#include "quakedef.h"
#include "image.h"
#include "jpeg.h"
#include "r_shadow.h"

int		image_width;
int		image_height;

#if 1
// written by LordHavoc in a readable way, optimized by Vic, further optimized by LordHavoc (the non-special index case), readable version preserved below this
void Image_CopyMux(qbyte *outpixels, const qbyte *inpixels, int inputwidth, int inputheight, qboolean inputflipx, qboolean inputflipy, qboolean inputflipdiagonal, int numoutputcomponents, int numinputcomponents, int *outputinputcomponentindices)
{
	int index, c, x, y;
	const qbyte *in, *line;
	int row_inc = (inputflipy ? -inputwidth : inputwidth) * numinputcomponents, col_inc = (inputflipx ? -1 : 1) * numinputcomponents;
	int row_ofs = (inputflipy ? (inputheight - 1) * inputwidth * numinputcomponents : 0), col_ofs = (inputflipx ? (inputwidth - 1) * numinputcomponents : 0);

	for (c = 0; c < numoutputcomponents; c++)
		if (outputinputcomponentindices[c] & 0x80000000)
			break;
	if (c < numoutputcomponents)
	{
		// special indices used
		if (inputflipdiagonal)
		{
			for (x = 0, line = inpixels + col_ofs; x < inputwidth; x++, line += col_inc)
				for (y = 0, in = line + row_ofs; y < inputheight; y++, in += row_inc, outpixels += numinputcomponents)
					for (c = 0; c < numoutputcomponents; c++)
						outpixels[c] = ((index = outputinputcomponentindices[c]) & 0x80000000) ? index : in[index];
		}
		else
		{
			for (y = 0, line = inpixels + row_ofs; y < inputheight; y++, line += row_inc)
				for (x = 0, in = line + col_ofs; x < inputwidth; x++, in += col_inc, outpixels += numinputcomponents)
					for (c = 0; c < numoutputcomponents; c++)
						outpixels[c] = ((index = outputinputcomponentindices[c]) & 0x80000000) ? index : in[index];
		}
	}
	else
	{
		// special indices not used
		if (inputflipdiagonal)
		{
			for (x = 0, line = inpixels + col_ofs; x < inputwidth; x++, line += col_inc)
				for (y = 0, in = line + row_ofs; y < inputheight; y++, in += row_inc, outpixels += numinputcomponents)
					for (c = 0; c < numoutputcomponents; c++)
						outpixels[c] = in[outputinputcomponentindices[c]];
		}
		else
		{
			for (y = 0, line = inpixels + row_ofs; y < inputheight; y++, line += row_inc)
				for (x = 0, in = line + col_ofs; x < inputwidth; x++, in += col_inc, outpixels += numinputcomponents)
					for (c = 0; c < numoutputcomponents; c++)
						outpixels[c] = in[outputinputcomponentindices[c]];
		}
	}
}
#else
// intentionally readable version
void Image_CopyMux(qbyte *outpixels, const qbyte *inpixels, int inputwidth, int inputheight, qboolean inputflipx, qboolean inputflipy, qboolean inputflipdiagonal, int numoutputcomponents, int numinputcomponents, int *outputinputcomponentindices)
{
	int index, c, x, y;
	const qbyte *in, *inrow, *incolumn;
	if (inputflipdiagonal)
	{
		for (x = 0;x < inputwidth;x++)
		{
			for (y = 0;y < inputheight;y++)
			{
				in = inpixels + ((inputflipy ? inputheight - 1 - y : y) * inputwidth + (inputflipx ? inputwidth - 1 - x : x)) * numinputcomponents;
				for (c = 0;c < numoutputcomponents;c++)
				{
					index = outputinputcomponentindices[c];
					if (index & 0x80000000)
						*outpixels++ = index;
					else
						*outpixels++ = in[index];
				}
			}
		}
	}
	else
	{
		for (y = 0;y < inputheight;y++)
		{
			for (x = 0;x < inputwidth;x++)
			{
				in = inpixels + ((inputflipy ? inputheight - 1 - y : y) * inputwidth + (inputflipx ? inputwidth - 1 - x : x)) * numinputcomponents;
				for (c = 0;c < numoutputcomponents;c++)
				{
					index = outputinputcomponentindices[c];
					if (index & 0x80000000)
						*outpixels++ = index;
					else
						*outpixels++ = in[index];
				}
			}
		}
	}
}
#endif

void Image_GammaRemapRGB(const qbyte *in, qbyte *out, int pixels, const qbyte *gammar, const qbyte *gammag, const qbyte *gammab)
{
	while (pixels--)
	{
		out[0] = gammar[in[0]];
		out[1] = gammag[in[1]];
		out[2] = gammab[in[2]];
		in += 3;
		out += 3;
	}
}

// note: pal must be 32bit color
void Image_Copy8bitRGBA(const qbyte *in, qbyte *out, int pixels, const unsigned int *pal)
{
	int *iout = (void *)out;
	while (pixels >= 8)
	{
		iout[0] = pal[in[0]];
		iout[1] = pal[in[1]];
		iout[2] = pal[in[2]];
		iout[3] = pal[in[3]];
		iout[4] = pal[in[4]];
		iout[5] = pal[in[5]];
		iout[6] = pal[in[6]];
		iout[7] = pal[in[7]];
		in += 8;
		iout += 8;
		pixels -= 8;
	}
	if (pixels & 4)
	{
		iout[0] = pal[in[0]];
		iout[1] = pal[in[1]];
		iout[2] = pal[in[2]];
		iout[3] = pal[in[3]];
		in += 4;
		iout += 4;
	}
	if (pixels & 2)
	{
		iout[0] = pal[in[0]];
		iout[1] = pal[in[1]];
		in += 2;
		iout += 2;
	}
	if (pixels & 1)
		iout[0] = pal[in[0]];
}

/*
=================================================================

  PCX Loading

=================================================================
*/

typedef struct
{
    char	manufacturer;
    char	version;
    char	encoding;
    char	bits_per_pixel;
    unsigned short	xmin,ymin,xmax,ymax;
    unsigned short	hres,vres;
    unsigned char	palette[48];
    char	reserved;
    char	color_planes;
    unsigned short	bytes_per_line;
    unsigned short	palette_type;
    char	filler[58];
} pcx_t;

/*
============
LoadPCX
============
*/
qbyte* LoadPCX (qbyte *f, int matchwidth, int matchheight)
{
	pcx_t pcx;
	qbyte *a, *b, *image_rgba, *pbuf;
	const qbyte *palette, *fin, *enddata;
	int x, y, x2, dataByte;

	if (fs_filesize < (int)sizeof(pcx) + 768)
	{
		Con_Print("Bad pcx file\n");
		return NULL;
	}

	fin = f;

	memcpy(&pcx, fin, sizeof(pcx));
	fin += sizeof(pcx);

