// This file is distributed under a BSD license. See LICENSE.txt for details.

ps.1.1

// constants / flag data layout
def c5,2,-1,0,0
def c6,0.5,-0.5,0,1
def c7,1,0,0,0.5

flags
	BaseFlags
	SpecialFlags
	LightFlags
	CodegenFlags
	TFlags[4]
	Combiner[13]
	_pad[2]
	AlphaCombiner
  TexUsed[4]
  AlphaOps[2]
  CombinerSrc[13]
  AlphaSrc[13]
endflags

// yes, this is rather hackish
temp r0,r1

// read textures used in combiners
for i = 0..3
  if TexUsed[i][0]
    tex   t0+i1
    iadd  i1,1
    iadd  i2,1
  endif
endfor

// read cube normalizers for lighting
if LightFlags[0..2] == 5 // bumpx 
  // texture reads
  tex   t0+i1 // light
  iadd  i1,1
  
  if !SpecialFlags[6] // specular?
    tex   t0+i1 // halfway
    iadd  i1,1
  endif
  
  tex   t0+i1 // attenuation
  iadd  i1,1
  
  // bump shader
  temp  normal
  
  if TexUsed[1][0]
    vmov  normal,t0
  else
    vmov  normal,c7
  endif
  
  if SpecialFlags[6] // diffuse only
    dp3_sat     r0.xyz,t0+i2_bx2,normal   // 1 diffuse dot
    +mov        r0.w,c5                   // 1 clear alpha
    mul         r1.xyz,t1+i2,c0           // 2 attenuation*light color
    mul         r0.xyz,r0,r1              // 3 final diffuse
  else // diffuse and specular
    dp3_sat     r0.xyz,t1+i2_bx2,normal   // 1 specular dot
    +mul        r0.w,c0,t2+i2             // 1 attenuation*light color
    dp3_sat     r1.xyz,t0+i2_bx2,normal   // 2 diffuse dot
    +mad_x4_sat r1.w,r0.z,c1.z,c1.w       // 2 specular pow 1
    mul         r0.xyz,c0,t2+i2           // 3 attenuation*light color
    +mul_x2     r1.w,r1,r1                // 3 specular pow 2
    mul_x2      t0+i2.xyz,r1.w,r1.w       // 4 specular pow 3
    // one free alpha operation here
    mul_x2_sat  t0+i2.xyz,t0+i2,t0+i2     // 5 specular pow 4
    
    if SpecialFlags[7] // specular map?
      +mul        r0.w,r0,t0              // 5 gloss map
    endif
    
    mul         r0.xyz,r0,r1              // 6 final diffuse
    +mul        r0.w,t0+i2.z,r0.w         // 6 final specular
  endif
else // combiner
  imov  i2,0 // "last was set" flag
    
  for i = 0..12 // combiner code!
    temp  src,asrc,rsrc
    vmov  src,CombinerSrc[i]
    vmov  asrc,AlphaSrc[i]
    
    imov  i3,0 // instruction already generated during alpha fade
    
    if Combiner[i][0..3] // alpha fade
      if Combiner[i][4..7] == 9 // lerp, do nothing
      elif Combiner[i][4..7] == 1 // set, change destination
        mul   r0.xyz,src,asrc
        imov  i3,1
      elif Combiner[i][4..7] == 2 // add, generate mad
        mad   r0.xyz,src,asrc,r0
        imov  i3,1
      else
        mul   r1.xyz,src,asrc
        vmov  src,r1
      endif
    endif
    
    if !i3
      if Combiner[i][4..7] // combiner operation
        imov  i2,0
        if Combiner[i][4..7] == 1 // set
          vmov    r0,src
          imov    i2,1
        elif Combiner[i][4..7] == 2 // add
          add     r0.xyz,r0,src
        elif Combiner[i][4..7] == 3 // sub
          sub     r0.xyz,r0,src
        elif Combiner[i][4..7] == 4 // mul
          mul     r0.xyz,r0,src
        elif Combiner[i][4..7] == 5 // mul2
          mul_x2  r0.xyz,r0,src
        elif Combiner[i][4..7] == 6 // mul4
          mul_x4  r0.xyz,r0,src
        elif Combiner[i][4..7] == 7 // mul8
          mul_x4  r0.xyz,r0,src
          add     r0.xyz,r0,r0
        elif Combiner[i][4..7] == 8 // smooth
          mad     r0.xyz,r0,1-src,src
        elif Combiner[i][4..7] == 9 // lerp
          lrp     r0.xyz,asrc,r0,src
        elif Combiner[i][4..7] == 10 // rsub
          sub     r0.xyz,src,r0
        elif Combiner[i][4..7] == 11 // dp3
          dp3     r0.xyz,r0,src
        elif Combiner[i][4..7] == 12 // muls
          mul_sat r0.xyz,r0,src
        endif
      endif
    endif
  endfor

  if i2 // if last operation was a set, force a move
    mov     r0.xyz,r0
  endif

  // alpha combiner implementation
  if AlphaCombiner[4..7] == 14 // mix01
    +mul    r0.w,t0,c0
    mad     r0.w,t1,c0,r0
  elif AlphaCombiner[4..7] // two-alpha-setup
    temp    src0,src1
    
    vmov    src0,AlphaOps[0]
    vmov    src1,AlphaOps[1]
    +mul    r0.w,src0,src1
  else
    temp    src
    
    vmov    src,AlphaOps[0]
    +mov    r0.w,src
  endif
endif