kolibrios/contrib/other/sdlquake-1.0.9/r_surf.c
CleverMouse 3cf7852e03 autobuild sdlquake
git-svn-id: svn://kolibrios.org@5131 a494cfbc-eb01-0410-851d-a64ba20cac60
2014-09-18 11:46:53 +00:00

679 lines
14 KiB
C

/*
Copyright (C) 1996-1997 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 "quakedef.h"
#include "r_local.h"
drawsurf_t r_drawsurf;
int lightleft, sourcesstep, blocksize, sourcetstep;
int lightdelta, lightdeltastep;
int lightright, lightleftstep, lightrightstep, blockdivshift;
unsigned blockdivmask;
void *prowdestbase;
unsigned char *pbasesource;
int surfrowbytes; // used by ASM files
unsigned *r_lightptr;
int r_stepback;
int r_lightwidth;
int r_numhblocks, r_numvblocks;
unsigned char *r_source, *r_sourcemax;
void R_DrawSurfaceBlock8_mip0 (void);
void R_DrawSurfaceBlock8_mip1 (void);
void R_DrawSurfaceBlock8_mip2 (void);
void R_DrawSurfaceBlock8_mip3 (void);
static void (*surfmiptable[4])(void) = {
R_DrawSurfaceBlock8_mip0,
R_DrawSurfaceBlock8_mip1,
R_DrawSurfaceBlock8_mip2,
R_DrawSurfaceBlock8_mip3
};
unsigned blocklights[18*18];
/*
===============
R_AddDynamicLights
===============
*/
void R_AddDynamicLights (void)
{
msurface_t *surf;
int lnum;
int sd, td;
float dist, rad, minlight;
vec3_t impact, local;
int s, t;
int i;
int smax, tmax;
mtexinfo_t *tex;
surf = r_drawsurf.surf;
smax = (surf->extents[0]>>4)+1;
tmax = (surf->extents[1]>>4)+1;
tex = surf->texinfo;
for (lnum=0 ; lnum<MAX_DLIGHTS ; lnum++)
{
if ( !(surf->dlightbits & (1<<lnum) ) )
continue; // not lit by this light
rad = cl_dlights[lnum].radius;
dist = DotProduct (cl_dlights[lnum].origin, surf->plane->normal) -
surf->plane->dist;
rad -= fabs(dist);
minlight = cl_dlights[lnum].minlight;
if (rad < minlight)
continue;
minlight = rad - minlight;
for (i=0 ; i<3 ; i++)
{
impact[i] = cl_dlights[lnum].origin[i] -
surf->plane->normal[i]*dist;
}
local[0] = DotProduct (impact, tex->vecs[0]) + tex->vecs[0][3];
local[1] = DotProduct (impact, tex->vecs[1]) + tex->vecs[1][3];
local[0] -= surf->texturemins[0];
local[1] -= surf->texturemins[1];
for (t = 0 ; t<tmax ; t++)
{
td = local[1] - t*16;
if (td < 0)
td = -td;
for (s=0 ; s<smax ; s++)
{
sd = local[0] - s*16;
if (sd < 0)
sd = -sd;
if (sd > td)
dist = sd + (td>>1);
else
dist = td + (sd>>1);
if (dist < minlight)
#ifdef QUAKE2
{
unsigned temp;
temp = (rad - dist)*256;
i = t*smax + s;
if (!cl_dlights[lnum].dark)
blocklights[i] += temp;
else
{
if (blocklights[i] > temp)
blocklights[i] -= temp;
else
blocklights[i] = 0;
}
}
#else
blocklights[t*smax + s] += (rad - dist)*256;
#endif
}
}
}
}
/*
===============
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;
size = smax*tmax;
lightmap = surf->samples;
if (r_fullbright.value || !cl.worldmodel->lightdata)
{
for (i=0 ; i<size ; i++)
blocklights[i] = 0;
return;
}
// clear to ambient
for (i=0 ; i<size ; i++)
blocklights[i] = r_refdef.ambientlight<<8;
// add all the lightmaps
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] * 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 = (255*256 - (int)blocklights[i]) >> (8 - VID_CBITS);
if (t < (1 << 6))
t = (1 << 6);
blocklights[i] = t;
}
}
/*
===============
R_TextureAnimation
Returns the proper texture for a given time and base texture
===============
*/
texture_t *R_TextureAnimation (texture_t *base)
{
int reletive;
int count;
if (currententity->frame)
{
if (base->alternate_anims)
base = base->alternate_anims;
}
if (!base->anim_total)
return base;
reletive = (int)(cl.time*10) % base->anim_total;
count = 0;
while (base->anim_min > reletive || base->anim_max <= reletive)
{
base = base->anim_next;
if (!base)
Sys_Error ("R_TextureAnimation: broken cycle");
if (++count > 100)
Sys_Error ("R_TextureAnimation: infinite cycle");
}
return base;
}
/*
===============
R_DrawSurface
===============
*/
void R_DrawSurface (void)
{
unsigned char *basetptr;
int smax, tmax, twidth;
int u;
int soffset, basetoffset, texwidth;
int horzblockstep;
unsigned char *pcolumndest;
void (*pblockdrawer)(void);
texture_t *mt;
// calculate the lightings
R_BuildLightMap ();
surfrowbytes = r_drawsurf.rowbytes;
