version 4.2

compiler/build.gradle

@@ -1,11 +1,11 @@
 buildscript {
     dependencies {
-        classpath "org.jetbrains.kotlin:kotlin-gradle-plugin:1.4.0"
+        classpath "org.jetbrains.kotlin:kotlin-gradle-plugin:1.4.10"
     }
 }
 
 plugins {
-    // id "org.jetbrains.kotlin.jvm" version "1.4.0"
+    // id "org.jetbrains.kotlin.jvm" version "1.4.10"
     id 'application'
     id 'org.jetbrains.dokka' version "0.9.18"
     id 'com.github.johnrengelman.shadow' version '5.2.0'

compiler/res/prog8lib/c64floats.asm

@@ -1,10 +1,15 @@
 ; --- low level floating point assembly routines for the C64
 
+FL_ONE_const	.byte  129     			; 1.0
+FL_ZERO_const	.byte  0,0,0,0,0		; 0.0
+
+floats_store_reg	.byte  0		; temp storage
+
 
 ub2float	.proc
 		; -- convert ubyte in SCRATCH_ZPB1 to float at address A/Y
 		;    clobbers A, Y
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		sta  P8ZP_SCRATCH_W2
 		sty  P8ZP_SCRATCH_W2+1
 		ldy  P8ZP_SCRATCH_B1
@@ -13,14 +18,14 @@ ub2float	.proc
 _fac_to_mem	ldx  P8ZP_SCRATCH_W2
 		ldy  P8ZP_SCRATCH_W2+1
 		jsr  MOVMF
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 		.pend
 
 b2float		.proc
 		; -- convert byte in SCRATCH_ZPB1 to float at address A/Y
 		;    clobbers A, Y
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		sta  P8ZP_SCRATCH_W2
 		sty  P8ZP_SCRATCH_W2+1
 		lda  P8ZP_SCRATCH_B1
@@ -30,7 +35,7 @@ b2float		.proc
 
 uw2float	.proc
 		; -- convert uword in SCRATCH_ZPWORD1 to float at address A/Y
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		sta  P8ZP_SCRATCH_W2
 		sty  P8ZP_SCRATCH_W2+1
 		lda  P8ZP_SCRATCH_W1
@@ -41,7 +46,7 @@ uw2float	.proc
 
 w2float		.proc
 		; -- convert word in SCRATCH_ZPWORD1 to float at address A/Y
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		sta  P8ZP_SCRATCH_W2
 		sty  P8ZP_SCRATCH_W2+1
 		ldy  P8ZP_SCRATCH_W1
@@ -54,7 +59,7 @@ stack_b2float	.proc
 		; -- b2float operating on the stack
 		inx
 		lda  P8ESTACK_LO,x
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  FREADSA
 		jmp  push_fac1_as_result
 		.pend
@@ -64,7 +69,7 @@ stack_w2float	.proc
 		inx
 		ldy  P8ESTACK_LO,x
 		lda  P8ESTACK_HI,x
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  GIVAYF
 		jmp  push_fac1_as_result
 		.pend
@@ -73,7 +78,7 @@ stack_ub2float	.proc
 		; -- ub2float operating on the stack
 		inx
 		lda  P8ESTACK_LO,x
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		tay
 		lda  #0
 		jsr  GIVAYF
@@ -85,16 +90,16 @@ stack_uw2float	.proc
 		inx
 		lda  P8ESTACK_LO,x
 		ldy  P8ESTACK_HI,x
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  GIVUAYFAY
 		jmp  push_fac1_as_result
 		.pend
 
 stack_float2w	.proc               ; also used for float2b
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  AYINT
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		lda  $64
 		sta  P8ESTACK_HI,x
 		lda  $65
@@ -105,9 +110,9 @@ stack_float2w	.proc               ; also used for float2b
 
 stack_float2uw	.proc               ; also used for float2ub
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  GETADR
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		sta  P8ESTACK_HI,x
 		tya
 		sta  P8ESTACK_LO,x
@@ -231,15 +236,15 @@ inc_var_f	.proc
 		; -- add 1 to float pointed to by A/Y
 		sta  P8ZP_SCRATCH_W1
 		sty  P8ZP_SCRATCH_W1+1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  MOVFM
-		lda  #<FL_FONE
-		ldy  #>FL_FONE
+		lda  #<FL_ONE_const
+		ldy  #>FL_ONE_const
 		jsr  FADD
 		ldx  P8ZP_SCRATCH_W1
 		ldy  P8ZP_SCRATCH_W1+1
 		jsr  MOVMF
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 		.pend
 
@@ -247,9 +252,9 @@ dec_var_f	.proc
 		; -- subtract 1 from float pointed to by A/Y
 		sta  P8ZP_SCRATCH_W1
 		sty  P8ZP_SCRATCH_W1+1
-		stx  P8ZP_SCRATCH_REG_X
-		lda  #<FL_FONE
-		ldy  #>FL_FONE
+		stx  P8ZP_SCRATCH_REG
+		lda  #<FL_ONE_const
+		ldy  #>FL_ONE_const
 		jsr  MOVFM
 		lda  P8ZP_SCRATCH_W1
 		ldy  P8ZP_SCRATCH_W1+1
@@ -257,7 +262,7 @@ dec_var_f	.proc
 		ldx  P8ZP_SCRATCH_W1
 		ldy  P8ZP_SCRATCH_W1+1
 		jsr  MOVMF
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 		.pend
 
@@ -286,7 +291,7 @@ push_fac1_as_result	.proc
 		jsr  MOVMF
 		lda  #<fmath_float1
 		ldy  #>fmath_float1
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		jmp  push_float
 		.pend
 
@@ -298,21 +303,21 @@ pow_f		.proc
 		lda  #<fmath_float1
 		ldy  #>fmath_float1
 		jsr  pop_float
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		lda  #<fmath_float1
 		ldy  #>fmath_float1
 		jsr  CONUPK		; fac2 = float1
 		lda  #<fmath_float2
 		ldy  #>fmath_float2
 		jsr  FPWR
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		jmp  push_fac1_as_result
 		.pend
 
 div_f		.proc
 		; -- push f1/f2 on stack
 		jsr  pop_2_floats_f2_in_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		lda  #<fmath_float1
 		ldy  #>fmath_float1
 		jsr  FDIV
@@ -322,7 +327,7 @@ div_f		.proc
 add_f		.proc
 		; -- push f1+f2 on stack
 		jsr  pop_2_floats_f2_in_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		lda  #<fmath_float1
 		ldy  #>fmath_float1
 		jsr  FADD
@@ -332,7 +337,7 @@ add_f		.proc
 sub_f		.proc
 		; -- push f1-f2 on stack
 		jsr  pop_2_floats_f2_in_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		lda  #<fmath_float1
 		ldy  #>fmath_float1
 		jsr  FSUB
@@ -342,7 +347,7 @@ sub_f		.proc
 mul_f		.proc
 		; -- push f1*f2 on stack
 		jsr  pop_2_floats_f2_in_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		lda  #<fmath_float1
 		ldy  #>fmath_float1
 		jsr  FMULT
@@ -352,7 +357,7 @@ mul_f		.proc
 neg_f		.proc
 		; -- push -flt back on stack
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  NEGOP
 		jmp  push_fac1_as_result
 		.pend
@@ -360,7 +365,7 @@ neg_f		.proc
 abs_f		.proc
 		; -- push abs(float) on stack (as float)
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  ABS
 		jmp  push_fac1_as_result
 		.pend
@@ -466,7 +471,7 @@ _return_true	lda  #1
 func_sin	.proc
 		; -- push sin(f) back onto stack
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  SIN
 		jmp  push_fac1_as_result
 		.pend
@@ -474,7 +479,7 @@ func_sin	.proc
 func_cos	.proc
 		; -- push cos(f) back onto stack
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  COS
 		jmp  push_fac1_as_result
 		.pend
@@ -482,7 +487,7 @@ func_cos	.proc
 func_tan	.proc
 		; -- push tan(f) back onto stack
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  TAN
 		jmp  push_fac1_as_result
 		.pend
@@ -490,7 +495,7 @@ func_tan	.proc
 func_atan	.proc
 		; -- push atan(f) back onto stack
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  ATN
 		jmp  push_fac1_as_result
 		.pend
@@ -498,7 +503,7 @@ func_atan	.proc
 func_ln		.proc
 		; -- push ln(f) back onto stack
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  LOG
 		jmp  push_fac1_as_result
 		.pend
@@ -506,7 +511,7 @@ func_ln		.proc
 func_log2	.proc
 		; -- push log base 2, ln(f)/ln(2), back onto stack
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  LOG
 		jsr  MOVEF
 		lda  #<c64.FL_LOG2
@@ -518,7 +523,7 @@ func_log2	.proc
 
 func_sqrt	.proc
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  SQR
 		jmp  push_fac1_as_result
 		.pend
@@ -526,7 +531,7 @@ func_sqrt	.proc
 func_rad	.proc
 		; -- convert degrees to radians (d * pi / 180)
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		lda  #<_pi_div_180
 		ldy  #>_pi_div_180
 		jsr  FMULT
@@ -537,7 +542,7 @@ _pi_div_180	.byte 123, 14, 250, 53, 18		; pi / 180
 func_deg	.proc
 		; -- convert radians to degrees (d * (1/ pi * 180))
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		lda  #<_one_over_pi_div_180
 		ldy  #>_one_over_pi_div_180
 		jsr  FMULT
@@ -547,7 +552,7 @@ _one_over_pi_div_180	.byte 134, 101, 46, 224, 211		; 1 / (pi * 180)
 
 func_round	.proc
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  FADDH
 		jsr  INT
 		jmp  push_fac1_as_result
@@ -555,7 +560,7 @@ func_round	.proc
 
 func_floor	.proc
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  INT
 		jmp  push_fac1_as_result
 		.pend
@@ -563,7 +568,7 @@ func_floor	.proc
 func_ceil	.proc
 		; -- ceil: tr = int(f); if tr==f -> return  else return tr+1
 		jsr  pop_float_fac1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		ldx  #<fmath_float1
 		ldy  #>fmath_float1
 		jsr  MOVMF
@@ -573,8 +578,8 @@ func_ceil	.proc
 		jsr  FCOMP
 		cmp  #0
 		beq  +
-		lda  #<FL_FONE
-		ldy  #>FL_FONE
+		lda  #<FL_ONE_const
+		ldy  #>FL_ONE_const
 		jsr  FADD
 +		jmp  push_fac1_as_result
 		.pend
@@ -630,7 +635,7 @@ func_max_f	.proc
 		ldy  #>_largest_neg_float
 _minmax_entry	jsr  MOVFM
 		jsr  prog8_lib.pop_array_and_lengthmin1Y
-		stx  P8ZP_SCRATCH_REG_X
+		stx  floats_store_reg
 -		sty  P8ZP_SCRATCH_REG
 		lda  P8ZP_SCRATCH_W1
 		ldy  P8ZP_SCRATCH_W1+1
@@ -650,6 +655,8 @@ _minmax_cmp	cmp  #255			; modified
 		dey
 		cpy  #255
 		bne  -
+		ldx  floats_store_reg
+		stx  P8ZP_SCRATCH_REG
 		jmp  push_fac1_as_result
 _largest_neg_float	.byte 255,255,255,255,255		; largest negative float -1.7014118345e+38
 		.pend
@@ -665,11 +672,11 @@ _largest_pos_float	.byte  255,127,255,255,255		; largest positive float
 		.pend
 
 func_sum_f	.proc
-		lda  #<ZERO
-		ldy  #>ZERO
+		lda  #<FL_ZERO_const
+		ldy  #>FL_ZERO_const
 		jsr  MOVFM
 		jsr  prog8_lib.pop_array_and_lengthmin1Y
-		stx  P8ZP_SCRATCH_REG_X
+		stx  floats_store_reg
 -		sty  P8ZP_SCRATCH_REG
 		lda  P8ZP_SCRATCH_W1
 		ldy  P8ZP_SCRATCH_W1+1
@@ -685,7 +692,9 @@ func_sum_f	.proc
 		bcc  -
 		inc  P8ZP_SCRATCH_W1+1
 		bne  -
-+		jmp  push_fac1_as_result
++		ldx  floats_store_reg
+		stx  P8ZP_SCRATCH_REG
+		jmp  push_fac1_as_result
 		.pend
 
 sign_f		.proc

compiler/res/prog8lib/c64flt.p8

@@ -4,14 +4,14 @@
 ;
 ; indent format: TABS, size=8
 
+%target c64
 %option enable_floats
 
 c64flt {
 	; ---- this block contains C-64 floating point related functions ----
 
-        const  float  PI        = 3.141592653589793
-        const  float  TWOPI	    = 6.283185307179586
-        const  float  ZERO      = 0.0
+        const float  PI     = 3.141592653589793
+        const float  TWOPI  = 6.283185307179586
 
 
 ; ---- C64 basic and kernal ROM float constants and functions ----
@@ -194,20 +194,20 @@ asmsub  GETADRAY  () clobbers(X) -> uword @ AY  {
 sub  print_f  (float value) {
 	; ---- prints the floating point value (without a newline).
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  floats_store_reg
 		lda  #<value
 		ldy  #>value
 		jsr  MOVFM		; load float into fac1
 		jsr  FOUT		; fac1 to string in A/Y
-		sta  P8ZP_SCRATCH_B1
-		sty  P8ZP_SCRATCH_REG
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
 		ldy  #0
--		lda  (P8ZP_SCRATCH_B1),y
+-		lda  (P8ZP_SCRATCH_W1),y
 		beq  +
 		jsr  c64.CHROUT
 		iny
 		bne  -
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  floats_store_reg
 +		rts
 	}}
 }

compiler/res/prog8lib/c64graphics.p8

@@ -1,3 +1,4 @@
+%target c64
 %import c64textio
 
 ; bitmap pixel graphics module for the C64
@@ -180,9 +181,7 @@ graphics {
 ;        @(addr) |= ormask[lsb(px) & 7]
 ;    }
 
-    ; TODO fix the use of X (or XY) as parameter so we can actually use this plot() routine
-    ;   calling it with a byte results in a compiler crash, calling it with word results in clobbering X register I think
-    asmsub  plotXXX(uword plotx @XY, ubyte ploty @A) {
+    asmsub  plot(uword plotx @XY, ubyte ploty @A) clobbers (A, X, Y) {
         %asm {{
             stx  internal_plotx
             sty  internal_plotx+1
@@ -190,12 +189,14 @@ graphics {
         }}
     }
 
+    ; for efficiency of internal algorithms here is the internal plot routine
+    ; that takes the plotx coordinate in a separate variable instead of the XY register pair:
+
     uword internal_plotx     ; 0..319        ; separate 'parameter' for internal_plot()
 
-    asmsub  internal_plot(ubyte ploty @A) {      ; internal_plotx is 16 bits 0 to 319... doesn't fit in a register
+    asmsub  internal_plot(ubyte ploty @A) clobbers (A, X, Y) {      ; internal_plotx is 16 bits 0 to 319... doesn't fit in a register
         %asm {{
         tay
-        stx  P8ZP_SCRATCH_REG_X
         lda  internal_plotx+1
         sta  P8ZP_SCRATCH_W2+1
         lsr  a            ; 0
@@ -219,8 +220,6 @@ graphics {
         lda  (P8ZP_SCRATCH_W2),y
         ora  _ormask,x
         sta  (P8ZP_SCRATCH_W2),y
-
-        ldx  P8ZP_SCRATCH_REG_X
         rts
 