	// LordHavoc: big-endian support ported from QF newtree
	pcx.xmax = LittleShort (pcx.xmax);
	pcx.xmin = LittleShort (pcx.xmin);
	pcx.ymax = LittleShort (pcx.ymax);
	pcx.ymin = LittleShort (pcx.ymin);
	pcx.hres = LittleShort (pcx.hres);
	pcx.vres = LittleShort (pcx.vres);
	pcx.bytes_per_line = LittleShort (pcx.bytes_per_line);
	pcx.palette_type = LittleShort (pcx.palette_type);

	image_width = pcx.xmax + 1 - pcx.xmin;
	image_height = pcx.ymax + 1 - pcx.ymin;
	if (pcx.manufacturer != 0x0a || pcx.version != 5 || pcx.encoding != 1 || pcx.bits_per_pixel != 8 || image_width > 4096 || image_height > 4096 || image_width <= 0 || image_height <= 0)
	{
		Con_Print("Bad pcx file\n");
		return NULL;
	}
	if ((matchwidth && image_width != matchwidth) || (matchheight && image_height != matchheight))
		return NULL;

	palette = f + fs_filesize - 768;

	image_rgba = Mem_Alloc(tempmempool, image_width*image_height*4);
	if (!image_rgba)
	{
		Con_Printf("LoadPCX: not enough memory for %i by %i image\n", image_width, image_height);
		return NULL;
	}
	pbuf = image_rgba + image_width*image_height*3;
	enddata = palette;

	for (y = 0;y < image_height && fin < enddata;y++)
	{
		a = pbuf + y * image_width;
		for (x = 0;x < image_width && fin < enddata;)
		{
			dataByte = *fin++;
			if(dataByte >= 0xC0)
			{
				if (fin >= enddata)
					break;
				x2 = x + (dataByte & 0x3F);
				dataByte = *fin++;
				if (x2 > image_width)
					x2 = image_width; // technically an error
				while(x < x2)
					a[x++] = dataByte;
			}
			else
				a[x++] = dataByte;
		}
		fin += pcx.bytes_per_line - image_width; // the number of bytes per line is always forced to an even number
		while(x < image_width)
			a[x++] = 0;
	}

	a = image_rgba;
	b = pbuf;

	for(x = 0;x < image_width*image_height;x++)
	{
		y = *b++ * 3;
		*a++ = palette[y];
		*a++ = palette[y+1];
		*a++ = palette[y+2];
		*a++ = 255;
	}

	return image_rgba;
}

/*
=========================================================

TARGA LOADING

=========================================================
*/

typedef struct _TargaHeader
{
	unsigned char 	id_length, colormap_type, image_type;
	unsigned short	colormap_index, colormap_length;
	unsigned char	colormap_size;
	unsigned short	x_origin, y_origin, width, height;
	unsigned char	pixel_size, attributes;
}
TargaHeader;

void PrintTargaHeader(TargaHeader *t)
{
	Con_Print("TargaHeader:\n");
	Con_Printf("uint8 id_length = %i;\n", t->id_length);
	Con_Printf("uint8 colormap_type = %i;\n", t->colormap_type);
	Con_Printf("uint8 image_type = %i;\n", t->image_type);
	Con_Printf("uint16 colormap_index = %i;\n", t->colormap_index);
	Con_Printf("uint16 colormap_length = %i;\n", t->colormap_length);
	Con_Printf("uint8 colormap_size = %i;\n", t->colormap_size);
	Con_Printf("uint16 x_origin = %i;\n", t->x_origin);
	Con_Printf("uint16 y_origin = %i;\n", t->y_origin);
	Con_Printf("uint16 width = %i;\n", t->width);
	Con_Printf("uint16 height = %i;\n", t->height);
	Con_Printf("uint8 pixel_size = %i;\n", t->pixel_size);
	Con_Printf("uint8 attributes = %i;\n", t->attributes);
}

/*
=============
LoadTGA
=============
*/
qbyte *LoadTGA (qbyte *f, int matchwidth, int matchheight)
{
	int x, y, row_inc, compressed, readpixelcount, red, green, blue, alpha, runlen, pindex;
	qbyte *pixbuf, *image_rgba;
	qbyte *fin, *enddata;
	TargaHeader targa_header;
	unsigned char palette[256*4], *p;

	if (fs_filesize < 19)
		return NULL;

	enddata = f + fs_filesize;

	targa_header.id_length = f[0];
	targa_header.colormap_type = f[1];
	targa_header.image_type = f[2];

	targa_header.colormap_index = f[3] + f[4] * 256;
	targa_header.colormap_length = f[5] + f[6] * 256;
	targa_header.colormap_size = f[7];
	targa_header.x_origin = f[8] + f[9] * 256;
	targa_header.y_origin = f[10] + f[11] * 256;
	targa_header.width = image_width = f[12] + f[13] * 256;
	targa_header.height = image_height = f[14] + f[15] * 256;
	if (image_width > 4096 || image_height > 4096 || image_width <= 0 || image_height <= 0)
	{
		Con_Print("LoadTGA: invalid size\n");
		PrintTargaHeader(&targa_header);
		return NULL;
	}
	if ((matchwidth && image_width != matchwidth) || (matchheight && image_height != matchheight))
		return NULL;
	targa_header.pixel_size = f[16];
	targa_header.attributes = f[17];

	fin = f + 18;
	if (targa_header.id_length != 0)
		fin += targa_header.id_length;  // skip TARGA image comment
	if (targa_header.image_type == 2 || targa_header.image_type == 10)
	{
		if (targa_header.pixel_size != 24 && targa_header.pixel_size != 32)
		{
			Con_Print("LoadTGA: only 24bit and 32bit pixel sizes supported for type 2 and type 10 images\n");
			PrintTargaHeader(&targa_header);
			return NULL;
		}
	}
	else if (targa_header.image_type == 1 || targa_header.image_type == 9)
	{
		if (targa_header.pixel_size != 8)
		{
			Con_Print("LoadTGA: only 8bit pixel size for type 1, 3, 9, and 11 images supported\n");
			PrintTargaHeader(&targa_header);
			return NULL;
		}
		if (targa_header.colormap_length > 256)
		{
			Con_Print("LoadTGA: only up to 256 colormap_length supported\n");
			PrintTargaHeader(&targa_header);
			return NULL;
		}
		if (targa_header.colormap_index)
		{
			Con_Print("LoadTGA: colormap_index not supported\n");
			PrintTargaHeader(&targa_header);
			return NULL;
		}
		if (targa_header.colormap_size == 24)
		{
			for (x = 0;x < targa_header.colormap_length;x++)
			{
				palette[x*4+2] = *fin++;
				palette[x*4+1] = *fin++;
				palette[x*4+0] = *fin++;
				palette[x*4+3] = 255;
			}
		}
		else if (targa_header.colormap_size == 32)
		{
			for (x = 0;x < targa_header.colormap_length;x++)
			{
				palette[x*4+2] = *fin++;
				palette[x*4+1] = *fin++;
				palette[x*4+0] = *fin++;
				palette[x*4+3] = *fin++;
			}
		}
		else
		{
			Con_Print("LoadTGA: Only 32 and 24 bit colormap_size supported\n");
			PrintTargaHeader(&targa_header);
			return NULL;
		}
	}
	else if (targa_header.image_type == 3 || targa_header.image_type == 11)
	{
		if (targa_header.pixel_size != 8)
		{
			Con_Print("LoadTGA: only 8bit pixel size for type 1, 3, 9, and 11 images supported\n");
			PrintTargaHeader(&targa_header);
			return NULL;
		}
	}
	else
	{
		Con_Printf("LoadTGA: Only type 1, 2, 3, 9, 10, and 11 targa RGB images supported, image_type = %i\n", targa_header.image_type);
		PrintTargaHeader(&targa_header);
		return NULL;
	}

	if (targa_header.attributes & 0x10)
	{
		Con_Print("LoadTGA: origin must be in top left or bottom left, top right and bottom right are not supported\n");
		return NULL;
	}

	image_rgba = Mem_Alloc(tempmempool, image_width * image_height * 4);
	if (!image_rgba)
	{
		Con_Printf("LoadTGA: not enough memory for %i by %i image\n", image_width, image_height);
		return NULL;
	}

