break;
	}
	case MODE_TRANS : {
	int alpha = globalBlend.alpha;
	SPG_EllipseBlend(getData().getSurface(), x, y, rx, ry, color, alpha);
	break;
	}
	}
	}

	void TranslucentBitmap::ellipse( int x, int y, int rx, int ry, int color ) const {
	int alpha = globalBlend.alpha;
	SPG_EllipseBlend(getData().getSurface(), x, y, rx, ry, color, alpha);
	}

	void Bitmap::ellipseFill( int x, int y, int rx, int ry, int color ) const {
	SPG_EllipseFilled(getData().getSurface(), x, y, rx, ry, color);
	}

	void Bitmap::light(int x, int y, int width, int height, int start_y, int focus_alpha, int edge_alpha, int focus_color, int edge_color) const {
	paintown_light16(getData().getSurface(), x, y, width, height, start_y, focus_alpha, edge_alpha, focus_color, edge_color);
	}

	void Bitmap::applyTrans(const int color) const {
	paintown_applyTrans16(getData().getSurface(), color);
	}

	void Bitmap::floodfill( const int x, const int y, const int color ) const {
	SPG_FloodFill(getData().getSurface(), x, y, color);
	}

	/*
	void Bitmap::horizontalLine( const int x1, const int y, const int x2, const int color ) const {
	SPG_LineH(getData().getSurface(), x1, y, x2, color);
	}
	*/

	void Bitmap::hLine( const int x1, const int y, const int x2, const int color ) const {
	SPG_LineH(getData().getSurface(), x1, y, x2, color);
	}

	void TranslucentBitmap::hLine( const int x1, const int y, const int x2, const int color ) const {
	int alpha = globalBlend.alpha;
	SPG_LineHBlend(getData().getSurface(), x1, y, x2, color, alpha);
	}

	void Bitmap::vLine( const int y1, const int x, const int y2, const int color ) const {
	SPG_LineV(getData().getSurface(), x, y1, y2, color);
	}

	void Bitmap::polygon( const int * verts, const int nverts, const int color ) const {
	SPG_Point * points = new SPG_Point[nverts];
	for (int i = 0; i < nverts; i++){
	points[i].x = verts[i*2];
	points[i].y = verts[i*2+1];
	}
	SPG_PolygonFilled(getData().getSurface(), nverts, points, color);
	delete[] points;
	}

	void Bitmap::arc(const int x, const int y, const double ang1, const double ang2, const int radius, const int color ) const {
	SPG_Arc(getData().getSurface(), x, y, radius, ang1, ang2, color);
	}

	void Bitmap::fill(int color) const {
	SDL_Rect area;
	area.x = 0;
	area.y = 0;
	area.w = getWidth();
	area.h = getHeight();
	SDL_FillRect(getData().getSurface(), &area, color);
	}

	/*
	void TranslucentBitmap::fill(int color) const {
	rectangleFill(0, 0, getWidth(), getHeight(), color);
	}
	*/

	int Bitmap::darken( int color, double factor ){
	int r = (int)((double)getRed(color) / factor);
	int g = (int)((double)getGreen(color) / factor);
	int b = (int)((double)getBlue(color) / factor);

	return makeColor(r, g, b);
	}

	void Bitmap::drawCharacter( const int x, const int y, const int color, const int background, const Bitmap & where ) const {
	/* TODO */
	}

	void Bitmap::drawRotate( const int x, const int y, const int angle, const Bitmap & where ){
	SDL_SetColorKey(getData().getSurface(), SDL_SRCCOLORKEY, MaskColor());
	SDL_Surface * src = getData().getSurface();
	SDL_Surface * dst = where.getData().getSurface();
	SPG_TransformX(src, dst, angle, 1, 1, 0, 0, x, y, SPG_TCOLORKEY);
	}

	void Bitmap::drawPivot( const int centerX, const int centerY, const int x, const int y, const int angle, const Bitmap & where ){
	SDL_SetColorKey(getData().getSurface(), SDL_SRCCOLORKEY, MaskColor());
	SDL_Surface * src = getData().getSurface();
	SDL_Surface * dst = where.getData().getSurface();
	SPG_TransformX(src, dst, angle, 1, 1, centerX, centerY, x, y, SPG_TCOLORKEY);
	}

	void Bitmap::drawPivot( const int centerX, const int centerY, const int x, const int y, const int angle, const double scale, const Bitmap & where ){
	SDL_SetColorKey(getData().getSurface(), SDL_SRCCOLORKEY, MaskColor());
	SDL_Surface * src = getData().getSurface();
	SDL_Surface * dst = where.getData().getSurface();
	SPG_TransformX(src, dst, angle, scale, scale, centerX, centerY, x, y, SPG_TCOLORKEY);
	}

	void Bitmap::replaceColor(int original, int replaced){
	paintown_replace16(getData().getSurface(), original, replaced);
	}

	Bitmap Bitmap::memoryPCX(unsigned char * const data, const int length, const bool mask){
	SDL_RWops * ops = SDL_RWFromConstMem(data, length);
	SDL_Surface * pcx = IMG_LoadPCX_RW(ops);
	SDL_FreeRW(ops);
	if (!pcx){
	std::ostringstream out;
	out << "Could not load PCX file at " << (void*) data << " length " << length;
	throw BitmapException(__FILE__, __LINE__, out.str());
	}
	SDL_Surface * display = optimizedSurface(pcx);
	Bitmap out(display, true);

	#ifndef PS3
	if (pcx->format->BitsPerPixel == 8){
	SDL_Color color = pcx->format->palette->colors[pcx->format->colorkey];
	int bad = makeColor(color.r, color.g, color.b);
	out.set8BitMaskColor(bad);
	// int mask = MaskColor();
	// out.replaceColor(bad, mask);
	}
	#endif

	SDL_FreeSurface(pcx);
	SDL_FreeSurface(display);
	// out.floodfill(0, 0, makeColor(255, 0, 255));

	return out;
	}

	int Bitmap::getPixel( const int x, const int y ) const {
	return SPG_GetPixel(getData().getSurface(), x, y);
	}

	void Bitmap::readLine( std::vector< int > & line, int y ){
	for (int x = 0; x < getWidth(); x++){
	line.push_back(getPixel(x, y));
	}
	}

	void Bitmap::StretchBy2( const Bitmap & where ){
	/* TODO */
	}

	void Bitmap::StretchBy4( const Bitmap & where ){
	/* TODO */
	}

	void Bitmap::draw(const int x, const int y, Filter * filter, const Bitmap & where) const {
	// paintown_draw_sprite_filter_ex16(where.getData().getSurface(), getData().getSurface(), x, y, filter);
	paintown_draw_sprite_ex16(where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_NORMAL, Bitmap::SPRITE_NO_FLIP, filter);
	}

	void LitBitmap::draw( const int x, const int y, const Bitmap & where ) const {
	paintown_draw_sprite_ex16( where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_LIT, Bitmap::SPRITE_NO_FLIP, NULL);
	}

	void LitBitmap::draw( const int x, const int y, Filter * filter, const Bitmap & where ) const {
	paintown_draw_sprite_ex16( where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_LIT, Bitmap::SPRITE_NO_FLIP, filter);
	}

	void LitBitmap::drawHFlip( const int x, const int y, const Bitmap & where ) const {
	paintown_draw_sprite_ex16( where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_LIT, Bitmap::SPRITE_H_FLIP, NULL);
	}

	void LitBitmap::drawHFlip( const int x, const int y, Filter * filter, const Bitmap & where ) const {
	paintown_draw_sprite_ex16( where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_LIT, Bitmap::SPRITE_H_FLIP, filter);
	}

	void LitBitmap::drawVFlip( const int x, const int y, const Bitmap & where ) const {
	paintown_draw_sprite_ex16( where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_LIT, Bitmap::SPRITE_V_FLIP, NULL);
	}

	void LitBitmap::drawVFlip( const int x, const int y, Filter * filter, const Bitmap & where ) const {
	paintown_draw_sprite_ex16( where.getData().getSurface(), getData().getSurface(), x, y, Bitmap::SPRITE_LIT, Bitmap::SPRITE_V_FLIP, filter);
	}

	void LitBitmap::drawHVFlip( const int x, const int y, const Bitmap & where ) const {
	paintown_draw_sprite_ex16( where.getData().getSurface(), getData().getSurface(), x, y, SPRITE_LIT, SPRITE_V_FLIP \| SPRITE_H_FLIP, NULL);
	}

	void LitBitmap::drawHVFlip( const int x, const int y, Filter * filter, const Bitmap & where ) const {
	paintown_draw_sprite_ex16( where.getData().getSurface(), getData().getSurface(), x, y, SPRITE_LIT, SPRITE_V_FLIP \| SPRITE_H_FLIP, filter);
	}

	/*
	#define PAINTOWN_DLS_BLENDER BLENDER_FUNC
	#define PAINTOWN_DTS_BLENDER BLENDER_FUNC
	#define PAINTOWN_MAKE_DLS_BLENDER(a) _blender_func16
	#define PAINTOWN_MAKE_DTS_BLENDER() _blender_func16
	#define PAINTOWN_PIXEL_PTR unsigned short*
	#define PAINTOWN_OFFSET_PIXEL_PTR(p,x) ((PAINTOWN_PIXEL_PTR) (p) + (x))
	#define PAINTOWN_INC_PIXEL_PTR(p) ((p)++)
	#define PAINTOWN_INC_PIXEL_PTR_EX(p,d) ((p) += d)
	#define PAINTOWN_GET_MEMORY_PIXEL(p) (*(p))
	#define PAINTOWN_IS_SPRITE_MASK(b,c) ((unsigned long) (c) == (unsigned long) MASK_COLOR)
	#define PAINTOWN_DLSX_BLEND(b,n) ((*(b))(_blender_col_16, (n), globalBlend.alpha))
	#define PAINTOWN_GET_PIXEL(p) ((unsigned short ) (p))
	#define PAINTOWN_DTS_BLEND(b,o,n) ((*(b))((n), (o), globalBlend.alpha))
	#define PAINTOWN_PUT_PIXEL(p,c) (((unsigned short ) p) = (c))
	#define PAINTOWN_PUT_MEMORY_PIXEL(p,c) (*(p) = (c))
	#define PAINTOWN_SET_ALPHA(a) (globalBlend.alpha = (a))
	*/

	static void paintown_applyTrans16(SDL_Surface * dst, const int color){
	int y1 = 0;
	int y2 = dst->h;
	int x1 = 0;
	int x2 = dst->w - 1;

	y1 = dst->clip_rect.y;
	y2 = dst->clip_rect.y + dst->clip_rect.h;
	x1 = dst->clip_rect.x;
	x2 = dst->clip_rect.x + dst->clip_rect.w;

	int bpp = dst->format->BytesPerPixel;
	unsigned int mask = Bitmap::makeColor(255, 0, 255);
	for (int y = y1; y < y2; y++) {
	Uint8 * sourceLine = computeOffset(dst, x1, y);

	for (int x = x2 - 1; x >= x1; sourceLine += bpp, x--) {
	unsigned long sourcePixel = (Uint16) sourceLine;
	if (!(sourcePixel == mask)){
	sourcePixel = globalBlend.currentBlender(color, sourcePixel, globalBlend.alpha);
	(Uint16 )sourceLine = sourcePixel;
	}
	}
	}
	}

	static void paintown_replace16(SDL_Surface * dst, const int original, const int replace){
	int y1 = 0;
	int y2 = dst->h;
	int x1 = 0;
	int x2 = dst->w - 1;

	y1 = dst->clip_rect.y;
	y2 = dst->clip_rect.y + dst->clip_rect.h;
	x1 = dst->clip_rect.x;
	x2 = dst->clip_rect.x + dst->clip_rect.w;

	int bpp = dst->format->BytesPerPixel;

	/* Attempted manual unrolling. Gcc can optimize the naive loop below better
	* than this unrolling attempt.
	*/
	/*
	for (int y = y1; y < y2; y += 4){
	switch (y2 - y1){
	case 1 : {
	Uint8 * source1 = computeOffset(dst, x1, y);
	for (int x = x2 - 1; x >= x1; source1 += bpp, x--) {
	unsigned long sourcePixel = (Uint16) source1;
	if ((int) sourcePixel == original){
	(Uint16 )source1 = replace;
	}
	}
	break;
	}
	case 2 : {
	Uint8 * source1 = computeOffset(dst, x1, y);
	Uint8 * source2 = computeOffset(dst, x1, y + 1);
	for (int x = x2 - 1; x >= x1; source1 += bpp, source2 += bpp, x--) {
	unsigned long sourcePixel = (Uint16) source1;
	unsigned long sourcePixel2 = (Uint16) source2;
	if ((int) sourcePixel == original){
	(Uint16 )source1 = replace;
	}
	if ((int) sourcePixel2 == original){
	(Uint16 )source2 = replace;
	}
	}
	break;
	}
	case 3 : {
	Uint8 * source1 = computeOffset(dst, x1, y);
	Uint8 * source2 = computeOffset(dst, x1, y + 1);
	Uint8 * source3 = computeOffset(dst, x1, y + 2);
	for (int x = x2 - 1; x >= x1; source1 += bpp, source2 += bpp, source3 += bpp, x--) {
	unsigned long sourcePixel = (Uint16) source1;
	unsigned long sourcePixel2 = (Uint16) source2;
	unsigned long sourcePixel3 = (Uint16) source3;
	if ((int) sourcePixel == original){
	(Uint16 )source1 = replace;
	}
	if ((int) sourcePixel2 == original){
	(Uint16 )source2 = replace;
	}
	if ((int) sourcePixel3 == original){
	(Uint16 )source3 = replace;
	}
	}
	break;
	}
	default:
	case 4 : {
	Uint8 * source1 = computeOffset(dst, x1, y);
	Uint8 * source2 = computeOffset(dst, x1, y + 1);
	Uint8 * source3 = computeOffset(dst, x1, y + 2);
	Uint8 * source4 = computeOffset(dst, x1, y + 3);
	for (int x = x2 - 1; x >= x1; source1 += bpp, source2 += bpp, source3 += bpp, source4 += bpp, x--) {
	unsigned long sourcePixel = (Uint16) source1;
	unsigned long sourcePixel2 = (Uint16) source2;
	unsigned long sourcePixel3 = (Uint16) source3;
	unsigned long sourcePixel4 = (Uint16) source4;
	if ((int) sourcePixel == original){
	(Uint16 )source1 = replace;
	}
	if ((int) sourcePixel2 == original){
	(Uint16 )source2 = replace;
	}
	if ((int) sourcePixel3 == original){
	(Uint16 )source3 = replace;
	}
	if ((int) sourcePixel4 == original){
	(Uint16 )source4 = replace;
	}
	}
	break;
	}
	}
	}
	*/
	if (SDL_MUSTLOCK(dst)){
	SDL_LockSurface(dst);
	}

	for (int y = y1; y < y2; y++) {
	Uint8 * sourceLine = computeOffset(dst, x1, y);

	for (int x = x2 - 1; x >= x1; sourceLine += bpp, x--) {
	unsigned long sourcePixel = (Uint16) sourceLine;
	if ((int) sourcePixel == original){
	(Uint16 )sourceLine = replace;
	}
	}
	}

	if (SDL_MUSTLOCK(dst)){
	SDL_UnlockSurface(dst);
	}
	}

	static void paintown_draw_sprite_filter_ex16(SDL_Surface * dst, SDL_Surface * src, int dx, int dy, Bitmap::Filter * filter){
	int x, y, w, h;
	int x_dir = 1, y_dir = 1;
	int dxbeg, dybeg;
	int sxbeg, sybeg;

	if (SDL_MUSTLOCK(dst)){
	SDL_LockSurface(dst);
	}

	if (true /* dst->clip*/ ) {
	int tmp;

	tmp = dst->clip_rect.x - dx;
	sxbeg = ((tmp < 0) ? 0 : tmp);
	dxbeg = sxbeg + dx;

	tmp = dst->clip_rect.x + dst->clip_rect.w - dx;
	w = ((tmp > src->w) ? src->w : tmp) - sxbeg;
	if (w <= 0)
	return;

	tmp = dst->clip_rect.y - dy;
	sybeg = ((tmp < 0) ? 0 : tmp);
	dybeg = sybeg + dy;

	tmp = dst->clip_rect.y + dst->clip_rect.h - dy;
	h = ((tmp > src->h) ? src->h : tmp) - sybeg;
	if (h <= 0)
	return;
	}

	unsigned int mask = Bitmap::makeColor(255, 0, 255);
	int bpp = src->format->BytesPerPixel;
	for (y = 0; y < h; y++) {
	Uint8 * sourceLine = computeOffset(src, sxbeg, sybeg + y);
	Uint8 * destLine = computeOffset(dst, dxbeg, dybeg + y * y_dir);

	for (x = w - 1; x >= 0; sourceLine += bpp, destLine += bpp * x_dir, x--) {
	unsigned long sourcePixel = (Uint16) sourceLine;
	if (!(sourcePixel == mask)){
	(Uint16 )destLine = filter->filter(sourcePixel);
	} else {
	(Uint16 )destLine = mask;
	}
	}
	}

	if (SDL_MUSTLOCK(dst)){
	SDL_UnlockSurface(dst);
	}
	}

	static void paintown_draw_sprite_ex16(SDL_Surface * dst, SDL_Surface * src, int dx, int dy, int mode, int flip, Bitmap::Filter * filter){
	int x, y, w, h;
	int x_dir = 1, y_dir = 1;
	int dxbeg, dybeg;
	int sxbeg, sybeg;
	/*
	PAINTOWN_DLS_BLENDER lit_blender;
	PAINTOWN_DTS_BLENDER trans_blender;
	*/

