/*
Copyright (C) 1997-2001 Id Software, Inc.

This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 2
of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.

*/
// r_surf.c: surface-related refresh code

#include "r_local.h"
#include "mod_local.h"

drawsurf_t	r_drawsurf;

uint				lightleft, sourcesstep, blocksize, sourcetstep;
uint				lightdelta, lightdeltastep;
uint				lightright, lightleftstep, lightrightstep, blockdivshift;
unsigned		blockdivmask;
void			*prowdestbase;
pixel_t	*pbasesource;
int				surfrowbytes;	// used by ASM files
unsigned		*r_lightptr;
int				r_stepback;
int				r_lightwidth;
int				r_numhblocks, r_numvblocks;
pixel_t	*r_source, *r_sourcemax;

void R_DrawSurfaceBlock8_mip0 (void);
void R_DrawSurfaceBlock8_mip1 (void);
void R_DrawSurfaceBlock8_mip2 (void);
void R_DrawSurfaceBlock8_mip3 (void);
void R_DrawSurfaceBlock8_Generic (void);
void R_DrawSurfaceBlock8_World (void);

static float worldlux_s, worldlux_t;

static void	(*surfmiptable[4])(void) = {
	R_DrawSurfaceBlock8_mip0,
	R_DrawSurfaceBlock8_mip1,
	R_DrawSurfaceBlock8_mip2,
	R_DrawSurfaceBlock8_mip3
};

//void R_BuildLightMap (void);
extern	unsigned		blocklights[10240];	// allow some very large lightmaps

float           surfscale;
qboolean        r_cache_thrash;         // set if surface cache is thrashing

int         sc_size;
surfcache_t	*sc_rover, *sc_base;

static int		rtable[MOD_FRAMES][MOD_FRAMES];

#if 1

static void R_BuildLightMap( void );
/*
===============
R_AddDynamicLights
===============
*/
void R_AddDynamicLights( msurface_t *surf )
{
	float		dist, rad, minlight;
	int		lnum, s, t, sd, td, smax, tmax;
	float		sl, tl, sacc, tacc;
	vec3_t		impact, origin_l;
	mextrasurf_t	*info = surf->info;
	int		sample_frac = 1.0;
	float		sample_size;
	mtexinfo_t	*tex;
	dlight_t		*dl;
	uint		*bl;

	// no dlighted surfaces here
	//if( !R_CountSurfaceDlights( surf )) return;

	sample_size = gEngfuncs.Mod_SampleSizeForFace( surf );
	smax = (info->lightextents[0] / sample_size) + 1;
	tmax = (info->lightextents[1] / sample_size) + 1;
	tex = surf->texinfo;

	if( FBitSet( tex->flags, TEX_WORLD_LUXELS ))
	{
		if( surf->texinfo->faceinfo )
			sample_frac = surf->texinfo->faceinfo->texture_step;
		else if( FBitSet( surf->texinfo->flags, TEX_EXTRA_LIGHTMAP ))
			sample_frac = LM_SAMPLE_EXTRASIZE;
		else sample_frac = LM_SAMPLE_SIZE;
	}

	for( lnum = 0; lnum < MAX_DLIGHTS; lnum++ )
	{
		if( !FBitSet( surf->dlightbits, BIT( lnum )))
			continue;	// not lit by this light

		dl = gEngfuncs.GetDynamicLight( lnum );

		// transform light origin to local bmodel space
		if( !tr.modelviewIdentity )
			Matrix4x4_VectorITransform( RI.objectMatrix, dl->origin, origin_l );
		else
			VectorCopy( dl->origin, origin_l );

		rad = dl->radius;
		dist = PlaneDiff( origin_l, surf->plane );
		rad -= fabs( dist );

		// rad is now the highest intensity on the plane
		minlight = dl->minlight;
		if( rad < minlight )
			continue;

		minlight = rad - minlight;

		if( surf->plane->type < 3 )
		{
			VectorCopy( origin_l, impact );
			impact[surf->plane->type] -= dist;
		}
		else VectorMA( origin_l, -dist, surf->plane->normal, impact );

		sl = DotProduct( impact, info->lmvecs[0] ) + info->lmvecs[0][3] - info->lightmapmins[0];
		tl = DotProduct( impact, info->lmvecs[1] ) + info->lmvecs[1][3] - info->lightmapmins[1];
		bl = blocklights;

		for( t = 0, tacc = 0; t < tmax; t++, tacc += sample_size )
		{
			td = (tl - tacc) * sample_frac;
			if( td < 0 ) td = -td;

			for( s = 0, sacc = 0; s < smax; s++, sacc += sample_size, bl += 1 )
			{
				sd = (sl - sacc) * sample_frac;
				if( sd < 0 ) sd = -sd;

				if( sd > td ) dist = sd + (td >> 1);
				else dist = td + (sd >> 1);

				if( dist < minlight )
				{
					//printf("dlight %f\n", dist);
					//*(void**)0 = 0;
					bl[0] +=  ((int)((rad - dist) * 256) * gEngfuncs.LightToTexGamma( (dl->color.r + dl->color.g + dl->color.b ) / 3) * 3) / 256;
					//bl[1] += ((int)((rad - dist) * 256) * 2.5) / 256;
					//bl[2] += ((int)((rad - dist) * 256) * 2.5) / 256;
				}
			}
		}
	}
}