	// If bit 5 of attributes isn't set, the image has been stored from bottom to top
	if ((targa_header.attributes & 0x20) == 0)
	{
		pixbuf = image_rgba + (image_height - 1)*image_width*4;
		row_inc = -image_width*4*2;
	}
	else
	{
		pixbuf = image_rgba;
		row_inc = 0;
	}

	compressed = targa_header.image_type == 9 || targa_header.image_type == 10 || targa_header.image_type == 11;
	x = 0;
	y = 0;
	red = green = blue = alpha = 255;
	while (y < image_height)
	{
		// decoder is mostly the same whether it's compressed or not
		readpixelcount = 1000000;
		runlen = 1000000;
		if (compressed && fin < enddata)
		{
			runlen = *fin++;
			// high bit indicates this is an RLE compressed run
			if (runlen & 0x80)
				readpixelcount = 1;
			runlen = 1 + (runlen & 0x7f);
		}

		while((runlen--) && y < image_height)
		{
			if (readpixelcount > 0)
			{
				readpixelcount--;
				red = green = blue = alpha = 255;
				if (fin < enddata)
				{
					switch(targa_header.image_type)
					{
					case 1:
					case 9:
						// colormapped
						pindex = *fin++;
						if (pindex >= targa_header.colormap_length)
							pindex = 0; // error
						p = palette + pindex * 4;
						red = p[0];
						green = p[1];
						blue = p[2];
						alpha = p[3];
						break;
					case 2:
					case 10:
						// BGR or BGRA
						blue = *fin++;
						if (fin < enddata)
							green = *fin++;
						if (fin < enddata)
							red = *fin++;
						if (targa_header.pixel_size == 32 && fin < enddata)
							alpha = *fin++;
						break;
					case 3:
					case 11:
						// greyscale
						red = green = blue = *fin++;
						break;
					}
				}
			}
			*pixbuf++ = red;
			*pixbuf++ = green;
			*pixbuf++ = blue;
			*pixbuf++ = alpha;
			x++;
			if (x == image_width)
			{
				// end of line, advance to next
				x = 0;
				y++;
				pixbuf += row_inc;
			}
		}
	}

	return image_rgba;
}

/*
============
LoadLMP
============
*/
qbyte *LoadLMP (qbyte *f, int matchwidth, int matchheight)
{
	qbyte *image_rgba;

	if (fs_filesize < 9)
	{
		Con_Print("LoadLMP: invalid LMP file\n");
		return NULL;
	}

	// parse the very complicated header *chuckle*
	image_width = f[0] + f[1] * 256 + f[2] * 65536 + f[3] * 16777216;
	image_height = f[4] + f[5] * 256 + f[6] * 65536 + f[7] * 16777216;
	if (image_width > 4096 || image_height > 4096 || image_width <= 0 || image_height <= 0)
	{
		Con_Printf("LoadLMP: invalid size %ix%i\n", image_width, image_height);
		return NULL;
	}
	if ((matchwidth && image_width != matchwidth) || (matchheight && image_height != matchheight))
		return NULL;

	if (fs_filesize < 8 + image_width * image_height)
	{
		Con_Print("LoadLMP: invalid LMP file\n");
		return NULL;
	}

	image_rgba = Mem_Alloc(tempmempool, image_width * image_height * 4);
	if (!image_rgba)
	{
		Con_Printf("LoadLMP: not enough memory for %i by %i image\n", image_width, image_height);
		return NULL;
	}
	Image_Copy8bitRGBA(f + 8, image_rgba, image_width * image_height, palette_complete);
	return image_rgba;
}

typedef struct
{
	char		name[32];
	unsigned	width, height;
	unsigned	offsets[MIPLEVELS];		// four mip maps stored
	char		animname[32];			// next frame in animation chain
	int			flags;
	int			contents;
	int			value;
} q2wal_t;

qbyte *LoadWAL (qbyte *f, int matchwidth, int matchheight)
{
	qbyte *image_rgba;
	const q2wal_t *inwal = (const void *)f;

	if (fs_filesize < (int) sizeof(q2wal_t))
	{
		Con_Print("LoadWAL: invalid WAL file\n");
		return NULL;
	}

	image_width = LittleLong(inwal->width);
	image_height = LittleLong(inwal->height);
	if (image_width > 4096 || image_height > 4096 || image_width <= 0 || image_height <= 0)
	{
		Con_Printf("LoadWAL: invalid size %ix%i\n", image_width, image_height);
		return NULL;
	}
	if ((matchwidth && image_width != matchwidth) || (matchheight && image_height != matchheight))
		return NULL;

	if ((int) fs_filesize < (int) sizeof(q2wal_t) + (int) LittleLong(inwal->offsets[0]) + image_width * image_height)
	{
		Con_Print("LoadWAL: invalid WAL file\n");
		return NULL;
	}

	image_rgba = Mem_Alloc(tempmempool, image_width * image_height * 4);
	if (!image_rgba)
	{
		Con_Printf("LoadLMP: not enough memory for %i by %i image\n", image_width, image_height);
		return NULL;
	}
	Image_Copy8bitRGBA(f + LittleLong(inwal->offsets[0]), image_rgba, image_width * image_height, palette_complete);
	return image_rgba;
}



/*
============
LoadLMP
============
*/
qbyte *LoadLMPAs8Bit (qbyte *f, int matchwidth, int matchheight)
{
	qbyte *image_8bit;

	if (fs_filesize < 9)
	{
		Con_Print("LoadLMPAs8Bit: invalid LMP file\n");
		return NULL;
	}