 _ormask     .byte 128, 64, 32, 16, 8, 4, 2, 1

compiler/res/prog8lib/c64lib.p8

@@ -5,6 +5,7 @@
 ;
 ; indent format: TABS, size=8
 
+%target c64
 
 c64 {
         &ubyte  TIME_HI         = $a0       ; software jiffy clock, hi byte
@@ -298,8 +299,6 @@ _irq_handler_init
 		sta  IRQ_SCRATCH_ZPB1
 		lda  P8ZP_SCRATCH_REG
 		sta  IRQ_SCRATCH_ZPREG
-		lda  P8ZP_SCRATCH_REG_X
-		sta  IRQ_SCRATCH_ZPREGX
 		lda  P8ZP_SCRATCH_W1
 		sta  IRQ_SCRATCH_ZPWORD1
 		lda  P8ZP_SCRATCH_W1+1
@@ -324,8 +323,6 @@ _irq_handler_end
 		sta  P8ZP_SCRATCH_B1
 		lda  IRQ_SCRATCH_ZPREG
 		sta  P8ZP_SCRATCH_REG
-		lda  IRQ_SCRATCH_ZPREGX
-		sta  P8ZP_SCRATCH_REG_X
 		lda  IRQ_SCRATCH_ZPWORD1
 		sta  P8ZP_SCRATCH_W1
 		lda  IRQ_SCRATCH_ZPWORD1+1
@@ -340,7 +337,6 @@ _irq_handler_end
 IRQ_X_REG		.byte  0
 IRQ_SCRATCH_ZPB1	.byte  0
 IRQ_SCRATCH_ZPREG	.byte  0
-IRQ_SCRATCH_ZPREGX	.byte  0
 IRQ_SCRATCH_ZPWORD1	.word  0
 IRQ_SCRATCH_ZPWORD2	.word  0
 

compiler/res/prog8lib/c64textio.p8

@@ -5,6 +5,7 @@
 ; indent format: TABS, size=8
 
 
+%target c64
 %import c64lib
 %import conv
 
@@ -74,7 +75,7 @@ asmsub  scroll_left_full  (ubyte alsocolors @ Pc) clobbers(A, Y)  {
 	;      Carry flag determines if screen color data must be scrolled too
 
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		bcs  +
 		jmp  _scroll_screen
 
@@ -102,7 +103,7 @@ _scroll_screen  ; scroll the screen memory
 		dey
 		bpl  -
 
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
@@ -112,7 +113,7 @@ asmsub  scroll_right_full  (ubyte alsocolors @ Pc) clobbers(A)  {
 	;      contents of the leftmost column are unchanged, you should clear/refill this yourself
 	;      Carry flag determines if screen color data must be scrolled too
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		bcs  +
 		jmp  _scroll_screen
 
@@ -136,7 +137,7 @@ _scroll_screen  ; scroll the screen memory
 		dex
 		bpl  -
 
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
@@ -146,7 +147,7 @@ asmsub  scroll_up_full  (ubyte alsocolors @ Pc) clobbers(A)  {
 	;      contents of the bottom row are unchanged, you should refill/clear this yourself
 	;      Carry flag determines if screen color data must be scrolled too
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		bcs  +
 		jmp  _scroll_screen
 
@@ -170,7 +171,7 @@ _scroll_screen  ; scroll the screen memory
 		dex
 		bpl  -
 
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
@@ -180,7 +181,7 @@ asmsub  scroll_down_full  (ubyte alsocolors @ Pc) clobbers(A)  {
 	;      contents of the top row are unchanged, you should refill/clear this yourself
 	;      Carry flag determines if screen color data must be scrolled too
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		bcs  +
 		jmp  _scroll_screen
 
@@ -204,7 +205,7 @@ _scroll_screen  ; scroll the screen memory
 		dex
 		bpl  -
 
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
@@ -230,7 +231,7 @@ asmsub  print (str text @ AY) clobbers(A,Y)  {
 asmsub  print_ub0  (ubyte value @ A) clobbers(A,Y)  {
 	; ---- print the ubyte in A in decimal form, with left padding 0s (3 positions total)
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  conv.ubyte2decimal
 		pha
 		tya
@@ -239,7 +240,7 @@ asmsub  print_ub0  (ubyte value @ A) clobbers(A,Y)  {
 		jsr  c64.CHROUT
 		txa
 		jsr  c64.CHROUT
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
@@ -247,7 +248,7 @@ asmsub  print_ub0  (ubyte value @ A) clobbers(A,Y)  {
 asmsub  print_ub  (ubyte value @ A) clobbers(A,Y)  {
 	; ---- print the ubyte in A in decimal form, without left padding 0s
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		jsr  conv.ubyte2decimal
 _print_byte_digits
 		pha
@@ -264,7 +265,7 @@ _print_byte_digits
         jsr  c64.CHROUT
 _ones   txa
 		jsr  c64.CHROUT
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
@@ -272,7 +273,7 @@ _ones   txa
 asmsub  print_b  (byte value @ A) clobbers(A,Y)  {
 	; ---- print the byte in A in decimal form, without left padding 0s
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		pha
 		cmp  #0
 		bpl  +
@@ -280,16 +281,14 @@ asmsub  print_b  (byte value @ A) clobbers(A,Y)  {
 		jsr  c64.CHROUT
 +		pla
 		jsr  conv.byte2decimal
-		jsr  print_ub._print_byte_digits
-		ldx  P8ZP_SCRATCH_REG_X
-		rts
+		jmp  print_ub._print_byte_digits
 	}}
 }
 
 asmsub  print_ubhex  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the ubyte in A in hex form (if Carry is set, a radix prefix '$' is printed as well)
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		bcc  +
 		pha
 		lda  #'$'
@@ -299,7 +298,7 @@ asmsub  print_ubhex  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
 		jsr  c64.CHROUT
 		tya
 		jsr  c64.CHROUT
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
@@ -307,7 +306,7 @@ asmsub  print_ubhex  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
 asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
 	; ---- print the ubyte in A in binary form (if Carry is set, a radix prefix '%' is printed as well)
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		sta  P8ZP_SCRATCH_B1
 		bcc  +
 		lda  #'%'
@@ -320,7 +319,7 @@ asmsub  print_ubbin  (ubyte value @ A, ubyte prefix @ Pc) clobbers(A,Y)  {
 +		jsr  c64.CHROUT
 		dey
 		bne  -
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
@@ -353,7 +352,7 @@ asmsub  print_uwhex  (uword value @ AY, ubyte prefix @ Pc) clobbers(A,Y)  {
 asmsub  print_uw0  (uword value @ AY) clobbers(A,Y)  {
 	; ---- print the uword in A/Y in decimal form, with left padding 0s (5 positions total)
 	%asm {{
-	    stx  P8ZP_SCRATCH_REG_X
+	    stx  P8ZP_SCRATCH_REG
 		jsr  conv.uword2decimal
 		ldy  #0
 -		lda  conv.uword2decimal.decTenThousands,y
@@ -361,7 +360,7 @@ asmsub  print_uw0  (uword value @ AY) clobbers(A,Y)  {
 		jsr  c64.CHROUT
 		iny
 		bne  -
-+		ldx  P8ZP_SCRATCH_REG_X
++		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }
@@ -369,9 +368,9 @@ asmsub  print_uw0  (uword value @ AY) clobbers(A,Y)  {
 asmsub  print_uw  (uword value @ AY) clobbers(A,Y)  {
 	; ---- print the uword in A/Y in decimal form, without left padding 0s
 	%asm {{
-	    stx  P8ZP_SCRATCH_REG_X
+	    stx  P8ZP_SCRATCH_REG
 		jsr  conv.uword2decimal
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		ldy  #0
 -		lda  conv.uword2decimal.decTenThousands,y
 		beq  _allzero
@@ -545,11 +544,11 @@ _colormod	sta  $ffff		; modified
 asmsub  plot  (ubyte col @ Y, ubyte row @ A) clobbers(A) {
 	; ---- safe wrapper around PLOT kernel routine, to save the X register.
 	%asm  {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		tax
 		clc
 		jsr  c64.PLOT
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 	}}
 }

compiler/res/prog8lib/conv.p8

@@ -213,7 +213,7 @@ asmsub  byte2decimal  (byte value @ A) -> ubyte @ Y, ubyte @ A, ubyte @ X  {
 asmsub  ubyte2hex  (ubyte value @ A) -> ubyte @ A, ubyte @ Y  {
 	; ---- A to hex petscii string in AY (first hex char in A, second hex char in Y)
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		pha
 		and  #$0f
 		tax
@@ -225,7 +225,7 @@ asmsub  ubyte2hex  (ubyte value @ A) -> ubyte @ A, ubyte @ Y  {
 		lsr  a
 		tax
 		lda  _hex_digits,x
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 
 _hex_digits	.text "0123456789abcdef"	; can probably be reused for other stuff as well

compiler/res/prog8lib/cx16flt.p8

@@ -4,14 +4,14 @@
 ;
 ; indent format: TABS, size=8
 
+%target cx16
 %option enable_floats
 
 c64flt {
 	; ---- this block contains C-64 floating point related functions ----
 
-        const  float  PI        = 3.141592653589793
-        const  float  TWOPI	    = 6.283185307179586
-        const  float  ZERO      = 0.0
+        const float  PI     = 3.141592653589793
+        const float  TWOPI  = 6.283185307179586
 
 
 ; ---- ROM float functions ----
@@ -130,20 +130,20 @@ asmsub  GETADRAY  () clobbers(X) -> uword @ AY  {
 sub  print_f  (float value) {
 	; ---- prints the floating point value (without a newline).
 	%asm {{
-		stx  P8ZP_SCRATCH_REG_X
+		phx
 		lda  #<value
 		ldy  #>value
 		jsr  MOVFM		; load float into fac1
 		jsr  FOUT		; fac1 to string in A/Y
-		sta  P8ZP_SCRATCH_B1
-		sty  P8ZP_SCRATCH_REG
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
 		ldy  #0
--		lda  (P8ZP_SCRATCH_B1),y
+-		lda  (P8ZP_SCRATCH_W1),y
 		beq  +
 		jsr  c64.CHROUT
 		iny
 		bne  -
-		ldx  P8ZP_SCRATCH_REG_X
+		plx
 +		rts
 	}}
 }

compiler/res/prog8lib/cx16lib.p8

@@ -5,6 +5,8 @@
 ;
 ; indent format: TABS, size=8
 
+%target cx16
+
 
 c64 {
 
@@ -14,42 +16,42 @@ c64 {
 ; CLEARSCR -> use screen.clear_screen
 ; HOMECRSR -> use screen.plot
 
-romsub $FF81 = CINT() clobbers(A,X,Y)                           ; (alias: SCINIT) initialize screen editor and video chip
-romsub $FF84 = IOINIT() clobbers(A, X)                          ; initialize I/O devices (CIA, SID, IRQ)
-romsub $FF87 = RAMTAS() clobbers(A,X,Y)                         ; initialize RAM, tape buffer, screen
-romsub $FF8A = RESTOR() clobbers(A,X,Y)                         ; restore default I/O vectors
-romsub $FF8D = VECTOR(uword userptr @ XY, ubyte dir @ Pc) clobbers(A,Y)     ; read/set I/O vector table
+romsub $FF81 = CINT()  clobbers(A,X,Y)                           ; (alias: SCINIT) initialize screen editor and video chip
+romsub $FF84 = IOINIT()  clobbers(A, X)                          ; initialize I/O devices (CIA, SID, IRQ)
+romsub $FF87 = RAMTAS()  clobbers(A,X,Y)                         ; initialize RAM, tape buffer, screen
+romsub $FF8A = RESTOR()  clobbers(A,X,Y)                         ; restore default I/O vectors
+romsub $FF8D = VECTOR(uword userptr @ XY, ubyte dir @ Pc)  clobbers(A,Y)     ; read/set I/O vector table
 romsub $FF90 = SETMSG(ubyte value @ A)                          ; set Kernal message control flag
-romsub $FF93 = SECOND(ubyte address @ A) clobbers(A)            ; (alias: LSTNSA) send secondary address after LISTEN
-romsub $FF96 = TKSA(ubyte address @ A) clobbers(A)              ; (alias: TALKSA) send secondary address after TALK
+romsub $FF93 = SECOND(ubyte address @ A)  clobbers(A)            ; (alias: LSTNSA) send secondary address after LISTEN
+romsub $FF96 = TKSA(ubyte address @ A)  clobbers(A)              ; (alias: TALKSA) send secondary address after TALK
 romsub $FF99 = MEMTOP(uword address @ XY, ubyte dir @ Pc) -> uword @ XY     ; read/set top of memory  pointer
 romsub $FF9C = MEMBOT(uword address @ XY, ubyte dir @ Pc) -> uword @ XY     ; read/set bottom of memory  pointer
-romsub $FF9F = SCNKEY() clobbers(A,X,Y)                         ; scan the keyboard
+romsub $FF9F = SCNKEY()  clobbers(A,X,Y)                         ; scan the keyboard
 romsub $FFA2 = SETTMO(ubyte timeout @ A)                        ; set time-out flag for IEEE bus
 romsub $FFA5 = ACPTR() -> ubyte @ A                             ; (alias: IECIN) input byte from serial bus
 romsub $FFA8 = CIOUT(ubyte databyte @ A)                        ; (alias: IECOUT) output byte to serial bus
-romsub $FFAB = UNTLK() clobbers(A)                              ; command serial bus device to UNTALK
-romsub $FFAE = UNLSN() clobbers(A)                              ; command serial bus device to UNLISTEN
-romsub $FFB1 = LISTEN(ubyte device @ A) clobbers(A)             ; command serial bus device to LISTEN
-romsub $FFB4 = TALK(ubyte device @ A) clobbers(A)               ; command serial bus device to TALK
+romsub $FFAB = UNTLK()  clobbers(A)                              ; command serial bus device to UNTALK
+romsub $FFAE = UNLSN()  clobbers(A)                              ; command serial bus device to UNLISTEN
+romsub $FFB1 = LISTEN(ubyte device @ A)  clobbers(A)             ; command serial bus device to LISTEN
+romsub $FFB4 = TALK(ubyte device @ A)  clobbers(A)               ; command serial bus device to TALK
 romsub $FFB7 = READST() -> ubyte @ A                            ; read I/O status word
 romsub $FFBA = SETLFS(ubyte logical @ A, ubyte device @ X, ubyte address @ Y)   ; set logical file parameters
 romsub $FFBD = SETNAM(ubyte namelen @ A, str filename @ XY)     ; set filename parameters
-romsub $FFC0 = OPEN() clobbers(A,X,Y)                           ; (via 794 ($31A)) open a logical file
-romsub $FFC3 = CLOSE(ubyte logical @ A) clobbers(A,X,Y)         ; (via 796 ($31C)) close a logical file
-romsub $FFC6 = CHKIN(ubyte logical @ X) clobbers(A,X)           ; (via 798 ($31E)) define an input channel
-romsub $FFC9 = CHKOUT(ubyte logical @ X) clobbers(A,X)          ; (via 800 ($320)) define an output channel
-romsub $FFCC = CLRCHN() clobbers(A,X)                           ; (via 802 ($322)) restore default devices
-romsub $FFCF = CHRIN() clobbers(Y) -> ubyte @ A                 ; (via 804 ($324)) input a character (for keyboard, read a whole line from the screen) A=byte read.
+romsub $FFC0 = OPEN()  clobbers(A,X,Y)                           ; (via 794 ($31A)) open a logical file
+romsub $FFC3 = CLOSE(ubyte logical @ A)  clobbers(A,X,Y)         ; (via 796 ($31C)) close a logical file
+romsub $FFC6 = CHKIN(ubyte logical @ X)  clobbers(A,X)           ; (via 798 ($31E)) define an input channel
+romsub $FFC9 = CHKOUT(ubyte logical @ X)  clobbers(A,X)          ; (via 800 ($320)) define an output channel
+romsub $FFCC = CLRCHN()  clobbers(A,X)                           ; (via 802 ($322)) restore default devices
+romsub $FFCF = CHRIN()  clobbers(Y) -> ubyte @ A                 ; (via 804 ($324)) input a character (for keyboard, read a whole line from the screen) A=byte read.
 romsub $FFD2 = CHROUT(ubyte char @ A)                           ; (via 806 ($326)) output a character
 romsub $FFD5 = LOAD(ubyte verify @ A, uword address @ XY) -> ubyte @Pc, ubyte @ A, ubyte @ X, ubyte @ Y     ; (via 816 ($330)) load from device
 romsub $FFD8 = SAVE(ubyte zp_startaddr @ A, uword endaddr @ XY) -> ubyte @ Pc, ubyte @ A                    ; (via 818 ($332)) save to a device
 romsub $FFDB = SETTIM(ubyte low @ A, ubyte middle @ X, ubyte high @ Y)      ; set the software clock
 romsub $FFDE = RDTIM() -> ubyte @ A, ubyte @ X, ubyte @ Y       ; read the software clock
-romsub $FFE1 = STOP() clobbers(A,X) -> ubyte @ Pz, ubyte @ Pc   ; (via 808 ($328)) check the STOP key
-romsub $FFE4 = GETIN() clobbers(X,Y) -> ubyte @ A               ; (via 810 ($32A)) get a character
-romsub $FFE7 = CLALL() clobbers(A,X)                            ; (via 812 ($32C)) close all files
-romsub $FFEA = UDTIM() clobbers(A,X)                            ; update the software clock
+romsub $FFE1 = STOP()  clobbers(A,X) -> ubyte @ Pz, ubyte @ Pc   ; (via 808 ($328)) check the STOP key
+romsub $FFE4 = GETIN()  clobbers(X,Y) -> ubyte @ A               ; (via 810 ($32A)) get a character
+romsub $FFE7 = CLALL()  clobbers(A,X)                            ; (via 812 ($32C)) close all files
+romsub $FFEA = UDTIM()  clobbers(A,X)                            ; update the software clock
 romsub $FFED = SCREEN() -> ubyte @ X, ubyte @ Y                 ; read number of screen rows and columns
 romsub $FFF0 = PLOT(ubyte col @ Y, ubyte row @ X, ubyte dir @ Pc) -> ubyte @ X, ubyte @ Y       ; read/set position of cursor on screen.  Use screen.plot for a 'safe' wrapper that preserves X.
 romsub $FFF3 = IOBASE() -> uword @ XY                           ; read base address of I/O devices
@@ -58,8 +60,11 @@ romsub $FFF3 = IOBASE() -> uword @ XY                           ; read base addr
 