/*
=================
R_BuildLightmap

Combine and scale multiple lightmaps into the floating
format in r_blocklights
=================
*/
static void R_BuildLightMap( void )
{
	int		smax, tmax;
	uint		*bl, scale;
	int		i, map, size, s, t;
	int		sample_size;
	msurface_t *surf = r_drawsurf.surf;
	mextrasurf_t	*info = surf->info;
	color24		*lm;
	qboolean dynamic = 0;

	sample_size = gEngfuncs.Mod_SampleSizeForFace( surf );
	smax = ( info->lightextents[0] / sample_size ) + 1;
	tmax = ( info->lightextents[1] / sample_size ) + 1;

	//smax = (surf->extents[0]>>4)+1;
	//tmax = (surf->extents[1]>>4)+1;

	size = smax * tmax;

	if( surf->flags & SURF_CONVEYOR )
	{
		smax = ( info->lightextents[0] * 3 / sample_size ) + 1;
		size = smax * tmax;
		memset( blocklights, 0xff, sizeof( uint ) * size );
		return;
	}

	lm = surf->samples;

	memset( blocklights, 0, sizeof( uint ) * size );

	// add all the lightmaps
	for( map = 0; map < MAXLIGHTMAPS && surf->styles[map] != 255; map++ )
	{
		scale = tr.lightstylevalue[surf->styles[map]];

		for( i = 0, bl = blocklights; i < size; i++, bl += 1, lm++ )
		{
			bl[0] += gEngfuncs.LightToTexGamma( lm->r ) * scale;
			bl[0] += gEngfuncs.LightToTexGamma( lm->g ) * scale;
			bl[0] += gEngfuncs.LightToTexGamma( lm->b ) * scale;

			//printf("test\n");
			//bl[1] += gEngfuncs.LightToTexGamma( lm->g ) * scale;
			//bl[2] += gEngfuncs.LightToTexGamma( lm->b ) * scale;
		}
	}

	// add all the dynamic lights
	if( surf->dlightframe == tr.framecount )
		R_AddDynamicLights( surf );

	// Put into texture format
	//stride -= (smax << 2);
	//bl = blocklights;

	/*for( t = 0; t < tmax; t++, dest += stride )
	{
		for( s = 0; s < smax; s++ )
		{
			dest[0] = Q_min((bl[0] >> 7), 255 );
			//dest[1] = Q_min((bl[1] >> 7), 255 );
			//dest[2] = Q_min((bl[2] >> 7), 255 );
			//dest[3] = 255;

			bl += 3;
			dest += 4;
		}
	}*/
	// bound, invert, and shift
		for (i=0 ; i<size ; i++)
		{
			t = (int)blocklights[i];
			if (t < 0)
				t = 0;
			if( t > 65535 * 3 )
				t = 65535 * 3;
			t = t / 2048 / 3;//(255*256 - t) >> (8 - VID_CBITS);

			//if (t < (1 << 6))
				//t = (1 << 6);
			t = t << 8;

			blocklights[i] = t;
		}
}
#else

/*
===============
R_BuildLightMap

Combine and scale multiple lightmaps into the 8.8 format in blocklights
===============
*/
void R_BuildLightMap (void)
{
	int			smax, tmax;
	int			t;
	int			i, size;
	byte		*lightmap;
	unsigned	scale;
	int			maps;
	msurface_t	*surf;

	surf = r_drawsurf.surf;

	//smax = (surf->extents[0]>>4)+1;
	//tmax = (surf->extents[1]>>4)+1;
	mextrasurf_t	*info = surf->info;
	int sample_size = gEngfuncs.Mod_SampleSizeForFace( surf );
	smax = ( info->lightextents[0] / sample_size ) + 1;
	tmax = ( info->lightextents[1] / sample_size ) + 1;

	size = smax*tmax;

	if (r_fullbright->value )
	{
		for (i=0 ; i<size ; i++)
			blocklights[i] = 0;
		return;
	}

// clear to no light
	for (i=0 ; i<size ; i++)
		blocklights[i] = 0;


// add all the lightmaps
	lightmap = surf->samples;
	if (lightmap)
		for (maps = 0 ; maps < MAXLIGHTMAPS && surf->styles[maps] != 255 ;
			 maps++)
		{
			scale = r_drawsurf.lightadj[maps];	// 8.8 fraction
			for (i=0 ; i<size ; i++)
			{
				blocklights[i] += lightmap[i*3] * 2.5; // * scale;
				blocklights[i] += lightmap[i*3+1] * 2.5; // * scale;
				blocklights[i] += lightmap[i*3+2] * 2.5; // * scale;
			}
			lightmap += size;	// skip to next lightmap
		}

// add all the dynamic lights
	//if (surf->dlightframe == r_framecount)
		//R_AddDynamicLights ();