	// parse the very complicated header *chuckle*
	image_width = f[0] + f[1] * 256 + f[2] * 65536 + f[3] * 16777216;
	image_height = f[4] + f[5] * 256 + f[6] * 65536 + f[7] * 16777216;
	if (image_width > 4096 || image_height > 4096 || image_width <= 0 || image_height <= 0)
	{
		Con_Printf("LoadLMPAs8Bit: invalid size %ix%i\n", image_width, image_height);
		return NULL;
	}
	if ((matchwidth && image_width != matchwidth) || (matchheight && image_height != matchheight))
		return NULL;

	if (fs_filesize < 8 + image_width * image_height)
	{
		Con_Print("LoadLMPAs8Bit: invalid LMP file\n");
		return NULL;
	}

	image_8bit = Mem_Alloc(tempmempool, image_width * image_height);
	if (!image_8bit)
	{
		Con_Printf("LoadLMPAs8Bit: not enough memory for %i by %i image\n", image_width, image_height);
		return NULL;
	}
	memcpy(image_8bit, f + 8, image_width * image_height);
	return image_8bit;
}

void Image_StripImageExtension (const char *in, char *out)
{
	const char *end, *temp;
	end = in + strlen(in);
	if ((end - in) >= 4)
	{
		temp = end - 4;
		if (strcmp(temp, ".tga") == 0
		 || strcmp(temp, ".pcx") == 0
		 || strcmp(temp, ".lmp") == 0
		 || strcmp(temp, ".png") == 0
		 || strcmp(temp, ".jpg") == 0)
			end = temp;
		while (in < end)
			*out++ = *in++;
		*out++ = 0;
	}
	else
		strcpy(out, in);
}

struct
{
	const char *formatstring;
	qbyte *(*loadfunc)(qbyte *f, int matchwidth, int matchheight);
}
imageformats[] =
{
	{"override/%s.tga", LoadTGA},
	{"override/%s.jpg", JPEG_LoadImage},
	{"textures/%s.tga", LoadTGA},
	{"textures/%s.jpg", JPEG_LoadImage},
	{"textures/%s.pcx", LoadPCX},
	{"textures/%s.wal", LoadWAL},
	{"%s.tga", LoadTGA},
	{"%s.jpg", JPEG_LoadImage},
	{"%s.pcx", LoadPCX},
	{"%s.lmp", LoadLMP},
	{NULL, NULL}
};

qbyte *loadimagepixels (const char *filename, qboolean complain, int matchwidth, int matchheight)
{
	int i;
	qbyte *f, *data = NULL;
	char basename[MAX_QPATH], name[MAX_QPATH], *c;
	if (developer_memorydebug.integer)
		Mem_CheckSentinelsGlobal();
	Image_StripImageExtension(filename, basename); // strip filename extensions to allow replacement by other types
	// replace *'s with #, so commandline utils don't get confused when dealing with the external files
	for (c = basename;*c;c++)
		if (*c == '*')
			*c = '#';
	for (i = 0;imageformats[i].formatstring;i++)
	{
		sprintf (name, imageformats[i].formatstring, basename);
		f = FS_LoadFile(name, tempmempool, true);
		if (f)
		{
			data = imageformats[i].loadfunc(f, matchwidth, matchheight);
			Mem_Free(f);
			if (data)
			{
				Con_DPrintf("loaded image %s (%dx%d)\n", name, image_width, image_height);
				if (developer_memorydebug.integer)
					Mem_CheckSentinelsGlobal();
				return data;
			}
		}
	}
	if (complain)
	{
		Con_Printf("Couldn't load %s using ", filename);
		for (i = 0;imageformats[i].formatstring;i++)
		{
			sprintf (name, imageformats[i].formatstring, basename);
			Con_Printf(i == 0 ? "\"%s\"" : (imageformats[i+1].formatstring ? ", \"%s\"" : " or \"%s\".\n"), imageformats[i].formatstring);
		}
	}
	if (developer_memorydebug.integer)
		Mem_CheckSentinelsGlobal();
	return NULL;
}

int image_makemask (const qbyte *in, qbyte *out, int size)
{
	int i, count;
	count = 0;
	for (i = 0;i < size;i++)
	{
		out[0] = out[1] = out[2] = 255;
		out[3] = in[3];
		if (in[3] != 255)
			count++;
		in += 4;
		out += 4;
	}
	return count;
}

qbyte* loadimagepixelsmask (const char *filename, qboolean complain, int matchwidth, int matchheight)
{
	qbyte *in, *data;
	in = data = loadimagepixels(filename, complain, matchwidth, matchheight);
	if (!data)
		return NULL;
	if (image_makemask(data, data, image_width * image_height))
		return data; // some transparency
	else
	{
		Mem_Free(data);
		return NULL; // all opaque
	}
}

rtexture_t *loadtextureimage (rtexturepool_t *pool, const char *filename, int matchwidth, int matchheight, qboolean complain, int flags)
{
	qbyte *data;
	rtexture_t *rt;
	if (!(data = loadimagepixels (filename, complain, matchwidth, matchheight)))
		return 0;
	rt = R_LoadTexture2D(pool, filename, image_width, image_height, data, TEXTYPE_RGBA, flags, NULL);
	Mem_Free(data);
	return rt;
}

rtexture_t *loadtextureimagemask (rtexturepool_t *pool, const char *filename, int matchwidth, int matchheight, qboolean complain, int flags)
{
	qbyte *data;
	rtexture_t *rt;
	if (!(data = loadimagepixelsmask (filename, complain, matchwidth, matchheight)))
		return 0;
	rt = R_LoadTexture2D(pool, filename, image_width, image_height, data, TEXTYPE_RGBA, flags, NULL);
	Mem_Free(data);
	return rt;
}

rtexture_t *image_masktex;
rtexture_t *image_nmaptex;
rtexture_t *loadtextureimagewithmask (rtexturepool_t *pool, const char *filename, int matchwidth, int matchheight, qboolean complain, int flags)
{
	qbyte *data;
	rtexture_t *rt;
	image_masktex = NULL;
	image_nmaptex = NULL;
	if (!(data = loadimagepixels (filename, complain, matchwidth, matchheight)))
		return 0;

	rt = R_LoadTexture2D(pool, filename, image_width, image_height, data, TEXTYPE_RGBA, flags, NULL);

	if (flags & TEXF_ALPHA && image_makemask(data, data, image_width * image_height))
		image_masktex = R_LoadTexture2D(pool, va("%s_mask", filename), image_width, image_height, data, TEXTYPE_RGBA, flags, NULL);

	Mem_Free(data);
	return rt;
}

rtexture_t *loadtextureimagewithmaskandnmap (rtexturepool_t *pool, const char *filename, int matchwidth, int matchheight, qboolean complain, int flags, float bumpscale)
{
	qbyte *data, *data2;
	rtexture_t *rt;
	image_masktex = NULL;
	image_nmaptex = NULL;
	if (!(data = loadimagepixels (filename, complain, matchwidth, matchheight)))
		return 0;

	data2 = Mem_Alloc(tempmempool, image_width * image_height * 4);

	rt = R_LoadTexture2D(pool, filename, image_width, image_height, data, TEXTYPE_RGBA, flags, NULL);

	Image_HeightmapToNormalmap(data, data2, image_width, image_height, (flags & TEXF_CLAMP) != 0, bumpscale);
	image_nmaptex = R_LoadTexture2D(pool, va("%s_nmap", filename), image_width, image_height, data2, TEXTYPE_RGBA, flags, NULL);

	if (flags & TEXF_ALPHA && image_makemask(data, data2, image_width * image_height))
		image_masktex = R_LoadTexture2D(pool, va("%s_mask", filename), image_width, image_height, data2, TEXTYPE_RGBA, flags, NULL);

	Mem_Free(data2);

	Mem_Free(data);
	return rt;
}

rtexture_t *loadtextureimagebumpasnmap (rtexturepool_t *pool, const char *filename, int matchwidth, int matchheight, qboolean complain, int flags, float bumpscale)
{
	qbyte *data, *data2;
	rtexture_t *rt;
	if (!(data = loadimagepixels (filename, complain, matchwidth, matchheight)))
		return 0;
	data2 = Mem_Alloc(tempmempool, image_width * image_height * 4);