 cx16 {
 
-; ---- Commander X-16 additions on top of C64 kernal routines ----
-; spelling of the names is taken from the Commander X-16 rom sources
+        ; 65c02 hardware vectors:
+        &uword  NMI_VEC         = $FFFA     ; 6502 nmi vector, determined by the kernal if banked in
+        &uword  RESET_VEC       = $FFFC     ; 6502 reset vector, determined by the kernal if banked in
+        &uword  IRQ_VEC         = $FFFE     ; 6502 interrupt vector, determined by the kernal if banked in
+
 
 ; the sixteen virtual 16-bit registers
     &uword r0  = $02
@@ -82,116 +87,159 @@ cx16 {
 ; VERA registers
 
     const uword VERA_BASE = $9F20
-    &uword VERA_ADDR_L   	 = VERA_BASE + $00
-    &uword VERA_ADDR_M   	 = VERA_BASE + $01
-    &uword VERA_ADDR_H   	 = VERA_BASE + $02
-    &uword VERA_DATA0        = VERA_BASE + $03
-    &uword VERA_DATA1        = VERA_BASE + $04
-    &uword VERA_CTRL         = VERA_BASE + $05
-    &uword VERA_IEN          = VERA_BASE + $06
-    &uword VERA_ISR          = VERA_BASE + $07
-    &uword VERA_IRQ_LINE_L   = VERA_BASE + $08
-    &uword VERA_DC_VIDEO     = VERA_BASE + $09
-    &uword VERA_DC_HSCALE    = VERA_BASE + $0A
-    &uword VERA_DC_VSCALE    = VERA_BASE + $0B
-    &uword VERA_DC_BORDER    = VERA_BASE + $0C
-    &uword VERA_DC_HSTART    = VERA_BASE + $09
-    &uword VERA_DC_HSTOP     = VERA_BASE + $0A
-    &uword VERA_DC_VSTART    = VERA_BASE + $0B
-    &uword VERA_DC_VSTOP     = VERA_BASE + $0C
-    &uword VERA_L0_CONFIG    = VERA_BASE + $0D
-    &uword VERA_L0_MAPBASE   = VERA_BASE + $0E
-    &uword VERA_L0_TILEBASE  = VERA_BASE + $0F
-    &uword VERA_L0_HSCROLL_L = VERA_BASE + $10
-    &uword VERA_L0_HSCROLL_H = VERA_BASE + $11
-    &uword VERA_L0_VSCROLL_L = VERA_BASE + $12
-    &uword VERA_L0_VSCROLL_H = VERA_BASE + $13
-    &uword VERA_L1_CONFIG    = VERA_BASE + $14
-    &uword VERA_L1_MAPBASE   = VERA_BASE + $15
-    &uword VERA_L1_TILEBASE  = VERA_BASE + $16
-    &uword VERA_L1_HSCROLL_L = VERA_BASE + $17
-    &uword VERA_L1_HSCROLL_H = VERA_BASE + $18
-    &uword VERA_L1_VSCROLL_L = VERA_BASE + $19
-    &uword VERA_L1_VSCROLL_H = VERA_BASE + $1A
-    &uword VERA_AUDIO_CTRL   = VERA_BASE + $1B
-    &uword VERA_AUDIO_RATE   = VERA_BASE + $1C
-    &uword VERA_AUDIO_DATA   = VERA_BASE + $1D
-    &uword VERA_SPI_DATA     = VERA_BASE + $1E
-    &uword VERA_SPI_CTRL     = VERA_BASE + $1F
+    &ubyte  VERA_ADDR_L   	  = VERA_BASE + $00
+    &ubyte  VERA_ADDR_M   	  = VERA_BASE + $01
+    &ubyte  VERA_ADDR_H   	  = VERA_BASE + $02
+    &ubyte  VERA_DATA0        = VERA_BASE + $03
+    &ubyte  VERA_DATA1        = VERA_BASE + $04
+    &ubyte  VERA_CTRL         = VERA_BASE + $05
+    &ubyte  VERA_IEN          = VERA_BASE + $06
+    &ubyte  VERA_ISR          = VERA_BASE + $07
+    &ubyte  VERA_IRQ_LINE_L   = VERA_BASE + $08
+    &ubyte  VERA_DC_VIDEO     = VERA_BASE + $09
+    &ubyte  VERA_DC_HSCALE    = VERA_BASE + $0A
+    &ubyte  VERA_DC_VSCALE    = VERA_BASE + $0B
+    &ubyte  VERA_DC_BORDER    = VERA_BASE + $0C
+    &ubyte  VERA_DC_HSTART    = VERA_BASE + $09
+    &ubyte  VERA_DC_HSTOP     = VERA_BASE + $0A
+    &ubyte  VERA_DC_VSTART    = VERA_BASE + $0B
+    &ubyte  VERA_DC_VSTOP     = VERA_BASE + $0C
+    &ubyte  VERA_L0_CONFIG    = VERA_BASE + $0D
+    &ubyte  VERA_L0_MAPBASE   = VERA_BASE + $0E
+    &ubyte  VERA_L0_TILEBASE  = VERA_BASE + $0F
+    &ubyte  VERA_L0_HSCROLL_L = VERA_BASE + $10
+    &ubyte  VERA_L0_HSCROLL_H = VERA_BASE + $11
+    &ubyte  VERA_L0_VSCROLL_L = VERA_BASE + $12
+    &ubyte  VERA_L0_VSCROLL_H = VERA_BASE + $13
+    &ubyte  VERA_L1_CONFIG    = VERA_BASE + $14
+    &ubyte  VERA_L1_MAPBASE   = VERA_BASE + $15
+    &ubyte  VERA_L1_TILEBASE  = VERA_BASE + $16
+    &ubyte  VERA_L1_HSCROLL_L = VERA_BASE + $17
+    &ubyte  VERA_L1_HSCROLL_H = VERA_BASE + $18
+    &ubyte  VERA_L1_VSCROLL_L = VERA_BASE + $19
+    &ubyte  VERA_L1_VSCROLL_H = VERA_BASE + $1A
+    &ubyte  VERA_AUDIO_CTRL   = VERA_BASE + $1B
+    &ubyte  VERA_AUDIO_RATE   = VERA_BASE + $1C
+    &ubyte  VERA_AUDIO_DATA   = VERA_BASE + $1D
+    &ubyte  VERA_SPI_DATA     = VERA_BASE + $1E
+    &ubyte  VERA_SPI_CTRL     = VERA_BASE + $1F
 ; VERA_PSG_BASE     = $1F9C0
 ; VERA_PALETTE_BASE = $1FA00
 ; VERA_SPRITES_BASE = $1FC00
 
+; I/O
+
+    const uword  via1  = $9f60                  ;VIA 6522 #1
+    &ubyte  d1prb	= via1+0
+    &ubyte  d1pra	= via1+1
+    &ubyte  d1ddrb	= via1+2
+    &ubyte  d1ddra	= via1+3
+    &ubyte  d1t1l	= via1+4
+    &ubyte  d1t1h	= via1+5
+    &ubyte  d1t1ll	= via1+6
+    &ubyte  d1t1lh	= via1+7
+    &ubyte  d1t2l	= via1+8
+    &ubyte  d1t2h	= via1+9
+    &ubyte  d1sr	= via1+10
+    &ubyte  d1acr	= via1+11
+    &ubyte  d1pcr	= via1+12
+    &ubyte  d1ifr	= via1+13
+    &ubyte  d1ier	= via1+14
+    &ubyte  d1ora	= via1+15
+
+    const uword  via2  = $9f70                  ;VIA 6522 #2
+    &ubyte  d2prb	=via2+0
+    &ubyte  d2pra	=via2+1
+    &ubyte  d2ddrb	=via2+2
+    &ubyte  d2ddra	=via2+3
+    &ubyte  d2t1l	=via2+4
+    &ubyte  d2t1h	=via2+5
+    &ubyte  d2t1ll	=via2+6
+    &ubyte  d2t1lh	=via2+7
+    &ubyte  d2t2l	=via2+8
+    &ubyte  d2t2h	=via2+9
+    &ubyte  d2sr	=via2+10
+    &ubyte  d2acr	=via2+11
+    &ubyte  d2pcr	=via2+12
+    &ubyte  d2ifr	=via2+13
+    &ubyte  d2ier	=via2+14
+    &ubyte  d2ora	=via2+15
+
+
+; ---- Commander X-16 additions on top of C64 kernal routines ----
+; spelling of the names is taken from the Commander X-16 rom sources
+
+; TODO specify the correct clobbers for alle these functions, for simplicity all 3 regs are marked clobbered atm
 
 ; supported C128 additions
-romsub $ff4a = close_all()
-romsub $ff59 = lkupla()
-romsub $ff5c = lkupsa()
-romsub $ff5f = screen_set_mode(ubyte mode @A) clobbers(A, X, Y) -> ubyte @Pc
-romsub $ff62 = screen_set_charset(ubyte charset @A, uword charsetptr @XY) clobbers(A,X,Y)      ; incompatible with C128  dlchr()
-romsub $ff65 = pfkey()
-romsub $ff6e = jsrfar()
-romsub $ff74 = fetch()
-romsub $ff77 = stash()
-romsub $ff7a = cmpare()
-romsub $ff7d = primm()
+romsub $ff4a = close_all()  clobbers(A,X,Y)
+romsub $ff59 = lkupla()  clobbers(A,X,Y)
+romsub $ff5c = lkupsa()  clobbers(A,X,Y)
+romsub $ff5f = screen_set_mode(ubyte mode @A)  clobbers(A, X, Y) -> ubyte @Pc
+romsub $ff62 = screen_set_charset(ubyte charset @A, uword charsetptr @XY)  clobbers(A,X,Y)      ; incompatible with C128  dlchr()
+romsub $ff65 = pfkey()  clobbers(A,X,Y)
+romsub $ff6e = jsrfar()  clobbers(A,X,Y)
+romsub $ff74 = fetch()  clobbers(A,X,Y)
+romsub $ff77 = stash()  clobbers(A,X,Y)
+romsub $ff7a = cmpare()  clobbers(A,X,Y)
+romsub $ff7d = primm()  clobbers(A,X,Y)
 
 ; X16 additions
-romsub $ff44 = macptr()
-romsub $ff47 = enter_basic(ubyte cold_or_warm @Pc)
-romsub $ff68 = mouse_config(ubyte shape @A, ubyte scale @X) clobbers (A, X, Y)
-romsub $ff6b = mouse_get(ubyte zpdataptr @X) clobbers(A)
-romsub $ff71 = mouse_scan() clobbers(A, X, Y)
-romsub $ff53 = joystick_scan() clobbers(A, X, Y)
+romsub $ff44 = macptr()  clobbers(A,X,Y)
+romsub $ff47 = enter_basic(ubyte cold_or_warm @Pc)  clobbers(A,X,Y)
+romsub $ff68 = mouse_config(ubyte shape @A, ubyte scale @X)  clobbers (A, X, Y)
+romsub $ff6b = mouse_get(ubyte zpdataptr @X)  clobbers(A)
+romsub $ff71 = mouse_scan()  clobbers(A, X, Y)
+romsub $ff53 = joystick_scan()  clobbers(A, X, Y)
 romsub $ff56 = joystick_get(ubyte joynr @A) -> ubyte @A, ubyte @X, ubyte @Y
-romsub $ff4d = clock_set_date_time() clobbers(A, X, Y)      ; args: r0, r1, r2, r3L
-romsub $ff50 = clock_get_date_time() clobbers(A)            ; outout args: r0, r1, r2, r3L
+romsub $ff4d = clock_set_date_time()  clobbers(A, X, Y)      ; args: r0, r1, r2, r3L
+romsub $ff50 = clock_get_date_time()  clobbers(A)            ; outout args: r0, r1, r2, r3L
 