mt = r_drawsurf.texture;
r_source = (byte *)mt + mt->offsets[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 = 16 >> r_drawsurf.surfmip;
blockdivshift = 4 - r_drawsurf.surfmip;
blockdivmask = (1 << blockdivshift) - 1;
r_lightwidth = (r_drawsurf.surf->extents[0]>>4)+1;
r_numhblocks = r_drawsurf.surfwidth >> blockdivshift;
r_numvblocks = r_drawsurf.surfheight >> blockdivshift;
//==============================
if (r_pixbytes == 1)
{
pblockdrawer = surfmiptable[r_drawsurf.surfmip];
// TODO: only needs to be set when there is a display settings change
horzblockstep = blocksize;
}
else
{
pblockdrawer = R_DrawSurfaceBlock16;
// TODO: only needs to be set when there is a display settings change
horzblockstep = blocksize << 1;
}
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);
soffset = r_drawsurf.surf->texturemins[0];
basetoffset = r_drawsurf.surf->texturemins[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;
}
}
//=============================================================================
#if !id386
/*
================
R_DrawSurfaceBlock8_mip0
================
*/
void R_DrawSurfaceBlock8_mip0 (void)
{
int v, i, b, lightstep, lighttemp, light;
unsigned char 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] = ((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, lightstep, lighttemp, light;
unsigned char 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] = ((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, lightstep, lighttemp, light;
unsigned char 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] = ((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, lightstep, lighttemp, light;
unsigned char 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] = ((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_DrawSurfaceBlock16
FIXME: make this work
================
*/
void R_DrawSurfaceBlock16 (void)
{
int k;
unsigned char *psource;
int lighttemp, lightstep, light;
unsigned short *prowdest;
prowdest = (unsigned short *)prowdestbase;
for (k=0 ; k<blocksize ; k++)
{
unsigned short *pdest;
unsigned char pix;
int b;
psource = pbasesource;
lighttemp = lightright - lightleft;
lightstep = lighttemp >> blockdivshift;
light = lightleft;
pdest = prowdest;
for (b=0; b<blocksize; b++)
{
pix = *psource;
*pdest = vid.colormap16[(light & 0xFF00) + pix];
psource += sourcesstep;
pdest++;
light += lightstep;
}
pbasesource += sourcetstep;
lightright += lightrightstep;
lightleft += lightleftstep;
prowdest = (unsigned short *)((long)prowdest + surfrowbytes);
}
prowdestbase = prowdest;
}
#endif
//============================================================================
/*
================
R_GenTurbTile
================
*/
void R_GenTurbTile (pixel_t *pbasetex, void *pdest)
{
int *turb;
int i, j, s, t;
byte *pd;
turb = sintable + ((int)(cl.time*SPEED)&(CYCLE-1));
pd = (byte *)pdest;
for (i=0 ; i<TILE_SIZE ; i++)
{
for (j=0 ; j<TILE_SIZE ; j++)
{
s = (((j << 16) + turb[i & (CYCLE-1)]) >> 16) & 63;
t = (((i << 16) + turb[j & (CYCLE-1)]) >> 16) & 63;
*pd++ = *(pbasetex + (t<<6) + s);
}
}
}
/*
================
R_GenTurbTile16
================
*/
void R_GenTurbTile16 (pixel_t *pbasetex, void *pdest)
{
int *turb;
int i, j, s, t;
unsigned short *pd;
turb = sintable + ((int)(cl.time*SPEED)&(CYCLE-1));
pd = (unsigned short *)pdest;
for (i=0 ; i<TILE_SIZE ; i++)
{
for (j=0 ; j<TILE_SIZE ; j++)
{
s = (((j << 16) + turb[i & (CYCLE-1)]) >> 16) & 63;
t = (((i << 16) + turb[j & (CYCLE-1)]) >> 16) & 63;
*pd++ = d_8to16table[*(pbasetex + (t<<6) + s)];
}
}
}
/*
================
R_GenTile
================
*/
void R_GenTile (msurface_t *psurf, void *pdest)
{
if (psurf->flags & SURF_DRAWTURB)
{
if (r_pixbytes == 1)
{
R_GenTurbTile ((pixel_t *)
((byte *)psurf->texinfo->texture + psurf->texinfo->texture->offsets[0]), pdest);
}
else
{
R_GenTurbTile16 ((pixel_t *)
((byte *)psurf->texinfo->texture + psurf->texinfo->texture->offsets[0]), pdest);
}
}
else if (psurf->flags & SURF_DRAWSKY)
{
if (r_pixbytes == 1)
{
R_GenSkyTile (pdest);
}
else
{
R_GenSkyTile16 (pdest);
}
}
else
{
Sys_Error ("Unknown tile type");
}
}