	Image_HeightmapToNormalmap(data, data2, image_width, image_height, (flags & TEXF_CLAMP) != 0, bumpscale);
	rt = R_LoadTexture2D(pool, filename, image_width, image_height, data2, TEXTYPE_RGBA, flags, NULL);

	Mem_Free(data2);
	Mem_Free(data);
	return rt;
}

qboolean Image_WriteTGARGB_preflipped (const char *filename, int width, int height, const qbyte *data, qbyte *buffer)
{
	qboolean ret;
	qbyte *out;
	const qbyte *in, *end;

	memset (buffer, 0, 18);
	buffer[2] = 2;		// uncompressed type
	buffer[12] = (width >> 0) & 0xFF;
	buffer[13] = (width >> 8) & 0xFF;
	buffer[14] = (height >> 0) & 0xFF;
	buffer[15] = (height >> 8) & 0xFF;
	buffer[16] = 24;	// pixel size

	// swap rgb to bgr
	in = data;
	out = buffer + 18;
	end = in + width*height*3;
	for (;in < end;in += 3)
	{
		*out++ = in[2];
		*out++ = in[1];
		*out++ = in[0];
	}
	ret = FS_WriteFile (filename, buffer, width*height*3 + 18 );

	return ret;
}

void Image_WriteTGARGB (const char *filename, int width, int height, const qbyte *data)
{
	int y;
	qbyte *buffer, *out;
	const qbyte *in, *end;

	buffer = Mem_Alloc(tempmempool, width*height*3 + 18);

	memset (buffer, 0, 18);
	buffer[2] = 2;		// uncompressed type
	buffer[12] = (width >> 0) & 0xFF;
	buffer[13] = (width >> 8) & 0xFF;
	buffer[14] = (height >> 0) & 0xFF;
	buffer[15] = (height >> 8) & 0xFF;
	buffer[16] = 24;	// pixel size

	// swap rgb to bgr and flip upside down
	out = buffer + 18;
	for (y = height - 1;y >= 0;y--)
	{
		in = data + y * width * 3;
		end = in + width * 3;
		for (;in < end;in += 3)
		{
			*out++ = in[2];
			*out++ = in[1];
			*out++ = in[0];
		}
	}
	FS_WriteFile (filename, buffer, width*height*3 + 18 );

	Mem_Free(buffer);
}

void Image_WriteTGARGBA (const char *filename, int width, int height, const qbyte *data)
{
	int y;
	qbyte *buffer, *out;
	const qbyte *in, *end;

	buffer = Mem_Alloc(tempmempool, width*height*4 + 18);

	memset (buffer, 0, 18);
	buffer[2] = 2;		// uncompressed type
	buffer[12] = (width >> 0) & 0xFF;
	buffer[13] = (width >> 8) & 0xFF;
	buffer[14] = (height >> 0) & 0xFF;
	buffer[15] = (height >> 8) & 0xFF;
	buffer[16] = 32;	// pixel size

	// swap rgba to bgra and flip upside down
	out = buffer + 18;
	for (y = height - 1;y >= 0;y--)
	{
		in = data + y * width * 4;
		end = in + width * 4;
		for (;in < end;in += 4)
		{
			*out++ = in[2];
			*out++ = in[1];
			*out++ = in[0];
			*out++ = in[3];
		}
	}
	FS_WriteFile (filename, buffer, width*height*4 + 18 );

	Mem_Free(buffer);
}

qboolean Image_CheckAlpha(const qbyte *data, int size, qboolean rgba)
{
	const qbyte *end;
	if (rgba)
	{
		// check alpha bytes
		for (end = data + size * 4, data += 3;data < end;data += 4)
			if (*data < 255)
				return 1;
	}
	else
	{
		// color 255 is transparent
		for (end = data + size;data < end;data++)
			if (*data == 255)
				return 1;
	}
	return 0;
}

static void Image_Resample32LerpLine (const qbyte *in, qbyte *out, int inwidth, int outwidth)
{
	int		j, xi, oldx = 0, f, fstep, endx, lerp;
	fstep = (int) (inwidth*65536.0f/outwidth);
	endx = (inwidth-1);
	for (j = 0,f = 0;j < outwidth;j++, f += fstep)
	{
		xi = f >> 16;
		if (xi != oldx)
		{
			in += (xi - oldx) * 4;
			oldx = xi;
		}
		if (xi < endx)
		{
			lerp = f & 0xFFFF;
			*out++ = (qbyte) ((((in[4] - in[0]) * lerp) >> 16) + in[0]);
			*out++ = (qbyte) ((((in[5] - in[1]) * lerp) >> 16) + in[1]);
			*out++ = (qbyte) ((((in[6] - in[2]) * lerp) >> 16) + in[2]);
			*out++ = (qbyte) ((((in[7] - in[3]) * lerp) >> 16) + in[3]);
		}
		else // last pixel of the line has no pixel to lerp to
		{
			*out++ = in[0];
			*out++ = in[1];
			*out++ = in[2];
			*out++ = in[3];
		}
	}
}

static void Image_Resample24LerpLine (const qbyte *in, qbyte *out, int inwidth, int outwidth)
{
	int		j, xi, oldx = 0, f, fstep, endx, lerp;
	fstep = (int) (inwidth*65536.0f/outwidth);
	endx = (inwidth-1);
	for (j = 0,f = 0;j < outwidth;j++, f += fstep)
	{
		xi = f >> 16;
		if (xi != oldx)
		{
			in += (xi - oldx) * 3;
			oldx = xi;
		}
		if (xi < endx)
		{
			lerp = f & 0xFFFF;
			*out++ = (qbyte) ((((in[3] - in[0]) * lerp) >> 16) + in[0]);
			*out++ = (qbyte) ((((in[4] - in[1]) * lerp) >> 16) + in[1]);
			*out++ = (qbyte) ((((in[5] - in[2]) * lerp) >> 16) + in[2]);
		}
		else // last pixel of the line has no pixel to lerp to
		{
			*out++ = in[0];
			*out++ = in[1];
			*out++ = in[2];
		}
	}
}

int resamplerowsize = 0;
qbyte *resamplerow1 = NULL;
qbyte *resamplerow2 = NULL;
mempool_t *resamplemempool = NULL;

