ref: 8f8b7a00d2b01f90e2d572515c1f0e4364aad476
dir: /Engine/src/draw.c/
// "Build Engine & Tools" Copyright (c) 1993-1997 Ken Silverman
// Ken Silverman's official web site: "http://www.advsys.net/ken"
// See the included license file "BUILDLIC.TXT" for license info.
// This file has been modified from Ken Silverman's original release
/* DDOI - This file is an attempt to reimplement a_nasm.asm in C */
/* FCS: However did that work: This is far from perfect but you have my eternal respect !!! */
#include "platform.h"
#include "build.h"
#include "draw.h"
#if RENDER_LIMIT_PIXELS
int64_t pixelsAllowed = 10000000000;
#endif
uint8_t *transluc = NULL;
static int transrev = 0;
#define shrd(a,b,c) (((b)<<(32-(c))) | ((a)>>(c)))
#define shld(a,b,c) (((b)>>(32-(c))) | ((a)<<(c)))
/* --------------- WALLS RENDERING METHOD (USED TO BE HIGHLY OPTIMIZED ASSEMBLY) ----------------------------*/
extern int32_t asm1;
extern intptr_t asm2;
extern uint8_t *asm3;
extern int32_t asm4;
static uint8_t machxbits_al;
static uint8_t bitsSetup;
static const uint8_t * textureSetup;
void sethlinesizes(int32_t i1, int32_t _bits, const uint8_t * textureAddress)
{
machxbits_al = i1;
bitsSetup = _bits;
textureSetup = textureAddress;
}
//FCS: Draw ceiling/floors
//Draw a line from destination in the framebuffer to framebuffer-numPixels
void hlineasm4(int32_t numPixels, int32_t shade, uint32_t i4, uint32_t i5, uint8_t *dest){
int32_t shifter = ((256-machxbits_al) & 0x1f);
uint32_t source;
const uint8_t * texture = textureSetup;
uint8_t bits = bitsSetup;
shade = shade & 0xffffff00;
numPixels++;
if (!RENDER_DRAW_CEILING_AND_FLOOR)
return;
while (numPixels) {
source = i5 >> shifter;
source = shld(source,i4,bits);
source = texture[source];
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
*dest = globalpalwritten[shade|source];
dest--;
i5 -= asm1;
i4 -= asm2;
numPixels--;
}
}
static int32_t rmach_eax;
static int32_t rmach_ebx;
static int32_t rmach_ecx;
static const uint8_t* rmach_edx;
static int32_t rmach_esi;
void setuprhlineasm4(int32_t i1, int32_t i2, int32_t i3, const uint8_t* i4, int32_t i5, int32_t i6)
{
rmach_eax = i1;
rmach_ebx = i2;
rmach_ecx = i3;
rmach_edx = i4;
rmach_esi = i5;
}
void rhlineasm4(int32_t i1, const uint8_t* texture, int32_t i3, uint32_t i4, uint32_t i5, uint8_t* dest)
{
uint32_t ebp = 0;
int32_t numPixels;
int32_t offset = i1 + 1;
if (i1 <= 0) return;
numPixels = i1;
do {
i3 = ((i3&0xffffff00)|(*texture));
i4 -= rmach_eax;
ebp = (((i4+rmach_eax) < i4) ? -1 : 0);
i5 -= rmach_ebx;
if ((i5 + rmach_ebx) < i5)
texture -= (rmach_ecx+1);
else
texture -= rmach_ecx;
ebp &= rmach_esi;
i1 = ((i1&0xffffff00)|rmach_edx[i3]);
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
dest[numPixels - offset] = (i1&0xff);
texture -= ebp;
numPixels--;
} while (numPixels);
}
static int32_t rmmach_eax;
static int32_t rmmach_ebx;
static int32_t rmmach_ecx;
static const uint8_t* rmmach_edx;
static int32_t setupTileHeight;
void setuprmhlineasm4(int32_t i1, int32_t i2, int32_t i3, const uint8_t* i4, int32_t tileHeight, int32_t i6)
{
rmmach_eax = i1;
rmmach_ebx = i2;
rmmach_ecx = i3;
rmmach_edx = i4;
setupTileHeight = tileHeight;
}
//FCS: ????