 ; high level graphics & fonts
-romsub $ff20 = GRAPH_init()             ; uses vectors=r0
-romsub $ff23 = GRAPH_clear()
-romsub $ff26 = GRAPH_set_window()       ; uses x=r0, y=r1, width=r2, height=r3
-romsub $ff29 = GRAPH_set_colors(ubyte stroke @A, ubyte fill @X, ubyte background @Y)
-romsub $ff2c = GRAPH_draw_line()        ; uses x1=r0, y1=r1, x2=r2, y2=r3
-romsub $ff2f = GRAPH_draw_rect(ubyte fill @Pc)        ; uses x=r0, y=r1, width=r2, height=r3, cornerradius=r4
-romsub $ff32 = GRAPH_move_rect()        ; uses sx=r0, sy=r1, tx=r2, ty=r3, width=r4, height=r5
-romsub $ff35 = GRAPH_draw_oval(ubyte fill @Pc)        ; uses x=r0, y=r1, width=r2, height=r3
-romsub $ff38 = GRAPH_draw_image()       ; uses x=r0, y=r1, ptr=r2, width=r3, height=r4
-romsub $ff3b = GRAPH_set_font()         ; uses ptr=r0
-romsub $ff3e = GRAPH_get_char_size(ubyte baseline @A, ubyte width @X, ubyte height_or_style @Y, ubyte is_control @Pc)
-romsub $ff41 = GRAPH_put_char(ubyte char @A)   ; uses x=r0, y=r1
+romsub $ff20 = GRAPH_init()  clobbers(A,X,Y)           ; uses vectors=r0
+romsub $ff23 = GRAPH_clear()  clobbers(A,X,Y)
+romsub $ff26 = GRAPH_set_window()  clobbers(A,X,Y)      ; uses x=r0, y=r1, width=r2, height=r3
+romsub $ff29 = GRAPH_set_colors(ubyte stroke @A, ubyte fill @X, ubyte background @Y)  clobbers (A,X,Y)
+romsub $ff2c = GRAPH_draw_line()  clobbers(A,X,Y)       ; uses x1=r0, y1=r1, x2=r2, y2=r3
+romsub $ff2f = GRAPH_draw_rect(ubyte fill @Pc)  clobbers(A,X,Y)   ; uses x=r0, y=r1, width=r2, height=r3, cornerradius=r4
+romsub $ff32 = GRAPH_move_rect()  clobbers(A,X,Y)       ; uses sx=r0, sy=r1, tx=r2, ty=r3, width=r4, height=r5
+romsub $ff35 = GRAPH_draw_oval(ubyte fill @Pc)  clobbers(A,X,Y)       ; uses x=r0, y=r1, width=r2, height=r3
+romsub $ff38 = GRAPH_draw_image()  clobbers(A,X,Y)      ; uses x=r0, y=r1, ptr=r2, width=r3, height=r4
+romsub $ff3b = GRAPH_set_font()  clobbers(A,X,Y)         ; uses ptr=r0
+romsub $ff3e = GRAPH_get_char_size(ubyte baseline @A, ubyte width @X, ubyte height_or_style @Y, ubyte is_control @Pc)  clobbers(A,X,Y)
+romsub $ff41 = GRAPH_put_char(ubyte char @A)  clobbers(A,X,Y)   ; uses x=r0, y=r1
 
 ; framebuffer
-romsub $fef6 = FB_init()
-romsub $fef9 = FB_get_info() -> byte @A    ; also outputs width=r0, height=r1
-romsub $fefc = FB_set_palette(ubyte index @A, ubyte bytecount @X)      ; also uses pointer=r0
-romsub $feff = FB_cursor_position()    ; uses x=r0, y=r1
-romsub $ff02 = FB_cursor_next_line()   ; uses x=r0
-romsub $ff05 = FB_get_pixel() -> ubyte @A
-romsub $ff08 = FB_get_pixels()         ; uses ptr=r0, count=r1
-romsub $ff0b = FB_set_pixel(ubyte color @A)
-romsub $ff0e = FB_set_pixels()         ; uses ptr=r0, count=r1
-romsub $ff11 = FB_set_8_pixels(ubyte pattern @A, ubyte color @X)
-romsub $ff14 = FB_set_8_pixels_opaque(ubyte pattern @A, ubyte color1 @X, ubyte color2 @Y)  ; also uses mask=r0L
-romsub $ff17 = FB_fill_pixels(ubyte color @A)   ; also uses count=r0, step=r1
-romsub $ff1a = FB_filter_pixels()      ; uses ptr=r0, count=r1
-romsub $ff1d = FB_move_pixels()        ; uses sx=r0, sy=r1, tx=r2, ty=r3, count=r4
+romsub $fef6 = FB_init()  clobbers(A,X,Y)
+romsub $fef9 = FB_get_info()  clobbers(X,Y) -> byte @A    ; also outputs width=r0, height=r1
+romsub $fefc = FB_set_palette(ubyte index @A, ubyte bytecount @X)  clobbers(A,X,Y)      ; also uses pointer=r0
+romsub $feff = FB_cursor_position()  clobbers(A,X,Y)    ; uses x=r0, y=r1
+romsub $ff02 = FB_cursor_next_line()  clobbers(A,X,Y)   ; uses x=r0
+romsub $ff05 = FB_get_pixel()  clobbers(X,Y) -> ubyte @A
+romsub $ff08 = FB_get_pixels()  clobbers(A,X,Y)         ; uses ptr=r0, count=r1
+romsub $ff0b = FB_set_pixel(ubyte color @A)  clobbers(A,X,Y)
+romsub $ff0e = FB_set_pixels()  clobbers(A,X,Y)         ; uses ptr=r0, count=r1
+romsub $ff11 = FB_set_8_pixels(ubyte pattern @A, ubyte color @X)  clobbers(A,X,Y)
+romsub $ff14 = FB_set_8_pixels_opaque(ubyte pattern @A, ubyte color1 @X, ubyte color2 @Y)  clobbers(A,X,Y)  ; also uses mask=r0L
+romsub $ff17 = FB_fill_pixels(ubyte color @A)  clobbers(A,X,Y)   ; also uses count=r0, step=r1
+romsub $ff1a = FB_filter_pixels()  clobbers(A,X,Y)      ; uses ptr=r0, count=r1
+romsub $ff1d = FB_move_pixels()  clobbers(A,X,Y)        ; uses sx=r0, sy=r1, tx=r2, ty=r3, count=r4
 
 ; misc
-romsub $fef0 = sprite_set_image(ubyte number @A, ubyte width @X, ubyte height @Y, ubyte apply_mask @Pc) -> ubyte @Pc  ; also uses pixels=r0, mask=r1, bpp=r2L
-romsub $fef3 = sprite_set_position(ubyte number @A)  ; also uses x=r0 and y=r1
-romsub $fee4 = memory_fill(ubyte value @A)      ; uses address=r0, num_bytes=r1
-romsub $fee7 = memory_copy()                    ; uses source=r0, target=r1, num_bytes=r2
-romsub $feea = memory_crc()                     ; uses address=r0, num_bytes=r1     result->r2
-romsub $feed = memory_decompress()              ; uses input=r0, output=r1     result->r1
-romsub $fedb = console_init()           ; uses x=r0, y=r1, width=r2, height=r3
-romsub $fede = console_put_char(ubyte char @A, ubyte wrapping @Pc)
-romsub $fee1 = console_get_char() -> ubyte @A
-romsub $fed8 = console_put_image()      ; uses ptr=r0, width=r1, height=r2
-romsub $fed5 = console_set_paging_message()     ; uses messageptr=r0
-romsub $fed2 = kbdbuf_put(ubyte key @A)
+romsub $fef0 = sprite_set_image(ubyte number @A, ubyte width @X, ubyte height @Y, ubyte apply_mask @Pc)  clobbers(A,X,Y) -> ubyte @Pc  ; also uses pixels=r0, mask=r1, bpp=r2L
+romsub $fef3 = sprite_set_position(ubyte number @A)  clobbers(A,X,Y)  ; also uses x=r0 and y=r1
+romsub $fee4 = memory_fill(ubyte value @A)  clobbers(A,X,Y)      ; uses address=r0, num_bytes=r1
+romsub $fee7 = memory_copy()  clobbers(A,X,Y)                    ; uses source=r0, target=r1, num_bytes=r2
+romsub $feea = memory_crc()  clobbers(A,X,Y)                     ; uses address=r0, num_bytes=r1     result->r2
+romsub $feed = memory_decompress()  clobbers(A,X,Y)              ; uses input=r0, output=r1     result->r1
+romsub $fedb = console_init()  clobbers(A,X,Y)           ; uses x=r0, y=r1, width=r2, height=r3
+romsub $fede = console_put_char(ubyte char @A, ubyte wrapping @Pc)  clobbers(A,X,Y)
+romsub $fee1 = console_get_char()  clobbers(X,Y) -> ubyte @A
+romsub $fed8 = console_put_image()  clobbers(A,X,Y)      ; uses ptr=r0, width=r1, height=r2
+romsub $fed5 = console_set_paging_message()  clobbers(A,X,Y)     ; uses messageptr=r0
+romsub $fed2 = kbdbuf_put(ubyte key @A)  clobbers(A,X,Y)
 romsub $fecf = entropy_get() -> ubyte @A, ubyte @X, ubyte @Y
-romsub $fecc = monitor()
+romsub $fecc = monitor()  clobbers(A,X,Y)
 
 
 
@@ -203,8 +251,11 @@ asmsub init_system()  {
     %asm {{
         sei
         cld
-        stz  $00
-        stz  $01
+        ;stz  $00
+        ;stz  $01
+        ;stz  d1prb      ; select rom bank 0
+        lda  #$80
+        sta  VERA_CTRL
         jsr  c64.IOINIT
         jsr  c64.RESTOR
         jsr  c64.CINT

compiler/res/prog8lib/cx16textio.p8

@@ -4,7 +4,7 @@
 ;
 ; indent format: TABS, size=8
 
-
+%target cx16
 %import cx16lib
 %import conv
 
@@ -54,6 +54,36 @@ asmsub  fill_screen (ubyte char @ A, ubyte txtcolor @ Y) clobbers(A)  {
 
 }
 
+asmsub  clear_screenchars (ubyte char @ A) clobbers(Y)  {
+	; ---- clear the character screen with the given fill character (leaves colors)
+	;      (assumes screen matrix is at the default address)
+	%asm {{
+        pha
+        phx
+        jsr  c64.SCREEN             ; get dimensions in X/Y
+        dex
+        dey
+        txa
+        asl  a
+        sta  P8ZP_SCRATCH_B1
+        pla
+-       ldx  P8ZP_SCRATCH_B1
+-       stz  cx16.VERA_ADDR_H
+        stx  cx16.VERA_ADDR_L
+        sty  cx16.VERA_ADDR_M
+        sta  cx16.VERA_DATA0
+        dex
+        dex
+        cpx  #254
+        bne  -
+        dey
+        bpl  --
+        plx
+        rts
+        }}
+}
+
+
 ubyte[16] color_to_charcode = [$90,$05,$1c,$9f,$9c,$1e,$1f,$9e,$81,$95,$96,$97,$98,$99,$9a,$9b]
 
 sub color (ubyte txtcol) {

compiler/res/prog8lib/math.asm

@@ -11,6 +11,10 @@
 ;	http://codebase64.org/doku.php?id=base:6502_6510_maths
 ;
 
+
+math_store_reg	.byte  0		; temporary storage
+
+
 multiply_bytes	.proc
 	; -- multiply 2 bytes A and Y, result as byte in A  (signed or unsigned)
 		sta  P8ZP_SCRATCH_B1         ; num1
@@ -31,7 +35,7 @@ multiply_bytes_16	.proc
 	; -- multiply 2 bytes A and Y, result as word in A/Y (unsigned)
 		sta  P8ZP_SCRATCH_B1
 		sty  P8ZP_SCRATCH_REG
-		stx  P8ZP_SCRATCH_REG_X
+		stx  math_store_reg
 		lda  #0
 		ldx  #8
 		lsr  P8ZP_SCRATCH_B1
@@ -44,7 +48,7 @@ multiply_bytes_16	.proc
 		bne  -
 		tay
 		lda  P8ZP_SCRATCH_B1
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  math_store_reg
 		rts
 		.pend
 
@@ -57,7 +61,7 @@ multiply_words	.proc
 
 		sta  P8ZP_SCRATCH_W2
 		sty  P8ZP_SCRATCH_W2+1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 
 mult16		lda  #0
 		sta  result+2	; clear upper bits of product
@@ -79,7 +83,7 @@ mult16		lda  #0
 		ror  result
 		dex
 		bne  -
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 
 result		.byte  0,0,0,0
@@ -124,7 +128,7 @@ divmod_ub_asm	.proc
 	;    division by zero will result in quotient = 255 and remainder = original number
 		sty  P8ZP_SCRATCH_REG
 		sta  P8ZP_SCRATCH_B1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  math_store_reg
 
 		lda  #0
 		ldx  #8
@@ -137,7 +141,7 @@ divmod_ub_asm	.proc
 		dex
 		bne  -
 		ldy  P8ZP_SCRATCH_B1
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  math_store_reg
 		rts
 		.pend
 
@@ -197,7 +201,7 @@ result = dividend ;save memory by reusing divident to store the result
 
 		sta  _divisor
 		sty  _divisor+1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		lda  #0	        	;preset remainder to 0
 		sta  remainder
 		sta  remainder+1
@@ -224,7 +228,7 @@ result = dividend ;save memory by reusing divident to store the result
 
 		lda  result
 		ldy  result+1
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		rts
 _divisor	.word 0
 		.pend
@@ -308,7 +312,8 @@ _seed		.word	$2c9e
 		.pend
 
 
-mul_byte_3	.proc
+; ----------- optimized multiplications (stack) : ---------
+stack_mul_byte_3	.proc
 		; X + X*2
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -318,7 +323,7 @@ mul_byte_3	.proc
 		rts
 		.pend
 
-mul_word_3	.proc
+stack_mul_word_3	.proc
 		; W*2 + W
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_REG
@@ -335,7 +340,7 @@ mul_word_3	.proc
 		.pend
 
 
-mul_byte_5	.proc
+stack_mul_byte_5	.proc
 		; X*4 + X
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -346,7 +351,7 @@ mul_byte_5	.proc
 		rts
 		.pend
 
-mul_word_5	.proc
+stack_mul_word_5	.proc
 		; W*4 + W
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_REG
@@ -365,7 +370,7 @@ mul_word_5	.proc
 		.pend
 
 
-mul_byte_6	.proc
+stack_mul_byte_6	.proc
 		; (X*2 + X)*2
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -376,7 +381,7 @@ mul_byte_6	.proc
 		rts
 		.pend
 
-mul_word_6	.proc
+stack_mul_word_6	.proc
 		; (W*2 + W)*2
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_REG
@@ -394,7 +399,7 @@ mul_word_6	.proc
 		rts
 		.pend
 
-mul_byte_7	.proc
+stack_mul_byte_7	.proc
 		; X*8 - X
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -406,7 +411,7 @@ mul_byte_7	.proc
 		rts
 		.pend
 
-mul_word_7	.proc
+stack_mul_word_7	.proc
 		; W*8 - W
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_REG
@@ -426,7 +431,7 @@ mul_word_7	.proc
 		rts
 		.pend
 
-mul_byte_9	.proc
+stack_mul_byte_9	.proc
 		; X*8 + X
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -438,7 +443,7 @@ mul_byte_9	.proc
 		rts
 		.pend
 