// bound, invert, and shift
	/*for (i=0 ; i<size ; i++)
	{
		t = (int)blocklights[i];
		if (t < 0)
			t = 0;
		t = (255*256 - t) >> (8 - VID_CBITS);

		if (t < (1 << 6))
			t = (1 << 6);

		blocklights[i] = t;
	}*/
	for (i=0 ; i<size ; i++)
	{
		t = (int)blocklights[i];
		if (t < 0)
			t = 0;
		if( t > 767 )
			t = 767;
		t = t * 31 / 256/3;//(255*256 - t) >> (8 - VID_CBITS);

		//if (t < (1 << 6))
			//t = (1 << 6);
		t = t << 8;

		blocklights[i] = t;
	}
}

#endif

void GL_InitRandomTable( void )
{
	int	tu, tv;

	for( tu = 0; tu < MOD_FRAMES; tu++ )
	{
		for( tv = 0; tv < MOD_FRAMES; tv++ )
		{
			rtable[tu][tv] = gEngfuncs.COM_RandomLong( 0, 0x7FFF );
		}
	}

	gEngfuncs.COM_SetRandomSeed( 0 );
}

/*
===============
R_TextureAnim

Returns the proper texture for a given time and base texture, do not process random tiling
===============
*/
texture_t *R_TextureAnim( texture_t *b )
{
	texture_t *base = b;
	int	count, reletive;

	if( RI.currententity->curstate.frame )
	{
		if( base->alternate_anims )
			base = base->alternate_anims;
	}

	if( !base->anim_total )
		return base;
	if( base->name[0] == '-' )
	{
		return b; // already tiled
	}
	else
	{
		int	speed;

		// Quake1 textures uses 10 frames per second
		if( FBitSet( R_GetTexture( base->gl_texturenum )->flags, TF_QUAKEPAL ))
			speed = 10;
		else speed = 20;

		reletive = (int)(gpGlobals->time * speed) % base->anim_total;
	}


	count = 0;

	while( base->anim_min > reletive || base->anim_max <= reletive )
	{
		base = base->anim_next;

		if( !base || ++count > MOD_FRAMES )
			return b;
	}

	return base;
}

/*
===============
R_TextureAnimation

Returns the proper texture for a given time and surface
===============
*/
texture_t *R_TextureAnimation( msurface_t *s )
{
	texture_t	*base = s->texinfo->texture;
	int	count, reletive;

	if( RI.currententity && RI.currententity->curstate.frame )
	{
		if( base->alternate_anims )
			base = base->alternate_anims;
	}

	if( !base->anim_total )
		return base;

	if( base->name[0] == '-' )
	{
		int	tx = (int)((s->texturemins[0] + (base->width << 16)) / base->width) % MOD_FRAMES;
		int	ty = (int)((s->texturemins[1] + (base->height << 16)) / base->height) % MOD_FRAMES;

		reletive = rtable[tx][ty] % base->anim_total;
	}
	else
	{
		int	speed;

		// Quake1 textures uses 10 frames per second
		if( FBitSet( R_GetTexture( base->gl_texturenum )->flags, TF_QUAKEPAL ))
			speed = 10;
		else speed = 20;

		reletive = (int)(gpGlobals->time * speed) % base->anim_total;
	}

	count = 0;

	while( base->anim_min > reletive || base->anim_max <= reletive )
	{
		base = base->anim_next;

		if( !base || ++count > MOD_FRAMES )
			return s->texinfo->texture;
	}

	return base;
}

/*
===============
R_DrawSurface
===============
*/
void R_DrawSurface (void)
{
	pixel_t	*basetptr;
	int				smax, tmax, twidth;
	int				u;
	int				soffset, basetoffset, texwidth;
	int				horzblockstep;
	pixel_t	*pcolumndest;
	void			(*pblockdrawer)(void);
	image_t			*mt;
	uint sample_size, sample_bits, sample_pot;

	surfrowbytes = r_drawsurf.rowbytes;

	sample_size = LM_SAMPLE_SIZE_AUTO(r_drawsurf.surf);
	if( sample_size == 16 )
		sample_bits = 4, sample_pot = sample_size;
	else
	{
		sample_bits = tr.sample_bits;

		if( sample_bits == -1 )
		{
			sample_bits = 0;
			for( sample_pot = 1; sample_pot < sample_size; sample_pot <<= 1, sample_bits++ );
		}
		else
			sample_pot = 1 << sample_bits;
	}
	mt = r_drawsurf.image;

	r_source = mt->pixels[r_drawsurf.surfmip];

// the fractional light values should range from 0 to (VID_GRADES - 1) << 16
// from a source range of 0 - 255

	texwidth = mt->width >> r_drawsurf.surfmip;

	blocksize = sample_pot >> r_drawsurf.surfmip;
	blockdivshift = sample_bits - r_drawsurf.surfmip;
	blockdivmask = (1 << blockdivshift) - 1;

	if( sample_size == 16 )
		r_lightwidth = ( r_drawsurf.surf->info->lightextents[0]>>4)+1;
	else
		r_lightwidth = ( r_drawsurf.surf->info->lightextents[0] / sample_size ) + 1;

	r_numhblocks = r_drawsurf.surfwidth >> blockdivshift;
	r_numvblocks = r_drawsurf.surfheight >> blockdivshift;