void rmhlineasm4(int32_t i1, const uint8_t* shade, int32_t colorIndex, int32_t i4, int32_t i5, uint8_t* dest)
{
uint32_t ebp = 0;
int32_t numPixels;
int32_t offset = i1 + 1;
if (i1 <= 0)
return;
numPixels = i1;
do {
colorIndex = ((colorIndex&0xffffff00)|(*((uint8_t *)shade)));
i4 -= rmmach_eax;
ebp = (((i4+rmmach_eax) < i4) ? -1 : 0);
i5 -= rmmach_ebx;
if ((i5 + rmmach_ebx) < i5)
shade -= (rmmach_ecx+1);
else
shade -= rmmach_ecx;
ebp &= setupTileHeight;
//Check if this colorIndex is the transparent color (255).
if ((colorIndex&0xff) != 255) {
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
{
i1 = ((i1&0xffffff00)|rmmach_edx[colorIndex]);
dest[numPixels - offset] = (i1 & 0xff);
}
}
shade -= ebp;
numPixels--;
} while (numPixels);
}
//Variable used to draw column.
//This is how much you have to skip in the framebuffer in order to be one pixel below.
static int32_t bytesperline;
void setBytesPerLine(int32_t _bytesperline)
{
bytesperline = _bytesperline;
}
static uint8_t mach3_al;
//FCS: RENDER TOP AND BOTTOM COLUMN
int32_t prevlineasm1(int32_t i1, const uint8_t* palette, int32_t i3, int32_t i4, const uint8_t *source, uint8_t *dest)
{
if (i3 == 0)
{
if (!RENDER_DRAW_TOP_AND_BOTTOM_COLUMN)
return 0;
i1 += i4;
i4 = ((uint32_t)i4) >> mach3_al;
i4 = (i4&0xffffff00) | source[i4];
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
*dest = palette[i4];
return i1;
} else {
return vlineasm1(i1,palette,i3,i4,source,dest);
}
}
//FCS: This is used to draw wall border vertical lines
int32_t vlineasm1(int32_t vince, const uint8_t* palookupoffse, int32_t numPixels, int32_t vplce, const uint8_t* texture, uint8_t* dest)
{
uint32_t temp;
if (!RENDER_DRAW_WALL_BORDERS)
return vplce;
numPixels++;
while (numPixels)
{
temp = ((uint32_t)vplce) >> mach3_al;
temp = texture[temp];
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
*dest = palookupoffse[temp];
vplce += vince;
dest += bytesperline;
numPixels--;
}
return vplce;
}
int32_t tvlineasm1(int32_t i1, const uint8_t *texture, int32_t numPixels, int32_t i4, const uint8_t *source, uint8_t *dest)
{
uint8_t shiftValue = (globalshiftval & 0x1f);
numPixels++;
while (numPixels)
{
uint32_t temp = i4;
temp >>= shiftValue;
temp = source[temp];
//255 is the index for transparent color index. Skip drawing this pixel.
if (temp != 255)
{
uint16_t colorIndex;
colorIndex = texture[temp];
colorIndex |= ((*dest)<<8);
if (transrev)
colorIndex = ((colorIndex>>8)|(colorIndex<<8));
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
*dest = transluc[colorIndex];
}
i4 += i1;
//We are drawing a column ?!