-mul_word_9	.proc
+stack_mul_word_9	.proc
 		; W*8 + W
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_REG
@@ -458,7 +463,7 @@ mul_word_9	.proc
 		rts
 		.pend
 
-mul_byte_10	.proc
+stack_mul_byte_10	.proc
 		; (X*4 + X)*2
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -470,7 +475,7 @@ mul_byte_10	.proc
 		rts
 		.pend
 
-mul_word_10	.proc
+stack_mul_word_10	.proc
 		; (W*4 + W)*2
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_REG
@@ -490,7 +495,7 @@ mul_word_10	.proc
 		rts
 		.pend
 
-mul_byte_11	.proc
+stack_mul_byte_11	.proc
 		; (X*2 + X)*4 - X
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -506,7 +511,7 @@ mul_byte_11	.proc
 
 ; mul_word_11 is skipped (too much code)
 
-mul_byte_12	.proc
+stack_mul_byte_12	.proc
 		; (X*2 + X)*4
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -518,7 +523,7 @@ mul_byte_12	.proc
 		rts
 		.pend
 
-mul_word_12	.proc
+stack_mul_word_12	.proc
 		; (W*2 + W)*4
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_REG
@@ -538,7 +543,7 @@ mul_word_12	.proc
 		rts
 		.pend
 
-mul_byte_13	.proc
+stack_mul_byte_13	.proc
 		; (X*2 + X)*4 + X
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -554,7 +559,7 @@ mul_byte_13	.proc
 
 ; mul_word_13 is skipped (too much code)
 
-mul_byte_14	.proc
+stack_mul_byte_14	.proc
 		; (X*8 - X)*2
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -569,7 +574,7 @@ mul_byte_14	.proc
 
 ; mul_word_14 is skipped (too much code)
 
-mul_byte_15	.proc
+stack_mul_byte_15	.proc
 		; X*16 - X
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -582,7 +587,7 @@ mul_byte_15	.proc
 		rts
 		.pend
 
-mul_word_15	.proc
+stack_mul_word_15	.proc
 		; W*16 - W
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_REG
@@ -604,7 +609,7 @@ mul_word_15	.proc
 		rts
 		.pend
 
-mul_byte_20	.proc
+stack_mul_byte_20	.proc
 		; (X*4 + X)*4
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -617,7 +622,7 @@ mul_byte_20	.proc
 		rts
 		.pend
 
-mul_word_20	.proc
+stack_mul_word_20	.proc
 		; (W*4 + W)*4
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_REG
@@ -639,7 +644,7 @@ mul_word_20	.proc
 		rts
 		.pend
 
-mul_byte_25	.proc
+stack_mul_byte_25	.proc
 		; (X*2 + X)*8 + X
 		lda  P8ESTACK_LO+1,x
 		asl  a
@@ -654,27 +659,26 @@ mul_byte_25	.proc
 		rts
 		.pend
 
-mul_word_25	.proc
-		; W + W*8 + W*16
+stack_mul_word_25	.proc
+		; W = (W*2 + W) *8 + W
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_W1+1
 		lda  P8ESTACK_LO+1,x
 		asl  a
 		rol  P8ZP_SCRATCH_W1+1
-		asl  a
-		rol  P8ZP_SCRATCH_W1+1
-		asl  a
-		rol  P8ZP_SCRATCH_W1+1
-		sta  P8ZP_SCRATCH_W1
 		clc
 		adc  P8ESTACK_LO+1,x
-		sta  P8ESTACK_LO+1,x
+		sta  P8ZP_SCRATCH_W1
 		lda  P8ZP_SCRATCH_W1+1
 		adc  P8ESTACK_HI+1,x
-		sta  P8ESTACK_HI+1,x
+		sta  P8ZP_SCRATCH_W1+1
 		lda  P8ZP_SCRATCH_W1
 		asl  a
 		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
 		clc
 		adc  P8ESTACK_LO+1,x
 		sta  P8ESTACK_LO+1,x
@@ -684,21 +688,16 @@ mul_word_25	.proc
 		rts
 		.pend
 
-mul_byte_40	.proc
-		; (X*4 + X)*8
+stack_mul_byte_40	.proc
 		lda  P8ESTACK_LO+1,x
-		asl  a
-		asl  a
-                clc
-		adc  P8ESTACK_LO+1,x
-		asl  a
-		asl  a
-		asl  a
+		and  #7
+		tay
+		lda  mul_byte_40._forties,y
 		sta  P8ESTACK_LO+1,x
 		rts
 		.pend
 
-mul_word_40	.proc
+stack_mul_word_40	.proc
 		; (W*4 + W)*8
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_REG
@@ -722,6 +721,529 @@ mul_word_40	.proc
 		rts
 		.pend
 
+stack_mul_byte_50	.proc
+		lda  P8ESTACK_LO+1,x
+		and  #7
+		tay
+		lda  mul_byte_50._fifties, y
+		sta  P8ESTACK_LO+1,x
+		rts
+		.pend
+
+stack_mul_word_50	.proc
+		; W = W * 25 * 2
+		jsr  stack_mul_word_25
+		asl  P8ESTACK_LO+1,x
+		rol  P8ESTACK_HI+1,x
+		rts
+		.pend
+
+stack_mul_byte_80	.proc
+		lda  P8ESTACK_LO+1,x
+		and  #3
+		tay
+		lda  mul_byte_80._eighties, y
+		sta  P8ESTACK_LO+1,x
+		rts
+		.pend
+
+stack_mul_word_80	.proc
+		; W = W * 40 * 2
+		jsr  stack_mul_word_40
+		asl  P8ESTACK_LO+1,x
+		rol  P8ESTACK_HI+1,x
+		rts
+		.pend
+
+stack_mul_byte_100	.proc
+		lda  P8ESTACK_LO+1,x
+		and  #3
+		tay
+		lda  mul_byte_100._hundreds, y
+		sta  P8ESTACK_LO+1,x
+		rts
+		.pend
+
+stack_mul_word_100	.proc
+		; W = W * 25 * 4
+		jsr  stack_mul_word_25
+		asl  P8ESTACK_LO+1,x
+		rol  P8ESTACK_HI+1,x
+		asl  P8ESTACK_LO+1,x
+		rol  P8ESTACK_HI+1,x
+		rts
+		.pend
+
+
+; ----------- optimized multiplications (in-place A (byte) and ?? (word)) : ---------
+mul_byte_3	.proc
+		; A = A + A*2
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+		clc
+		adc  P8P_P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+mul_word_3	.proc
+		; AY = AY*2 + AY
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
+		sta  P8ZP_SCRATCH_W2
+		sty  P8ZP_SCRATCH_W2+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		clc
+		adc  P8ZP_SCRATCH_W2
+		sta  P8ZP_SCRATCH_W1
+		lda  P8ZP_SCRATCH_W1+1
+		adc  P8ZP_SCRATCH_W2+1
+		tay
+		lda  P8ZP_SCRATCH_W1
+		rts
+		.pend
+
+
+mul_byte_5	.proc
+		; A = A*4 + A
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+		clc
+		adc  P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+mul_word_5	.proc
+		; AY = AY*4 + AY
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
+		sta  P8ZP_SCRATCH_W2
+		sty  P8ZP_SCRATCH_W2+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		clc
+		adc  P8ZP_SCRATCH_W2
+		sta  P8ZP_SCRATCH_W1
+		lda  P8ZP_SCRATCH_W1+1
+		adc  P8ZP_SCRATCH_W2+1
+		tay
+		lda  P8ZP_SCRATCH_W1
+		rts
+		.pend
+
+
+mul_byte_6	.proc
+		; A = (A*2 + A)*2
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+                clc
+                adc  P8ZP_SCRATCH_REG
+		asl  a
+		rts
+		.pend
+
+mul_word_6	.proc
+		; AY = (AY*2 + AY)*2
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
+		sta  P8ZP_SCRATCH_W2
+		sty  P8ZP_SCRATCH_W2+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		clc
+		adc  P8ZP_SCRATCH_W2
+		sta  P8ZP_SCRATCH_W1
+		tay
+		lda  P8ZP_SCRATCH_W1+1
+		adc  P8ZP_SCRATCH_W2+1
+		sta  P8ZP_SCRATCH_W1+1
+		tya
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		ldy  P8ZP_SCRATCH_W1+1
+		rts
+		.pend
+
+mul_byte_7	.proc
+		; A = A*8 - A
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+		asl  a
+		sec
+		sbc  P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+mul_word_7	.proc
+		; AY = AY*8 - AY
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
+		sta  P8ZP_SCRATCH_W2
+		sty  P8ZP_SCRATCH_W2+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		sec
+		sbc  P8ZP_SCRATCH_W2
+		sta  P8ZP_SCRATCH_W1
+		lda  P8ZP_SCRATCH_W1+1
+		sbc  P8ZP_SCRATCH_W2+1
+		tay
+		lda  P8ZP_SCRATCH_W1
+		rts
+		.pend
+
+mul_byte_9	.proc
+		; A = A*8 + A
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+		asl  a
+		clc
+		adc  P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+mul_word_9	.proc
+		; AY = AY*8 + AY
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
+		sta  P8ZP_SCRATCH_W2
+		sty  P8ZP_SCRATCH_W2+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		clc
+		adc  P8ZP_SCRATCH_W2
+		sta  P8ZP_SCRATCH_W1
+		lda  P8ZP_SCRATCH_W1+1
+		adc  P8ZP_SCRATCH_W2+1
+		tay
+		lda  P8ZP_SCRATCH_W1
+		rts
+		rts
+		.pend
+
+mul_byte_10	.proc
+		; A=(A*4 + A)*2
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+		clc
+		adc  P8ZP_SCRATCH_REG
+		asl  a
+		rts
+		.pend
+
+mul_word_10	.proc
+		; AY=(AY*4 + AY)*2
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
+		sta  P8ZP_SCRATCH_W2
+		sty  P8ZP_SCRATCH_W2+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		clc
+		adc  P8ZP_SCRATCH_W2
+		sta  P8ZP_SCRATCH_W1
+		lda  P8ZP_SCRATCH_W1+1
+		adc  P8ZP_SCRATCH_W2+1
+		sta  P8ZP_SCRATCH_W1+1
+		lda  P8ZP_SCRATCH_W1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		ldy  P8ZP_SCRATCH_W1+1
+		rts
+		.pend
+
+mul_byte_11	.proc
+		; A=(A*2 + A)*4 - A
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+		clc
+		adc  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+		sec
+		sbc  P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+; mul_word_11 is skipped (too much code)
+
+mul_byte_12	.proc
+		; A=(A*2 + A)*4
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+		clc
+		adc  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+		rts
+		.pend
+
+mul_word_12	.proc
+		; AY=(AY*2 + AY)*4
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
+		sta  P8ZP_SCRATCH_W2
+		sty  P8ZP_SCRATCH_W2+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		clc
+		adc  P8ZP_SCRATCH_W2
+		sta  P8ZP_SCRATCH_W1
+		lda  P8ZP_SCRATCH_W1+1
+		adc  P8ZP_SCRATCH_W2+1
+		sta  P8ZP_SCRATCH_W1+1
+		lda  P8ZP_SCRATCH_W1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		ldy  P8ZP_SCRATCH_W1+1
+		rts
+		.pend
+
+mul_byte_13	.proc
+		; A=(A*2 + A)*4 + A
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+                clc
+		adc  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+                clc
+		adc  P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+; mul_word_13 is skipped (too much code)
+
+mul_byte_14	.proc
+		; A=(A*8 - A)*2
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+		asl  a
+                sec
+		sbc  P8ZP_SCRATCH_REG
+                asl  a
+		rts
+		.pend
+
+; mul_word_14 is skipped (too much code)
+
+mul_byte_15	.proc
+		; A=A*16 - A
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+		asl  a
+		asl  a
+		sec
+		sbc  P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+mul_word_15	.proc
+		; AY = AY * 16 - AY
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
+		sta  P8ZP_SCRATCH_W2
+		sty  P8ZP_SCRATCH_W2+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		sec
+		sbc  P8ZP_SCRATCH_W2
+		sta  P8ZP_SCRATCH_W1
+		lda  P8ZP_SCRATCH_W1+1
+		sbc  P8ZP_SCRATCH_W2+1
+		tay
+		lda  P8ZP_SCRATCH_W1
+		rts
+		.pend
+
+mul_byte_20	.proc
+		; A=(A*4 + A)*4
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+		clc
+		adc  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+		rts
+		.pend
+
+mul_word_20	.proc
+		; AY = AY * 10 * 2
+		jsr  mul_word_10
+		sty  P8ZP_SCRATCH_REG
+		asl  a
+		rol  P8ZP_SCRATCH_REG
+		ldy  P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+mul_byte_25	.proc
+		; A=(A*2 + A)*8 + A
+		sta  P8ZP_SCRATCH_REG
+		asl  a
+		clc
+		adc  P8ZP_SCRATCH_REG
+		asl  a
+		asl  a
+		asl  a
+		clc
+		adc  P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+mul_word_25	.proc
+		; AY = (AY*2 + AY) *8 + AY
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
+		sta  P8ZP_SCRATCH_W2
+		sty  P8ZP_SCRATCH_W2+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		clc
+		adc  P8ZP_SCRATCH_W2
+		sta  P8ZP_SCRATCH_W1
+		lda  P8ZP_SCRATCH_W1+1
+		adc  P8ZP_SCRATCH_W2+1
+		sta  P8ZP_SCRATCH_W1+1
+		lda  P8ZP_SCRATCH_W1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		clc
+		adc  P8ZP_SCRATCH_W2
+		sta  P8ZP_SCRATCH_W1
+		lda  P8ZP_SCRATCH_W1+1
+		adc  P8ZP_SCRATCH_W2+1
+		tay
+		lda  P8ZP_SCRATCH_W1
+		rts
+		.pend
+
+mul_byte_40	.proc
+		and  #7
+		tay
+		lda  _forties,y
+		rts
+_forties	.byte  0*40, 1*40, 2*40, 3*40, 4*40, 5*40, 6*40, 7*40 & 255
+		.pend
+
+mul_word_40	.proc
+		; AY = (AY*4 + AY)*8
+		sta  P8ZP_SCRATCH_W1
+		sty  P8ZP_SCRATCH_W1+1
+		sta  P8ZP_SCRATCH_W2
+		sty  P8ZP_SCRATCH_W2+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		asl  a
+		rol  P8ZP_SCRATCH_W1+1
+		clc
+		adc  P8ZP_SCRATCH_W2
+		sta  P8ZP_SCRATCH_W1
+		lda  P8ZP_SCRATCH_W1+1
+		adc  P8ZP_SCRATCH_W2+1
+		asl  P8ZP_SCRATCH_W1
+		rol  a
+		asl  P8ZP_SCRATCH_W1
+		rol  a
+		asl  P8ZP_SCRATCH_W1
+		rol  a
+		asl  P8ZP_SCRATCH_W1
+		rol  a
+		tay
+		lda  P8ZP_SCRATCH_W1
+		rts
+		.pend
+
+mul_byte_50	.proc
+		and  #7
+		tay
+		lda  _fifties, y
+		rts
+_fifties	.byte  0*50, 1*50, 2*50, 3*50, 4*50, 5*50, 6*50 & 255, 7*50 & 255
+		.pend
+
+mul_word_50	.proc
+		; AY = AY * 25 * 2
+		jsr  mul_word_25
+		sty  P8ZP_SCRATCH_REG
+		asl  a
+		rol  P8ZP_SCRATCH_REG
+		ldy  P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+mul_byte_80	.proc
+		and  #3
+		tay
+		lda  _eighties, y
+		rts
+_eighties	.byte  0*80, 1*80, 2*80, 3*80
+		.pend
+
+mul_word_80	.proc
+		; AY = AY * 40 * 2
+		jsr  mul_word_40
+		sty  P8ZP_SCRATCH_REG
+		asl  a
+		rol  P8ZP_SCRATCH_REG
+		ldy  P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+mul_byte_100	.proc
+		and  #3
+		tay
+		lda  _hundreds, y
+		rts
+_hundreds	.byte  0*100, 1*100, 2*100, 3*100 & 255
+		.pend
+
+mul_word_100	.proc
+		; AY = AY * 25 * 4
+		jsr  mul_word_25
+		sty  P8ZP_SCRATCH_REG
+		asl  a
+		rol  P8ZP_SCRATCH_REG
+		asl  a
+		rol  P8ZP_SCRATCH_REG
+		ldy  P8ZP_SCRATCH_REG
+		rts
+		.pend
+
+; ----------- end optimized multiplications -----------
+
+
 sign_b		.proc
 		lda  P8ESTACK_LO+1,x
 		beq  _sign_zero