//==============================

	if( sample_size == 16 )
		pblockdrawer = surfmiptable[r_drawsurf.surfmip];
	else
		pblockdrawer = R_DrawSurfaceBlock8_Generic;

// TODO: only needs to be set when there is a display settings change
	horzblockstep = blocksize;

	smax = mt->width >> r_drawsurf.surfmip;
	twidth = texwidth;
	tmax = mt->height >> r_drawsurf.surfmip;
	sourcetstep = texwidth;
	r_stepback = tmax * twidth;

	r_sourcemax = r_source + (tmax * smax);

	// glitchy and slow way to draw some lightmap
	if( r_drawsurf.surf->texinfo->flags & TEX_WORLD_LUXELS )
	{
		worldlux_s = r_drawsurf.surf->extents[0] / r_drawsurf.surf->info->lightextents[0];
		worldlux_t = r_drawsurf.surf->extents[1] / r_drawsurf.surf->info->lightextents[1];
		if( worldlux_s == 0 )
			worldlux_s = 1;
		if( worldlux_t == 0 )
			worldlux_t = 1;

		soffset = r_drawsurf.surf->texturemins[0];
		basetoffset = r_drawsurf.surf->texturemins[1];
		//soffset =  r_drawsurf.surf->info->lightmapmins[0] * worldlux_s;
		//basetoffset = r_drawsurf.surf->info->lightmapmins[1] * worldlux_t;
		// << 16 components are to guarantee positive values for %
		soffset = ((soffset >> r_drawsurf.surfmip) + (smax << 16)) % smax;
		basetptr = &r_source[((((basetoffset >> r_drawsurf.surfmip)
			+ (tmax << 16)) % tmax) * twidth)];

		pcolumndest = r_drawsurf.surfdat;

		for (u=0 ; u<r_numhblocks; u++)
		{
			r_lightptr = blocklights + (int)(u/ (worldlux_s+0.5f));

			prowdestbase = pcolumndest;

			pbasesource = basetptr + soffset;

			R_DrawSurfaceBlock8_World();

			soffset = soffset + blocksize;
			if (soffset >= smax)
				soffset = 0;

			pcolumndest += horzblockstep;
		}
		return;
	}

	soffset =  r_drawsurf.surf->info->lightmapmins[0];
	basetoffset = r_drawsurf.surf->info->lightmapmins[1];

// << 16 components are to guarantee positive values for %
	soffset = ((soffset >> r_drawsurf.surfmip) + (smax << 16)) % smax;
	basetptr = &r_source[((((basetoffset >> r_drawsurf.surfmip)
		+ (tmax << 16)) % tmax) * twidth)];

	pcolumndest = r_drawsurf.surfdat;

	for (u=0 ; u<r_numhblocks; u++)
	{
		r_lightptr = blocklights + u;

		prowdestbase = pcolumndest;

		pbasesource = basetptr + soffset;

		(*pblockdrawer)();

		soffset = soffset + blocksize;
		if (soffset >= smax)
			soffset = 0;

		pcolumndest += horzblockstep;
	}
    // test what if we have very slow cache building
    //usleep(10000);
}


//=============================================================================

#if	!id386
#define BLEND_LM(pix, light) vid.colormap[(pix >> 3) | ((light & 0x1f00) << 5)] | ( pix & 7 );

/*
================
R_DrawSurfaceBlock8_World

Does not draw lightmap correclty, but scale it correctly. Better than nothing
================
*/
void R_DrawSurfaceBlock8_World (void)
{
	int				v, i, b;
	uint lightstep, lighttemp, light;
	pixel_t	pix, *psource, *prowdest;
	int lightpos = 0;

	psource = pbasesource;
	prowdest = prowdestbase;

	for (v=0 ; v<r_numvblocks ; v++)
	{
	// FIXME: make these locals?
	// FIXME: use delta rather than both right and left, like ASM?
		lightleft = r_lightptr[(lightpos/r_lightwidth) * r_lightwidth];
		lightright = r_lightptr[(lightpos/r_lightwidth) * r_lightwidth+1];
		lightpos += r_lightwidth / worldlux_s;
		lightleftstep = (r_lightptr[(lightpos/r_lightwidth) * r_lightwidth] - lightleft) >> (4-r_drawsurf.surfmip);
		lightrightstep =(r_lightptr[(lightpos/r_lightwidth) * r_lightwidth+1] - lightright) >> (4-r_drawsurf.surfmip);

		for (i=0 ; i<blocksize ; i++)
		{
			lighttemp = lightleft - lightright;
			lightstep = lighttemp >> (4-r_drawsurf.surfmip);

			light = lightright;

			for (b=blocksize-1; b>=0; b--)
			{
				//pix = psource[(uint)(b * worldlux_s)];
				pix = psource[b];
				prowdest[b] = BLEND_LM(pix, light);
				if( pix == TRANSPARENT_COLOR )
					prowdest[b] = TRANSPARENT_COLOR;
						//((unsigned char *)vid.colormap)
						//[(light & 0xFF00) + pix];
				light += lightstep;
			}

			psource += sourcetstep;
			lightright += lightrightstep;
			lightleft += lightleftstep;
			prowdest += surfrowbytes;
		}