#define LERPBYTE(i) r = resamplerow1[i];out[i] = (qbyte) ((((resamplerow2[i] - r) * lerp) >> 16) + r)
void Image_Resample32Lerp(const void *indata, int inwidth, int inheight, void *outdata, int outwidth, int outheight)
{
	int i, j, r, yi, oldy, f, fstep, lerp, endy = (inheight-1), inwidth4 = inwidth*4, outwidth4 = outwidth*4;
	qbyte *out;
	const qbyte *inrow;
	out = outdata;
	fstep = (int) (inheight*65536.0f/outheight);

	inrow = indata;
	oldy = 0;
	Image_Resample32LerpLine (inrow, resamplerow1, inwidth, outwidth);
	Image_Resample32LerpLine (inrow + inwidth4, resamplerow2, inwidth, outwidth);
	for (i = 0, f = 0;i < outheight;i++,f += fstep)
	{
		yi = f >> 16;
		if (yi < endy)
		{
			lerp = f & 0xFFFF;
			if (yi != oldy)
			{
				inrow = (qbyte *)indata + inwidth4*yi;
				if (yi == oldy+1)
					memcpy(resamplerow1, resamplerow2, outwidth4);
				else
					Image_Resample32LerpLine (inrow, resamplerow1, inwidth, outwidth);
				Image_Resample32LerpLine (inrow + inwidth4, resamplerow2, inwidth, outwidth);
				oldy = yi;
			}
			j = outwidth - 4;
			while(j >= 0)
			{
				LERPBYTE( 0);
				LERPBYTE( 1);
				LERPBYTE( 2);
				LERPBYTE( 3);
				LERPBYTE( 4);
				LERPBYTE( 5);
				LERPBYTE( 6);
				LERPBYTE( 7);
				LERPBYTE( 8);
				LERPBYTE( 9);
				LERPBYTE(10);
				LERPBYTE(11);
				LERPBYTE(12);
				LERPBYTE(13);
				LERPBYTE(14);
				LERPBYTE(15);
				out += 16;
				resamplerow1 += 16;
				resamplerow2 += 16;
				j -= 4;
			}
			if (j & 2)
			{
				LERPBYTE( 0);
				LERPBYTE( 1);
				LERPBYTE( 2);
				LERPBYTE( 3);
				LERPBYTE( 4);
				LERPBYTE( 5);
				LERPBYTE( 6);
				LERPBYTE( 7);
				out += 8;
				resamplerow1 += 8;
				resamplerow2 += 8;
			}
			if (j & 1)
			{
				LERPBYTE( 0);
				LERPBYTE( 1);
				LERPBYTE( 2);
				LERPBYTE( 3);
				out += 4;
				resamplerow1 += 4;
				resamplerow2 += 4;
			}
			resamplerow1 -= outwidth4;
			resamplerow2 -= outwidth4;
		}
		else
		{
			if (yi != oldy)
			{
				inrow = (qbyte *)indata + inwidth4*yi;
				if (yi == oldy+1)
					memcpy(resamplerow1, resamplerow2, outwidth4);
				else
					Image_Resample32LerpLine (inrow, resamplerow1, inwidth, outwidth);
				oldy = yi;
			}
			memcpy(out, resamplerow1, outwidth4);
		}
	}
}

void Image_Resample32Nearest(const void *indata, int inwidth, int inheight, void *outdata, int outwidth, int outheight)
{
	int i, j;
	unsigned frac, fracstep;
	// relies on int being 4 bytes
	int *inrow, *out;
	out = outdata;

	fracstep = inwidth*0x10000/outwidth;
	for (i = 0;i < outheight;i++)
	{
		inrow = (int *)indata + inwidth*(i*inheight/outheight);
		frac = fracstep >> 1;
		j = outwidth - 4;
		while (j >= 0)
		{
			out[0] = inrow[frac >> 16];frac += fracstep;
			out[1] = inrow[frac >> 16];frac += fracstep;
			out[2] = inrow[frac >> 16];frac += fracstep;
			out[3] = inrow[frac >> 16];frac += fracstep;
			out += 4;
			j -= 4;
		}
		if (j & 2)
		{
			out[0] = inrow[frac >> 16];frac += fracstep;
			out[1] = inrow[frac >> 16];frac += fracstep;
			out += 2;
		}
		if (j & 1)
		{
			out[0] = inrow[frac >> 16];frac += fracstep;
			out += 1;
		}
	}
}

void Image_Resample24Lerp(const void *indata, int inwidth, int inheight, void *outdata, int outwidth, int outheight)
{
	int i, j, r, yi, oldy, f, fstep, lerp, endy = (inheight-1), inwidth3 = inwidth * 3, outwidth3 = outwidth * 3;
	qbyte *out;
	const qbyte *inrow;
	out = outdata;
	fstep = (int) (inheight*65536.0f/outheight);

	inrow = indata;
	oldy = 0;
	Image_Resample24LerpLine (inrow, resamplerow1, inwidth, outwidth);
	Image_Resample24LerpLine (inrow + inwidth3, resamplerow2, inwidth, outwidth);
	for (i = 0, f = 0;i < outheight;i++,f += fstep)
	{
		yi = f >> 16;
		if (yi < endy)
		{
			lerp = f & 0xFFFF;
			if (yi != oldy)
			{
				inrow = (qbyte *)indata + inwidth3*yi;
				if (yi == oldy+1)
					memcpy(resamplerow1, resamplerow2, outwidth3);
				else
					Image_Resample24LerpLine (inrow, resamplerow1, inwidth, outwidth);
				Image_Resample24LerpLine (inrow + inwidth3, resamplerow2, inwidth, outwidth);
				oldy = yi;
			}
			j = outwidth - 4;
			while(j >= 0)
			{
				LERPBYTE( 0);
				LERPBYTE( 1);
				LERPBYTE( 2);
				LERPBYTE( 3);
				LERPBYTE( 4);
				LERPBYTE( 5);
				LERPBYTE( 6);
				LERPBYTE( 7);
				LERPBYTE( 8);
				LERPBYTE( 9);
				LERPBYTE(10);
				LERPBYTE(11);
				out += 12;
				resamplerow1 += 12;
				resamplerow2 += 12;
				j -= 4;
			}
			if (j & 2)
			{
				LERPBYTE( 0);
				LERPBYTE( 1);
				LERPBYTE( 2);
				LERPBYTE( 3);
				LERPBYTE( 4);
				LERPBYTE( 5);
				out += 6;
				resamplerow1 += 6;
				resamplerow2 += 6;
			}
			if (j & 1)
			{
				LERPBYTE( 0);
				LERPBYTE( 1);
				LERPBYTE( 2);
				out += 3;
				resamplerow1 += 3;
				resamplerow2 += 3;
			}
			resamplerow1 -= outwidth3;
			resamplerow2 -= outwidth3;
		}
		else
		{
			if (yi != oldy)
			{
				inrow = (qbyte *)indata + inwidth3*yi;
				if (yi == oldy+1)
					memcpy(resamplerow1, resamplerow2, outwidth3);
				else
					Image_Resample24LerpLine (inrow, resamplerow1, inwidth, outwidth);
				oldy = yi;
			}
			memcpy(out, resamplerow1, outwidth3);
		}
	}
}

void Image_Resample24Nolerp(const void *indata, int inwidth, int inheight, void *outdata, int outwidth, int outheight)
{
	int i, j, f, inwidth3 = inwidth * 3;
	unsigned frac, fracstep;
	qbyte *inrow, *out;
	out = outdata;