compiler/res/prog8lib/prog8lib.asm

@@ -686,7 +686,7 @@ func_sqrt16	.proc
 		sta  P8ZP_SCRATCH_W2
 		lda  P8ESTACK_HI+1,x
 		sta  P8ZP_SCRATCH_W2+1
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		ldy  #$00    ; r = 0
 		ldx  #$07
 		clc         ; clear bit 16 of m
@@ -721,7 +721,7 @@ _skip1
 _skip2
 		iny         ; r = r or d (d is 1 here)
 _skip3
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		tya
 		sta  P8ESTACK_LO+1,x
 		lda  #0
@@ -1216,7 +1216,7 @@ func_rndw	.proc
 func_memcopy	.proc
 	; note: clobbers A,Y
 		inx
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		lda  P8ESTACK_LO+2,x
 		sta  P8ZP_SCRATCH_W1
 		lda  P8ESTACK_HI+2,x
@@ -1233,7 +1233,7 @@ func_memcopy	.proc
 		iny
 		dex
 		bne  -
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		inx
 		inx
 		rts
@@ -1242,7 +1242,7 @@ func_memcopy	.proc
 func_memset	.proc
 	; note: clobbers A,Y
 		inx
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		lda  P8ESTACK_LO+2,x
 		sta  P8ZP_SCRATCH_W1
 		lda  P8ESTACK_HI+2,x
@@ -1253,7 +1253,7 @@ func_memset	.proc
 		lda  P8ESTACK_LO,x
 		ldx  P8ZP_SCRATCH_B1
 		jsr  memset
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		inx
 		inx
 		rts
@@ -1264,7 +1264,7 @@ func_memsetw	.proc
 		; -- fill memory from (SCRATCH_ZPWORD1) number of words in SCRATCH_ZPWORD2, with word value in AY.
 
 		inx
-		stx  P8ZP_SCRATCH_REG_X
+		stx  P8ZP_SCRATCH_REG
 		lda  P8ESTACK_LO+2,x
 		sta  P8ZP_SCRATCH_W1
 		lda  P8ESTACK_HI+2,x
@@ -1276,7 +1276,7 @@ func_memsetw	.proc
 		lda  P8ESTACK_LO,x
 		ldy  P8ESTACK_HI,x
 		jsr  memsetw
-		ldx  P8ZP_SCRATCH_REG_X
+		ldx  P8ZP_SCRATCH_REG
 		inx
 		inx
 		rts
@@ -1328,9 +1328,9 @@ memset          .proc
 	; -- fill memory from (SCRATCH_ZPWORD1), length XY, with value in A.
 	;    clobbers X, Y
 		stx  P8ZP_SCRATCH_B1
-		sty  P8ZP_SCRATCH_REG
+		sty  _save_reg
 		ldy  #0
-		ldx  P8ZP_SCRATCH_REG
+		ldx  _save_reg
 		beq  _lastpage
 
 _fullpage	sta  (P8ZP_SCRATCH_W1),y
@@ -1347,6 +1347,7 @@ _lastpage	ldy  P8ZP_SCRATCH_B1
 		bne  -
 
 +           	rts
+_save_reg	.byte  0
 		.pend
 
 

compiler/res/version.txt

@@ -1 +1 @@
-4.1
+4.2

compiler/src/prog8/CompilerMain.kt

@@ -4,13 +4,10 @@ import kotlinx.cli.*
 import prog8.ast.base.AstException
 import prog8.compiler.CompilationResult
 import prog8.compiler.compileProgram
+import prog8.compiler.target.C64Target
+import prog8.compiler.target.Cx16Target
 import prog8.compiler.target.CompilationTarget
-import prog8.compiler.target.c64.C64MachineDefinition
-import prog8.compiler.target.c64.Petscii
-import prog8.compiler.target.c64.codegen.AsmGen
-import prog8.compiler.target.cx16.CX16MachineDefinition
 import prog8.parser.ParsingFailedError
-import java.io.IOException
 import java.nio.file.FileSystems
 import java.nio.file.Path
 import java.nio.file.StandardWatchEventKinds
@@ -41,7 +38,8 @@ private fun compileMain(args: Array<String>) {
     val dontWriteAssembly by cli.flagArgument("-noasm", "don't create assembly code")
     val dontOptimize by cli.flagArgument("-noopt", "don't perform any optimizations")
     val watchMode by cli.flagArgument("-watch", "continuous compilation mode (watches for file changes), greatly increases compilation speed")
-    val compilationTarget by cli.flagValueArgument("-target", "compilertarget", "target output of the compiler, currently 'c64' and 'cx16' available", "c64")
+    val compilationTarget by cli.flagValueArgument("-target", "compilertarget",
+            "target output of the compiler, currently '${C64Target.name}' and '${Cx16Target.name}' available", C64Target.name)
     val moduleFiles by cli.positionalArgumentsList("modules", "main module file(s) to compile", minArgs = 1)
 
     try {
@@ -50,39 +48,6 @@ private fun compileMain(args: Array<String>) {
         exitProcess(1)
     }
 
-    when(compilationTarget) {
-        "c64" -> {
-            with(CompilationTarget) {
-                name = "Commodore-64"
-                machine = C64MachineDefinition
-                encodeString = { str, altEncoding ->
-                    if(altEncoding) Petscii.encodeScreencode(str, true) else Petscii.encodePetscii(str, true)
-                }
-                decodeString = { bytes, altEncoding ->
-                    if(altEncoding) Petscii.decodeScreencode(bytes, true) else Petscii.decodePetscii(bytes, true)
-                }
-                asmGenerator = ::AsmGen
-            }
-        }
-        "cx16" -> {
-            with(CompilationTarget) {
-                name = "Commander X16"
-                machine = CX16MachineDefinition
-                encodeString = { str, altEncoding ->
-                    if(altEncoding) Petscii.encodeScreencode(str, true) else Petscii.encodePetscii(str, true)
-                }
-                decodeString = { bytes, altEncoding ->
-                    if(altEncoding) Petscii.decodeScreencode(bytes, true) else Petscii.decodePetscii(bytes, true)
-                }
-                asmGenerator = ::AsmGen
-            }
-        }
-        else -> {
-            System.err.println("invalid compilation target. Available are: c64, cx16")
-            exitProcess(1)
-        }
-    }
-
     val outputPath = pathFrom(outputDir)
     if(!outputPath.toFile().isDirectory) {
         System.err.println("Output path doesn't exist")
@@ -97,7 +62,7 @@ private fun compileMain(args: Array<String>) {
             println("Continuous watch mode active. Main module: $filepath")
 
             try {
-                val compilationResult = compileProgram(filepath, !dontOptimize, !dontWriteAssembly, outputDir=outputPath)
+                val compilationResult = compileProgram(filepath, !dontOptimize, !dontWriteAssembly, compilationTarget, outputPath)
                 println("Imported files (now watching:)")
                 for (importedFile in compilationResult.importedFiles) {
                     print("  ")
@@ -122,7 +87,7 @@ private fun compileMain(args: Array<String>) {
             val filepath = pathFrom(filepathRaw).normalize()
             val compilationResult: CompilationResult
             try {
-                compilationResult = compileProgram(filepath, !dontOptimize, !dontWriteAssembly, outputDir=outputPath)
+                compilationResult = compileProgram(filepath, !dontOptimize, !dontWriteAssembly, compilationTarget, outputPath)
                 if(!compilationResult.success)
                     exitProcess(1)
             } catch (x: ParsingFailedError) {
@@ -135,7 +100,7 @@ private fun compileMain(args: Array<String>) {
                 if (compilationResult.programName.isEmpty())
                     println("\nCan't start emulator because no program was assembled.")
                 else if(startEmulator) {
-                    CompilationTarget.machine.launchEmulator(compilationResult.programName)
+                    CompilationTarget.instance.machine.launchEmulator(compilationResult.programName)
                 }
             }
         }

compiler/src/prog8/ast/AstToSourceCode.kt

@@ -140,7 +140,7 @@ class AstToSourceCode(val output: (text: String) -> Unit, val program: Program):
                             param.second.stack -> "stack"
                             param.second.registerOrPair!=null -> param.second.registerOrPair.toString()
                             param.second.statusflag!=null -> param.second.statusflag.toString()
-                            else -> "?????1"
+                            else -> "?????"
                         }
                 output("${datatypeString(param.first.type)} ${param.first.name} @$reg")
                 if(param.first!==subroutine.parameters.last())

compiler/src/prog8/ast/antlr/Antr2Kotlin.kt

@@ -254,7 +254,7 @@ private fun prog8Parser.Asmsub_declContext.toAst(): AsmsubDecl {
     val clobbers = asmsub_clobbers()?.clobber()?.toAst() ?: emptySet()
     val normalParameters = params.map { SubroutineParameter(it.name, it.type, it.position) }
     val normalReturntypes = returns.map { it.type }
-    val paramRegisters = params.map { RegisterOrStatusflag(it.registerOrPair, it.statusflag, it.stack) }
+    val paramRegisters = params.map { RegisterOrStatusflag(it.registerOrPair, it.statusflag, false) }
     val returnRegisters = returns.map { RegisterOrStatusflag(it.registerOrPair, it.statusflag, it.stack) }
     return AsmsubDecl(name, normalParameters, normalReturntypes, paramRegisters, returnRegisters, clobbers)
 }
@@ -263,7 +263,7 @@ private class AsmSubroutineParameter(name: String,
                                      type: DataType,
                                      val registerOrPair: RegisterOrPair?,
                                      val statusflag: Statusflag?,
-                                     val stack: Boolean,
+                                     // TODO implement: val stack: Boolean,
                                      position: Position) : SubroutineParameter(name, type, position)
 
 private class AsmSubroutineReturn(val type: DataType,
@@ -305,8 +305,7 @@ private fun prog8Parser.Asmsub_paramsContext.toAst(): List<AsmSubroutineParamete
             else -> throw FatalAstException("invalid register or status flag '$name'")
         }
     }
-    AsmSubroutineParameter(vardecl.varname.text, datatype, registerorpair, statusregister,
-            !it.stack?.text.isNullOrEmpty(), toPosition())
+    AsmSubroutineParameter(vardecl.varname.text, datatype, registerorpair, statusregister, toPosition())
 }
 
 private fun prog8Parser.Functioncall_stmtContext.toAst(): Statement {
@@ -472,7 +471,7 @@ private fun prog8Parser.ExpressionContext.toAst() : Expression {
                 litval.charliteral()!=null -> {
                     try {
                         val cc=litval.charliteral()
-                        NumericLiteralValue(DataType.UBYTE, CompilationTarget.encodeString(
+                        NumericLiteralValue(DataType.UBYTE, CompilationTarget.instance.encodeString(
                                 unescape(litval.charliteral().SINGLECHAR().text, litval.toPosition()),
                                 litval.charliteral().ALT_STRING_ENCODING()!=null)[0], litval.toPosition())
                     } catch (ce: CharConversionException) {

compiler/src/prog8/ast/base/Base.kt

@@ -56,8 +56,8 @@ enum class DataType {
         return when(this) {
             in ByteDatatypes -> 1
             in WordDatatypes -> 2
-            FLOAT -> CompilationTarget.machine.FLOAT_MEM_SIZE
-            in PassByReferenceDatatypes -> CompilationTarget.machine.POINTER_MEM_SIZE
+            FLOAT -> CompilationTarget.instance.machine.FLOAT_MEM_SIZE
+            in PassByReferenceDatatypes -> CompilationTarget.instance.machine.POINTER_MEM_SIZE
             else -> -9999999
         }
     }

compiler/src/prog8/ast/expressions/AstExpressions.kt

@@ -671,8 +671,8 @@ class RangeExpr(var from: Expression,
         val toString = to as? StringLiteralValue
         if(fromString!=null && toString!=null ) {
             // string range -> int range over character values
-            fromVal = CompilationTarget.encodeString(fromString.value, fromString.altEncoding)[0].toInt()
-            toVal = CompilationTarget.encodeString(toString.value, fromString.altEncoding)[0].toInt()
+            fromVal = CompilationTarget.instance.encodeString(fromString.value, fromString.altEncoding)[0].toInt()
+            toVal = CompilationTarget.instance.encodeString(toString.value, fromString.altEncoding)[0].toInt()
         } else {
             val fromLv = from as? NumericLiteralValue
             val toLv = to as? NumericLiteralValue

compiler/src/prog8/ast/processing/AstChecker.kt

@@ -7,7 +7,9 @@ import prog8.ast.base.*
 import prog8.ast.expressions.*
 import prog8.ast.statements.*
 import prog8.compiler.CompilationOptions
+import prog8.compiler.target.C64Target
 import prog8.compiler.target.CompilationTarget
+import prog8.compiler.target.Cx16Target
 import prog8.functions.BuiltinFunctions
 import java.io.File
 
@@ -471,14 +473,11 @@ internal class AstChecker(private val program: Program,
     }
 
     override fun visit(decl: VarDecl) {
-        fun err(msg: String, position: Position?=null) {
-            errors.err(msg, position ?: decl.position)
-        }
+        fun err(msg: String, position: Position?=null) = errors.err(msg, position ?: decl.position)
 
         // the initializer value can't refer to the variable itself (recursive definition)
-        if(decl.value?.referencesIdentifiers(decl.name) == true || decl.arraysize?.index?.referencesIdentifiers(decl.name) == true) {
+        if(decl.value?.referencesIdentifiers(decl.name) == true || decl.arraysize?.index?.referencesIdentifiers(decl.name) == true)
             err("recursive var declaration")
-        }
 
         // CONST can only occur on simple types (byte, word, float)
         if(decl.type== VarDeclType.CONST) {
@@ -486,10 +485,12 @@ internal class AstChecker(private val program: Program,
                 err("const modifier can only be used on numeric types (byte, word, float)")
         }
 