		if (psource >= r_sourcemax)
			psource -= r_stepback;
	}
}


/*
================
R_DrawSurfaceBlock8_Generic
================
*/
void R_DrawSurfaceBlock8_Generic (void)
{
	int				v, i, b;
	uint lightstep, lighttemp, light;
	pixel_t	pix, *psource, *prowdest;

	psource = pbasesource;
	prowdest = prowdestbase;

	for (v=0 ; v<r_numvblocks ; v++)
	{
	// FIXME: make these locals?
	// FIXME: use delta rather than both right and left, like ASM?
		lightleft = r_lightptr[0];
		lightright = r_lightptr[1];
		r_lightptr += r_lightwidth;
		lightleftstep = (r_lightptr[0] - lightleft) >> (4-r_drawsurf.surfmip);
		lightrightstep = (r_lightptr[1] - lightright) >> (4-r_drawsurf.surfmip);

		for (i=0 ; i<blocksize ; i++)
		{
			lighttemp = lightleft - lightright;
			lightstep = lighttemp >> (4-r_drawsurf.surfmip);

			light = lightright;

			for (b=blocksize-1; b>=0; b--)
			{
				pix = psource[b];
				prowdest[b] = BLEND_LM(pix, light);
				if( pix == TRANSPARENT_COLOR )
					prowdest[b] = TRANSPARENT_COLOR;
						//((unsigned char *)vid.colormap)
						//[(light & 0xFF00) + pix];
				light += lightstep;
			}

			psource += sourcetstep;
			lightright += lightrightstep;
			lightleft += lightleftstep;
			prowdest += surfrowbytes;
		}

		if (psource >= r_sourcemax)
			psource -= r_stepback;
	}
}


/*
================
R_DrawSurfaceBlock8_mip0
================
*/
void R_DrawSurfaceBlock8_mip0 (void)
{
	int				v, i, b;
	uint lightstep, lighttemp, light;
	pixel_t	pix, *psource, *prowdest;

	psource = pbasesource;
	prowdest = prowdestbase;

	for (v=0 ; v<r_numvblocks ; v++)
	{
	// FIXME: make these locals?
	// FIXME: use delta rather than both right and left, like ASM?
		lightleft = r_lightptr[0];
		lightright = r_lightptr[1];
		r_lightptr += r_lightwidth;
		lightleftstep = (r_lightptr[0] - lightleft) >> 4;
		lightrightstep = (r_lightptr[1] - lightright) >> 4;

		for (i=0 ; i<16 ; i++)
		{
			lighttemp = lightleft - lightright;
			lightstep = lighttemp >> 4;

			light = lightright;

			for (b=15; b>=0; b--)
			{
				pix = psource[b];
				prowdest[b] = BLEND_LM(pix, light);
				if( pix == TRANSPARENT_COLOR )
					prowdest[b] = TRANSPARENT_COLOR;

						// pix;
						//((unsigned char *)vid.colormap)
						//[(light & 0xFF00) + pix];
				light += lightstep;
			}

			psource += sourcetstep;
			lightright += lightrightstep;
			lightleft += lightleftstep;
			prowdest += surfrowbytes;
		}

		if (psource >= r_sourcemax)
			psource -= r_stepback;
	}
}


/*
================
R_DrawSurfaceBlock8_mip1
================
*/
void R_DrawSurfaceBlock8_mip1 (void)
{
	int				v, i, b;
	uint lightstep, lighttemp, light;
	pixel_t	pix, *psource, *prowdest;

	psource = pbasesource;
	prowdest = prowdestbase;

	for (v=0 ; v<r_numvblocks ; v++)
	{
	// FIXME: make these locals?
	// FIXME: use delta rather than both right and left, like ASM?
		lightleft = r_lightptr[0];
		lightright = r_lightptr[1];
		r_lightptr += r_lightwidth;
		lightleftstep = (r_lightptr[0] - lightleft) >> 3;
		lightrightstep = (r_lightptr[1] - lightright) >> 3;

		for (i=0 ; i<8 ; i++)
		{
			lighttemp = lightleft - lightright;
			lightstep = lighttemp >> 3;

			light = lightright;

			for (b=7; b>=0; b--)
			{
				pix = psource[b];
				prowdest[b] = BLEND_LM(pix, light);
						//((unsigned char *)vid.colormap)
						//[(light & 0xFF00) + pix];
				light += lightstep;
			}

			psource += sourcetstep;
			lightright += lightrightstep;
			lightleft += lightleftstep;
			prowdest += surfrowbytes;
		}

		if (psource >= r_sourcemax)
			psource -= r_stepback;
	}
}


/*
================
R_DrawSurfaceBlock8_mip2
================
*/
void R_DrawSurfaceBlock8_mip2 (void)
{
	int				v, i, b;
	uint lightstep, lighttemp, light;
	pixel_t	pix, *psource, *prowdest;

	psource = pbasesource;
	prowdest = prowdestbase;