	/*
	ASSERT(dst);
	ASSERT(src);
	*/

	if (SDL_MUSTLOCK(src)){
	SDL_LockSurface(src);
	}

	if (SDL_MUSTLOCK(dst)){
	SDL_LockSurface(dst);
	}

	if ( flip & Bitmap::SPRITE_V_FLIP ){
	y_dir = -1;
	}
	if ( flip & Bitmap::SPRITE_H_FLIP ){
	x_dir = -1;
	}

	if (true /* dst->clip*/ ) {
	int tmp;

	tmp = dst->clip_rect.x - dx;
	sxbeg = ((tmp < 0) ? 0 : tmp);
	dxbeg = sxbeg + dx;

	tmp = dst->clip_rect.x + dst->clip_rect.w - dx;
	w = ((tmp > src->w) ? src->w : tmp) - sxbeg;
	if (w <= 0)
	return;

	if ( flip & Bitmap::SPRITE_H_FLIP ){
	/* use backward drawing onto dst */
	sxbeg = src->w - (sxbeg + w);
	dxbeg += w - 1;
	}

	tmp = dst->clip_rect.y - dy;
	sybeg = ((tmp < 0) ? 0 : tmp);
	dybeg = sybeg + dy;

	tmp = dst->clip_rect.y + dst->clip_rect.h - dy;
	h = ((tmp > src->h) ? src->h : tmp) - sybeg;
	if (h <= 0)
	return;

	if ( flip & Bitmap::SPRITE_V_FLIP ){
	/* use backward drawing onto dst */
	sybeg = src->h - (sybeg + h);
	dybeg += h - 1;
	}
	} else {
	w = src->w;
	h = src->h;
	sxbeg = 0;
	sybeg = 0;
	dxbeg = dx;
	if ( flip & Bitmap::SPRITE_H_FLIP ){
	dxbeg = dx + w - 1;
	}
	dybeg = dy;
	if ( flip & Bitmap::SPRITE_V_FLIP ){
	dybeg = dy + h - 1;
	}
	}

	/*
	lit_blender = PAINTOWN_MAKE_DLS_BLENDER(0);
	trans_blender = PAINTOWN_MAKE_DTS_BLENDER();
	*/

	#if 0
	if (dst->id & (BMP_ID_VIDEO \| BMP_ID_SYSTEM)) {
	bmp_select(dst);

	switch (mode){
	case Bitmap::SPRITE_NORMAL : {
	for (y = 0; y < h; y++) {
	PAINTOWN_PIXEL_PTR s = PAINTOWN_OFFSET_PIXEL_PTR(src->line[sybeg + y], sxbeg);
	PAINTOWN_PIXEL_PTR d = PAINTOWN_OFFSET_PIXEL_PTR(bmp_write_line(dst, dybeg + y * y_dir), dxbeg);
	for (x = w - 1; x >= 0; PAINTOWN_INC_PIXEL_PTR(s), PAINTOWN_INC_PIXEL_PTR_EX(d,x_dir), x--) {

	unsigned long c = PAINTOWN_GET_MEMORY_PIXEL(s);
	if (!PAINTOWN_IS_SPRITE_MASK(src, c)) {
	PAINTOWN_PUT_PIXEL(d, c);
	}
	}
	}

	break;
	}
	case Bitmap::SPRITE_LIT : {
	for (y = 0; y < h; y++) {
	PAINTOWN_PIXEL_PTR s = PAINTOWN_OFFSET_PIXEL_PTR(src->line[sybeg + y], sxbeg);
	PAINTOWN_PIXEL_PTR d = PAINTOWN_OFFSET_PIXEL_PTR(bmp_write_line(dst, dybeg + y * y_dir), dxbeg);
	for (x = w - 1; x >= 0; PAINTOWN_INC_PIXEL_PTR(s), PAINTOWN_INC_PIXEL_PTR_EX(d,x_dir), x--) {

	unsigned long c = PAINTOWN_GET_MEMORY_PIXEL(s);
	if (!PAINTOWN_IS_SPRITE_MASK(src, c)) {
	c = PAINTOWN_DLSX_BLEND(lit_blender, c);
	PAINTOWN_PUT_PIXEL(d, c);
	}
	}
	}
	break;
	}
	case Bitmap::SPRITE_TRANS : {
	for (y = 0; y < h; y++) {
	PAINTOWN_PIXEL_PTR s = PAINTOWN_OFFSET_PIXEL_PTR(src->line[sybeg + y], sxbeg);
	PAINTOWN_PIXEL_PTR d = PAINTOWN_OFFSET_PIXEL_PTR(bmp_write_line(dst, dybeg + y * y_dir), dxbeg);
	for (x = w - 1; x >= 0; PAINTOWN_INC_PIXEL_PTR(s), PAINTOWN_INC_PIXEL_PTR_EX(d,x_dir), x--) {

	unsigned long c = PAINTOWN_GET_MEMORY_PIXEL(s);
	if (!PAINTOWN_IS_SPRITE_MASK(src, c)) {
	c = PAINTOWN_DTS_BLEND(trans_blender, PAINTOWN_GET_PIXEL(d), c);
	PAINTOWN_PUT_PIXEL(d, c);
	}
	}
	}
	break;
	}
	}
	#xendif

	for (y = 0; y < h; y++) {
	PAINTOWN_PIXEL_PTR s = PAINTOWN_OFFSET_PIXEL_PTR(src->line[sybeg + y], sxbeg);

	/* flipped if y_dir is -1 */
	PAINTOWN_PIXEL_PTR d = PAINTOWN_OFFSET_PIXEL_PTR(bmp_write_line(dst, dybeg + y * y_dir), dxbeg);

	/* d is incremented by x_dir, -1 if flipped */
	for (x = w - 1; x >= 0; PAINTOWN_INC_PIXEL_PTR(s), PAINTOWN_INC_PIXEL_PTR_EX(d,x_dir), x--) {
	unsigned long c = PAINTOWN_GET_MEMORY_PIXEL(s);
	if (!PAINTOWN_IS_SPRITE_MASK(src, c)) {
	switch( mode ){
	case Bitmap::SPRITE_NORMAL : break;
	case Bitmap::SPRITE_LIT : {
	c = PAINTOWN_DLSX_BLEND(lit_blender, c);
	break;
	}
	case Bitmap::SPRITE_TRANS : {
	c = PAINTOWN_DTS_BLEND(trans_blender, PAINTOWN_GET_PIXEL(d), c);
	break;
	}
	}
	PAINTOWN_PUT_PIXEL(d, c);
	}
	}
	}

	bmp_unwrite_line(dst);
	}
	else {
	#endif
	{

	switch (mode){
	case Bitmap::SPRITE_NORMAL : {
	unsigned int mask = Bitmap::makeColor(255, 0, 255);
	int bpp = src->format->BytesPerPixel;
	for (y = 0; y < h; y++) {
	Uint8 * sourceLine = computeOffset(src, sxbeg, sybeg + y);
	Uint8 * destLine = computeOffset(dst, dxbeg, dybeg + y * y_dir);

	for (x = w - 1; x >= 0; sourceLine += bpp, destLine += bpp * x_dir, x--) {
	unsigned long sourcePixel = (Uint16) sourceLine;
	if (!(sourcePixel == mask)){
	// unsigned int destPixel = (Uint16) destLine;
	// sourcePixel = globalBlend.currentBlender(destPixel, sourcePixel, globalBlend.alpha);
	if (filter != NULL){
	(Uint16 )destLine = filter->filter(sourcePixel);
	} else {
	(Uint16 )destLine = sourcePixel;
	}
	}
	}
	}
	break;
	}
	case Bitmap::SPRITE_LIT : {
	int bpp = src->format->BytesPerPixel;
	int litColor = Bitmap::makeColor(globalBlend.red, globalBlend.green, globalBlend.blue);
	unsigned int mask = Bitmap::makeColor(255, 0, 255);
	for (y = 0; y < h; y++) {
	Uint8 * sourceLine = computeOffset(src, sxbeg, sybeg + y);
	Uint8 * destLine = computeOffset(dst, dxbeg, dybeg + y * y_dir);

	for (x = w - 1; x >= 0; sourceLine += bpp, destLine += bpp * x_dir, x--) {
	unsigned long sourcePixel = (Uint16) sourceLine;
	if (!(sourcePixel == mask)){
	// unsigned int destPixel = (Uint16) destLine;
	if (filter != NULL){
	sourcePixel = globalBlend.currentBlender(litColor, filter->filter(sourcePixel), globalBlend.alpha);
	(Uint16 )destLine = sourcePixel;
	} else {
	sourcePixel = globalBlend.currentBlender(litColor, sourcePixel, globalBlend.alpha);
	(Uint16 )destLine = sourcePixel;
	}
	}
	}
	}
	break;
	}
	case Bitmap::SPRITE_TRANS : {
	int bpp = src->format->BytesPerPixel;
	unsigned int mask = Bitmap::makeColor(255, 0, 255);
	for (y = 0; y < h; y++) {
	Uint8 * sourceLine = computeOffset(src, sxbeg, sybeg + y);
	Uint8 * destLine = computeOffset(dst, dxbeg, dybeg + y * y_dir);

	for (x = w - 1; x >= 0; sourceLine += bpp, destLine += bpp * x_dir, x--) {
	unsigned long sourcePixel = (Uint16) sourceLine;
	if (!(sourcePixel == mask)){
	unsigned int destPixel = (Uint16) destLine;
	if (filter != NULL){
	sourcePixel = globalBlend.currentBlender(filter->filter(sourcePixel), destPixel, globalBlend.alpha);
	(Uint16 )destLine = sourcePixel;
	} else {
	sourcePixel = globalBlend.currentBlender(sourcePixel, destPixel, globalBlend.alpha);
	(Uint16 )destLine = sourcePixel;
	}
	}
	}
	}
	break;
	}
	default : { break; }
	}

	#if 0
	for (y = 0; y < h; y++) {
	PAINTOWN_PIXEL_PTR s = PAINTOWN_OFFSET_PIXEL_PTR(src->line[sybeg + y], sxbeg);
	PAINTOWN_PIXEL_PTR d = PAINTOWN_OFFSET_PIXEL_PTR(bmp_write_line(dst, dybeg + y * y_dir), dxbeg);

	for (x = w - 1; x >= 0; PAINTOWN_INC_PIXEL_PTR(s), PAINTOWN_INC_PIXEL_PTR_EX(d,x_dir), x--) {
	unsigned long c = PAINTOWN_GET_MEMORY_PIXEL(s);
	if (!PAINTOWN_IS_SPRITE_MASK(src, c)) {
	switch( mode ){
	case Bitmap::SPRITE_NORMAL : break;
	case Bitmap::SPRITE_LIT : {
	c = PAINTOWN_DLSX_BLEND(lit_blender, c);
	break;
	}
	case Bitmap::SPRITE_TRANS : {
	c = PAINTOWN_DTS_BLEND(trans_blender, PAINTOWN_GET_PIXEL(d), c);
	break;
	}
	}
	PAINTOWN_PUT_MEMORY_PIXEL(d, c);
	}
	}
	}
	#endif
	}

	if (SDL_MUSTLOCK(src)){
	SDL_UnlockSurface(src);
	}

	if (SDL_MUSTLOCK(dst)){
	SDL_UnlockSurface(dst);
	}
	}

	/* ultra special-case for drawing a light (like from a lamp).
	* center of light is x,y and shines in a perfect isosolese triangle.
	*/
	static void paintown_light16(SDL_Surface * dst, const int x, const int y, int width, int height, const int start_y, const int focus_alpha, const int edge_alpha, const int focus_color, const int edge_color){

	if (width > dst->w){
	width = dst->w;
	}

	if (height > dst->h){
	height = dst->h;
	}

	int dxbeg = x - width;
	int x_dir = 1;
	unsigned char * alphas = new unsigned char[width];
	int * colors = new int[width];
	for (int i = 0; i < width; i++){
	alphas[i] = (unsigned char)((double)(edge_alpha - focus_alpha) * (double)i / (double)width + focus_alpha);
	}
	Util::blend_palette(colors, width, focus_color, edge_color);

	if (SDL_MUSTLOCK(dst)){
	SDL_LockSurface(dst);
	}

	int min_y, max_y, min_x, max_x;
	min_y = dst->clip_rect.y;
	max_y = dst->clip_rect.y + dst->clip_rect.h - 1;
	min_x = dst->clip_rect.x;
	max_x = dst->clip_rect.x + dst->clip_rect.w - 1;

	int dybeg = y;

	/* tan(theta) = y / x */
	double xtan = (double) height / (double) width;
	int bpp = dst->format->BytesPerPixel;
	for (int sy = start_y; sy < height; sy++) {
	if (dybeg + sy < min_y \|\| dybeg + sy > max_y){
	continue;
	}
	/* x = y / tan(theta) */
	int top_width = (int)((double) sy / xtan);
	if (top_width == 0){
	continue;
	}

	dxbeg = x - top_width;
	Uint8* line = computeOffset(dst, dxbeg, dybeg + y);

	for (int sx = -top_width; sx <= top_width; line += x_dir * bpp, sx++) {
	if (sx + x < min_x \|\| sx + x > max_x){
	continue;
	}

	unsigned long c = (Uint16) line;

	/* TODO:
	* converting to a double and calling fabs is overkill, just
	* write an integer abs() function.
	*/
	int sx_abs = (int) fabs((double) sx);
	int alphaUse = alphas[sx_abs];
	int color = colors[sx_abs];

	c = globalBlend.currentBlender(color, c, alphaUse);

	(Uint16) line = c;
	}
	}
	delete[] alphas;
	delete[] colors;