	fracstep = inwidth*0x10000/outwidth;
	for (i = 0;i < outheight;i++)
	{
		inrow = (qbyte *)indata + inwidth3*(i*inheight/outheight);
		frac = fracstep >> 1;
		j = outwidth - 4;
		while (j >= 0)
		{
			f = (frac >> 16)*3;*out++ = inrow[f+0];*out++ = inrow[f+1];*out++ = inrow[f+2];frac += fracstep;
			f = (frac >> 16)*3;*out++ = inrow[f+0];*out++ = inrow[f+1];*out++ = inrow[f+2];frac += fracstep;
			f = (frac >> 16)*3;*out++ = inrow[f+0];*out++ = inrow[f+1];*out++ = inrow[f+2];frac += fracstep;
			f = (frac >> 16)*3;*out++ = inrow[f+0];*out++ = inrow[f+1];*out++ = inrow[f+2];frac += fracstep;
			j -= 4;
		}
		if (j & 2)
		{
			f = (frac >> 16)*3;*out++ = inrow[f+0];*out++ = inrow[f+1];*out++ = inrow[f+2];frac += fracstep;
			f = (frac >> 16)*3;*out++ = inrow[f+0];*out++ = inrow[f+1];*out++ = inrow[f+2];frac += fracstep;
			out += 2;
		}
		if (j & 1)
		{
			f = (frac >> 16)*3;*out++ = inrow[f+0];*out++ = inrow[f+1];*out++ = inrow[f+2];frac += fracstep;
			out += 1;
		}
	}
}

/*
================
Image_Resample
================
*/
void Image_Resample (const void *indata, int inwidth, int inheight, int indepth, void *outdata, int outwidth, int outheight, int outdepth, int bytesperpixel, int quality)
{
	if (indepth != 1 || outdepth != 1)
		Sys_Error("Image_Resample: 3D resampling not supported\n");
	if (resamplerowsize < outwidth*4)
	{
		if (resamplerow1)
			Mem_Free(resamplerow1);
		resamplerowsize = outwidth*4;
		if (!resamplemempool)
			resamplemempool = Mem_AllocPool("Image Scaling Buffer");
		resamplerow1 = Mem_Alloc(resamplemempool, resamplerowsize*2);
		resamplerow2 = resamplerow1 + resamplerowsize;
	}
	if (bytesperpixel == 4)
	{
		if (quality)
			Image_Resample32Lerp(indata, inwidth, inheight, outdata, outwidth, outheight);
		else
			Image_Resample32Nearest(indata, inwidth, inheight, outdata, outwidth, outheight);
	}
	else if (bytesperpixel == 3)
	{
		if (quality)
			Image_Resample24Lerp(indata, inwidth, inheight, outdata, outwidth, outheight);
		else
			Image_Resample24Nolerp(indata, inwidth, inheight, outdata, outwidth, outheight);
	}
	else
		Sys_Error("Image_Resample: unsupported bytesperpixel %i\n", bytesperpixel);
}

// in can be the same as out
void Image_MipReduce(const qbyte *in, qbyte *out, int *width, int *height, int *depth, int destwidth, int destheight, int destdepth, int bytesperpixel)
{
	int x, y, nextrow;
	if (*depth != 1 || destdepth != 1)
		Sys_Error("Image_Resample: 3D resampling not supported\n");
	nextrow = *width * bytesperpixel;
	if (*width > destwidth)
	{
		*width >>= 1;
		if (*height > destheight)
		{
			// reduce both
			*height >>= 1;
			if (bytesperpixel == 4)
			{
				for (y = 0;y < *height;y++)
				{
					for (x = 0;x < *width;x++)
					{
						out[0] = (qbyte) ((in[0] + in[4] + in[nextrow  ] + in[nextrow+4]) >> 2);
						out[1] = (qbyte) ((in[1] + in[5] + in[nextrow+1] + in[nextrow+5]) >> 2);
						out[2] = (qbyte) ((in[2] + in[6] + in[nextrow+2] + in[nextrow+6]) >> 2);
						out[3] = (qbyte) ((in[3] + in[7] + in[nextrow+3] + in[nextrow+7]) >> 2);
						out += 4;
						in += 8;
					}
					in += nextrow; // skip a line
				}
			}
			else if (bytesperpixel == 3)
			{
				for (y = 0;y < *height;y++)
				{
					for (x = 0;x < *width;x++)
					{
						out[0] = (qbyte) ((in[0] + in[3] + in[nextrow  ] + in[nextrow+3]) >> 2);
						out[1] = (qbyte) ((in[1] + in[4] + in[nextrow+1] + in[nextrow+4]) >> 2);
						out[2] = (qbyte) ((in[2] + in[5] + in[nextrow+2] + in[nextrow+5]) >> 2);
						out += 3;
						in += 6;
					}
					in += nextrow; // skip a line
				}
			}
			else
				Sys_Error("Image_MipReduce: unsupported bytesperpixel %i\n", bytesperpixel);
		}
		else
		{
			// reduce width
			if (bytesperpixel == 4)
			{
				for (y = 0;y < *height;y++)
				{
					for (x = 0;x < *width;x++)
					{
						out[0] = (qbyte) ((in[0] + in[4]) >> 1);
						out[1] = (qbyte) ((in[1] + in[5]) >> 1);
						out[2] = (qbyte) ((in[2] + in[6]) >> 1);
						out[3] = (qbyte) ((in[3] + in[7]) >> 1);
						out += 4;
						in += 8;
					}
				}
			}
			else if (bytesperpixel == 3)
			{
				for (y = 0;y < *height;y++)
				{
					for (x = 0;x < *width;x++)
					{
						out[0] = (qbyte) ((in[0] + in[3]) >> 1);
						out[1] = (qbyte) ((in[1] + in[4]) >> 1);
						out[2] = (qbyte) ((in[2] + in[5]) >> 1);
						out += 3;
						in += 6;
					}
				}
			}
			else
				Sys_Error("Image_MipReduce: unsupported bytesperpixel %i\n", bytesperpixel);
		}
	}
	else
	{
		if (*height > destheight)
		{
			// reduce height
			*height >>= 1;
			if (bytesperpixel == 4)
			{
				for (y = 0;y < *height;y++)
				{
					for (x = 0;x < *width;x++)
					{
						out[0] = (qbyte) ((in[0] + in[nextrow  ]) >> 1);
						out[1] = (qbyte) ((in[1] + in[nextrow+1]) >> 1);
						out[2] = (qbyte) ((in[2] + in[nextrow+2]) >> 1);
						out[3] = (qbyte) ((in[3] + in[nextrow+3]) >> 1);
						out += 4;
						in += 4;
					}
					in += nextrow; // skip a line
				}
			}
			else if (bytesperpixel == 3)
			{
				for (y = 0;y < *height;y++)
				{
					for (x = 0;x < *width;x++)
					{
						out[0] = (qbyte) ((in[0] + in[nextrow  ]) >> 1);
						out[1] = (qbyte) ((in[1] + in[nextrow+1]) >> 1);
						out[2] = (qbyte) ((in[2] + in[nextrow+2]) >> 1);
						out += 3;
						in += 3;
					}
					in += nextrow; // skip a line
				}
			}
			else
				Sys_Error("Image_MipReduce: unsupported bytesperpixel %i\n", bytesperpixel);
		}
		else
			Sys_Error("Image_MipReduce: desired size already achieved\n");
	}
}