dest += bytesperline;
numPixels--;
}
return i4;
} /* tvlineasm1 */
static uint8_t tran2shr;
static const uint8_t* tran2pal_ebx;
static const uint8_t* tran2pal_ecx;
void setuptvlineasm2(int32_t i1, const uint8_t* i2, const uint8_t* i3)
{
tran2shr = (i1&0x1f);
tran2pal_ebx = i2;
tran2pal_ecx = i3;
} /* */
void tvlineasm2(uint32_t i1, uint32_t i2, uintptr_t i3, uintptr_t i4, uint32_t i5, uintptr_t i6)
{
uint32_t ebp = i1;
uint32_t tran2inca = i2;
uint32_t tran2incb = asm1;
uintptr_t tran2bufa = i3;
uintptr_t tran2bufb = i4;
uintptr_t tran2edi = asm2;
uintptr_t tran2edi1 = asm2 + 1;
i6 -= asm2;
do {
i1 = i5 >> tran2shr;
i2 = ebp >> tran2shr;
i5 += tran2inca;
ebp += tran2incb;
i3 = ((uint8_t *)tran2bufa)[i1];
i4 = ((uint8_t *)tran2bufb)[i2];
if (i3 == 255) { // skipdraw1
if (i4 != 255) { // skipdraw3
uint16_t val;
val = tran2pal_ecx[i4];
val |= (((uint8_t *)i6)[tran2edi1]<<8);
if (transrev)
val = ((val>>8)|(val<<8));
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
((uint8_t *)i6)[tran2edi1] = transluc[val];
}
} else if (i4 == 255) { // skipdraw2
uint16_t val;
val = tran2pal_ebx[i3];
val |= (((uint8_t *)i6)[tran2edi]<<8);
if (transrev)
val = ((val>>8)|(val<<8));
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
((uint8_t *)i6)[tran2edi] = transluc[val];
} else {
uint16_t l = ((uint8_t *)i6)[tran2edi]<<8;
uint16_t r = ((uint8_t *)i6)[tran2edi1]<<8;
l |= tran2pal_ebx[i3];
r |= tran2pal_ecx[i4];
if (transrev) {
l = ((l>>8)|(l<<8));
r = ((r>>8)|(r<<8));
}
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
{
((uint8_t *)i6)[tran2edi] = transluc[l];
((uint8_t *)i6)[tran2edi1] =transluc[r];
#if RENDER_LIMIT_PIXELS
pixelsAllowed--;
#endif
}
}
i6 += bytesperline;
} while (i6 > i6 - bytesperline);
asm1 = i5;
asm2 = ebp;
}
static uint8_t machmv;
int32_t mvlineasm1(int32_t vince, const uint8_t* palookupoffse, int32_t i3, int32_t vplce, const uint8_t* texture, uint8_t *dest)
{
uint32_t temp;
for(;i3>=0;i3--)
{
temp = ((uint32_t)vplce) >> machmv;
temp = texture[temp];
if (temp != 255)
{
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
*dest = palookupoffse[temp];
}
vplce += vince;
dest += bytesperline;
}
return vplce;
}
void setupvlineasm(int32_t i1)
{
mach3_al = (i1&0x1f);
}
//FCS This is used to fill the inside of a wall (so it draws VERTICAL column, always).
void vlineasm4(int32_t columnIndex, uint8_t* framebuffer)
{
if (!RENDER_DRAW_WALL_INSIDE)
return;
int i;
uint32_t temp;
uint32_t index = 0;
uint32_t length = ylookup[columnIndex];
do {
for (i = 0; i < 4; i++)
{
temp = ((uint32_t)vplce[i]) >> mach3_al;
temp = (((uint8_t*)(bufplce[i]))[temp]);
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
framebuffer[index + i] = palookupoffse[i][temp];
vplce[i] += vince[i];
}
index += bytesperline;
} while (index < length);
}
void setupmvlineasm(int32_t i1)
{
//Only keep 5 first bits
machmv = (i1&0x1f);
}
void mvlineasm4(int32_t columnIndex, uint8_t* framebuffer)
{
int i;
uint32_t temp;
uint32_t index = 0;
uint32_t length = ylookup[columnIndex];
do {
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed <= 0)
return;
#endif
for (i = 0; i < 4; i++)
{
temp = ((uint32_t)vplce[i]) >> machmv;
temp = (((uint8_t *)(bufplce[i]))[temp]);
if (temp != 255)
{
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
framebuffer[index+i] = palookupoffse[i][temp];
}
vplce[i] += vince[i];
}
index += bytesperline;
} while (index < length);
}
/* END --------------- WALLS RENDERING METHOD (USED TO BE HIGHLY OPTIMIZED ASSEMBLY) ----------------------------*/
/* --------------- SPRITE RENDERING METHOD (USED TO BE HIGHLY OPTIMIZED ASSEMBLY) ----------------------------*/
const uint8_t * tspal;
uint32_t tsmach_eax1;
uint32_t adder;
uint32_t tsmach_eax3;
uint32_t tsmach_ecx;
void tsetupspritevline(const uint8_t * palette, int32_t i2, int32_t i3, int32_t i4, int32_t i5)
{
tspal = palette;
tsmach_eax1 = i5 << 16;
adder = (i5 >> 16) + i2;
tsmach_eax3 = adder + i4;
tsmach_ecx = i3;
}
/*
FCS: Draw a sprite vertical line of pixels.