	if (SDL_MUSTLOCK(dst)){
	SDL_UnlockSurface(dst);
	}
	}
	diff --git a/util/sdl/stretch/SDL_stretch.h b/util/sdl/stretch/SDL_stretch.h
	new file mode 100644
	index 00000000..5c0ea6de
	--- /dev/null
	+++ b/util/sdl/stretch/SDL_stretch.h
	@@ -0,0 +1,88 @@
	+/*
	+ SDL_stretch - Stretch Functions For The Simple DirectMedia Layer
	+ Copyright (C) 2003 Guido Draheim
	+
	+ This library is free software; you can redistribute it and/or
	+ modify it under the terms of the GNU Library General Public
	+ License as published by the Free Software Foundation; either
	+ version 2 of the License, or (at your option) any later version.
	+
	+ This library 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
	+ Library General Public License for more details.
	+
	+ You should have received a copy of the GNU Library General Public
	+ License along with this library; if not, write to the Free
	+ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
	+
	+ Guido Draheim, guidod@gmx.de
	+*/
	+
	+#ifndef _SDL_stretch_h
	+#define _SDL_stretch_h
	+
	+#include "SDL_video.h"
	+#ifndef _SDL_video_h
	+#include <SDL/SDL_video.h>
	+#endif
	+
	+#ifdef __cplusplus
	+extern "C" {
	+#endif
	+
	+#ifndef DECLSPEC
	+#define DECLSPEC
	+#endif
	+
	+/** Perform a stretch blit between two surfaces of the same format.
	+ * NOTE: This function is not safe to call from multiple threads!
	+ *
	+ * This function will stretch a srcrect smoothly to the full area of
	+ * the dst surface. If no srcrect is given then the full area of the
	+ * src surface is stretched smoothly to the full dst surface. The
	+ * dstrect is ignored always.
	+ *
	+ * This function will also watch for a clip rectangle on the src
	+ * surface. This may speed up handling in your programs by creating
	+ * a larger src surface with an associated viewframe, and the srcrect
	+ * argument needs not be recomputed.
	+ */
	+extern DECLSPEC int
	+SDL_StretchSurfaceRect(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect);
	+/** Perform a stretch blit between two surfaces of the same format.
	+ * NOTE: This function is not safe to call from multiple threads!
	+ *
	+ * This function will stretch a srcrect smoothly to the full area of
	+ * the dst surface. If no srcrect is given then the full area of the
	+ * src surface is stretched smoothly to the full dst surface. The
	+ * dstrect is ignored always.
	+ *
	+ * Remember that this is the inverse meaning, the main SDL lib will
	+ * ignore the srcrect and only watch for the dstrect if any.
	+ */
	+extern DECLSPEC int
	+SDL_StretchSurfaceBlit(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect);
	+/**
	+ * This function will stretch to 150%. This is not only a fast function
	+ * but it is also safe to call from multiple threads. If the srcrect
	+ * is given then only that rect is copied. Otherwise the full src
	+ * surface is copied to the full dst surface. The dstrect is ignored.
	+ */
	+extern DECLSPEC int
	+SDL_StretchSurface_23(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect);
	+
	+/**
	+ * return some informative information.
	+ */
	+extern DECLSPEC char*
	+SDL_StretchInfo(void);
	+
	+
	+#ifdef __cplusplus
	+}
	+#endif
	+#endif
	diff --git a/util/sdl/stretch/SDL_stretchasm.h b/util/sdl/stretch/SDL_stretchasm.h
	new file mode 100644
	index 00000000..d2441456
	--- /dev/null
	+++ b/util/sdl/stretch/SDL_stretchasm.h
	@@ -0,0 +1,102 @@
	+/*
	+ SDL_stretch - Stretch Functions For The Simple DirectMedia Layer
	+ Copyright (C) 2003 Guido Draheim
	+
	+ This library is free software; you can redistribute it and/or
	+ modify it under the terms of the GNU Library General Public
	+ License as published by the Free Software Foundation; either
	+ version 2 of the License, or (at your option) any later version.
	+
	+ This library 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
	+ Library General Public License for more details.
	+
	+ You should have received a copy of the GNU Library General Public
	+ License along with this library; if not, write to the Free
	+ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
	+
	+ Guido Draheim, guidod@gmx.de
	+*/
	+
	+#ifndef _SDL_STRETCHASM_IMPLEMENTATION_
	+#define _SDL_STRETCHASM_IMPLEMENTATION_
	+
	+/* development header - do not use */
	+
	+#if (defined(WIN32) && !defined(_M_ALPHA) && !defined(_WIN32_WCE))
	+#define SDL_STRETCH_I386
	+#define SDL_STRETCH_MASM
	+#define SDL_STRETCH_USE_ASM
	+#endif
	+
	+#if defined(__i386__) && defined(__GNUC__)
	+#define SDL_STRETCH_I386
	+#define SDL_STRETCH_GAS
	+#define SDL_STRETCH_USE_ASM
	+#endif
	+
	+#if defined(__x86_64__) && defined(__GNUC__)
	+#define SDL_STRETCH_X86_64
	+#define SDL_STRETCH_GAS
	+#define SDL_STRETCH_USE_ASM
	+#endif
	+
	+#if defined SDL_STRETCH_I386 && defined SDL_STRETCH_MASM
	+#define SDL_STRETCH_CALL(code, srcp, dstp) \
	+__asm {\
	+ push edi ;\
	+ push esi ;\
	+ mov edi, dstp ;\
	+ mov esi, srcp ;\
	+ call dword ptr code ;\
	+ pop esi ;\
	+ pop edi ;\
	+}
	+#endif
	+
	+#if defined SDL_STRETCH_X86_64 && defined SDL_STRETCH_MASM
	+#define SDL_STRETCH_CALL(code, srcp, dstp) \
	+__asm {\
	+ push rdi ;\
	+ push rsi ;\
	+ mov rdi, dstp ;\
	+ mov rsi, srcp ;\
	+ call qword ptr code ;\
	+ pop rsi ;\
	+ pop rdi ;\
	+}
	+#endif
	+
	+#if defined SDL_STRETCH_I386 && defined SDL_STRETCH_GAS
	+#define SDL_STRETCH_CALL(code, srcp, dstp) \
	+__asm__ __volatile__ ("call *%%eax" \
	+ :: "S" (dstp), "D" (srcp), "a" (code));
	+#endif
	+
	+#if defined SDL_STRETCH_X86_64 && defined SDL_STRETCH_GAS
	+#define SDL_STRETCH_CALL(code, srcp, dstp) \
	+__asm__ __volatile__ ("call *%%rax" \
	+ :: "S" (dstp), "D" (srcp), "a" (code));
	+#endif
	+
	+#if defined USE_ASM_STRETCH && ! defined SDL_STRETCH_USE_ASM
	+#define SDL_STRETCH_USE_ASM 1
	+#endif
	+
	+#if defined SDL_STRETCH_DISABLE_ASM
	+#undef SDL_STRETCH_USE_ASM
	+#undef SDL_STRETCH_CALL
	+#endif
	+
	+#if defined SDL_STRETCH_USE_INTERPRETED_CODE
	+#undef SDL_STRETCH_CALL
	+#endif
	+
	+#if defined SDL_STRETCH_I386 \|\| defined WIN32 \|\| !defined SDL_STRETCH_CALL
	+#define SDL_STRETCH_DATA_EXECUTABLE
	+#else
	+#define SDL_STRETCH_DATA_NOEXEC
	+#endif
	+
	+#endif
	diff --git a/util/sdl/stretch/SDL_stretchcode.h b/util/sdl/stretch/SDL_stretchcode.h
	new file mode 100644
	index 00000000..f8d06c3d
	--- /dev/null
	+++ b/util/sdl/stretch/SDL_stretchcode.h
	@@ -0,0 +1,103 @@
	+/*
	+ SDL_stretch - Stretch Functions For The Simple DirectMedia Layer
	+ Copyright (C) 2003 Guido Draheim
	+
	+ This library is free software; you can redistribute it and/or
	+ modify it under the terms of the GNU Library General Public
	+ License as published by the Free Software Foundation; either
	+ version 2 of the License, or (at your option) any later version.
	+
	+ This library 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
	+ Library General Public License for more details.
	+
	+ You should have received a copy of the GNU Library General Public
	+ License along with this library; if not, write to the Free
	+ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
	+
	+ Guido Draheim, guidod@gmx.de
	+*/
	+
	+#ifndef _SDL_STRETCHCODE_IMPLEMENTATION_
	+#define _SDL_STRETCHCODE_IMPLEMENTATION_
	+
	+#ifdef __cplusplus
	+extern "C" {
	+#endif
	+
	+/* development header - do not use */
	+
	+#define SDL_STRETCHCODE_BUFFERSIZE 8192
	+
	+/**
	+ * TheRowStretchCode is a shared buffer between Stretch-routines that
	+ * use no extra buffer-argument. You should call SDL_SetRowStretchCode
	+ * to fill this internal buffer and set a "call"-operation for your target
	+ * cpu to execute this static buffer. That is done for effiency
	+ * as the RowStretch is often called in a tight loop for each Row
	+ * in a rectengular stretch and it is best to not use a variable
	+ * argument with an indirect call or a function call that would
	+ * build up a callframe and release that callframe later.
	+ *
	+ * If you do not need that effiency, use PutRowStretchCode and RunRowStretchCode
	+ * which are also good in a multithreading environment. To allocate a new buffer
	+ * for usage with Put/Run you can use the NewRowStretchCode routine which is
	+ * also used on NX machines (e.g. AMD64) where the data segment is set to be
	+ * not-executable (in that case it will allocate from heap and use mprotect).
	+ * if the argument is 0 then a buffer of the default size is allocated. If the
	+ * buffer allocation (or mprotect) fails it will return NULL and SDL_SetError.
	+ */
	+extern unsigned char* SDL_GetRowStretchCode(void);
	+
	+/** => SDL_GetRowStretchCode */
	+extern unsigned char* SDL_NewRowStretchCode(unsigned size);
	+
	+/**
	+ * The SetRowStretchCode is a wrapper around PutRowStretchCode that
	+ * uses the Adress and Size of the shared SDL_TheRowStretchCode buffer.
	+ * The PutRowStretchCode will fill the given buffer with an assembler
	+ * stream that should be called with SDL_RunRowStretchCode. The
	+ * assembler stream is usually faster as all the scale decisions are
	+ * done in advance of the execution. This helps when a RunCode is
	+ * done multiple times with the same src_w/dst_w/bpp pair. All the
	+ * pixel-get and pixel-set calls are unrolled in that buffer. Therefore,
	+ * the buffer should be big enough - as a rule of thumb use a buffer
	+ * of size (src_w+dst_w)*5
	+ *
	+ * If PutCode or SetCode fails, a NULL is returned and SDL_SetError.
	+ * Otherwise, the start adress of the machine code buffer is returned,
	+ * which is also the input argument of PutCode and RunCode.
	+ */
	+unsigned char* SDL_SetRowStretchCode(int src_w, int dst_w, int bpp);
	+/** => SDL_SetRowStretchCode */
	+unsigned char* SDL_PutRowStretchCode(unsigned char* buffer, int buflen,
	+ int src_w, int dst_w, int bpp);
	+/** => SDL_SetRowStretchCode */
	+void SDL_RunRowStretchCode(unsigned char* buffer,
	+ unsigned char* src,
	+ unsigned char* dst);
	+
	+/** => SDL_SetRowStretchCode
	+ * If SDL_SetRowStretchCode fails, this function must be used instead.
	+ * This function and its cousins are singular routines that work in
	+ * a tight loop to scale a single row. The number specifies the
	+ * byte-width of each pixel (it is not a bit-width!).
	+ */
	+void SDL_StretchRow1(Uint8 src, int src_w, Uint8 dst, int dst_w);
	+/** => SDL_SetRowStretchCode */
	+void SDL_StretchRow2(Uint16 src, int src_w, Uint16 dst, int dst_w);
	+/** => SDL_SetRowStretchCode */
	+void SDL_StretchRow3(Uint8 src, int src_w, Uint8 dst, int dst_w);
	+/** => SDL_SetRowStretchCode */
	+void SDL_StretchRow4(Uint32 src, int src_w, Uint32 dst, int dst_w);
	+
	+/**
	+ * return some informative information.
	+ */
	+extern char* SDL_StretchRowInfo(void);
	+
	+#ifdef __cplusplus
	+}
	+#endif
	+#endif
	diff --git a/util/sdl/stretch/config.h b/util/sdl/stretch/config.h
	new file mode 100644
	index 00000000..3f811f54
	--- /dev/null
	+++ b/util/sdl/stretch/config.h
	@@ -0,0 +1,60 @@
	+/* src/config.h. Generated from config.h.in by configure. */
	+/* src/config.h.in. Generated from configure.ac by autoheader. */
	+
	+/* Define to 1 if you have the <dlfcn.h> header file. */
	+#define HAVE_DLFCN_H 1
	+
	+/* Define to 1 if you have the <inttypes.h> header file. */
	+#define HAVE_INTTYPES_H 1
	+
	+/* Define to 1 if you have the <memory.h> header file. */
	+#define HAVE_MEMORY_H 1
	+
	+/* Define to 1 if you have the <stdint.h> header file. */
	+#define HAVE_STDINT_H 1
	+
	+/* Define to 1 if you have the <stdlib.h> header file. */
	+#define HAVE_STDLIB_H 1
	+
	+/* Define to 1 if you have the <strings.h> header file. */
	+#define HAVE_STRINGS_H 1
	+
	+/* Define to 1 if you have the <string.h> header file. */
	+#define HAVE_STRING_H 1
	+
	+/* Define to 1 if you have the <sys/stat.h> header file. */
	+#define HAVE_SYS_STAT_H 1
	+
	+/* Define to 1 if you have the <sys/types.h> header file. */
	+#define HAVE_SYS_TYPES_H 1
	+
	+/* Define to 1 if you have the <unistd.h> header file. */
	+#define HAVE_UNISTD_H 1
	+
	+/* Define to the sub-directory in which libtool stores uninstalled libraries.
	+ */
	+#define LT_OBJDIR ".libs/"
	+
	+/* Name of package */
	+#define PACKAGE "SDL_stretch"
	+
	+/* Define to the address where bug reports for this package should be sent. */
	+#define PACKAGE_BUGREPORT ""
	+
	+/* Define to the full name of this package. */
	+#define PACKAGE_NAME ""
	+
	+/* Define to the full name and version of this package. */
	+#define PACKAGE_STRING ""
	+
	+/* Define to the one symbol short name of this package. */
	+#define PACKAGE_TARNAME ""
	+
	+/* Define to the version of this package. */
	+#define PACKAGE_VERSION ""
	+
	+/* Define to 1 if you have the ANSI C header files. */
	+#define STDC_HEADERS 1
	+
	+/* Version number of package */
	+#define VERSION "0.3.1"
	diff --git a/util/sdl/stretch/sdlscreen.c b/util/sdl/stretch/sdlscreen.c
	new file mode 100644
	index 00000000..a1df98fe
	--- /dev/null
	+++ b/util/sdl/stretch/sdlscreen.c
	@@ -0,0 +1,376 @@
	+/*
	+ SDL_stretch - Stretch Functions For The Simple DirectMedia Layer
	+ Copyright (C) 2003 Guido Draheim
	+
	+ This library is free software; you can redistribute it and/or
	+ modify it under the terms of the GNU Library General Public
	+ License as published by the Free Software Foundation; either
	+ version 2 of the License, or (at your option) any later version.
	+
	+ This library 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
	+ Library General Public License for more details.
	+
	+ You should have received a copy of the GNU Library General Public
	+ License along with this library; if not, write to the Free
	+ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
	+
	+ Guido Draheim, guidod@gmx.de
	+*/
	+
	+/* these will fill the target surface completely */