void Image_HeightmapToNormalmap(const unsigned char *inpixels, unsigned char *outpixels, int width, int height, int clamp, float bumpscale)
{
	int x, y;
	const unsigned char *p0, *p1, *p2;
	unsigned char *out;
	float iwidth, iheight, ibumpscale, n[3];
	iwidth = 1.0f / width;
	iheight = 1.0f / height;
	ibumpscale = (255.0f * 3.0f) / bumpscale;
	out = outpixels;
	for (y = 0;y < height;y++)
	{
		for (x = 0;x < width;x++)
		{
			p0 = inpixels + (y * width + x) * 4;
			if (x == width - 1)
			{
				if (clamp)
					p1 = inpixels + (y * width + x) * 4;
				else
					p1 = inpixels + (y * width) * 4;
			}
			else
				p1 = inpixels + (y * width + x + 1) * 4;
			if (y == height - 1)
			{
				if (clamp)
					p2 = inpixels + (y * width + x) * 4;
				else
					p2 = inpixels + x * 4;
			}
			else
				p2 = inpixels + ((y + 1) * width + x) * 4;
			/*
			dv[0][0] = iwidth;
			dv[0][1] = 0;
			dv[0][2] = ((p0[0] + p0[1] + p0[2]) * ibumpscale) - ((p1[0] + p1[1] + p1[2]) * ibumpscale);
			dv[1][0] = 0;
			dv[1][1] = iheight;
			dv[1][2] = ((p2[0] + p2[1] + p2[2]) * ibumpscale) - ((p0[0] + p0[1] + p0[2]) * ibumpscale);
			n[0] = dv[0][1]*dv[1][2]-dv[0][2]*dv[1][1];
			n[1] = dv[0][2]*dv[1][0]-dv[0][0]*dv[1][2];
			n[2] = dv[0][0]*dv[1][1]-dv[0][1]*dv[1][0];
			*/
			n[0] = ((p1[0] + p1[1] + p1[2]) - (p0[0] + p0[1] + p0[2]));
			n[1] = ((p0[0] + p0[1] + p0[2]) - (p2[0] + p2[1] + p2[2]));
			n[2] = ibumpscale;
			VectorNormalize(n);
			/*
			// this should work for the bottom right triangle if anyone wants
			// code for that for some reason
			n[0] = ((p3[0] + p3[1] + p3[2]) - (p1[0] + p1[1] + p1[2]));
			n[1] = ((p2[0] + p2[1] + p2[2]) - (p3[0] + p3[1] + p3[2]));
			n[2] = ibumpscale;
			VectorNormalize(n);
			*/
			out[0] = 128.0f + n[0] * 127.0f;
			out[1] = 128.0f + n[1] * 127.0f;
			out[2] = 128.0f + n[2] * 127.0f;
			out[3] = 255;
			out += 4;
		}
	}
}

int image_loadskin(imageskin_t *s, char *shadername)
{
	int j;
	qbyte *bumppixels;
	int bumppixels_width, bumppixels_height;
	char name[MAX_QPATH];
	Image_StripImageExtension(shadername, name);
	memset(s, 0, sizeof(*s));
	s->basepixels = loadimagepixels(name, false, 0, 0);
	if (s->basepixels == NULL)
		return false;
	s->basepixels_width = image_width;
	s->basepixels_height = image_height;

	bumppixels = NULL;bumppixels_width = 0;bumppixels_height = 0;
	if (Image_CheckAlpha(s->basepixels, s->basepixels_width * s->basepixels_height, true))
	{
		s->maskpixels = Mem_Alloc(loadmodel->mempool, s->basepixels_width * s->basepixels_height * 4);
		s->maskpixels_width = s->basepixels_width;
		s->maskpixels_height = s->basepixels_height;
		memcpy(s->maskpixels, s->basepixels, s->maskpixels_width * s->maskpixels_height * 4);
		for (j = 0;j < s->basepixels_width * s->basepixels_height * 4;j += 4)
		{
			s->maskpixels[j+0] = 255;
			s->maskpixels[j+1] = 255;
			s->maskpixels[j+2] = 255;
		}
	}

	// _luma is supported for tenebrae compatibility
	// (I think it's a very stupid name, but oh well)
	if ((s->glowpixels = loadimagepixels(va("%s_glow", name), false, 0, 0)) != NULL
	 || (s->glowpixels = loadimagepixels(va("%s_luma", name), false, 0, 0)) != NULL)
	{
		s->glowpixels_width = image_width;
		s->glowpixels_height = image_height;
	}
	// _norm is the name used by tenebrae
	// (I don't like the name much)
	if ((s->nmappixels = loadimagepixels(va("%s_norm", name), false, 0, 0)) != NULL)
	{
		s->nmappixels_width = image_width;
		s->nmappixels_height = image_height;
	}
	else if ((bumppixels = loadimagepixels(va("%s_bump", name), false, 0, 0)) != NULL)
	{
		bumppixels_width = image_width;
		bumppixels_height = image_height;
	}
	if ((s->glosspixels = loadimagepixels(va("%s_gloss", name), false, 0, 0)) != NULL)
	{
		s->glosspixels_width = image_width;
		s->glosspixels_height = image_height;
	}
	if ((s->pantspixels = loadimagepixels(va("%s_pants", name), false, 0, 0)) != NULL)
	{
		s->pantspixels_width = image_width;
		s->pantspixels_height = image_height;
	}
	if ((s->shirtpixels = loadimagepixels(va("%s_shirt", name), false, 0, 0)) != NULL)
	{
		s->shirtpixels_width = image_width;
		s->shirtpixels_height = image_height;
	}

	if (s->nmappixels == NULL)
	{
		if (bumppixels != NULL)
		{
			if (r_shadow_bumpscale_bumpmap.value > 0)
			{
				s->nmappixels = Mem_Alloc(loadmodel->mempool, bumppixels_width * bumppixels_height * 4);
				s->nmappixels_width = bumppixels_width;
				s->nmappixels_height = bumppixels_height;
				Image_HeightmapToNormalmap(bumppixels, s->nmappixels, s->nmappixels_width, s->nmappixels_height, false, r_shadow_bumpscale_bumpmap.value);
			}
		}
		else
		{
			if (r_shadow_bumpscale_basetexture.value > 0)
			{
				s->nmappixels = Mem_Alloc(loadmodel->mempool, s->basepixels_width * s->basepixels_height * 4);
				s->nmappixels_width = s->basepixels_width;
				s->nmappixels_height = s->basepixels_height;
				Image_HeightmapToNormalmap(s->basepixels, s->nmappixels, s->nmappixels_width, s->nmappixels_height, false, r_shadow_bumpscale_basetexture.value);
			}
		}
	}
	if (bumppixels != NULL)
		Mem_Free(bumppixels);
	return true;
}

void image_freeskin(imageskin_t *s)
{
	if (s->basepixels)
		Mem_Free(s->basepixels);
	if (s->maskpixels)
		Mem_Free(s->maskpixels);
	if (s->nmappixels)
		Mem_Free(s->nmappixels);
	if (s->glowpixels)
		Mem_Free(s->glowpixels);
	if (s->glosspixels)
		Mem_Free(s->glosspixels);
	if (s->pantspixels)
		Mem_Free(s->pantspixels);
	if (s->shirtpixels)
		Mem_Free(s->shirtpixels);
	memset(s, 0, sizeof(*s));
}