	for (v=0 ; v<r_numvblocks ; v++)
	{
	// FIXME: make these locals?
	// FIXME: use delta rather than both right and left, like ASM?
		lightleft = r_lightptr[0];
		lightright = r_lightptr[1];
		r_lightptr += r_lightwidth;
		lightleftstep = (r_lightptr[0] - lightleft) >> 2;
		lightrightstep = (r_lightptr[1] - lightright) >> 2;

		for (i=0 ; i<4 ; i++)
		{
			lighttemp = lightleft - lightright;
			lightstep = lighttemp >> 2;

			light = lightright;

			for (b=3; b>=0; b--)
			{
				pix = psource[b];
				prowdest[b] = BLEND_LM(pix, light);;
						//((unsigned char *)vid.colormap)
						//[(light & 0xFF00) + pix];
				light += lightstep;
			}

			psource += sourcetstep;
			lightright += lightrightstep;
			lightleft += lightleftstep;
			prowdest += surfrowbytes;
		}

		if (psource >= r_sourcemax)
			psource -= r_stepback;
	}
}


/*
================
R_DrawSurfaceBlock8_mip3
================
*/
void R_DrawSurfaceBlock8_mip3 (void)
{
	int				v, i, b;
	uint lightstep, lighttemp, light;
	pixel_t	pix, *psource, *prowdest;

	psource = pbasesource;
	prowdest = prowdestbase;

	for (v=0 ; v<r_numvblocks ; v++)
	{
	// FIXME: make these locals?
	// FIXME: use delta rather than both right and left, like ASM?
		lightleft = r_lightptr[0];
		lightright = r_lightptr[1];
		r_lightptr += r_lightwidth;
		lightleftstep = (r_lightptr[0] - lightleft) >> 1;
		lightrightstep = (r_lightptr[1] - lightright) >> 1;

		for (i=0 ; i<2 ; i++)
		{
			lighttemp = lightleft - lightright;
			lightstep = lighttemp >> 1;

			light = lightright;

			for (b=1; b>=0; b--)
			{
				pix = psource[b];
				prowdest[b] = BLEND_LM(pix, light);;
						//((unsigned char *)vid.colormap)
						//[(light & 0xFF00) + pix];
				light += lightstep;
			}

			psource += sourcetstep;
			lightright += lightrightstep;
			lightleft += lightleftstep;
			prowdest += surfrowbytes;
		}

		if (psource >= r_sourcemax)
			psource -= r_stepback;
	}
}

#endif


//============================================================================


/*
================
R_InitCaches

================
*/
void R_InitCaches (void)
{
	int		size;
	int		pix;

	// calculate size to allocate
	if (sw_surfcacheoverride.value)
	{
		size = sw_surfcacheoverride.value;
	}
	else
	{
		size = SURFCACHE_SIZE_AT_320X240 * 2;

		pix = vid.width * vid.height * 2;
		if (pix > 64000)
			size += (pix-64000)*3;
	}

	// round up to page size
	size = (size + 8191) & ~8191;

	gEngfuncs.Con_Printf ("%s surface cache\n", Q_memprint(size));

	sc_size = size;
	if( sc_base )
	{
		D_FlushCaches(  );
		Mem_Free( sc_base );
	}
	sc_base = (surfcache_t *)Mem_Calloc(r_temppool,size);
	sc_rover = sc_base;

	sc_base->next = NULL;
	sc_base->owner = NULL;
	sc_base->size = sc_size;
}


/*
==================
D_FlushCaches
==================
*/
void D_FlushCaches( void )
{
	surfcache_t     *c;

	// if newmap, surfaces already freed
	if( !tr.map_unload )
	{
		for(c = sc_base ; c ; c = c->next )
		{
			if ( c->owner )
				*c->owner = NULL;
		}
	}

	sc_rover = sc_base;
	sc_base->next = NULL;
	sc_base->owner = NULL;
	sc_base->size = sc_size;
}

/*
=================
D_SCAlloc
=================
*/
surfcache_t     *D_SCAlloc (int width, int size)
{
	surfcache_t             *new;
	qboolean                wrapped_this_time;

	if ((width < 0) )// || (width > 256))
		gEngfuncs.Host_Error ("D_SCAlloc: bad cache width %d\n", width);

	if ((size <= 0) || (size > 0x10000000))
		gEngfuncs.Host_Error ("D_SCAlloc: bad cache size %d\n", size);

	size = (int)&((surfcache_t *)0)->data[size];
	size = (size + 3) & ~3;
	if (size > sc_size)
		gEngfuncs.Host_Error ("D_SCAlloc: %i > cache size of %i",size, sc_size);

// if there is not size bytes after the rover, reset to the start
	wrapped_this_time = false;

	if ( !sc_rover || (byte *)sc_rover - (byte *)sc_base > sc_size - size)
	{
		if (sc_rover)
		{
			wrapped_this_time = true;
		}
		sc_rover = sc_base;
	}