	+
	+/****************************************************
	+ *
	+ * WARNING: these routines are not used anymore.
	+ * (they will not be part of the normal library)
	+ *
	+ ****************************************************
	+ */
	+
	+#ifdef SAVE_RCSID
	+static char rcsid =
	+ "@(#) $Id: sdstretch.c $";
	+#endif
	+
	+#include "SDL/SDL_error.h"
	+#include "SDL/SDL_video.h"
	+#include "SDL_stretch.h"
	+#include "SDL_stretchcode.h"
	+#include "SDL_stretchasm.h"
	+#include "SDL_stretchtest.hh"
	+
	+#define PRERUN 0
	+
	+
	+/* Perform a stretch blit between two surfaces of the same format.
	+ NOTE: This function is not safe to call from multiple threads!
	+*/
	+int SDL_StretchSurfaceBlitTo(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect)
	+{
	+ int i = 0;
	+ int src_row, dst_row;
	+ SDL_Rect full_src;
	+ /* SDL_Rect full_dst; */
	+# if defined SDL_STRETCH_USE_ASM
	+ unsigned char* code = 0;
	+# endif
	+ const int bpp = dst->format->BytesPerPixel;
	+ auto int dest_w = srcrect->w * dst->w / src->w;
	+ auto int dest_x = srcrect->x * dst->w / src->w;
	+
	+ if ( src->format->BitsPerPixel != dst->format->BitsPerPixel ) {
	+ SDL_SetError("Only works with same format surfaces");
	+ return(-1);
	+ }
	+
	+ /* Verify the blit rectangles */
	+ if ( srcrect ) {
	+ if ( (srcrect->x < 0) \|\| (srcrect->y < 0) \|\|
	+ ((srcrect->x+srcrect->w) > src->w) \|\|
	+ ((srcrect->y+srcrect->h) > src->h) ) {
	+ SDL_SetError("Invalid source blit rectangle");
	+ return(-1);
	+ }
	+ } else {
	+ full_src.x = 0;
	+ full_src.y = 0;
	+ full_src.w = src->w;
	+ full_src.h = src->h;
	+ srcrect = &full_src;
	+ }
	+
	+# ifdef SDL_STRETCH_USE_ASM
	+ /* Write the opcodes for this stretch */
	+ /* if ( (bpp != 3) */
	+ code = SDL_SetRowStretchCode(srcrect->w, dest_w, bpp);
	+# endif
	+
	+
	+ if (PRERUN) /* let the compiler do the dead-code removal */
	+ { /* Pre-Run the stretch blit (the invisible lines) */
	+ src_row = 0;
	+ dst_row = 0;
	+ while (1)
	+ {
	+ if (src_row >= srcrect->y) break;
	+ dst_row++; i += src->h;
	+ if (i < dst->h) continue;
	+ do { i -= dst->h; src_row++; } while (i >= dst->h);
	+ }
	+ }else
	+ { /* or compute the resulting dst_row and i-fraction directly: */
	+ src_row = srcrect->y;
	+ dst_row = ((src_row * dst->h)+(src->h-1)) / src->h;
	+ i = (dst_row * src->h - src_row * dst->h ) % dst->h;
	+ if (i < 0) i += dst->h;
	+ }
	+
	+ if (dstrect) { /returnvalue/
	+ dstrect->y = dst_row;
	+ dstrect->x = dest_x;
	+ dstrect->w = dest_w;
	+ }
	+
	+
	+ while (src_row < srcrect->y + srcrect->h)
	+ {
	+ Uint8 srcp, dstp;
	+ srcp = (Uint8 )src->pixels + (src_rowsrc->pitch)
	+ + (srcrect->x*bpp);
	+ draw:
	+ dstp = (Uint8 )dst->pixels + (dst_rowdst->pitch)
	+ + (dest_x*bpp);
	+# ifdef SDL_STRETCH_USE_ASM
	+ if (code)
	+ {
	+# if defined SDL_STRETCH_CALL
	+ SDL_STRETCH_CALL(code, srcp, dstp);
	+# else
	+ SDL_RunRowStretchCode(code, srcp, dstp);
	+# endif
	+ }else
	+# endif /USE_ASM/
	+ {
	+ switch (bpp) {
	+ case 1:
	+ SDL_StretchRow1(srcp, srcrect->w, dstp, dest_w);
	+ break;
	+ case 2:
	+ SDL_StretchRow2((Uint16 *)srcp, srcrect->w,
	+ (Uint16 *)dstp, dest_w);
	+ break;
	+ case 3:
	+ SDL_StretchRow3(srcp, srcrect->w, dstp, dest_w);
	+ break;
	+ case 4:
	+ SDL_StretchRow4((Uint32 *)srcp, srcrect->w,
	+ (Uint32 *)dstp, dest_w);
	+ break;
	+ }
	+ }
	+
	+ dst_row++; i += src->h;
	+ if (dst_row == dst->h) break; /oops/
	+ if (i < dst->h) goto draw; /* draw the line again */
	+ do { i -= dst->h; src_row++; } while (i >= dst->h);
	+ }
	+
	+ if (dstrect) dstrect->h = dst_row - dstrect->y; /returnvalue/
	+
	+ return(0);
	+}
	+
	+/* ------------------------------------------------------------------------ */
	+
	+#define OPTIMIZED 1
	+
	+#if defined SDL_STRETCH_USE_ASM && defined OPTIMIZED
	+static int SDL_StretchFastBlitTo(
	+ unsigned char* code,
	+ int dst_row, int src_row,
	+ SDL_Surface* dst, SDL_Surface* src,
	+ SDL_Rect* rect, SDL_Rect* clip, int dest_x, int i)
	+{
	+ const int bpp = src->format->BytesPerPixel;
	+ Uint8 srcp = (Uint8 )src->pixels + (src_rowsrc->pitch) + (rect->xbpp);
	+ Uint8 srcE = (Uint8 )src->pixels + ((rect->y + rect->h)
	+ src->pitch) + (rect->xbpp);
	+ Uint8 dstp = (Uint8 )dst->pixels + (dst_rowdst->pitch) + (dest_xbpp);
	+// Uint8* dstE = (Uint8 )dst->pixels + (dst->hdst->pitch) + (dest_x*bpp);
	+
	+ while (srcp < srcE)
	+ {
	+ draws:
	+# if defined SDL_STRETCH_CALL
	+ SDL_STRETCH_CALL(code, srcp, dstp)
	+# else
	+ SDL_RunRowStretchCode(code, srcp, dstp);
	+# endif
	+ dstp += dst->pitch; i += clip->h;
	+ // if (dstp == dstE) break; /oops/
	+ if (i < dst->h) goto draws; /* draw the line again */
	+ do { i -= dst->h; srcp += src->pitch; } while (i >= dst->h);
	+ }
	+ return dst_row;
	+}
	+#endif
	+
	+#define DEBUGSIZES 0
	+#define DEBUGSIZES_ (src->h != dst->h)
	+
	+/* Perform a stretch blit between two surfaces of the same format.
	+ NOTE: This function is not safe to call from multiple threads!
	+*/
	+int SDL_StretchSurfaceRectTo(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect)
	+{
	+ int i = 0;
	+ int src_row, dst_row;
	+ int dest_x, dest_w;
	+ SDL_Rect rect;
	+ SDL_Rect clip;
	+# if defined SDL_STRETCH_USE_ASM
	+ unsigned char* code = 0;
	+# endif
	+ const int bpp = src->format->BytesPerPixel;
	+
	+ if ( src->format->BitsPerPixel != dst->format->BitsPerPixel ) {
	+ SDL_SetError("Only works with same format surfaces");
	+ return(-1);
	+ }
	+
	+ /* Verify the blit rectangles */
	+ if ( srcrect ) {
	+ rect = *srcrect;
	+#if 0
	+ clip.x = clip.y = 0; clip.h = src->h; clip.w = src->w;
	+#else
	+ clip = src->clip_rect;
	+
	+ if (rect.x > clip.x + clip.w \|\|
	+ rect.x + rect.w < clip.x \|\|
	+ rect.y > clip.y + clip.h \|\|
	+ rect.y + rect.h < clip.y) return 1;
	+ if (rect.x < clip.x) { rect.w -= clip.x-rect.x; rect.x = clip.x; }
	+ if (rect.w > clip.x + clip.w - rect.x)
	+ rect.w = clip.x + clip.w - rect.x;
	+ if (rect.y < clip.y) { rect.h -= clip.y-rect.y; rect.y = clip.y; }
	+ if (rect.h > clip.y + clip.h - rect.y)
	+ rect.h = clip.y + clip.h - rect.y;
	+#endif
	+ if (! rect.h \|\| !rect.w) return 1;
	+ } else {
	+ rect.x = 0;
	+ rect.y = 0;
	+ rect.w = src->w;
	+ rect.h = src->h;
	+ clip = rect;
	+ }
	+
	+ if (DEBUGSIZES)
	+ fprintf (stderr, "[%i/%i][%i+%i/%i+%i][%i+%i/%i+%i]\n",
	+ src->w,src->h,
	+ clip.x,clip.w,clip.y,clip.h,
	+ rect.x,rect.w,rect.y,rect.h);
	+
	+ dest_w = rect.w * dst->w / clip.w;
	+ dest_x = (rect.x-clip.x) * dst->w / clip.w;
	+
	+# ifdef SDL_STRETCH_USE_ASM
	+ /* Write the opcodes for this stretch */
	+ /* if ( (bpp != 3) */
	+ code = SDL_SetRowStretchCode(rect.w, dest_w, bpp);
	+# endif
	+
	+ if (PRERUN) /* let the compiler do the dead-code removal */
	+ { /* Pre-Run the stretch blit (the invisible lines) */
	+ src_row = clip.y;
	+ dst_row = 0;
	+ while (1)
	+ {
	+ if (src_row >= rect.y) break;
	+ dst_row++; i += clip.h;
	+ if (i < dst->h) continue;
	+ do { i -= dst->h; src_row++; } while (i >= dst->h);
	+ }
	+ }else
	+ { /* or compute the resulting dst_row and i-fraction directly: */
	+ src_row = rect.y - clip.y;
	+ dst_row = ((src_row * dst->h)+(clip.h-1)) / clip.h;
	+ i = (dst_row * clip.h - src_row * dst->h ) % dst->h;
	+ if (i < 0) i += dst->h;
	+ }
	+
	+ if (dstrect) {
	+ dstrect->y = dst_row;
	+ dstrect->x = dest_x;
	+ dstrect->w = dest_w;
	+ }
	+
	+ if (DEBUGSIZES)
	+ fprintf (stderr, "[%i/%i][%i+%i/%i+%i][%i+%i/%i+%i]"
	+ " -> [%i/%i][%i+%i/%i+?]\n",
	+ src->w,src->h,
	+ clip.x,clip.w,clip.y,clip.h,
	+ rect.x,rect.w,rect.y,rect.h,
	+ dst->w,dst->h,
	+ dest_x,dest_w,dst_row);
	+
	+# ifdef SDL_STRETCH_USE_ASM
	+ if (code)
	+ {
	+# ifdef OPTIMIZED
	+ dst_row = SDL_StretchFastBlitTo(code, dst_row, src_row, dst, src, &rect, &clip, dest_x,i);
	+# else
	+ while (src_row < rect.y + rect.h)
	+ {
	+ Uint8 srcp, dstp;
	+ srcp = (Uint8 )src->pixels + (src_rowsrc->pitch)
	+ + (rect.x*bpp);
	+ draws:
	+ dstp = (Uint8 )dst->pixels + (dst_rowdst->pitch)
	+ + (dest_x*bpp);
	+# ifdef SDL_STRETCH_CALL
	+ SDL_STRETCH_CALL(code, srcp, dstp);
	+# else
	+ SDL_RunRowStretchCode(code, srcp, dstp);
	+# endif
	+ dst_row++; i += clip.h;
	+ if (dst_row == dst->h) break; /oops/
	+ if (i < dst->h) goto draws; /* draw the line again */
	+ do { i -= dst->h; src_row++; } while (i >= dst->h);
	+ }
	+# endif
	+ }else
	+# endif /USE_ASM/
	+ {
	+ while (src_row < rect.y + rect.h)
	+ {
	+ Uint8 srcp, dstp;
	+ srcp = (Uint8 )src->pixels + (src_rowsrc->pitch)
	+ + (rect.x*bpp);
	+ draw:
	+ dstp = (Uint8 )dst->pixels + (dst_rowdst->pitch)
	+ + (dest_x*bpp);
	+
	+ if (DEBUGSIZES) fprintf(stderr, ".");
	+ switch (bpp) {
	+ case 1:
	+ SDL_StretchRow1(srcp, rect.w, dstp, dest_w);
	+ break;
	+ case 2:
	+ SDL_StretchRow2((Uint16 *)srcp, rect.w,
	+ (Uint16 *)dstp, dest_w);
	+ break;
	+ case 3:
	+ SDL_StretchRow3(srcp, rect.w, dstp, dest_w);
	+ break;
	+ case 4:
	+ SDL_StretchRow4((Uint32 *)srcp, rect.w,
	+ (Uint32 *)dstp, dest_w);
	+ break;
	+ }
	+
	+ dst_row++; i += clip.h;
	+ if (dst_row == dst->h) break; /oops/
	+ if (i < dst->h) goto draw; /* draw the line again */
	+ do { i -= dst->h; src_row++; } while (i >= dst->h);
	+ }
	+ }
	+
	+ if (dstrect) dstrect->h = dst_row - dstrect->y; /returnvalue/
	+
	+ if (DEBUGSIZES)
	+ fprintf (stderr, "[%i/%i][%i+%i/%i+%i][%i+%i/%i+%i]"
	+ " -> [%i/%i][%i+%i/?..%i]\n",
	+ src->w,src->h,
	+ clip.x,clip.w,clip.y,clip.h,
	+ rect.x,rect.w,rect.y,rect.h,
	+ dst->w,dst->h,
	+ dest_x,dest_w,dst_row);
	+
	+ return(0);
	+}
	diff --git a/util/sdl/stretch/sdlstretch.c b/util/sdl/stretch/sdlstretch.c
	new file mode 100644
	index 00000000..5701691a
	--- /dev/null
	+++ b/util/sdl/stretch/sdlstretch.c
	@@ -0,0 +1,382 @@
	+/*
	+ SDL_stretch - Stretch Functions For The Simple DirectMedia Layer
	+ Copyright (C) 2003 Guido Draheim
	+
	+ This library is free software; you can redistribute it and/or
	+ modify it under the terms of the GNU Library General Public
	+ License as published by the Free Software Foundation; either
	+ version 2 of the License, or (at your option) any later version.
	+
	+ This library 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
	+ Library General Public License for more details.
	+
	+ You should have received a copy of the GNU Library General Public
	+ License along with this library; if not, write to the Free
	+ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
	+
	+ Guido Draheim, guidod@gmx.de
	+*/
	+
	+#ifdef SAVE_RCSID
	+static char rcsid =
	+ "@(#) $Id: sdstretch.c $";
	+#endif
	+
	+#include "SDL/SDL_error.h"
	+#include "SDL/SDL_video.h"
	+#include "SDL_stretch.h"
	+#include "SDL_stretchcode.h"
	+#include "SDL_stretchasm.h"
	+
	+#define PRERUN 0
	+
	+
	+/* Perform a stretch blit between two surfaces of the same format.
	+ NOTE: This function is not safe to call from multiple threads!
	+*/
	+int SDL_StretchSurfaceBlit(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect)
	+{
	+ int i = 0;
	+ int src_row, dst_row;
	+ SDL_Rect rect;
	+# if defined SDL_STRETCH_USE_ASM
	+ unsigned char* code = 0;
	+# endif
	+ const int bpp = dst->format->BytesPerPixel;
	+ auto int dest_w;
	+ auto int dest_x;
	+
	+ if ( src->format->BitsPerPixel != dst->format->BitsPerPixel ) {
	+ SDL_SetError("Only works with same format surfaces");
	+ return(-1);
	+ }
	+
	+ /* Verify the blit rectangles */
	+ if ( srcrect ) {
	+ if ( (srcrect->x < 0) \|\| (srcrect->y < 0) \|\|
	+ ((srcrect->x+srcrect->w) > src->w) \|\|
	+ ((srcrect->y+srcrect->h) > src->h) ) {
	+ SDL_SetError("Invalid source blit rectangle");
	+ return(-1);
	+ }
	+ } else {
	+ rect.x = 0;
	+ rect.y = 0;
	+ rect.w = src->w;
	+ rect.h = src->h;
	+ srcrect = &rect;
	+ }
	+
	+ dest_w = srcrect->w * dst->w / src->w;
	+ dest_x = srcrect->x * dst->w / src->w;
	+
	+# ifdef SDL_STRETCH_USE_ASM
	+ /* Write the opcodes for this stretch */
	+ /* if ( (bpp != 3) */
	+ code = SDL_SetRowStretchCode(srcrect->w, dest_w, bpp);
	+# endif
	+
	+
	+ if (PRERUN) /* let the compiler do the dead-code removal */
	+ { /* Pre-Run the stretch blit (the invisible lines) */
	+ src_row = 0;
	+ dst_row = 0;
	+ while (1)
	+ {
	+ if (src_row >= srcrect->y) break;
	+ dst_row++; i += src->h;
	+ if (i < dst->h) continue;
	+ do { i -= dst->h; src_row++; } while (i >= dst->h);
	+ }
	+ }else
	+ { /* or compute the resulting dst_row and i-fraction directly: */
	+ src_row = srcrect->y;
	+ dst_row = ((src_row * dst->h)+(src->h-1)) / src->h;
	+ i = (dst_row * src->h - src_row * dst->h ) % dst->h;
	+ if (i < 0) i += dst->h;
	+ }
	+
	+ if (dstrect) { /returnvalue/
	+ dstrect->y = dst_row;
	+ dstrect->x = dest_x;
	+ dstrect->w = dest_w;
	+ }
	+
	+
	+ while (src_row < srcrect->y + srcrect->h)
	+ {
	+ Uint8 srcp, dstp;
	+ srcp = (Uint8 *)src->pixels
	+ + (src_row*src->pitch)
	+ + (srcrect->x*bpp);
	+ draw:
	+ dstp = (Uint8 *)dst->pixels
	+ + (dst_row*dst->pitch)
	+ + (dest_x*bpp);
	+# ifdef SDL_STRETCH_USE_ASM
	+ if (code)
	+ {
	+# ifdef SDL_STRETCH_CALL
	+ SDL_STRETCH_CALL(code, srcp, dstp);
	+# else
	+ SDL_RunRowStretchCode(code, srcp, dstp);
	+# endif
	+ }else
	+# endif /USE_ASM/
	+ {
	+ switch (bpp) {
	+ case 1:
	+ SDL_StretchRow1(srcp, srcrect->w, dstp, dest_w);
	+ break;
	+ case 2:
	+ SDL_StretchRow2((Uint16 *)srcp, srcrect->w,
	+ (Uint16 *)dstp, dest_w);
	+ break;
	+ case 3:
	+ SDL_StretchRow3(srcp, srcrect->w, dstp, dest_w);
	+ break;
	+ case 4:
	+ SDL_StretchRow4((Uint32 *)srcp, srcrect->w,
	+ (Uint32 *)dstp, dest_w);
	+ break;
	+ }
	+ }
	+
	+ dst_row++; i += src->h;
	+ if (dst_row == dst->h) break; /oops/
	+ if (i < dst->h) goto draw; /* draw the line again */
	+ do { i -= dst->h; src_row++; } while (i >= dst->h);
	+ }
	+
	+ if (dstrect) dstrect->h = dst_row - dstrect->y; /returnvalue/
	+
	+ return(0);
	+}
	+