*/
void DrawSpriteVerticalLine(int32_t i2, int32_t numPixels, uint32_t i4, const uint8_t * texture, uint8_t * dest)
{
uint8_t colorIndex;
while (numPixels)
{
numPixels--;
if (numPixels != 0)
{
i4 += tsmach_ecx;
if (i4 < (i4 - tsmach_ecx))
adder = tsmach_eax3;
colorIndex = *texture;
i2 += tsmach_eax1;
if (i2 < (i2 - tsmach_eax1))
texture++;
texture += adder;
//255 is the index of the transparent color: Do not draw it.
if (colorIndex != 255)
{
uint16_t val;
val = tspal[colorIndex];
val |= (*dest)<<8;
if (transrev)
val = ((val>>8)|(val<<8));
colorIndex = transluc[val];
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
*dest = colorIndex;
}
//Move down one pixel on the framebuffer
dest += bytesperline;
}
}
}
/* END--------------- SPRITE RENDERING METHOD (USED TO BE HIGHLY OPTIMIZED ASSEMBLY) ----------------------------*/
/* --------------- FLOOR/CEILING RENDERING METHOD (USED TO BE HIGHLY OPTIMIZED ASSEMBLY) ----------------------------*/
void settrans(int32_t type){
transrev = type;
}
static uint8_t * textureData;
static uint8_t * mmach_asm3;
static int32_t mmach_asm1;
static int32_t mmach_asm2;
void mhline(uint8_t * texture, int32_t i2, int32_t numPixels, int32_t i4, int32_t i5, uint8_t* dest)
{
textureData = texture;
mmach_asm3 = asm3;
mmach_asm1 = asm1;
mmach_asm2 = asm2;
mhlineskipmodify(i2,numPixels>>16,i5,dest);
}
static uint8_t mshift_al = 26;
static uint8_t mshift_bl = 6;
void mhlineskipmodify( uint32_t i2, int32_t numPixels, int32_t i5, uint8_t* dest)
{
uint32_t ebx;
int32_t colorIndex;
while (numPixels >= 0)
{
ebx = i2 >> mshift_al;
ebx = shld (ebx, (uint32_t)i5, mshift_bl);
colorIndex = textureData[ebx];
//Skip transparent color.
if ((colorIndex&0xff) != 0xff){
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
*dest = mmach_asm3[colorIndex];
}
i2 += mmach_asm1;
i5 += mmach_asm2;
dest++;
numPixels--;
}
}
void msethlineshift(int32_t i1, int32_t i2)
{
i1 = 256-i1;
mshift_al = (i1&0x1f);
mshift_bl = (i2&0x1f);
} /* msethlineshift */
static uint8_t * tmach_eax;
static uint8_t * tmach_asm3;
static int32_t tmach_asm1;
static int32_t tmach_asm2;
void thline(uint8_t * i1, int32_t i2, int32_t i3, int32_t i4, int32_t i5, uint8_t * i6)
{
tmach_eax = i1;
tmach_asm3 = asm3;
tmach_asm1 = asm1;
tmach_asm2 = asm2;
thlineskipmodify(asm2,i2,i3,i4,i5,i6);
}
static uint8_t tshift_al = 26;
static uint8_t tshift_bl = 6;
void thlineskipmodify(int32_t i1, uint32_t i2, uint32_t i3, int32_t i4, int32_t i5, uint8_t * i6)
{
uint32_t ebx;
int counter = (i3>>16);
while (counter >= 0)
{
ebx = i2 >> tshift_al;
ebx = shld (ebx, (uint32_t)i5, tshift_bl);
i1 = tmach_eax[ebx];
if ((i1&0xff) != 0xff)
{
uint16_t val = tmach_asm3[i1];
val |= (*i6)<<8;
if (transrev)
val = ((val>>8)|(val<<8));
