; Uniforms
.fvec projection[4], modelView[4], meshColor
.fvec lightVec[2], lightClr[3], shininess

; Constants
.constf myconst(0.0, 1.0, -1.0, -0.5)

; Outputs
.out outpos position
.out outtc0 texcoord0
.out outclr color

; Inputs
.alias inpos v0
.alias innrm v1
.alias intex v2

.proc main
	; Prepare constants in usable temp regs
	mov r15.x, myconst.x ; 0.0
	mov r15.y, myconst.y ; 1.0
	mov r15.z, myconst.z ; -1.0

	; Force the w component of inpos to be 1.0
	mov r0.xyz, inpos
	mov r0.w,   r15.y

	; r1 = modelView * inpos
	dp4 r1.x, modelView[0], r0
	dp4 r1.y, modelView[1], r0
	dp4 r1.z, modelView[2], r0
	dp4 r1.w, modelView[3], r0

	; outpos = projection * r1
	dp4 outpos.x, projection[0], r1
	dp4 outpos.y, projection[1], r1
	dp4 outpos.z, projection[2], r1
	dp4 outpos.w, projection[3], r1

	; outtex = intex
	mov outtc0, intex
	mov outtc0.zw, myconst.xy

	; Transform normal
	mov r2.xyz, innrm
	mov r2.w,   r15.x
	dp4 r3.x, modelView[0], r2
	dp4 r3.y, modelView[1], r2
	dp4 r3.z, modelView[2], r2
	mov r3.w,   r15.x
	dp3 r4.x, r3, r3
	rsq r4.x, r4.x
	mul r3, r4.xxxx, r3  ; r3 = normalized normal

	; Normalize lightVec[0]
	mov r5, lightVec[0]
	dp3 r6.x, r5, r5
	rsq r6.x, r6.x
	mul r5, r6.xxxx, r5

	; dot(normal, lightVec[0])
	dp3 r6.x, r3, r5
	max r6.x, r6.x, r15.xxxx

	; Normalize lightVec[1]
	mov r7, lightVec[1]
	dp3 r8.x, r7, r7
	rsq r8.x, r8.x
	mul r7, r8.xxxx, r7

	; dot(normal, lightVec[1])
	dp3 r6.y, r3, r7
	max r6.y, r6.y, r15.xxxx

	; Load lightClr
	mov r8, lightClr[2] ; ambient
	mov r9, lightClr[0] ; point
	mov r10, lightClr[1] ; directional

	; diffuse = ambient + (lightClr[0] * dot0) + (lightClr[1] * dot1)
	mul r11, r9, r6.xxxx
	add r8, r8, r11
	mul r11, r10, r6.yyyy
	add r8, r8, r11 ; r8 = diffuse

	; Check if shininess > 0
	mov r12, shininess
	slt r13.x, r15.x, r12.x

	; viewVec = normalize(-position.xyz)
	mov r14.xyz, r1.xyz
	mul r14.xyz, r14.xyz, r15.zzz
	dp3 r4.x, r14, r14
	rsq r4.x, r4.x
	mul r14, r4.xxxx, r14

	; H = normalize(view + lightVec[1])
	add r11, r14, r7
	dp3 r4.x, r11, r11
	rsq r4.x, r4.x
	mul r11, r4.xxxx, r11

	; dot(normal, H)
	dp3 r4.x, r3, r11
	max r4.x, r4.x, r15.x

	; Approximate pow(dotNH, 10) by repeated multiplication
	mul r5.x, r4.x, r4.x      ; dotNH^2
	mul r5.x, r5.x, r5.x      ; dotNH^4
	mul r5.x, r5.x, r5.x      ; dotNH^8
	mul r4.x, r5.x, r4.x      ; dotNH^9
	mul r4.x, r4.x, r4.x      ; dotNH^10

	; Multiply by shininess > 0 flag
	mul r4.x, r4.x, r13.x

	; specular = lightClr[1] * spec
	mul r5, r10, r4.xxxx

	; final = diffuse * meshColor + specular * lightClr[1]
	mov r9, meshColor
	mul r6, r8, r9      ; diffuse * meshColor
	add r7.xyz, r6.xyz, r5.xyz  ; add specular (already multiplied by lightClr)
	min r7.xyz, r7.xyz, r15.yyyy

	mov outclr.xyz, r7.xyz
	mov outclr.w, meshColor.w

	end
.end