	+/* ------------------------------------------------------------------------ */
	+
	+#define OPTIMIZED 1
	+
	+#if defined SDL_STRETCH_USE_ASM && defined OPTIMIZED
	+static int SDL_StretchFastBlit(unsigned char* code,
	+ int dst_row, int src_row,
	+ SDL_Surface* dst, SDL_Surface* src,
	+ SDL_Rect* rect, SDL_Rect* clip, int dest_x, int i)
	+{
	+ const int bpp = src->format->BytesPerPixel;
	+ Uint8 srcp = (Uint8 )src->pixels
	+ + (src_row*src->pitch)
	+ + (rect->x*bpp);
	+ Uint8 srcE = (Uint8 )src->pixels
	+ + ((rect->y + rect->h) * src->pitch)
	+ + (rect->x*bpp);
	+ Uint8 dstp = (Uint8 )dst->pixels
	+ + (dst_row*dst->pitch)
	+ + (dest_x*bpp);
	+// Uint8* dstE = (Uint8 )dst->pixels + (dst->hdst->pitch) + (dest_x*bpp);
	+
	+ while (srcp < srcE)
	+ {
	+ draws:
	+# ifdef SDL_STRETCH_CALL
	+ SDL_STRETCH_CALL(code, srcp, dstp);
	+# else
	+ SDL_RunRowStretchCode(code, srcp, dstp);
	+# endif
	+ dstp += dst->pitch; i += clip->h;
	+ // if (dstp == dstE) break; /oops/
	+ if (i < dst->h) goto draws; /* draw the line again */
	+ do { i -= dst->h; srcp += src->pitch; } while (i >= dst->h);
	+ }
	+ return dst_row;
	+}
	+#endif
	+
	+#define DEBUGSIZES 0
	+#define DEBUGSIZES_ (src->h != dst->h)
	+
	+/* Perform a stretch blit between two surfaces of the same format.
	+ NOTE: This function is not safe to call from multiple threads!
	+*/
	+int SDL_StretchSurfaceRect(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect)
	+{
	+ int i = 0;
	+ int src_row, dst_row;
	+ int dest_x, dest_w;
	+ SDL_Rect rect;
	+ SDL_Rect clip;
	+# if defined SDL_STRETCH_USE_ASM
	+ unsigned char* code = 0;
	+# endif
	+ const int bpp = src->format->BytesPerPixel;
	+
	+ if ( src->format->BitsPerPixel != dst->format->BitsPerPixel ) {
	+ SDL_SetError("Only works with same format surfaces");
	+ return(-1);
	+ }
	+
	+ /* Verify the blit rectangles */
	+ if ( srcrect ) {
	+ rect = *srcrect;
	+#if 0
	+ clip.x = clip.y = 0; clip.h = src->h; clip.w = src->w;
	+#else
	+ clip = src->clip_rect;
	+
	+ if (rect.x > clip.x + clip.w \|\|
	+ rect.x + rect.w < clip.x \|\|
	+ rect.y > clip.y + clip.h \|\|
	+ rect.y + rect.h < clip.y) return 1;
	+ if (rect.x < clip.x) { rect.w -= clip.x-rect.x; rect.x = clip.x; }
	+ if (rect.w > clip.x + clip.w - rect.x)
	+ rect.w = clip.x + clip.w - rect.x;
	+ if (rect.y < clip.y) { rect.h -= clip.y-rect.y; rect.y = clip.y; }
	+ if (rect.h > clip.y + clip.h - rect.y)
	+ rect.h = clip.y + clip.h - rect.y;
	+#endif
	+ if (! rect.h \|\| !rect.w) return 1;
	+ } else {
	+ rect.x = 0;
	+ rect.y = 0;
	+ rect.w = src->w;
	+ rect.h = src->h;
	+ clip = rect;
	+ }
	+
	+ if (DEBUGSIZES)
	+ fprintf (stderr, "[%i/%i][%i+%i/%i+%i][%i+%i/%i+%i]\n",
	+ src->w,src->h,
	+ clip.x,clip.w,clip.y,clip.h,
	+ rect.x,rect.w,rect.y,rect.h);
	+
	+ dest_w = rect.w * dst->w / clip.w;
	+ dest_x = (rect.x-clip.x) * dst->w / clip.w;
	+
	+# ifdef SDL_STRETCH_USE_ASM
	+ /* Write the opcodes for this stretch */
	+ /* if ( (bpp != 3) */
	+ code = SDL_SetRowStretchCode(rect.w, dest_w, bpp);
	+# endif
	+
	+ if (PRERUN) /* let the compiler do the dead-code removal */
	+ { /* Pre-Run the stretch blit (the invisible lines) */
	+ src_row = clip.y;
	+ dst_row = 0;
	+ while (1)
	+ {
	+ if (src_row >= rect.y) break;
	+ dst_row++; i += clip.h;
	+ if (i < dst->h) continue;
	+ do { i -= dst->h; src_row++; } while (i >= dst->h);
	+ }
	+ }else
	+ { /* or compute the resulting dst_row and i-fraction directly: */
	+ src_row = rect.y - clip.y;
	+ dst_row = ((src_row * dst->h)+(clip.h-1)) / clip.h;
	+ i = (dst_row * clip.h - src_row * dst->h ) % dst->h;
	+ if (i < 0) i += dst->h;
	+ }
	+
	+ if (dstrect) {
	+ dstrect->y = dst_row;
	+ dstrect->x = dest_x;
	+ dstrect->w = dest_w;
	+ }
	+
	+ if (DEBUGSIZES)
	+ fprintf (stderr, "[%i/%i][%i+%i/%i+%i][%i+%i/%i+%i]"
	+ " -> [%i/%i][%i+%i/%i+?]\n",
	+ src->w,src->h,
	+ clip.x,clip.w,clip.y,clip.h,
	+ rect.x,rect.w,rect.y,rect.h,
	+ dst->w,dst->h,
	+ dest_x,dest_w,dst_row);
	+
	+#ifdef SDL_STRETCH_USE_ASM
	+ if (code)
	+ {
	+# ifdef OPTIMIZED
	+ dst_row = SDL_StretchFastBlit(code, dst_row, src_row, dst, src, &rect, &clip, dest_x,i);
	+# else
	+ while (src_row < rect.y + rect.h)
	+ {
	+ Uint8 srcp, dstp;
	+ srcp = (Uint8 *)src->pixels
	+ + (src_row*src->pitch)
	+ + (rect.x*bpp);
	+ draws:
	+ dstp = (Uint8 *)dst->pixels
	+ + (dst_row*dst->pitch)
	+ + (dest_x*bpp);
	+
	+# ifdef SDL_STRETCH_CALL
	+ SDL_STRETCH_CALL(code, srcp, dstp);
	+# else
	+ SDL_RunRowStretchCode(code, srcp, dstp);
	+# endif
	+ dst_row++; i += clip.h;
	+ if (dst_row == dst->h) break; /oops/
	+ if (i < dst->h) goto draws; /* draw the line again */
	+ do { i -= dst->h; src_row++; } while (i >= dst->h);
	+ }
	+# endif
	+ }else
	+# endif /USE_ASM/
	+ {
	+ while (src_row < rect.y + rect.h)
	+ {
	+ Uint8 srcp, dstp;
	+ srcp = (Uint8 *)src->pixels
	+ + (src_row*src->pitch)
	+ + (rect.x*bpp);
	+ draw:
	+ dstp = (Uint8 *)dst->pixels
	+ + (dst_row*dst->pitch)
	+ + (dest_x*bpp);
	+
	+ if (DEBUGSIZES) fprintf(stderr, ".");
	+ switch (bpp) {
	+ case 1:
	+ SDL_StretchRow1(srcp, rect.w, dstp, dest_w);
	+ break;
	+ case 2:
	+ SDL_StretchRow2((Uint16 *)srcp, rect.w,
	+ (Uint16 *)dstp, dest_w);
	+ break;
	+ case 3:
	+ SDL_StretchRow3(srcp, rect.w, dstp, dest_w);
	+ break;
	+ case 4:
	+ SDL_StretchRow4((Uint32 *)srcp, rect.w,
	+ (Uint32 *)dstp, dest_w);
	+ break;
	+ }
	+
	+ dst_row++; i += clip.h;
	+ if (dst_row == dst->h) break; /oops/
	+ if (i < dst->h) goto draw; /* draw the line again */
	+ do { i -= dst->h; src_row++; } while (i >= dst->h);
	+ }
	+ }
	+
	+ if (dstrect) dstrect->h = dst_row - dstrect->y; /returnvalue/
	+
	+ if (DEBUGSIZES)
	+ fprintf (stderr, "[%i/%i][%i+%i/%i+%i][%i+%i/%i+%i]"
	+ " -> [%i/%i][%i+%i/?..%i]\n",
	+ src->w,src->h,
	+ clip.x,clip.w,clip.y,clip.h,
	+ rect.x,rect.w,rect.y,rect.h,
	+ dst->w,dst->h,
	+ dest_x,dest_w,dst_row);
	+
	+ return(0);
	+}
	+
	+char* SDL_StretchInfo(void) {
	+ return SDL_StretchRowInfo();
	+}
	diff --git a/util/sdl/stretch/sdlstretch23.c b/util/sdl/stretch/sdlstretch23.c
	new file mode 100644
	index 00000000..d30868dc
	--- /dev/null
	+++ b/util/sdl/stretch/sdlstretch23.c
	@@ -0,0 +1,2219 @@
	+/*
	+ SDL_stretch - Stretch Functions For The Simple DirectMedia Layer
	+ Copyright (C) 2003 Guido Draheim
	+
	+ This library is free software; you can redistribute it and/or
	+ modify it under the terms of the GNU Library General Public
	+ License as published by the Free Software Foundation; either
	+ version 2 of the License, or (at your option) any later version.
	+
	+ This library 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
	+ Library General Public License for more details.
	+
	+ You should have received a copy of the GNU Library General Public
	+ License along with this library; if not, write to the Free
	+ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
	+
	+ Guido Draheim, guidod@gmx.de
	+*/
	+
	+/* routines that stretch by constant factor (3 / 2) = 150% */
	+
	+#ifdef SAVE_RCSID
	+static char rcsid =
	+ "@(#) $Id: sdstretch.c $";
	+#endif
	+
	+#include "SDL/SDL_error.h"
	+#include "SDL/SDL_video.h"
	+#include "SDL_stretch.h"
	+#include "SDL_stretchasm.h"
	+
	+#if 0
	+/* the first part is for documentation only */
	+
	+/* Perform a stretch blit between two surfaces of the same format.
	+ NOTE: This function is not safe to call from multiple threads!
	+*/
	+int SDL_StretchSurface_23(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect)
	+{
	+ union { Uint8* b; Uint16* w; Uint32* x; } srcp, dstp, endp;
	+ unsigned int srci, dsti;
	+ SDL_Rect rect;
	+ SDL_Rect clip;
	+ const unsigned bpp = src->format->BytesPerPixel;
	+
	+ if ( src->format->BitsPerPixel != dst->format->BitsPerPixel ) {
	+ SDL_SetError("Only works with same format surfaces");
	+ return(-1);
	+ }
	+
	+ /* Verify the source blit rectangle - copy good values to "rect",
	+ * note that the input srcrect is clipped with the src->clip_rect
	+ * and not just the src-surface borders.
	+ */
	+ if ( srcrect ) {
	+ rect = *srcrect;
	+#if 0 /* for debugging, you might want to assume good srcrect values */
	+ clip.x = clip.y = 0; clip.h = src->h; clip.w = src->w;
	+#else /* otherwise we test with the clprect as well */
	+ clip = src->clip_rect;
	+
	+ if (rect.x > clip.x + clip.w \|\|
	+ rect.x + rect.w < clip.x \|\|
	+ rect.y > clip.y + clip.h \|\|
	+ rect.y + rect.h < clip.y) return 1;
	+ if (rect.x < clip.x) { rect.w -= clip.x-rect.x; rect.x = clip.x; }
	+ if (rect.w > clip.x + clip.w - rect.x)
	+ rect.w = clip.x + clip.w - rect.x;
	+ if (rect.y < clip.y) { rect.h -= clip.y-rect.y; rect.y = clip.y; }
	+ if (rect.h > clip.y + clip.h - rect.y)
	+ rect.h = clip.y + clip.h - rect.y;
	+#endif
	+ if (! rect.h \|\| !rect.w) return 1;
	+ } else
	+ { /* in the default however, we use srcrect == cliprect == src-screen */
	+ rect.x = 0;
	+ rect.y = 0;
	+ rect.w = src->w;
	+ rect.h = src->h;
	+ clip = rect;
	+ }
	+
	+ { /* and now compute the dest rect area - starting off with fixscaled */
	+ int dest_x = (rect.x - clip.x) * 3 / 2;
	+ int dest_y = (rect.y - clip.y) * 3 / 2;
	+ int dest_w = rect.w * 3 / 2;
	+ int dest_h = rect.h * 3 / 2;
	+
	+ /* great, so we have the dest measure from src according to the scale
	+ * but we need to check if that is still inside the dst->clip_rect
	+ * and possibly adapt the source rect for it to work. Items in
	+ * comments represent what a dest-only oriented routine would do.
	+ */
	+
	+ if (dest_x >= dst->clip_rect.x + dst->clip_rect.w \|\|
	+ dest_y >= dst->clip_rect.y + dst->clip_rect.h \|\|
	+ dest_x + dest_w <= dst->clip_rect.x \|\|
	+ dest_y + dest_h <= dst->clip_rect.y) return 1;
	+ if (dest_x < dst->clip_rect.x)
	+ { /dest_w -= dst->clip_rect.x - dest_x; dest_x = dst->clip_rect.x; /
	+ rect.w = (dest_w - dst->clip_rect.x - dest_x) / 3;
	+ rect.x += (dst->clip_rect.x - dest_x) / 3 * 2;
	+ dest_w = rect.w * 3; rect.w *= 2; dest_x = dst->clip_rect.x; }
	+ if (dest_w > dst->clip_rect.x + dst->clip_rect.w - dest_x)
	+ { /dest_w = dst->clip_rect.x + dst->clip_rect.w - dest_x; /
	+ rect.w = (dst->clip_rect.x + dst->clip_rect.w - dest_x) / 3;
	+ dest_w = rect.w * 3; rect.w *= 2; }
	+ if (dest_y < dst->clip_rect.y)
	+ { /dest_h -= dst->clip_rect.y - dest_y; dest_y = dst->clip_rect.y; /
	+ rect.h = (dest_h - dst->clip_rect.y - dest_y) / 3;
	+ rect.y += (dst->clip_rect.y - dest_y) / 3 * 2;
	+ dest_h = rect.h * 3; rect.h *= 2; dest_y = dst->clip_rect.y; }
	+ if (dest_h > dst->clip_rect.y + dst->clip_rect.h - dest_y)
	+ { /dest_h = dst->clip_rect.y + dst->clip_rect.h - dest_y; /
	+ rect.h = (dst->clip_rect.y + dst->clip_rect.h - dest_y) / 3;
	+ dest_h = rect.h * 3; rect.h *= 2; }
	+ if (!dest_h \|\| !dest_w \|\| !rect.h \|\| !rect.w) return 1;
	+
	+ if (dstrect) { /* inform the caller of the */
	+ dstrect->y = dest_y; /* computed results - if the */
	+ dstrect->x = dest_x; /* "dst" surface is actually */
	+ dstrect->w = dest_w; /* the user screen then one can */
	+ dstrect->h = dest_h; /* just use Update(dst,dstret) */
	+ }
	+
	+ /* these computations happen to be the same for all bitdepths */
	+ dstp.b = (Uint8*)dst->pixels +
	+ (dst->pitchdest_y) + (bppdest_x);
	+ srcp.b = (Uint8*)src->pixels +
	+ (src->pitchrect.y) + (bpprect.x);
	+ endp.b = srcp.b +
	+ src->pitch*rect.h;
	+ srci = src->pitch - (bpp*rect.w);
	+ dsti = dst->pitch - (bpp*dest_w);
	+ }
	+
	+ switch (bpp) /* the copy operations shall be as tight as possible */
	+ { /* to let all values be in cpu registers and the */
	+ case 4: /* complete series in the pipeline instruction cache */
	+ do {
	+ int count = rect.w / 2;
	+ do {
	+ dstp.x++ = srcp.x;
	+ dstp.x++ = srcp.x++;
	+ dstp.x++ = srcp.x++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ } while (srcp.b != endp.b);
	+ break;
	+ case 3:
	+ do {
	+ int count = rect.w / 2;
	+ do {
	+ dstp.b++ = srcp.b++;
	+ dstp.w++ = srcp.w; srcp.b--;
	+ dstp.w++ = srcp.w++;
	+ dstp.x++ = srcp.x++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ } while (srcp.b != endp.b);
	+ break;
	+ case 2:
	+ do {
	+ int count = rect.w / 2;
	+ do {
	+ dstp.w++ = srcp.w;
	+ dstp.w++ = srcp.w++;
	+ dstp.w++ = srcp.w++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ } while (srcp.b != endp.b);
	+ break;
	+ case 1:
	+ do {
	+ int count = rect.w / 2;
	+ do {
	+ dstp.b++ = srcp.b;
	+ dstp.b++ = srcp.b++;
	+ dstp.b++ = srcp.b++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ } while (srcp.b != endp.b);
	+ break;
	+ }
	+ return(0);
	+}
	+
	+#elif 0
	+/* the second part shows a nice optimization:
	+ *
	+ * we keep the rect.w and rect.h at halfsize - consequently,
	+ * dest.w == 3 * rect.w and dest.h == 3 * src.w always - the
	+ * copy-loops go in two src-lines -> three dst-lines anyway.
	+ *
	+ * and it does the fix the wrong behavior of the previous one,
	+ * if you try it then you will see that it does scale correctly
	+ * in the horizontal direction but the vertical does sometimes
	+ * seem to get messed up. That is because we did not yet unroll
	+ * the vertical part into being XY->XXY and instead it was kept
	+ * as XYz->XYz i.e. it would copy a bit too much out of the
	+ * source area instead of stretching it.
	+ */
	+
	+
	+/* Perform a stretch blit between two surfaces of the same format.
	+ NOTE: This function is not safe to call from multiple threads!
	+*/
	+int SDL_StretchSurface_23(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect)
	+{
	+ union { Uint8* b; Uint16* w; Uint32* x; } srcp, dstp, endp;
	+ unsigned int srci, dsti;
	+ SDL_Rect rect;
	+ SDL_Rect clip;
	+ const unsigned bpp = src->format->BytesPerPixel;
	+
	+ if ( src->format->BitsPerPixel != dst->format->BitsPerPixel ) {
	+ SDL_SetError("Only works with same format surfaces");
	+ return(-1);
	+ }
	+