// colect and free surfcache_t blocks until the rover block is large enough
	new = sc_rover;
	if (sc_rover->owner)
		*sc_rover->owner = NULL;

	while (new->size < size)
	{
	// free another
		sc_rover = sc_rover->next;
		if (!sc_rover)
			gEngfuncs.Host_Error ("D_SCAlloc: hit the end of memory");
		if (sc_rover->owner)
			*sc_rover->owner = NULL;

		new->size += sc_rover->size;
		new->next = sc_rover->next;
	}

// create a fragment out of any leftovers
	if (new->size - size > 256)
	{
		sc_rover = (surfcache_t *)( (byte *)new + size);
		sc_rover->size = new->size - size;
		sc_rover->next = new->next;
		sc_rover->width = 0;
		sc_rover->owner = NULL;
		new->next = sc_rover;
		new->size = size;
	}
	else
		sc_rover = new->next;

	new->width = width;
// DEBUG
	if (width > 0)
		new->height = (size - sizeof(*new) + sizeof(new->data)) / width;

	new->owner = NULL;              // should be set properly after return

	if (d_roverwrapped)
	{
		if (wrapped_this_time || (sc_rover >= d_initial_rover))
			r_cache_thrash = true;
	}
	else if (wrapped_this_time)
	{
		d_roverwrapped = true;
	}

	return new;
}

//=============================================================================
void R_DecalComputeBasis( msurface_t *surf, int flags, vec3_t textureSpaceBasis[3] );
void R_DrawSurfaceDecals( void )
{
	msurface_t *fa = r_drawsurf.surf;
	decal_t *p;

	for( p = fa->pdecals; p; p = p->pnext)
	{
			pixel_t	*dest, *source;
			vec4_t		textureU, textureV;
			image_t *tex = R_GetTexture( p->texture );
			int s1 = 0,t1 = 0, s2 = tex->width, t2 = tex->height;
			unsigned int				height;
			unsigned int				f, fstep;
			int				skip;
			pixel_t *buffer;
			qboolean transparent;
			int x, y, u,v, sv, w, h;
			vec3_t basis[3];

			Vector4Copy( fa->texinfo->vecs[0], textureU );
			Vector4Copy( fa->texinfo->vecs[1], textureV );

			R_DecalComputeBasis( fa, 0, basis );

			w = fabs( tex->width  * DotProduct( textureU, basis[0] )) +
					fabs( tex->height * DotProduct( textureU, basis[1] ));
			h = fabs( tex->width  * DotProduct( textureV, basis[0] )) +
					fabs( tex->height * DotProduct( textureV, basis[1] ));

			// project decal center into the texture space of the surface
			x = DotProduct( p->position, textureU ) + textureU[3] - fa->texturemins[0] - w/2;
			y = DotProduct( p->position, textureV ) + textureV[3] - fa->texturemins[1] - h/2;

			x = x >> r_drawsurf.surfmip;
			y = y >> r_drawsurf.surfmip;
			w = w >> r_drawsurf.surfmip;
			h = h >> r_drawsurf.surfmip;

			if( w < 1 || h < 1 )
				continue;

			if( x < 0 )
			{
				s1 += (-x)*(s2-s1) / w;
				x = 0;
			}
			if( x + w > r_drawsurf.surfwidth )
			{
				s2 -= (x + w - r_drawsurf.surfwidth) * (s2 - s1)/ w ;
				w = r_drawsurf.surfwidth - x;
			}
			if( y + h > r_drawsurf.surfheight )
			{
				t2 -= (y + h - r_drawsurf.surfheight) * (t2 - t1) / h;
				h = r_drawsurf.surfheight - y;
			}

			if( s1 < 0 )
				s1 = 0;
			if( t1 < 0 )
				t1 = 0;

			if( s2 > tex->width )
				s2 = tex->width;
			if( t2 > tex->height )
				t2 = tex->height;

			if( !tex->pixels[0] || s1 >= s2 || t1 >= t2 || !w )
				continue;

			if( tex->alpha_pixels )
			{
				buffer = tex->alpha_pixels;
				transparent = true;
			}
			else
				buffer = tex->pixels[0];

			height = h;
			if (y < 0)
			{
				skip = -y;
				height += y;
				y = 0;
			}
			else
				skip = 0;

			dest = ((pixel_t*)r_drawsurf.surfdat) + y * r_drawsurf.rowbytes + x;

			for (v=0 ; v<height ; v++)
			{
				//int alpha1 = vid.alpha;
				sv = (skip + v)*(t2-t1)/h + t1;
				source = buffer + sv*tex->width + s1;

				{
					f = 0;
					fstep = (s2-s1)*0x10000/w;
					if( w == s2 - s1 )
						fstep = 0x10000;

					for (u=0 ; u<w ; u++)
					{
						pixel_t src = source[f>>16];
						int alpha = 7;
						f += fstep;

						if( transparent )
						{
							alpha &= src >> (16 - 3);
							src = src << 3;
						}

						if( alpha <= 0 )
							continue;

						if( alpha < 7) // && (vid.rendermode == kRenderTransAlpha || vid.rendermode == kRenderTransTexture ) )
						{
							pixel_t screen = dest[u]; //  | 0xff & screen & src ;
							if( screen == TRANSPARENT_COLOR )
								continue;
							dest[u] = BLEND_ALPHA( alpha, src, screen);

						}
						else
							dest[u] = src;