-        // FLOATS
-        if(!compilerOptions.floats && decl.datatype in setOf(DataType.FLOAT, DataType.ARRAY_F) && decl.type!= VarDeclType.MEMORY) {
+        // FLOATS enabled?
+        if(!compilerOptions.floats && decl.datatype in setOf(DataType.FLOAT, DataType.ARRAY_F) && decl.type!= VarDeclType.MEMORY)
             err("floating point used, but that is not enabled via options")
-        }
+
+        if(decl.datatype == DataType.FLOAT && (decl.zeropage==ZeropageWish.REQUIRE_ZEROPAGE || decl.zeropage==ZeropageWish.PREFER_ZEROPAGE))
+            errors.warn("floating point values won't be placed in Zeropage due to size constraints", decl.position)
 
         // ARRAY without size specifier MUST have an iterable initializer value
         if(decl.isArray && decl.arraysize==null) {
@@ -558,9 +559,11 @@ internal class AstChecker(private val program: Program,
                         checkValueTypeAndRange(decl.datatype, decl.value as NumericLiteralValue)
                     }
                     else -> {
-                        err("var/const declaration needs a compile-time constant initializer value, or range, instead found: ${decl.value!!.javaClass.simpleName}")
-                        super.visit(decl)
-                        return
+                        if(decl.type==VarDeclType.CONST) {
+                            err("const declaration needs a compile-time constant initializer value, or range")
+                            super.visit(decl)
+                            return
+                        }
                     }
                 }
             }
@@ -709,6 +712,14 @@ internal class AstChecker(private val program: Program,
                 else if(directive.args.map{it.name in setOf("enable_floats", "force_output")}.any { !it })
                     err("invalid option directive argument(s)")
             }
+            "%target" -> {
+                if(directive.parent !is Block && directive.parent !is Module)
+                    err("this directive may only occur in a block or at module level")
+                if(directive.args.size != 1)
+                    err("directive requires one argument")
+                if(directive.args.single().name !in setOf(C64Target.name, Cx16Target.name))
+                    err("invalid compilation target")
+            }
             else -> throw SyntaxError("invalid directive ${directive.directive}", directive.position)
         }
         super.visit(directive)
@@ -1156,7 +1167,7 @@ internal class AstChecker(private val program: Program,
 
                     // check if the floating point values are all within range
                     val doubles = value.value.map {it.constValue(program)?.number!!.toDouble()}.toDoubleArray()
-                    if(doubles.any { it < CompilationTarget.machine.FLOAT_MAX_NEGATIVE || it > CompilationTarget.machine.FLOAT_MAX_POSITIVE })
+                    if(doubles.any { it < CompilationTarget.instance.machine.FLOAT_MAX_NEGATIVE || it > CompilationTarget.instance.machine.FLOAT_MAX_POSITIVE })
                         return err("floating point value overflow")
                     return true
                 }

compiler/src/prog8/ast/processing/AstIdentifiersChecker.kt

@@ -22,7 +22,7 @@ internal class AstIdentifiersChecker(private val program: Program, private val e
     }
 
     override fun visit(block: Block) {
-        if(block.name in CompilationTarget.machine.opcodeNames)
+        if(block.name in CompilationTarget.instance.machine.opcodeNames)
             errors.err("can't use a cpu opcode name as a symbol: '${block.name}'", block.position)
 
         val existing = blocks[block.name]
@@ -40,7 +40,7 @@ internal class AstIdentifiersChecker(private val program: Program, private val e
         if(decl.name in BuiltinFunctions)
             errors.err("builtin function cannot be redefined", decl.position)
 
-        if(decl.name in CompilationTarget.machine.opcodeNames)
+        if(decl.name in CompilationTarget.instance.machine.opcodeNames)
             errors.err("can't use a cpu opcode name as a symbol: '${decl.name}'", decl.position)
 
         if(decl.datatype==DataType.STRUCT) {
@@ -74,7 +74,7 @@ internal class AstIdentifiersChecker(private val program: Program, private val e
     }
 
     override fun visit(subroutine: Subroutine) {
-        if(subroutine.name in CompilationTarget.machine.opcodeNames) {
+        if(subroutine.name in CompilationTarget.instance.machine.opcodeNames) {
             errors.err("can't use a cpu opcode name as a symbol: '${subroutine.name}'", subroutine.position)
         } else if(subroutine.name in BuiltinFunctions) {
             // the builtin functions can't be redefined
@@ -119,7 +119,7 @@ internal class AstIdentifiersChecker(private val program: Program, private val e
     }
 
     override fun visit(label: Label) {
-        if(label.name in CompilationTarget.machine.opcodeNames)
+        if(label.name in CompilationTarget.instance.machine.opcodeNames)
             errors.err("can't use a cpu opcode name as a symbol: '${label.name}'", label.position)
 
         if(label.name in BuiltinFunctions) {

compiler/src/prog8/ast/processing/AstVariousTransforms.kt

@@ -12,6 +12,7 @@ internal class AstVariousTransforms(private val program: Program) : AstWalker()
 
     override fun after(functionCallStatement: FunctionCallStatement, parent: Node): Iterable<IAstModification> {
         if(functionCallStatement.target.nameInSource == listOf("swap")) {
+            // TODO don't replace swap(), let the code generator figure this all out
             // if x and y are both just identifiers, do not rewrite (there should be asm generation for that)
             // otherwise:
             // rewrite swap(x,y) as follows:

compiler/src/prog8/ast/processing/StatementReorderer.kt

@@ -71,23 +71,6 @@ internal class StatementReorderer(val program: Program) : AstWalker() {
         return noModifications
     }
 
-    override fun after(decl: VarDecl, parent: Node): Iterable<IAstModification> {
-        val declValue = decl.value
-        if(declValue!=null && decl.type== VarDeclType.VAR && decl.datatype in NumericDatatypes) {
-            val declConstValue = declValue.constValue(program)
-            if(declConstValue==null) {
-                // move the vardecl (without value) to the scope and replace this with a regular assignment
-                decl.value = null
-                val target = AssignTarget(IdentifierReference(listOf(decl.name), decl.position), null, null, decl.position)
-                val assign = Assignment(target, declValue, decl.position)
-                return listOf(
-                        IAstModification.ReplaceNode(decl, assign, parent),
-                        IAstModification.InsertFirst(decl, decl.definingScope() as Node)
-                )
-            }
-        }
-        return noModifications
-    }
 
     override fun after(whenStatement: WhenStatement, parent: Node): Iterable<IAstModification> {
         val choices = whenStatement.choiceValues(program).sortedBy {

compiler/src/prog8/ast/processing/TypecastsAdder.kt

@@ -18,6 +18,21 @@ class TypecastsAdder(val program: Program, val errors: ErrorReporter) : AstWalke
 
     private val noModifications = emptyList<IAstModification>()
 
+    override fun after(decl: VarDecl, parent: Node): Iterable<IAstModification> {
+        val declValue = decl.value
+        if(decl.type==VarDeclType.VAR && declValue!=null && decl.struct==null) {
+            val valueDt = declValue.inferType(program)
+            if(!valueDt.istype(decl.datatype)) {
+                return listOf(IAstModification.ReplaceNode(
+                        declValue,
+                        TypecastExpression(declValue, decl.datatype, true, declValue.position),
+                        decl
+                ))
+            }
+        }
+        return noModifications
+    }
+
     override fun after(expr: BinaryExpression, parent: Node): Iterable<IAstModification> {
         val leftDt = expr.left.inferType(program)
         val rightDt = expr.right.inferType(program)
@@ -139,6 +154,7 @@ class TypecastsAdder(val program: Program, val errors: ErrorReporter) : AstWalke
                                         call.args[arg.second.index],
                                         TypecastExpression(arg.second.value, possibleType, true, arg.second.value.position),
                                         call as Node)
+                                break
                             }
                         }
                     }

compiler/src/prog8/compiler/BeforeAsmGenerationAstChanger.kt

@@ -1,5 +1,6 @@
 package prog8.compiler
 
+import prog8.ast.IFunctionCall
 import prog8.ast.Node
 import prog8.ast.Program
 import prog8.ast.base.*
@@ -14,7 +15,7 @@ internal class BeforeAsmGenerationAstChanger(val program: Program, val errors: E
     private val noModifications = emptyList<IAstModification>()
 
     override fun after(decl: VarDecl, parent: Node): Iterable<IAstModification> {
-        if (decl.value == null && decl.type == VarDeclType.VAR && decl.datatype in NumericDatatypes) {
+        if (decl.value == null && !decl.autogeneratedDontRemove && decl.type == VarDeclType.VAR && decl.datatype in NumericDatatypes) {
             // a numeric vardecl without an initial value is initialized with zero.
             decl.value = decl.zeroElementValue()
         }
@@ -54,7 +55,8 @@ internal class BeforeAsmGenerationAstChanger(val program: Program, val errors: E
                 }
             }
             if (!conflicts) {
-                val numericVarsWithValue = decls.filter { it.value != null && it.datatype in NumericDatatypes }
+                // move vardecls of the scope into the upper scope. Make sure the order remains the same!
+                val numericVarsWithValue = decls.filter { it.value != null && it.datatype in NumericDatatypes }.reversed()
                 return numericVarsWithValue.map {
                     val initValue = it.value!!  // assume here that value has always been set by now
                     it.value = null     // make sure no value init assignment for this vardecl will be created later (would be superfluous)
@@ -114,6 +116,18 @@ internal class BeforeAsmGenerationAstChanger(val program: Program, val errors: E
         // that the types of assignment values and their target are the same,
         // and that the types of both operands of a binaryexpression node are the same.
         // So, it is not easily possible to remove the typecasts that are there to make these conditions true.
+        // The only place for now where we can do this is for:
+        //    asmsub register pair parameter.
+
+        if(typecast.type in WordDatatypes) {
+            val fcall = typecast.parent as? IFunctionCall
+            if (fcall != null) {
+                val sub = fcall.target.targetStatement(program.namespace) as? Subroutine
+                if (sub != null && sub.isAsmSubroutine) {
+                    return listOf(IAstModification.ReplaceNode(typecast, typecast.expression, parent))
+                }
+            }
+        }
 
         if(sourceDt in PassByReferenceDatatypes) {
             if(typecast.type==DataType.UWORD) {

compiler/src/prog8/compiler/Compiler.kt

@@ -27,7 +27,8 @@ data class CompilationOptions(val output: OutputType,
                               val launcher: LauncherType,
                               val zeropage: ZeropageType,
                               val zpReserved: List<IntRange>,
-                              val floats: Boolean)
+                              val floats: Boolean,
+                              val compilationTarget: String?)
 
 
 class CompilerException(message: String?) : Exception(message)

compiler/src/prog8/compiler/Main.kt

@@ -4,7 +4,9 @@ import prog8.ast.AstToSourceCode
 import prog8.ast.Program
 import prog8.ast.base.*
 import prog8.ast.statements.Directive
+import prog8.compiler.target.C64Target
 import prog8.compiler.target.CompilationTarget
+import prog8.compiler.target.Cx16Target
 import prog8.optimizer.UnusedCodeRemover
 import prog8.optimizer.constantFold
 import prog8.optimizer.optimizeStatements
@@ -13,6 +15,7 @@ import prog8.parser.ModuleImporter
 import prog8.parser.ParsingFailedError
 import prog8.parser.moduleName
 import java.nio.file.Path
+import kotlin.system.exitProcess
 import kotlin.system.measureTimeMillis
 
 
@@ -25,12 +28,22 @@ class CompilationResult(val success: Boolean,
 fun compileProgram(filepath: Path,
                    optimize: Boolean,
                    writeAssembly: Boolean,
+                   compilationTarget: String,
                    outputDir: Path): CompilationResult {
     var programName = ""
     lateinit var programAst: Program
     lateinit var importedFiles: List<Path>
     val errors = ErrorReporter()
 
+    when(compilationTarget) {
+        C64Target.name -> CompilationTarget.instance = C64Target
+        Cx16Target.name -> CompilationTarget.instance = Cx16Target
+        else -> {
+            System.err.println("invalid compilation target")
+            exitProcess(1)
+        }
+    }
+
     try {
         val totalTime = measureTimeMillis {
             // import main module and everything it needs
@@ -90,8 +103,8 @@ private fun parseImports(filepath: Path, errors: ErrorReporter): Triple<Program,
     if (compilerOptions.launcher == LauncherType.BASIC && compilerOptions.output != OutputType.PRG)
         throw ParsingFailedError("${programAst.modules.first().position} BASIC launcher requires output type PRG.")
 
-    // depending on the mach9ine and compiler options we may have to include some libraries
-    CompilationTarget.machine.importLibs(compilerOptions, importer, programAst)
+    // depending on the machine and compiler options we may have to include some libraries
+    CompilationTarget.instance.machine.importLibs(compilerOptions, importer, programAst)
 
     // always import prog8lib and math
     importer.importLibraryModule(programAst, "math")
@@ -129,16 +142,26 @@ private fun determineCompilationOptions(program: Program): CompilationOptions {
             .map { it[0].int!!..it[1].int!! }
             .toList()
 
+    var target = (mainModule.statements.singleOrNull { it is Directive && it.directive == "%target" }
+            as? Directive)?.args?.single()?.name
+
+    when(target) {
+        C64Target.name -> CompilationTarget.instance = C64Target
+        Cx16Target.name -> CompilationTarget.instance = Cx16Target
+        null -> target = CompilationTarget.instance.name
+        else -> throw FatalAstException("invalid target")
+    }
+
     return CompilationOptions(
             if (outputType == null) OutputType.PRG else OutputType.valueOf(outputType),
             if (launcherType == null) LauncherType.BASIC else LauncherType.valueOf(launcherType),
-            zpType, zpReserved, floatsEnabled
+            zpType, zpReserved, floatsEnabled, target
     )
 }
 
 private fun processAst(programAst: Program, errors: ErrorReporter, compilerOptions: CompilationOptions) {
     // perform initial syntax checks and processings
-    println("Processing...")
+    println("Processing for target ${CompilationTarget.instance.name}...")
     programAst.checkIdentifiers(errors)
     errors.handle()
     programAst.constantFold(errors)
@@ -191,11 +214,14 @@ private fun writeAssembly(programAst: Program, errors: ErrorReporter, outputDir:
 
     // printAst(programAst)
 