	+ /* Verify the blit rectangles */
	+ if ( srcrect ) {
	+ rect = *srcrect;
	+#if 0
	+ clip.x = clip.y = 0; clip.h = src->h; clip.w = src->w;
	+#else
	+ clip = src->clip_rect;
	+
	+ if (rect.x > clip.x + clip.w \|\|
	+ rect.x + rect.w < clip.x \|\|
	+ rect.y > clip.y + clip.h \|\|
	+ rect.y + rect.h < clip.y) return 1;
	+ if (rect.x < clip.x) { rect.w -= clip.x-rect.x; rect.x = clip.x; }
	+ if (rect.w > clip.x + clip.w - rect.x)
	+ rect.w = clip.x + clip.w - rect.x;
	+ if (rect.y < clip.y) { rect.h -= clip.y-rect.y; rect.y = clip.y; }
	+ if (rect.h > clip.y + clip.h - rect.y)
	+ rect.h = clip.y + clip.h - rect.y;
	+#endif
	+ if (! rect.h \|\| !rect.w) return 1;
	+ rect.x /= 2; rect.w /= 2;
	+ rect.y /= 2; rect.h /= 2;
	+ clip.x /= 2; clip.w /= 2;
	+ clip.y /= 2; clip.h /= 2;
	+
	+ } else {
	+ rect.x = 0;
	+ rect.y = 0;
	+ rect.w = src->w / 2;
	+ rect.h = src->h / 2;
	+ clip = rect;
	+ }
	+
	+ {
	+ int dest_x = (rect.x - clip.x) * 3;
	+ int dest_y = (rect.y - clip.y) * 3;
	+ int dest_w = rect.w * 3;
	+ int dest_h = rect.h * 3;
	+
	+ if (dest_x >= dst->clip_rect.x + dst->clip_rect.w \|\|
	+ dest_y >= dst->clip_rect.y + dst->clip_rect.h \|\|
	+ dest_x + dest_w <= dst->clip_rect.x \|\|
	+ dest_y + dest_h <= dst->clip_rect.y) return 1;
	+ if (dest_x < dst->clip_rect.x)
	+ {
	+ rect.w = (dest_w - dst->clip_rect.x - dest_x) / 3;
	+ rect.x += (dst->clip_rect.x - dest_x) / 3;
	+ dest_w = rect.w * 3; dest_x = dst->clip_rect.x; }
	+ if (dest_w > dst->clip_rect.x + dst->clip_rect.w - dest_x)
	+ {
	+ rect.w = (dst->clip_rect.x + dst->clip_rect.w - dest_x) / 3;
	+ dest_w = rect.w * 3; }
	+ if (dest_y < dst->clip_rect.y)
	+ {
	+ rect.h = (dest_h - dst->clip_rect.y - dest_y) / 3;
	+ rect.y += (dst->clip_rect.y - dest_y) / 3;
	+ dest_h = rect.h * 3; dest_y = dst->clip_rect.y; }
	+ if (dest_h > dst->clip_rect.y + dst->clip_rect.h - dest_y)
	+ {
	+ rect.h = (dst->clip_rect.y + dst->clip_rect.h - dest_y) / 3;
	+ dest_h = rect.h * 3; }
	+ if (!dest_h \|\| !dest_w \|\| !rect.h \|\| !rect.w) return 1;
	+
	+ if (dstrect) { /returnvalue/
	+ dstrect->y = dest_y;
	+ dstrect->x = dest_x;
	+ dstrect->w = dest_w;
	+ dstrect->h = dest_h;
	+ }
	+
	+ dstp.b = (Uint8*)dst->pixels +
	+ (dst->pitchdest_y) + (bppdest_x);
	+ srcp.b = (Uint8*)src->pixels +
	+ (src->pitchrect.y2) + (bpprect.x2);
	+ endp.b = srcp.b + src->pitchrect.h2;
	+ srci = src->pitch - (bpprect.w2);
	+ dsti = dst->pitch - (bpp*dest_w);
	+ }
	+
	+ switch (bpp) /* the copy operations shall be as tight as possible */
	+ { /* to let all values be in cpu registers and the */
	+ case 4: /* complete series in the pipeline instruction cache */
	+ do {
	+ int count = rect.w;
	+ do {
	+ dstp.x++ = srcp.x;
	+ dstp.x++ = srcp.x++;
	+ dstp.x++ = srcp.x++;
	+ } while (--count != 0);
	+ srcp.b -= rect.w24;
	+ dstp.b += dsti;
	+ count = rect.w;
	+ do {
	+ dstp.x++ = srcp.x;
	+ dstp.x++ = srcp.x++;
	+ dstp.x++ = srcp.x++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ count = rect.w;
	+ do {
	+ dstp.x++ = srcp.x;
	+ dstp.x++ = srcp.x++;
	+ dstp.x++ = srcp.x++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ } while (srcp.b != endp.b);
	+ break;
	+ case 3:
	+ do {
	+ int count = rect.w;
	+ do {
	+ dstp.b++ = srcp.b++;
	+ dstp.w++ = srcp.w; srcp.b--;
	+ dstp.w++ = srcp.w++;
	+ dstp.x++ = srcp.x++;
	+ } while (--count != 0);
	+ srcp.b -= rect.w23;
	+ dstp.b += dsti;
	+ count = rect.w;
	+ do {
	+ dstp.b++ = srcp.b++;
	+ dstp.w++ = srcp.w; srcp.b--;
	+ dstp.w++ = srcp.w++;
	+ dstp.x++ = srcp.x++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ count = rect.w;
	+ do {
	+ dstp.b++ = srcp.b++;
	+ dstp.w++ = srcp.w; srcp.b--;
	+ dstp.w++ = srcp.w++;
	+ dstp.x++ = srcp.x++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ } while (srcp.b != endp.b);
	+ break;
	+ case 2:
	+ do {
	+ int count = rect.w;
	+ do {
	+ dstp.w++ = srcp.w;
	+ dstp.w++ = srcp.w++;
	+ dstp.w++ = srcp.w++;
	+ } while (--count != 0);
	+ srcp.b -= rect.w22;
	+ dstp.b += dsti;
	+ count = rect.w;
	+ do {
	+ dstp.w++ = srcp.w;
	+ dstp.w++ = srcp.w++;
	+ dstp.w++ = srcp.w++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ count = rect.w;
	+ do {
	+ dstp.w++ = srcp.w;
	+ dstp.w++ = srcp.w++;
	+ dstp.w++ = srcp.w++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ } while (srcp.b != endp.b);
	+ break;
	+ case 1:
	+ do {
	+ int count = rect.w;
	+ do {
	+ dstp.b++ = srcp.b;
	+ dstp.b++ = srcp.b++;
	+ dstp.b++ = srcp.b++;
	+ } while (--count != 0);
	+ srcp.b -= rect.w21;
	+ dstp.b += dsti;
	+ count = rect.w;
	+ do {
	+ dstp.b++ = srcp.b;
	+ dstp.b++ = srcp.b++;
	+ dstp.b++ = srcp.b++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ count = rect.w;
	+ do {
	+ dstp.b++ = srcp.b;
	+ dstp.b++ = srcp.b++;
	+ dstp.b++ = srcp.b++;
	+ } while (--count != 0);
	+ srcp.b += srci;
	+ dstp.b += dsti;
	+ } while (srcp.b != endp.b);
	+ break;
	+ }
	+ return(0);
	+}
	+
	+
	+#elif 0
	+/* the third part shows two extra optimizations:
	+ *
	+ * (1) the i386 is a desperatly register-starved architecture
	+ * and for the case of GNUC we say now where to store the
	+ * copyloop variables and we even cut in some assembler snippets.
	+ * (2) for that we better use copies of the srcp/dstp/endp pointers and
	+ * therefore the initial computations can be done on Byte-pointers
	+ * instead of the union-of-pointers as shown above for clarification.
	+ *
	+ * enough for UAE
	+ */
	+
	+
	+/* Perform a stretch blit between two surfaces of the same format.
	+ NOTE: This function is not safe to call from multiple threads!
	+*/
	+int SDL_StretchSurface_23(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect)
	+{
	+ Uint8 srcp, dstp, *endp;
	+ unsigned int srci, dsti;
	+ SDL_Rect rect;
	+ SDL_Rect clip;
	+ const unsigned bpp = src->format->BytesPerPixel;
	+
	+ if ( src->format->BitsPerPixel != dst->format->BitsPerPixel ) {
	+ SDL_SetError("Only works with same format surfaces");
	+ return(-1);
	+ }
	+
	+ /* Verify the blit rectangles */
	+ if ( srcrect ) {
	+ rect = *srcrect;
	+#if 0
	+ clip.x = clip.y = 0; clip.h = src->h; clip.w = src->w;
	+#else
	+ clip = src->clip_rect;
	+
	+ if (rect.x > clip.x + clip.w \|\|
	+ rect.x + rect.w < clip.x \|\|
	+ rect.y > clip.y + clip.h \|\|
	+ rect.y + rect.h < clip.y) return 1;
	+ if (rect.x < clip.x) { rect.w -= clip.x-rect.x; rect.x = clip.x; }
	+ if (rect.w > clip.x + clip.w - rect.x)
	+ rect.w = clip.x + clip.w - rect.x;
	+ if (rect.y < clip.y) { rect.h -= clip.y-rect.y; rect.y = clip.y; }
	+ if (rect.h > clip.y + clip.h - rect.y)
	+ rect.h = clip.y + clip.h - rect.y;
	+#endif
	+ if (! rect.h \|\| !rect.w) return 1;
	+ rect.x /= 2; rect.w /= 2;
	+ rect.y /= 2; rect.h /= 2;
	+ clip.x /= 2; clip.w /= 2;
	+ clip.y /= 2; clip.h /= 2;
	+
	+ } else {
	+ rect.x = 0;
	+ rect.y = 0;
	+ rect.w = src->w / 2;
	+ rect.h = src->h / 2;
	+ clip = rect;
	+ }
	+
	+ {
	+ int dest_x = (rect.x - clip.x) * 3;
	+ int dest_y = (rect.y - clip.y) * 3;
	+ int dest_w = rect.w * 3;
	+ int dest_h = rect.h * 3;
	+
	+ if (dest_x >= dst->clip_rect.x + dst->clip_rect.w \|\|
	+ dest_y >= dst->clip_rect.y + dst->clip_rect.h \|\|
	+ dest_x + dest_w <= dst->clip_rect.x \|\|
	+ dest_y + dest_h <= dst->clip_rect.y) return 1;
	+ if (dest_x < dst->clip_rect.x)
	+ {
	+ rect.w = (dest_w - dst->clip_rect.x - dest_x) / 3;
	+ rect.x += (dst->clip_rect.x - dest_x) / 3;
	+ dest_w = rect.w * 3; dest_x = dst->clip_rect.x; }
	+ if (dest_w > dst->clip_rect.x + dst->clip_rect.w - dest_x)
	+ {
	+ rect.w = (dst->clip_rect.x + dst->clip_rect.w - dest_x) / 3;
	+ dest_w = rect.w * 3; }
	+ if (dest_y < dst->clip_rect.y)
	+ {
	+ rect.h = (dest_h - dst->clip_rect.y - dest_y) / 3;
	+ rect.y += (dst->clip_rect.y - dest_y) / 3;
	+ dest_h = rect.h * 3; dest_y = dst->clip_rect.y; }
	+ if (dest_h > dst->clip_rect.y + dst->clip_rect.h - dest_y)
	+ {
	+ rect.h = (dst->clip_rect.y + dst->clip_rect.h - dest_y) / 3;
	+ dest_h = rect.h * 3; }
	+ if (!dest_h \|\| !dest_w \|\| !rect.h \|\| !rect.w) return 1;
	+
	+ if (dstrect) { /returnvalue/
	+ dstrect->y = dest_y;
	+ dstrect->x = dest_x;
	+ dstrect->w = dest_w;
	+ dstrect->h = dest_h;
	+ }
	+
	+ dstp = (Uint8*)dst->pixels +
	+ (dst->pitchdest_y) + (bppdest_x);
	+ srcp = (Uint8*)src->pixels +
	+ (src->pitchrect.y2) + (bpprect.x2);
	+ endp = srcp + src->pitchrect.h2;
	+ srci = src->pitch - (bpprect.w2);
	+ dsti = dst->pitch - (bpp*dest_w);
	+ }
	+
	+#if defined __i386_ \|\| defined __x86_64_
	+#define ASM_i386
	+#endif
	+
	+# if defined __GNUC__ && defined ASM_i386
	+# define ECX __asm__("%ecx")
	+# define ESI __asm__("%esi")
	+# define EDI __asm__("%edi")
	+# else
	+# define ECX
	+# define ESI
	+# define EDI
	+# endif
	+
	+ switch (bpp)
	+ {
	+ case 4:
	+ {
	+ register Uint32* dstP EDI = (Uint32*) dstp;
	+ register Uint32* srcP ESI = (Uint32*) srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/4;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 3:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w23;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 2:
	+ {
	+ register Uint16* dstP EDI = (Uint16*) dstp;
	+ register Uint16* srcP ESI = (Uint16*) srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/2;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 1:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp -= rect.w*2;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ }
	+ return 0;
	+}
	+
	+
	+#elif 1
	+/* a pentium-specific one: using PREFETCH - use with Athlon or Piii only.
	+ *
	+ * PREFETCHn prefetch into cache (Intel 24547112.pdf, page 649)
	+ * MOVNTI store dword using nontemporal hint (page 516)
	+ * generally, noticable effect is only seen with PREFETCHnta and MOVNTI
	+ * for memcpy implementations. However, in SDL we do often blit to a
	+ * surface that will also be used right away, so it is fine to keep it
	+ * in a (distant cache).
	+ */
	+
	+
	+/* Perform a stretch blit between two surfaces of the same format.
	+ NOTE: This function is not safe to call from multiple threads!
	+*/
	+int SDL_StretchSurface_23(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect)
	+{
	+ Uint8 srcp, dstp, *endp;
	+ unsigned int srci, dsti;
	+ SDL_Rect rect;
	+ SDL_Rect clip;
	+ const unsigned bpp = src->format->BytesPerPixel;
	+
	+ if ( src->format->BitsPerPixel != dst->format->BitsPerPixel ) {
	+ SDL_SetError("Only works with same format surfaces");
	+ return(-1);
	+ }
	+
	+ /* Verify the blit rectangles */
	+ if ( srcrect ) {
	+ rect = *srcrect;
	+#if 0
	+ clip.x = clip.y = 0; clip.h = src->h; clip.w = src->w;
	+#else
	+ clip = src->clip_rect;
	+
	+ if (rect.x > clip.x + clip.w \|\|
	+ rect.x + rect.w < clip.x \|\|
	+ rect.y > clip.y + clip.h \|\|
	+ rect.y + rect.h < clip.y) return 1;
	+ if (rect.x < clip.x) { rect.w -= clip.x-rect.x; rect.x = clip.x; }
	+ if (rect.w > clip.x + clip.w - rect.x)
	+ rect.w = clip.x + clip.w - rect.x;
	+ if (rect.y < clip.y) { rect.h -= clip.y-rect.y; rect.y = clip.y; }
	+ if (rect.h > clip.y + clip.h - rect.y)
	+ rect.h = clip.y + clip.h - rect.y;
	+#endif
	+ if (! rect.h \|\| !rect.w) return 1;
	+ rect.x /= 2; rect.w /= 2;
	+ rect.y /= 2; rect.h /= 2;
	+ clip.x /= 2; clip.w /= 2;
	+ clip.y /= 2; clip.h /= 2;
	+
	+ } else {
	+ rect.x = 0;
	+ rect.y = 0;
	+ rect.w = src->w / 2;
	+ rect.h = src->h / 2;
	+ clip = rect;
	+ }
	+
	+ {
	+ int dest_x = (rect.x - clip.x) * 3;
	+ int dest_y = (rect.y - clip.y) * 3;
	+ int dest_w = rect.w * 3;
	+ int dest_h = rect.h * 3;
	+
	+ if (dest_x >= dst->clip_rect.x + dst->clip_rect.w \|\|
	+ dest_y >= dst->clip_rect.y + dst->clip_rect.h \|\|
	+ dest_x + dest_w <= dst->clip_rect.x \|\|
	+ dest_y + dest_h <= dst->clip_rect.y) return 1;
	+ if (dest_x < dst->clip_rect.x)
	+ {
	+ rect.w = (dest_w - dst->clip_rect.x - dest_x) / 3;
	+ rect.x += (dst->clip_rect.x - dest_x) / 3;
	+ dest_w = rect.w * 3; dest_x = dst->clip_rect.x; }
	+ if (dest_w > dst->clip_rect.x + dst->clip_rect.w - dest_x)
	+ {
	+ rect.w = (dst->clip_rect.x + dst->clip_rect.w - dest_x) / 3;
	+ dest_w = rect.w * 3; }
	+ if (dest_y < dst->clip_rect.y)
	+ {
	+ rect.h = (dest_h - dst->clip_rect.y - dest_y) / 3;
	+ rect.y += (dst->clip_rect.y - dest_y) / 3;
	+ dest_h = rect.h * 3; dest_y = dst->clip_rect.y; }
	+ if (dest_h > dst->clip_rect.y + dst->clip_rect.h - dest_y)
	+ {
	+ rect.h = (dst->clip_rect.y + dst->clip_rect.h - dest_y) / 3;
	+ dest_h = rect.h * 3; }
	+ if (!dest_h \|\| !dest_w \|\| !rect.h \|\| !rect.w) return 1;
	+
	+ if (dstrect) { /returnvalue/
	+ dstrect->y = dest_y;
	+ dstrect->x = dest_x;
	+ dstrect->w = dest_w;
	+ dstrect->h = dest_h;
	+ }
	+
	+ dstp = (Uint8*)dst->pixels +
	+ (dst->pitchdest_y) + (bppdest_x);
	+ srcp = (Uint8*)src->pixels +
	+ (src->pitchrect.y2) + (bpprect.x2);
	+ endp = srcp + src->pitchrect.h2;
	+ srci = src->pitch - (bpprect.w2);
	+ dsti = dst->pitch - (bpp*dest_w);
	+ }
	+
	+# if defined __GNUC__ && defined ASM_i386
	+# define ECX __asm__("%ecx")
	+# define ESI __asm__("%esi")
	+# define EDI __asm__("%edi")
	+# else
	+# define ECX
	+# define ESI
	+# define EDI
	+# endif
	+
	+ switch (bpp)
	+ {
	+ case 4:
	+ {
	+ register Uint32* dstP EDI = (Uint32*) dstp;
	+ register Uint32* srcP ESI = (Uint32*) srcp;
	+ do {
	+ register int count ECX = rect.w;
	+# if defined __GNUC__ && defined ASM_i386
	+ if (rect.w^(rect.w&255))
	+ __asm__ __volatile__ ("prefetchnta 320(%0)\n"
	+ :: "r" (srcP));
	+# endif
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/4;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 3:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w23;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 2:
	+ {
	+ register Uint16* dstP EDI = (Uint16*) dstp;
	+ register Uint16* srcP ESI = (Uint16*) srcp;
	+# if defined __GNUC__ && defined ASM_i386