					}
				}
				dest += r_drawsurf.rowbytes;
			}
	}

}

/*
================
D_CacheSurface
================
*/
surfcache_t *D_CacheSurface (msurface_t *surface, int miplevel)
{
	surfcache_t     *cache;
	int maps;
//
// if the surface is animating or flashing, flush the cache
//
	r_drawsurf.image = R_GetTexture(R_TextureAnimation (surface)->gl_texturenum);

	// does not support conveyors with world luxels now
	if( surface->texinfo->flags & TEX_WORLD_LUXELS )
		surface->flags &= ~SURF_CONVEYOR;

	if( surface->flags & SURF_CONVEYOR)
	{
		if( miplevel >= 1)
		{
			surface->extents[0] = surface->info->lightextents[0] * LM_SAMPLE_SIZE_AUTO( r_drawsurf.surf ) * 2 ;
			surface->info->lightmapmins[0] = -surface->info->lightextents[0] * LM_SAMPLE_SIZE_AUTO( r_drawsurf.surf );
		}
		else
		{
			surface->extents[0] = surface->info->lightextents[0] * LM_SAMPLE_SIZE_AUTO( r_drawsurf.surf ) ;
			surface->info->lightmapmins[0] = -surface->info->lightextents[0] * LM_SAMPLE_SIZE_AUTO( r_drawsurf.surf )/2;
		}
	}
	/// todo: port this
	//r_drawsurf.lightadj[0] = r_newrefdef.lightstyles[surface->styles[0]].white*128;
	//r_drawsurf.lightadj[1] = r_newrefdef.lightstyles[surface->styles[1]].white*128;
	//r_drawsurf.lightadj[2] = r_newrefdef.lightstyles[surface->styles[2]].white*128;
	//r_drawsurf.lightadj[3] = r_newrefdef.lightstyles[surface->styles[3]].white*128;

//
// see if the cache holds apropriate data
//
	cache = CACHESPOT(surface)[miplevel];

	// check for lightmap modification
	for( maps = 0; maps < MAXLIGHTMAPS && surface->styles[maps] != 255; maps++ )
	{
		if( tr.lightstylevalue[surface->styles[maps]] != surface->cached_light[maps] )
		{
			surface->dlightframe = tr.framecount;
		}
	}


	if (cache && !cache->dlight && surface->dlightframe != tr.framecount
			&& cache->image == r_drawsurf.image
			&& cache->lightadj[0] == r_drawsurf.lightadj[0]
			&& cache->lightadj[1] == r_drawsurf.lightadj[1]
			&& cache->lightadj[2] == r_drawsurf.lightadj[2]
			&& cache->lightadj[3] == r_drawsurf.lightadj[3] )
		return cache;

	if( surface->dlightframe == tr.framecount )
	{
		int i;
		// invalidate dlight cache
		for( i = 0; i < 4; i++)
		{
			if( CACHESPOT(surface)[i] )
				CACHESPOT(surface)[i]->image = NULL;
		}
	}
//
// determine shape of surface
//
	surfscale = 1.0 / (1<<miplevel);
	r_drawsurf.surfmip = miplevel;
	if( surface->flags & SURF_CONVEYOR )
		r_drawsurf.surfwidth = surface->extents[0] >> miplevel;
	else
		r_drawsurf.surfwidth = surface->info->lightextents[0] >> miplevel;
	r_drawsurf.rowbytes = r_drawsurf.surfwidth;
	r_drawsurf.surfheight = surface->info->lightextents[1] >> miplevel;

	// use texture space if world luxels used
	if( surface->texinfo->flags & TEX_WORLD_LUXELS )
	{
		r_drawsurf.surfwidth = surface->extents[0] >> miplevel;
		r_drawsurf.rowbytes = r_drawsurf.surfwidth;
		r_drawsurf.surfheight = surface->extents[1] >> miplevel;
	}


//
// allocate memory if needed
//
	if (!cache)     // if a texture just animated, don't reallocate it
	{
		cache = D_SCAlloc (r_drawsurf.surfwidth,
						   r_drawsurf.surfwidth * r_drawsurf.surfheight * 2);
		CACHESPOT(surface)[miplevel] = cache;
		cache->owner = &CACHESPOT(surface)[miplevel];
		cache->mipscale = surfscale;
	}

	if (surface->dlightframe == tr.framecount)
		cache->dlight = 1;
	else
		cache->dlight = 0;

	r_drawsurf.surfdat = (pixel_t *)cache->data;

	cache->image = r_drawsurf.image;
	cache->lightadj[0] = r_drawsurf.lightadj[0];
	cache->lightadj[1] = r_drawsurf.lightadj[1];
	cache->lightadj[2] = r_drawsurf.lightadj[2];
	cache->lightadj[3] = r_drawsurf.lightadj[3];
	for( maps = 0; maps < MAXLIGHTMAPS && surface->styles[maps] != 255; maps++ )
	{
		surface->cached_light[maps] = tr.lightstylevalue[surface->styles[maps]];
	}
//
// draw and light the surface texture
//
	r_drawsurf.surf = surface;

	//c_surf++;

	// calculate the lightings
	R_BuildLightMap ( );

	// rasterize the surface into the cache
	R_DrawSurface ();
	R_DrawSurfaceDecals();

	return cache;
}