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
*i6 = transluc[val];
}
i2 += tmach_asm1;
i5 += tmach_asm2;
i6++;
counter--;
}
}
void tsethlineshift(int32_t i1, int32_t i2)
{
i1 = 256-i1;
tshift_al = (i1&0x1f);
tshift_bl = (i2&0x1f);
}
static intptr_t slopemach_ebx;
static int32_t slopemach_ecx;
static int32_t slopemach_edx;
static uint8_t slopemach_ah1;
static uint8_t slopemach_ah2;
static float asm2_f;
typedef union { unsigned int i; float f; } bitwisef2i;
void setupslopevlin(int32_t i1, intptr_t i2, int32_t i3)
{
bitwisef2i c;
slopemach_ebx = i2;
slopemach_ecx = i3;
slopemach_edx = (1<<(i1&0x1f)) - 1;
slopemach_edx <<= ((i1&0x1f00)>>8);
slopemach_ah1 = 32-((i1&0x1f00)>>8);
slopemach_ah2 = (slopemach_ah1 - (i1&0x1f)) & 0x1f;
c.f = asm2_f = (float)asm1;
asm2 = c.i;
}
extern int32_t reciptable[2048];
extern int32_t globalx3, globaly3;
extern int32_t fpuasm;
#define low32(a) (((a)&0xffffffff))
#define high32(a) ((int)(((int64_t)(a)&(int64_t)0xffffffff00000000)>>32))
//FCS: Render RENDER_SLOPPED_CEILING_AND_FLOOR
void slopevlin(intptr_t i1, uint32_t i2, intptr_t* i3, uint32_t index, int32_t i4, int32_t i5, int32_t i6)
{
bitwisef2i c;
uintptr_t ecx, eax, ebx, edx, esi;
uint32_t edi;
//This is so bad to cast asm3 to int then float :( !!!
float a = (float)(int32_t) asm3 + asm2_f;
i1 -= slopemach_ecx;
esi = i5 + low32((int64_t)globalx3 * (int64_t)(i2<<3));
edi = i6 + low32((int64_t)globaly3 * (int64_t)(i2<<3));
ebx = i4;
if (!RENDER_SLOPPED_CEILING_AND_FLOOR)
return;
do {
// -------------
// All this is calculating a fixed point approx. of 1/a
c.f = a;
fpuasm = eax = c.i;
edx = (((int32_t)eax) < 0) ? 0xffffffff : 0;
eax = eax << 1;
ecx = (eax>>24); // exponent
eax = ((eax&0xffe000)>>11);
ecx = ((ecx&0xffffff00)|((ecx-2)&0xff));
eax = reciptable[eax/4];
eax >>= (ecx&0x1f);
eax ^= edx;
// -------------
edx = i2;
i2 = eax;
eax -= edx;
ecx = low32((int64_t)globalx3 * (int64_t)eax);
eax = low32((int64_t)globaly3 * (int64_t)eax);
a += asm2_f;
asm4 = ebx;
ecx = ((ecx&0xffffff00)|(ebx&0xff));
if (ebx >= 8) ecx = ((ecx&0xffffff00)|8);
ebx = esi;
edx = edi;
while ((ecx&0xff))
{
ebx >>= slopemach_ah2;
esi += ecx;
edx >>= slopemach_ah1;
ebx &= slopemach_edx;
edi += eax;
i1 += slopemach_ecx;
edx = ((edx&0xffffff00)|((((uint8_t *)(ebx+edx))[slopemach_ebx])));
ebx = i3[index];
index--;
eax = ((eax & 0xffffff00) | (*((uint8_t*)(ebx + edx))));
ebx = esi;
#if RENDER_LIMIT_PIXELS
if (pixelsAllowed-- > 0)
#endif
*((uint8_t *)i1) = (eax&0xff);
edx = edi;
ecx = ((ecx&0xffffff00)|((ecx-1)&0xff));
}
ebx = asm4;
ebx -= 8; // BITSOFPRECISIONPOW
} while ((int32_t)ebx > 0);
}
/* END --------------- FLOOR/CEILING RENDERING METHOD (USED TO BE HIGHLY OPTIMIZED ASSEMBLY) ----------------------------*/