	+ if (rect.w^(rect.w&255))
	+ __asm__ __volatile__ ("prefetchnta 512(%0)\n"
	+ :: "r" (srcP));
	+# endif
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/2;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 1:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp -= rect.w*2;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ }
	+ return 0;
	+}
	+
	+#elif 1
	+
	+/* that routine is not a perfect stretch either - hopefully you have
	+ * already noticed it: we have implicitly cut everything to div-2,
	+ * the src-start and src-width are all on even columns and rows and
	+ * the stretch algorithm will always do XY -> XXY and never the way
	+ * of the odd rows/columns with a XY -> XYY copy operation. If we
	+ * want to fix it but still want to keep the copy-loop as short as
	+ * ever possible then we need to make an outer switch of 4x4 = 16 cases
	+ * given that src-x/src-y can be each even/odd (four combinations)
	+ * and that src-w/src-h be each even/odd as well (four combinations).
	+ * Along with four bytedepths, we would come out at 64 copy-loop
	+ * definitions but in reality there are 128 copy-loops. Guess why.
	+ * oddX
	+ * only
	+ */
	+
	+/* Perform a stretch blit between two surfaces of the same format.
	+ NOTE: This function is not safe to call from multiple threads!
	+*/
	+int SDL_StretchSurface_23(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect)
	+{
	+ Uint8 srcp, dstp, *endp;
	+ unsigned int srci, dsti;
	+ SDL_Rect rect;
	+ SDL_Rect clip;
	+ const unsigned bpp = src->format->BytesPerPixel;
	+ Uint8 oddX = 0;
	+
	+ if ( src->format->BitsPerPixel != dst->format->BitsPerPixel ) {
	+ SDL_SetError("Only works with same format surfaces");
	+ return(-1);
	+ }
	+
	+ /* Verify the blit rectangles */
	+ if ( srcrect ) {
	+ rect = *srcrect;
	+#if 0
	+ clip.x = clip.y = 0; clip.h = src->h; clip.w = src->w;
	+#else
	+ clip = src->clip_rect;
	+
	+ if (rect.x > clip.x + clip.w \|\|
	+ rect.x + rect.w < clip.x \|\|
	+ rect.y > clip.y + clip.h \|\|
	+ rect.y + rect.h < clip.y) return 1;
	+ if (rect.x < clip.x) { rect.w -= clip.x-rect.x; rect.x = clip.x; }
	+ if (rect.w > clip.x + clip.w - rect.x)
	+ rect.w = clip.x + clip.w - rect.x;
	+ if (rect.y < clip.y) { rect.h -= clip.y-rect.y; rect.y = clip.y; }
	+ if (rect.h > clip.y + clip.h - rect.y)
	+ rect.h = clip.y + clip.h - rect.y;
	+#endif
	+ if (! rect.h \|\| !rect.w) return 1;
	+ oddX = rect.x&1;
	+ rect.x /= 2;
	+ rect.y /= 2;
	+ rect.w /= 2;
	+ rect.h /= 2;
	+ } else {
	+ rect.x = 0;
	+ rect.y = 0;
	+ rect.w = src->w / 2;
	+ rect.h = src->h / 2;
	+ clip = rect;
	+ }
	+
	+ {
	+ int dest_x = (rect.x - clip.x/2) * 3 + oddX;
	+ int dest_y = (rect.y - clip.y/2) * 3;
	+ int dest_w = rect.w * 3;
	+ int dest_h = rect.h * 3;
	+
	+ if (dest_x >= dst->clip_rect.x + dst->clip_rect.w \|\|
	+ dest_y >= dst->clip_rect.y + dst->clip_rect.h \|\|
	+ dest_x + dest_w <= dst->clip_rect.x \|\|
	+ dest_y + dest_h <= dst->clip_rect.y) return 1;
	+ if (dest_x < dst->clip_rect.x)
	+ {
	+ rect.w = (dest_w - dst->clip_rect.x - dest_x) / 3;
	+ rect.x += (dst->clip_rect.x - dest_x) / 3; oddX = 0;
	+ dest_w = rect.w * 3; dest_x = dst->clip_rect.x; }
	+ if (dest_w > dst->clip_rect.x + dst->clip_rect.w - dest_x)
	+ {
	+ rect.w = (dst->clip_rect.x + dst->clip_rect.w - dest_x) / 3;
	+ dest_w = rect.w * 3; }
	+ if (dest_y < dst->clip_rect.y)
	+ {
	+ rect.h = (dest_h - dst->clip_rect.y - dest_y) / 3;
	+ rect.y += (dst->clip_rect.y - dest_y) / 3;
	+ dest_h = rect.h * 3; dest_y = dst->clip_rect.y; }
	+ if (dest_h > dst->clip_rect.y + dst->clip_rect.h - dest_y)
	+ {
	+ rect.h = (dst->clip_rect.y + dst->clip_rect.h - dest_y) / 3;
	+ dest_h = rect.h * 3; }
	+ if (!dest_h \|\| !dest_w \|\| !rect.h \|\| !rect.w) return 1;
	+
	+ if (dstrect) { /returnvalue/
	+ dstrect->y = dest_y;
	+ dstrect->x = dest_x;
	+ dstrect->w = dest_w;
	+ dstrect->h = dest_h;
	+ }
	+
	+ dstp = (Uint8*)dst->pixels +
	+ (dst->pitchdest_y) + (bppdest_x);
	+ srcp = (Uint8*)src->pixels +
	+ (src->pitchrect.y2) + (bpp(rect.x2+oddX));
	+ endp = srcp + src->pitchrect.h2;
	+ srci = src->pitch - (bpprect.w2);
	+ dsti = dst->pitch - (bpp*dest_w);
	+ }
	+
	+# if defined __GNUC__ && defined ASM_i386
	+# define ECX __asm__("%ecx")
	+# define ESI __asm__("%esi")
	+# define EDI __asm__("%edi")
	+# else
	+# define ECX
	+# define ESI
	+# define EDI
	+# endif
	+
	+ if (oddX) goto odd_X_even_Y;
	+
	+ switch (bpp)
	+ {
	+ case 4:
	+ {
	+ register Uint32* dstP EDI = (Uint32*) dstp;
	+ register Uint32* srcP ESI = (Uint32*) srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/4;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 3:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w23;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 2:
	+ {
	+ register Uint16* dstP EDI = (Uint16*) dstp;
	+ register Uint16* srcP ESI = (Uint16*) srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/2;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 1:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp -= rect.w*2;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ }
	+ return 0;
	+
	+odd_X_even_Y:
	+ switch (bpp)
	+ {
	+ case 4:
	+ {
	+ register Uint32* dstP EDI = (Uint32*) dstp;
	+ register Uint32* srcP ESI = (Uint32*) srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsl\n" "lodsl\n" "stosl\n" "stosl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/4;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsl\n" "lodsl\n" "stosl\n" "stosl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsl\n" "lodsl\n" "stosl\n" "stosl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 3:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ } while (--count != 0);
	+ srcP -= rect.w23;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 2:
	+ {
	+ register Uint16* dstP EDI = (Uint16*) dstp;
	+ register Uint16* srcP ESI = (Uint16*) srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "lodsw\n" "stosw\n" "stosw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/2;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "lodsw\n" "stosw\n" "stosw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "lodsw\n" "stosw\n" "stosw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 1:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsb\n" "lodsb\n" "stosb\n" "stosb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp -= rect.w*2;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsb\n" "lodsb\n" "stosb\n" "stosb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsb\n" "lodsb\n" "stosb\n" "stosb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ }
	+ return 0;
	+}
	+
	+#elif 1
	+
	+/* that routine is not a perfect stretch either - hopefully you have
	+ * already noticed it: we have implicitly cut everything to div-2,
	+ * the src-start and src-width are all on even columns and rows and
	+ * the stretch algorithm will always do XY -> XXY and never the way
	+ * of the odd rows/columns with a XY -> XYY copy operation. If we
	+ * want to fix it but still want to keep the copy-loop as short as
	+ * ever possible then we need to make an outer switch of 4x4 = 16 cases
	+ * given that src-x/src-y can be each even/odd (four combinations)
	+ * and that src-w/src-h be each even/odd as well (four combinations).
	+ * Along with four bytedepths, we would come out at 64 copy-loop
	+ * definitions but in reality there are 128 copy-loops. Guess why.
	+ * [[ ALL ]]
	+ * unfinished
	+ */
	+
	+/* Perform a stretch blit between two surfaces of the same format.
	+ NOTE: This function is not safe to call from multiple threads!
	+*/
	+int SDL_StretchSurface_23(SDL_Surface src, SDL_Rect srcrect,
	+ SDL_Surface dst, SDL_Rect dstrect)
	+{
	+ Uint8 srcp, dstp, *endp;
	+ unsigned int srci, dsti;
	+ SDL_Rect rect;
	+ SDL_Rect clip;
	+ const unsigned bpp = src->format->BytesPerPixel;
	+ Uint8 kind = 0;
	+
	+ if ( src->format->BitsPerPixel != dst->format->BitsPerPixel ) {
	+ SDL_SetError("Only works with same format surfaces");
	+ return(-1);
	+ }
	+
	+ if (0) fprintf(stderr,
	+ "[%4i +%4i /%4i +%4i ] ->"
	+ "[%4i+%4i/%4i+%4i]\n",
	+ srcrect->x, srcrect->w,
	+ srcrect->y, srcrect->h,
	+ dst->clip_rect.x, dst->clip_rect.w,
	+ dst->clip_rect.y, dst->clip_rect.h);
	+
	+ /* Verify the blit rectangles */
	+ if ( srcrect ) {
	+ rect = *srcrect;
	+#if 0
	+ clip.x = clip.y = 0; clip.h = src->h; clip.w = src->w;
	+#else
	+ clip = src->clip_rect;
	+
	+ if (rect.x > clip.x + clip.w \|\|
	+ rect.x + rect.w < clip.x \|\|
	+ rect.y > clip.y + clip.h \|\|
	+ rect.y + rect.h < clip.y) return 1;
	+ if (rect.x < clip.x) { rect.w -= clip.x-rect.x; rect.x = clip.x; }
	+ if (rect.w > clip.x + clip.w - rect.x)
	+ rect.w = clip.x + clip.w - rect.x;
	+ if (rect.y < clip.y) { rect.h -= clip.y-rect.y; rect.y = clip.y; }
	+ if (rect.h > clip.y + clip.h - rect.y)
	+ rect.h = clip.y + clip.h - rect.y;
	+#endif
	+ if (! rect.h \|\| !rect.w) return 1;
	+ kind \|= rect.x&1; rect.x /= 2; kind<<=1;
	+ kind \|= rect.y&1; rect.y /= 2; kind<<=1;
	+ kind \|= rect.w&1; rect.w /= 2; kind<<=1;
	+ kind \|= rect.h&1; rect.h /= 2;
	+ } else {
	+ rect.x = 0;
	+ rect.y = 0; kind \|= src->w&1; kind<<=1;
	+ rect.w = src->w / 2; kind \|= src->h&1;
	+ rect.h = src->h / 2;
	+ clip = rect;
	+ }
	+
	+/* bitmasks */
	+# define _H 1
	+# define _W 2
	+# define _Y 4
	+# define _X 8
	+/* shiftvals */
	+# define _H_ 0
	+# define _W_ 1
	+# define _Y_ 2
	+# define _X_ 3
	+
	+ {
	+ int dest_x = (rect.x - clip.x/2) * 3;
	+ int dest_y = (rect.y - clip.y/2) * 3;
	+ int dest_w = rect.w * 3;
	+ int dest_h = rect.h * 3;
	+ if (0) fprintf(stderr,
	+ "[%3i2%c+%3i2%c/%3i2%c+%3i2%c] ->"
	+ "[%4i+%4i/%4i+%4i]\n",
	+ rect.x,(kind&_X)?'>':'.',rect.w,(kind&_W)?'>':'.',
	+ rect.y,(kind&_Y)?'>':'.',rect.h,(kind&_H)?'>':'.',
	+ dest_x,dest_w, dest_y,dest_h);
	+ switch(kind) /* a switch statement is not an if() - it turns */
	+ { /* out to be a table lookup operation actually */
	+ case _X\|_Y\|_W\|_H: dest_h ++; />>/
	+ case _X\|_Y\|_W: dest_w ++; />>/
	+ case _X\|_Y: dest_y ++; />>/
	+ case _X: dest_x ++; break;
	+ case _Y\|_W\|_H: dest_h ++; />>/
	+ case _Y\|_W: dest_w ++; />>/
	+ case _Y: dest_x ++; break;
	+ case _W\|_X: dest_x ++; goto case__W;
	+ case _W\|_H\|_X: dest_x ++; />>/
	+ case _W\|_H: dest_h ++; />>/
	+ case__W:
	+ case _W: dest_w ++; break;
	+ case _H\|_Y: dest_y ++; goto case__H;
	+ case _H\|_X\|_Y: dest_y ++; />>/
	+ case _H\|_X: dest_x ++; />>/
	+ case__H:
	+ case _H: dest_h ++; break;
	+ case 0: break;
	+ }
	+ if (0) fprintf(stderr,
	+ "[%3i2%c+%3i2%c/%3i2%c+%3i2%c] ->"
	+ "[%4i+%4i/%4i+%4i] !\n",
	+ rect.x,(kind&_X)?'>':'.',rect.w,(kind&_W)?'>':'.',
	+ rect.y,(kind&_Y)?'>':'.',rect.h,(kind&_H)?'>':'.',
	+ dest_x,dest_w, dest_y,dest_h);
	+
	+ if (dest_x >= dst->clip_rect.x + dst->clip_rect.w \|\|
	+ dest_y >= dst->clip_rect.y + dst->clip_rect.h \|\|
	+ dest_x + dest_w <= dst->clip_rect.x \|\|
	+ dest_y + dest_h <= dst->clip_rect.y) return 1;
	+ if (dest_x < dst->clip_rect.x)
	+ {
	+ rect.w = (dest_w - dst->clip_rect.x - dest_x) / 3;
	+ rect.x += (dst->clip_rect.x - dest_x) / 3;
	+ dest_w = rect.w * 3; dest_x = dst->clip_rect.x;
	+ kind &= _X\|_Y\|_H; }
	+ if (dest_w > dst->clip_rect.x + dst->clip_rect.w - dest_x)
	+ {
	+ rect.w = (dst->clip_rect.x + dst->clip_rect.w - dest_x) / 3;
	+ dest_w = rect.w * 3;
	+ kind &= _X\|_Y\|_H; }
	+ if (dest_y < dst->clip_rect.y)
	+ {
	+ rect.h = (dest_h - dst->clip_rect.y - dest_y) / 3;
	+ rect.y += (dst->clip_rect.y - dest_y) / 3;
	+ dest_h = rect.h * 3; dest_y = dst->clip_rect.y;
	+ kind &= _X\|_Y\|_W; }
	+ if (dest_h > dst->clip_rect.y + dst->clip_rect.h - dest_y)
	+ {
	+ rect.h = (dst->clip_rect.y + dst->clip_rect.h - dest_y) / 3;
	+ dest_h = rect.h * 3;
	+ kind &= _X\|_Y\|_W; }
	+ if (!dest_h \|\| !dest_w \|\| !rect.h \|\| !rect.w) return 1;
	+
	+ if (dstrect) { /returnvalue/
	+ dstrect->y = dest_y;
	+ dstrect->x = dest_x;
	+ dstrect->w = dest_w;
	+ dstrect->h = dest_h;
	+ }
	+
	+ if (0) fprintf(stderr,
	+ "[%3i2%c+%3i2%c/%3i2%c+%3i2%c] ->"
	+ "[%4i+%4i/%4i+%4i] !!\n",
	+ rect.x,(kind&_X)?'>':'.',rect.w,(kind&_W)?'>':'.',
	+ rect.y,(kind&_Y)?'>':'.',rect.h,(kind&_H)?'>':'.',
	+ dest_x,dest_w, dest_y,dest_h);
	+
	+ dstp = (Uint8*)dst->pixels +
	+ (dst->pitchdest_y) + (bppdest_x);
	+ srcp = (Uint8*)src->pixels +
	+ (src->pitchrect.y2) + (bpprect.x2);
	+ endp = srcp + src->pitchrect.h2;
	+ srci = src->pitch - (bpprect.w2);
	+ dsti = dst->pitch - (bpp*dest_w);
	+ }
	+
	+# if defined __GNUC__ && defined ASM_i386
	+# define ECX __asm__("%ecx")
	+# define ESI __asm__("%esi")
	+# define EDI __asm__("%edi")
	+# else
	+# define ECX
	+# define ESI
	+# define EDI
	+# endif
	+
	+ if (kind&_X) goto odd_X_even_Y;
	+
	+ switch (bpp)
	+ {
	+ case 4:
	+ {
	+ register Uint32* dstP EDI = (Uint32*) dstp;
	+ register Uint32* srcP ESI = (Uint32*) srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/4;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsl\n" "stosl\n" "stosl\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 3:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w23;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "movsl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 2:
	+ {
	+ register Uint16* dstP EDI = (Uint16*) dstp;
	+ register Uint16* srcP ESI = (Uint16*) srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/2;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsw\n" "stosw\n" "stosw\n" "movsw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 1:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp -= rect.w*2;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("lodsb\n" "stosb\n" "stosb\n" "movsb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ }
	+ return 0;
	+
	+odd_X_even_Y:
	+ switch (bpp)
	+ {
	+ case 4:
	+ {
	+ register Uint32* dstP EDI = (Uint32*) dstp;
	+ register Uint32* srcP ESI = (Uint32*) srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsl\n" "lodsl\n" "stosl\n" "stosl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/4;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsl\n" "lodsl\n" "stosl\n" "stosl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsl\n" "lodsl\n" "stosl\n" "stosl\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/4; /* gcc is clever enough to make */
	+ dstP += dsti/4; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 3:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ } while (--count != 0);
	+ srcP -= rect.w23;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ count = rect.w;
	+ do {
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP[0] = srcP[0];
	+ dstP[1] = srcP[1];
	+ dstP[2] = srcP[2];
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ dstP++ = srcP++;
	+ } while (--count != 0);
	+ srcP += srci;
	+ dstP += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 2:
	+ {
	+ register Uint16* dstP EDI = (Uint16*) dstp;
	+ register Uint16* srcP ESI = (Uint16*) srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "lodsw\n" "stosw\n" "stosw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP -= rect.w*2;
	+ dstP += dsti/2;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "lodsw\n" "stosw\n" "stosw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsw\n" "lodsw\n" "stosw\n" "stosw\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcP += srci/2; /* gcc is clever enough to make */
	+ dstP += dsti/2; /* it really a "ADD dsti, dstP" */
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ case 1:
	+ {
	+ register Uint8* dstP EDI = dstp;
	+ register Uint8* srcP ESI = srcp;
	+ do {
	+ register int count ECX = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsb\n" "lodsb\n" "stosb\n" "stosb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp -= rect.w*2;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsb\n" "lodsb\n" "stosb\n" "stosb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ count = rect.w;
	+ do {
	+# if defined __GNUC__ && defined ASM_i386
	+ __asm__ __volatile__ ("movsb\n" "lodsb\n" "stosb\n" "stosb\n"
	+ :: "S" (srcP), "D" (dstP) : "%eax");
	+# else
	+ dstP++ = srcP++;
	+ dstP++ = srcP;
	+ dstP++ = srcP++;
	+# endif
	+ } while (--count != 0);
	+ srcp += srci;
	+ dstp += dsti;
	+ } while ((Uint8*)srcP != endp);
	+ } break;
	+ }
	+ return 0;
	+}
	+
	+#endif
	+
	+/* --------------------------------------------------------------------- */
	+/* RowCode */
	+/* --------------------------------------------------------------------- */
	diff --git a/util/sdl/stretch/sdlstretchcode.c b/util/sdl/stretch/sdlstretchcode.c
	new file mode 100644
	index 00000000..bf929128
	--- /dev/null
	+++ b/util/sdl/stretch/sdlstretchcode.c
	@@ -0,0 +1,381 @@
	+/*
	+ SDL_stretch - Stretch Functions For The Simple DirectMedia Layer
	+ Copyright (C) 2003 Guido Draheim
	+
	+ This library is free software; you can redistribute it and/or
	+ modify it under the terms of the GNU Library General Public
	+ License as published by the Free Software Foundation; either
	+ version 2 of the License, or (at your option) any later version.
	+
	+ This library 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
	+ Library General Public License for more details.
	+
	+ You should have received a copy of the GNU Library General Public
	+ License along with this library; if not, write to the Free
	+ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
	+
	+ Guido Draheim, guidod@gmx.de
	+*/
	+
	+#ifdef SAVE_RCSID
	+static char rcsid =
	+ "@(#) $Id: sdlcode.c $";
	+#endif
	+
	+#include "config.h"
	+#include "SDL/SDL_error.h"
	+#include "SDL/SDL_video.h"
	+#include "SDL_stretchcode.h"
	+#include "SDL_stretchasm.h"
	+
	+#ifdef SDL_STRETCH_DATA_NOEXEC
	+#ifdef __unix__
	+#include <unistd.h>
	+#include <sys/mman.h>
	+#endif
	+#include <malloc.h>
	+#endif
	+
	+#ifdef SDL_STRETCH_CALL
	+#if defined SDL_STRETCH_I386 \|\| defined SDL_STRETCH_X86_64
	+/*
	+ * in AMD64 "long mode" all traditional opcodes stay 32bit
	+ * and 64bit cousins have to be encoded via REX prefixes.
	+ * The only exception are PUSH/POP/RET opcodes that will
	+ * implicitly be 64bit. This makes it so the following
	+ * opcode settings work in both 32bit and 64bit mode but
	+ * for one case: the "INCSI" bit series is dropped and used
	+ * for the "REX" prefix opcodes in AMD64 long mode.
	+ */
	+#define iLOADSB 0xAC
	+#define iLOADSW 0xAD
	+#define iSTOSB 0xAA
	+#define iSTOSW 0xAB
	+#define iPUSHCX 0x51
	+#define iPOPCX 0x59
	+#define iINCSI 0x46
	+#define iXCHGCXAX 0x91
	+#define iPREFIX16 0x66
	+#define iRET 0xC3
	+/*
	+ * does XCHG yield a fast register renaming on P >= P3 ?
	+ */
	+# define PREP_1BYTE(dst)
	+# define PREP_2BYTE(dst)
	+# define PREP_3BYTE(dst) { *dst++ = iPUSHCX; }
	+# define PREP_4BYTE(dst)
	+# define LOAD_1BYTE(dst) { *dst++ = iLOADSB; }
	+# define LOAD_2BYTE(dst) { dst++ = iPREFIX16; dst++ = iLOADSW; }
	+# define LOAD_3BYTE(dst) { dst++ = iLOADSB; dst++ = iXCHGCXAX; \
	+ dst++ = iPREFIX16; dst++ = iLOADSW; }
	+# define LOAD_4BYTE(dst) { *dst++ = iLOADSW; }
	+# define STORE_1BYTE(dst) { *dst++ = iSTOSB; }
	+# define STORE_2BYTE(dst) { dst++ = iPREFIX16; dst++ = iSTOSW; }
	+# define STORE_3BYTE(dst) { dst++ = iXCHGCXAX; dst++ = iSTOSB; *dst++ = iXCHGCXAX; \
	+ dst++ = iPREFIX16; dst++ = iSTOSW; }
	+# define STORE_4BYTE(dst) { *dst++ = iSTOSW; }
	+# define RETURN_1BYTE(dst) { *dst++ = iRET; }
	+# define RETURN_2BYTE(dst) { *dst++ = iRET; }
	+# define RETURN_3BYTE(dst) { dst++ = iPOPCX; dst++ = iRET; }
	+# define RETURN_4BYTE(dst) { *dst++ = iRET; }
	+# ifdef SDL_STRETCH_I386
	+# define SKIP_1BYTE(dst) { *dst++ = iINCSI; }
	+# define SKIP_2BYTE(dst) { dst++ = iINCSI; dst++ = iINCSI; }
	+# define SKIP_3BYTE(dst) { dst++ = iINCSI; dst++ = iINCSI; *dst++ = iINCSI; }
	+#endif
	+#define HAVE_NATIVE_CODE
	+#endif
	+#endif
	+
	+/* ASM i386 */
	+#ifndef HAVE_NATIVE_CODE
	+#define PREP_1BYTE(dst)
	+#define PREP_2BYTE(dst)
	+#define PREP_3BYTE(dst)
	+#define PREP_4BYTE(dst)
	+#define LOAD_1BYTE(dst) { *dst++ = 001; }
	+#define LOAD_2BYTE(dst) { *dst++ = 002; }
	+#define LOAD_3BYTE(dst) { *dst++ = 003; }
	+#define LOAD_4BYTE(dst) { *dst++ = 004; }
	+#define STORE_1BYTE(dst) { *dst++ = 011; }
	+#define STORE_2BYTE(dst) { *dst++ = 012; }
	+#define STORE_3BYTE(dst) { *dst++ = 013; }
	+#define STORE_4BYTE(dst) { *dst++ = 014; }
	+#define RETURN_1BYTE(dst) { *dst++ = 0; }
	+#define RETURN_2BYTE(dst) { *dst++ = 0; }
	+#define RETURN_3BYTE(dst) { *dst++ = 0; }
	+#define RETURN_4BYTE(dst) { *dst++ = 0; }
	+#define HAVE_INTERPRETED_CODE
	+#endif
	+
	+#ifndef SKIP_1BYTE
	+#define SKIP_1BYTE(dst) LOAD_1BYTE(dst)
	+#endif
	+#ifndef SKIP_2BYTE
	+#define SKIP_2BYTE(dst) LOAD_2BYTE(dst)
	+#endif
	+#ifndef SKIP_3BYTE
	+#define SKIP_3BYTE(dst) LOAD_3BYTE(dst)
	+#endif
	+#ifndef SKIP_4BYTE
	+#define SKIP_4BYTE(dst) LOAD_4BYTE(dst)
	+#endif
	+
	+#ifdef SDL_STRETCH_DATA_NOEXEC
	+unsigned char* SDL_TheRowStretchCode = NULL;
	+#else
	+unsigned char SDL_TheRowStretchCode[SDL_STRETCHCODE_BUFFERSIZE];
	+#endif
	+
	+unsigned char* SDL_NewRowStretchCode (unsigned size) {
	+ if (size == 0) size = SDL_STRETCHCODE_BUFFERSIZE;
	+# if defined SDL_STRETCH_DATA_NOEXEC && defined __unix__
	+ unsigned char* buffer = memalign(getpagesize(), SDL_STRETCHCODE_BUFFERSIZE);
	+ if (-1 == mprotect(buffer, SDL_STRETCHCODE_BUFFERSIZE, PROT_READ\|PROT_WRITE\|PROT_EXEC)) {
	+ free (buffer);
	+ SDL_SetError("SDL_TheRowStretchCode - can not allocate and make executable");
	+ return 0;
	+ }
	+# else
	+ unsigned char* buffer = malloc(SDL_STRETCHCODE_BUFFERSIZE);
	+# endif
	+ return buffer;
	+}
	+
	+unsigned char* SDL_GetRowStretchCode (void) {
	+#ifdef SDL_STRETCH_DATA_NOEXEC
	+ if (SDL_TheRowStretchCode == NULL) {
	+ SDL_TheRowStretchCode = SDL_NewRowStretchCode(0);
	+ }
	+#endif
	+ return SDL_TheRowStretchCode;
	+}
	+
	+unsigned char* SDL_SetRowStretchCode (int src_w, int dst_w, int bpp)
	+{
	+#ifdef SDL_STRETCH_DATA_NOEXEC
	+ if (SDL_TheRowStretchCode == NULL) {
	+ SDL_TheRowStretchCode = SDL_GetRowStretchCode();
	+ }
	+#endif
	+ return SDL_PutRowStretchCode(SDL_TheRowStretchCode,
	+ sizeof(SDL_TheRowStretchCode),
	+ src_w, dst_w, bpp);
	+}
	+
	+/*
	+ * The computed stretch code represents a kind of Bresenham x/y line
	+ * algorithm where our stretching factor matches with the steepness.
	+ */
	+unsigned char* SDL_PutRowStretchCode(unsigned char* buf_ptr, int buf_len,
	+ int src_w, int dst_w, int bpp)
	+{
	+ static struct {
	+ int bpp;
	+ int src_w;
	+ int dst_w;
	+ } last;
	+
	+ int i = 0; int w = dst_w; int src = 0;
	+ unsigned char *buf = buf_ptr;
	+ unsigned char *buf_end = buf + buf_len;
	+
	+ if (last.src_w == src_w && last.dst_w == dst_w && last.bpp == bpp)
	+ { /* so, we do not need to regenerate the code buffer */
	+ return 0;
	+ }else{
	+ last.src_w = src_w; last.dst_w = dst_w; last.bpp = bpp;
	+ }
	+
	+ switch (bpp)
	+ {
	+ case 4:
	+ PREP_4BYTE(buf);
	+ while(src<src_w) {
	+ LOAD_4BYTE(buf);
	+ draw4:
	+ STORE_4BYTE(buf);
	+ --w; i += src_w;
	+ if (! w) break;
	+ if (buf >= buf_end) goto overflow;
	+ if (i < dst_w) goto draw4;
	+ i -= dst_w; src++;
	+ while (i >= dst_w) { i -= dst_w; src++; SKIP_4BYTE(buf); }
	+ };
	+ RETURN_4BYTE(buf);
	+ break;
	+ case 3:
	+ PREP_3BYTE(buf);
	+ while(src<src_w) {
	+ LOAD_3BYTE(buf);
	+ draw3:
	+ STORE_3BYTE(buf);
	+ --w; i += src_w;
	+ if (! w) break;
	+ if (buf >= buf_end) goto overflow;
	+ if (i < dst_w) goto draw3;
	+ i -= dst_w; src++;
	+ while (i >= dst_w) { i -= dst_w; src++; SKIP_3BYTE(buf); }
	+ };
	+ RETURN_3BYTE(buf);
	+ break;
	+ case 2:
	+ PREP_2BYTE(buf);
	+ while(src<src_w) {
	+ LOAD_2BYTE(buf);
	+ draw2:
	+ STORE_2BYTE(buf);
	+ --w; i += src_w;
	+ if (! w) break;
	+ if (buf >= buf_end) goto overflow;
	+ if (i < dst_w) goto draw2;
	+ i -= dst_w; src++;
	+ while (i >= dst_w) { i -= dst_w; src++; SKIP_2BYTE(buf); }
	+ };
	+ RETURN_2BYTE(buf);
	+ break;
	+ case 1:
	+ PREP_1BYTE(buf);
	+ while(src<src_w) {
	+ LOAD_1BYTE(buf);
	+ draw1:
	+ STORE_1BYTE(buf);
	+ --w; i += src_w;
	+ if (! w) break;
	+ if (buf >= buf_end) goto overflow;
	+ if (i < dst_w) goto draw1;
	+ i -= dst_w; src++;
	+ while (i >= dst_w) { i -= dst_w; src++; SKIP_1BYTE(buf); }
	+ };
	+ RETURN_1BYTE(buf);
	+ break;
	+ default:
	+ SDL_SetError("ASM stretch of %d bytes isn't supported\n", bpp);
	+ return 0;
	+ }
	+# if 0 /* debugging - gives a nice visual representation of the code */
	+ *buf = 0; buf = SDL_TheRowStretchCode; fprintf (stderr,"\n");
	+ while (*buf) {
	+ fprintf(stderr,"%c", (*buf++)&0x7F);
	+ if (! (((long)buf)&0x3F)) fprintf (stderr, "\n");
	+ } fprintf(stderr,"(%iB=(%i->%i))\n", buf-SDL_TheRowStretchCode,
	+ src_w, dst_w);
	+# endif
	+ return buf_ptr; /* SUCCESS */
	+
	+ overflow:
	+ /* so, we did overflow - too late anyway */
	+ SDL_SetError("Copy buffer overflow");
	+ return 0;
	+}
	+
	+
	+void SDL_RunRowStretchCode(unsigned char* code,
	+ unsigned char* srcp,
	+ unsigned char* dstp)
	+{
	+# if defined HAVE_NATIVE_CODE && defined SDL_STRETCH_CALL
	+ SDL_STRETCH_CALL(code, srcp, dstp);
	+# elif defined HAVE_INTERPRETED_CODE
	+ unsigned char a = 0, b = 0, c = 0, d= 0;
	+ while (1) {
	+ switch (*code++) {
	+ case 014: *dstp++ = a;
	+ case 013: *dstp++ = b;
	+ case 012: *dstp++ = c;
	+ case 011: *dstp++ = d;
	+ break;
	+ case 004: a = *srcp++;
	+ case 003: b = *srcp++;
	+ case 002: c = *srcp++;
	+ case 001: d = *srcp++;
	+ break;
	+ default:
	+ return;
	+ }
	+ }
	+# else
	+# error Need inline assembly for this compiler
	+# endif
	+}
	+
	+#define DEFINE_COPY_ROW(name, type) \
	+void name(type src, int src_w, type dst, int dst_w) \
	+{ \
	+ int i = 0; \
	+ int w = dst_w; \
	+ type pixel = 0; \
	+ type* src_max = src + src_w; \
	+ \
	+ while(src < src_max) { \
	+ pixel = *src; \
	+ draw: \
	+ *dst++ = pixel; \
	+ --w; i += src_w; \
	+ if (! w) break; \
	+ if (i < dst_w) goto draw; \
	+ do { i -= dst_w; src++; } while (i >= dst_w);\
	+ }; return; \
	+}
	+DEFINE_COPY_ROW(SDL_StretchRow1, Uint8)
	+DEFINE_COPY_ROW(SDL_StretchRow2, Uint16)
	+DEFINE_COPY_ROW(SDL_StretchRow4, Uint32)
	+
	+/* The ASM code has a hard time to handle 24-bpp stretch blits */
	+void SDL_StretchRow3(Uint8 src, int src_w, Uint8 dst, int dst_w)
	+{
	+ int i = 0;
	+ int w = dst_w;
	+ Uint8 pixel [3];
	+ Uint8* src_max = src + src_w;
	+
	+ while(src < src_max) {
	+ pixel[0] = *src++;
	+ pixel[1] = *src++;
	+ pixel[2] = *src++;
	+ draw:
	+ *dst++ = pixel[0];
	+ *dst++ = pixel[1];
	+ *dst++ = pixel[2];
	+ --w ; i += src_w;
	+ if (! w) break;
	+ if (i < dst_w) goto draw; /* draw same pixel again */
	+ do { i -= dst_w; src+=3; } while (i >= dst_w);
	+ }; return;
	+}
	+
	+char* SDL_StretchRowInfo(void) {
	+ return "sdlstretch"
	+#ifdef SDL_STRETCH_I386
	+ "-i386"
	+#endif
	+#ifdef SDL_STRETCH_X86_64
	+ "-x86_64"
	+#endif
	+#ifdef HAVE_NATIVE_CODE
	+ "-native"
	+#endif
	+#ifdef HAVE_INTERPRETED_CODE
	+ "-interpreted"
	+#endif
	+#ifdef SDL_STRETCH_CALL
	+ "-call"
	+#endif
	+#ifdef SDL_STRETCH_DATA_NOEXEC
	+ "-noexec"
	+#endif
	+#ifdef SDL_STRETCH_DISABLE_ASM
	+ "-disabledasm"
	+#endif
	+#ifdef SDL_STRETCH_USE_ASM
	+ "-autodetectedasm"
	+#endif
	+#ifdef VERSION
	+ "-" VERSION
	+#else
	+ "-X"
	+#endif
	+ ;
	+}

File Metadata

Mime Type: text/x-diff
Expires: Thu, Jun 11, 11:34 AM (3 w, 4 d ago)
Storage Engine: blob
Storage Format: Raw Data
Storage Handle: 68793
Default Alt Text: (177 KB)

No OneTemporaryActions

View Options

File Metadata

Event Timeline

No OneTemporary
Actions