-    CompilationTarget.machine.initializeZeropage(compilerOptions)
-    val assembly = CompilationTarget.asmGenerator(
+    if(compilerOptions.compilationTarget!=null && compilerOptions.compilationTarget != CompilationTarget.instance.name)
+        throw AssemblyError("program's compilation target differs from configured target")
+
+    CompilationTarget.instance.machine.initializeZeropage(compilerOptions)
+    val assembly = CompilationTarget.instance.asmGenerator(
             programAst,
             errors,
-            CompilationTarget.machine.zeropage,
+            CompilationTarget.instance.machine.zeropage,
             compilerOptions,
             outputDir).compileToAssembly(optimize)
     assembly.assemble(compilerOptions)

compiler/src/prog8/compiler/Zeropage.kt

@@ -10,7 +10,6 @@ abstract class Zeropage(protected val options: CompilationOptions) {
 
     abstract val SCRATCH_B1 : Int      // temp storage for a single byte
     abstract val SCRATCH_REG : Int     // temp storage for a register
-    abstract val SCRATCH_REG_X : Int   // temp storage for register X (the evaluation stack pointer)
     abstract val SCRATCH_W1 : Int      // temp storage 1 for a word  $fb+$fc
     abstract val SCRATCH_W2 : Int      // temp storage 2 for a word  $fb+$fc
 

compiler/src/prog8/compiler/target/CompilationTarget.kt

@@ -4,15 +4,50 @@ import prog8.ast.Program
 import prog8.ast.base.ErrorReporter
 import prog8.compiler.CompilationOptions
 import prog8.compiler.Zeropage
+import prog8.compiler.target.c64.C64MachineDefinition
+import prog8.compiler.target.c64.Petscii
+import prog8.compiler.target.c64.codegen.AsmGen
+import prog8.compiler.target.cx16.CX16MachineDefinition
 import java.nio.file.Path
 
 
 internal interface CompilationTarget {
+    val name: String
+    val machine: IMachineDefinition
+    fun encodeString(str: String, altEncoding: Boolean): List<Short>
+    fun decodeString(bytes: List<Short>, altEncoding: Boolean): String
+    fun asmGenerator(program: Program, errors: ErrorReporter, zp: Zeropage, options: CompilationOptions, path: Path): IAssemblyGenerator
+    val asmForSystemReset: String
+    val initProcName: String?
+
     companion object {
-        lateinit var name: String
-        lateinit var machine: IMachineDefinition
-        lateinit var encodeString: (str: String, altEncoding: Boolean) -> List<Short>
-        lateinit var decodeString: (bytes: List<Short>, altEncoding: Boolean) -> String
-        lateinit var asmGenerator: (Program, ErrorReporter, Zeropage, CompilationOptions, Path) -> IAssemblyGenerator
+        lateinit var instance: CompilationTarget
     }
 }
+
+
+internal object C64Target: CompilationTarget {
+    override val name = "c64"
+    override val machine = C64MachineDefinition
+    override fun encodeString(str: String, altEncoding: Boolean) =
+            if(altEncoding) Petscii.encodeScreencode(str, true) else Petscii.encodePetscii(str, true)
+    override fun decodeString(bytes: List<Short>, altEncoding: Boolean) =
+            if(altEncoding) Petscii.decodeScreencode(bytes, true) else Petscii.decodePetscii(bytes, true)
+    override fun asmGenerator(program: Program, errors: ErrorReporter, zp: Zeropage, options: CompilationOptions, path: Path) =
+            AsmGen(program, errors, zp, options, path)
+    override val asmForSystemReset = "  sei |  lda  #14 |  sta  $1 |  jmp  (c64.RESET_VEC)"
+    override val initProcName = "c64.init_system"
+}
+
+internal object Cx16Target: CompilationTarget {
+    override val name = "cx16"
+    override val machine = CX16MachineDefinition
+    override fun encodeString(str: String, altEncoding: Boolean) =
+            if(altEncoding) Petscii.encodeScreencode(str, true) else Petscii.encodePetscii(str, true)
+    override fun decodeString(bytes: List<Short>, altEncoding: Boolean) =
+            if(altEncoding) Petscii.decodeScreencode(bytes, true) else Petscii.decodePetscii(bytes, true)
+    override fun asmGenerator(program: Program, errors: ErrorReporter, zp: Zeropage, options: CompilationOptions, path: Path) =
+            AsmGen(program, errors, zp, options, path)
+    override val asmForSystemReset = "  sei |  stz  cx16.d1prb |  jmp  (cx16.RESET_VEC)"
+    override val initProcName = "cx16.init_system"
+}

compiler/src/prog8/compiler/target/IMachineDefinition.kt

@@ -28,7 +28,6 @@ internal interface IMachineDefinition {
 
     val opcodeNames: Set<String>
     var zeropage: Zeropage
-    val initSystemProcname: String
     val cpu: CpuType
 
     fun initializeZeropage(compilerOptions: CompilationOptions)

compiler/src/prog8/compiler/target/c64/AssemblyProgram.kt

@@ -2,6 +2,7 @@ package prog8.compiler.target.c64
 
 import prog8.compiler.CompilationOptions
 import prog8.compiler.OutputType
+import prog8.compiler.target.CompilationTarget
 import prog8.compiler.target.IAssemblyProgram
 import prog8.compiler.target.generatedLabelPrefix
 import java.nio.file.Path
@@ -22,12 +23,12 @@ class AssemblyProgram(override val name: String, outputDir: Path) : IAssemblyPro
         val outFile = when (options.output) {
             OutputType.PRG -> {
                 command.add("--cbm-prg")
-                println("\nCreating prg.")
+                println("\nCreating prg for target ${CompilationTarget.instance.name}.")
                 prgFile
             }
             OutputType.RAW -> {
                 command.add("--nostart")
-                println("\nCreating raw binary.")
+                println("\nCreating raw binary for target ${CompilationTarget.instance.name}.")
                 binFile
             }
         }

compiler/src/prog8/compiler/target/c64/C64MachineDefinition.kt

@@ -29,7 +29,6 @@ internal object C64MachineDefinition: IMachineDefinition {
     override val ESTACK_HI = 0xcf00        //  $ce00-$ceff inclusive
 
     override lateinit var zeropage: Zeropage
-    override val initSystemProcname = "c64.init_system"
 
     override fun getFloat(num: Number) = Mflpt5.fromNumber(num)
 
@@ -109,7 +108,6 @@ internal object C64MachineDefinition: IMachineDefinition {
 
         override val SCRATCH_B1 = 0x02      // temp storage for a single byte
         override val SCRATCH_REG = 0x03     // temp storage for a register
-        override val SCRATCH_REG_X = 0xfa   // temp storage for register X (the evaluation stack pointer)
         override val SCRATCH_W1 = 0xfb      // temp storage 1 for a word  $fb+$fc
         override val SCRATCH_W2 = 0xfd      // temp storage 2 for a word  $fb+$fc
 
@@ -127,7 +125,7 @@ internal object C64MachineDefinition: IMachineDefinition {
             if (options.zeropage == ZeropageType.FULL) {
                 free.addAll(0x04..0xf9)
                 free.add(0xff)
-                free.removeAll(listOf(SCRATCH_B1, SCRATCH_REG, SCRATCH_REG_X, SCRATCH_W1, SCRATCH_W1 + 1, SCRATCH_W2, SCRATCH_W2 + 1))
+                free.removeAll(listOf(SCRATCH_B1, SCRATCH_REG, SCRATCH_W1, SCRATCH_W1 + 1, SCRATCH_W2, SCRATCH_W2 + 1))
                 free.removeAll(listOf(0xa0, 0xa1, 0xa2, 0x91, 0xc0, 0xc5, 0xcb, 0xf5, 0xf6))        // these are updated by IRQ
             } else {
                 if (options.zeropage == ZeropageType.KERNALSAFE || options.zeropage == ZeropageType.FLOATSAFE) {
@@ -170,7 +168,6 @@ internal object C64MachineDefinition: IMachineDefinition {
             }
             require(SCRATCH_B1 !in free)
             require(SCRATCH_REG !in free)
-            require(SCRATCH_REG_X !in free)
             require(SCRATCH_W1 !in free)
             require(SCRATCH_W2 !in free)
 

compiler/src/prog8/compiler/target/c64/codegen/AsmGen.kt

@@ -35,6 +35,10 @@ internal class AsmGen(private val program: Program,
                       val options: CompilationOptions,
                       private val outputDir: Path): IAssemblyGenerator {
 
+    // for expressions and augmented assignments:
+    val optimizedByteMultiplications = setOf(3,5,6,7,9,10,11,12,13,14,15,20,25,40,50,80,100)
+    val optimizedWordMultiplications = setOf(3,5,6,7,9,10,12,15,20,25,40,50,80,100)
+
     private val assemblyLines = mutableListOf<String>()
     private val globalFloatConsts = mutableMapOf<Double, String>()     // all float values in the entire program (value -> varname)
     private val allocatedZeropageVariables = mutableMapOf<String, Pair<Int, DataType>>()
@@ -79,7 +83,7 @@ internal class AsmGen(private val program: Program,
 
     private fun header() {
         val ourName = this.javaClass.name
-        val cpu = when(CompilationTarget.machine.cpu) {
+        val cpu = when(CompilationTarget.instance.machine.cpu) {
             CpuType.CPU6502 -> "6502"
             CpuType.CPU65c02 -> "65c02"
             else -> "unsupported"
@@ -94,18 +98,16 @@ internal class AsmGen(private val program: Program,
         program.actualLoadAddress = program.definedLoadAddress
         if (program.actualLoadAddress == 0)   // fix load address
             program.actualLoadAddress = if (options.launcher == LauncherType.BASIC)
-                CompilationTarget.machine.BASIC_LOAD_ADDRESS else CompilationTarget.machine.RAW_LOAD_ADDRESS
+                CompilationTarget.instance.machine.BASIC_LOAD_ADDRESS else CompilationTarget.instance.machine.RAW_LOAD_ADDRESS
 
         // the global prog8 variables needed
-        val zp = CompilationTarget.machine.zeropage
-        val initproc = CompilationTarget.machine.initSystemProcname
+        val zp = CompilationTarget.instance.machine.zeropage
         out("P8ZP_SCRATCH_B1 = ${zp.SCRATCH_B1}")
         out("P8ZP_SCRATCH_REG = ${zp.SCRATCH_REG}")
-        out("P8ZP_SCRATCH_REG_X = ${zp.SCRATCH_REG_X}")
         out("P8ZP_SCRATCH_W1 = ${zp.SCRATCH_W1}    ; word")
         out("P8ZP_SCRATCH_W2 = ${zp.SCRATCH_W2}    ; word")
-        out("P8ESTACK_LO = ${CompilationTarget.machine.ESTACK_LO.toHex()}")
-        out("P8ESTACK_HI = ${CompilationTarget.machine.ESTACK_HI.toHex()}")
+        out("P8ESTACK_LO = ${CompilationTarget.instance.machine.ESTACK_LO.toHex()}")
+        out("P8ESTACK_HI = ${CompilationTarget.instance.machine.ESTACK_HI.toHex()}")
 
         when {
             options.launcher == LauncherType.BASIC -> {
@@ -120,16 +122,16 @@ internal class AsmGen(private val program: Program,
                 out("_prog8_entrypoint\t; assembly code starts here\n")
                 out("  tsx")
                 out("  stx  prog8_lib.orig_stackpointer")
-                if(!initproc.isNullOrEmpty())
-                    out("  jsr  $initproc")
+                if(!CompilationTarget.instance.initProcName.isNullOrEmpty())
+                    out("  jsr  ${CompilationTarget.instance.initProcName}")
             }
             options.output == OutputType.PRG -> {
                 out("; ---- program without basic sys call ----")
                 out("* = ${program.actualLoadAddress.toHex()}\n")
                 out("  tsx")
                 out("  stx  prog8_lib.orig_stackpointer")
-                if(!initproc.isNullOrEmpty())
-                    out("  jsr  $initproc")
+                if(!CompilationTarget.instance.initProcName.isNullOrEmpty())
+                    out("  jsr  ${CompilationTarget.instance.initProcName}")
             }
             options.output == OutputType.RAW -> {
                 out("; ---- raw assembler program ----")
@@ -161,7 +163,7 @@ internal class AsmGen(private val program: Program,
             }
             Zeropage.ExitProgramStrategy.SYSTEM_RESET -> {
                 out("  jsr  main.start\t; call program entrypoint")
-                out("  jmp  (c64.RESET_VEC)\t; cold reset")
+                out(CompilationTarget.instance.asmForSystemReset)
             }
         }
     }
@@ -170,7 +172,7 @@ internal class AsmGen(private val program: Program,
         // the global list of all floating point constants for the whole program
         out("; global float constants")
         for (flt in globalFloatConsts) {
-            val floatFill = CompilationTarget.machine.getFloat(flt.key).makeFloatFillAsm()
+            val floatFill = CompilationTarget.instance.machine.getFloat(flt.key).makeFloatFillAsm()
             val floatvalue = flt.key
             out("${flt.value}\t.byte  $floatFill  ; float $floatvalue")
         }
@@ -340,7 +342,7 @@ internal class AsmGen(private val program: Program,
                         }
                 val floatFills = array.map {
                     val number = (it as NumericLiteralValue).number
-                    CompilationTarget.machine.getFloat(number).makeFloatFillAsm()
+                    CompilationTarget.instance.machine.getFloat(number).makeFloatFillAsm()
                 }
                 out(name)
                 for (f in array.zip(floatFills))
@@ -474,7 +476,7 @@ internal class AsmGen(private val program: Program,
     }
 
     internal fun getFloatAsmConst(number: Double): String {
-        var asmName = CompilationTarget.machine.getFloatRomConst(number)
+        var asmName = CompilationTarget.instance.machine.getFloatRomConst(number)
         if(asmName.isNullOrEmpty()) {
             // no ROM float const for this value, create our own
             asmName = globalFloatConsts[number]
@@ -548,16 +550,35 @@ internal class AsmGen(private val program: Program,
 
     internal fun fixNameSymbols(name: String) = name.replace("<", "prog8_").replace(">", "")     // take care of the autogenerated invalid (anon) label names
 
+
+    private val saveRegisterLabels = Stack<String>();
+
     internal fun saveRegister(register: CpuRegister) {
         when(register) {
             CpuRegister.A -> out("  pha")
             CpuRegister.X -> {
-                if(CompilationTarget.machine.cpu == CpuType.CPU65c02)
-                    out("  phx")
-                else
-                    out("  stx  P8ZP_SCRATCH_REG_X")
+                if (CompilationTarget.instance.machine.cpu == CpuType.CPU65c02) out("  phx")
+                else {
+                    val save = makeLabel("saveX")
+                    saveRegisterLabels.push(save)
+                    out("""
+            stx  $save
+            jmp  +
+$save       .byte 0
++""")
+                }
+            }
+            CpuRegister.Y -> {
+                if (CompilationTarget.instance.machine.cpu == CpuType.CPU65c02) out("  phy")
+                else {
+                    val save = makeLabel("saveY")
+                    out("""
+            sty  $save
+            jmp  +
+$save       .byte 0
++""")
+                }
             }
-            CpuRegister.Y -> out("  tya | pha")
         }
     }
 
@@ -565,15 +586,23 @@ internal class AsmGen(private val program: Program,
         when(register) {
             CpuRegister.A -> out("  pla")
             CpuRegister.X -> {
-                if(CompilationTarget.machine.cpu == CpuType.CPU65c02)
-                    out("  plx")
-                else
-                    out("  ldx  P8ZP_SCRATCH_REG_X")
+                if (CompilationTarget.instance.machine.cpu == CpuType.CPU65c02) out("  plx")
+                else {
+                    val save = saveRegisterLabels.pop()
+                    out("  ldx  $save")
+                }
+            }
+            CpuRegister.Y -> {
+                if (CompilationTarget.instance.machine.cpu == CpuType.CPU65c02) out("  ply")
+                else {
+                    val save = saveRegisterLabels.pop()
+                    out("  ldy  $save")
+                }
             }
-            CpuRegister.Y -> out("  pla | tay")
         }
     }
 
+
     internal fun translate(stmt: Statement) {
         outputSourceLine(stmt)
         when(stmt) {
@@ -684,7 +713,7 @@ internal class AsmGen(private val program: Program,
 +""")
                 when(register) {
                     CpuRegister.A -> out("  inx |  lda  P8ESTACK_LO,x")
-                    CpuRegister.X -> throw AssemblyError("can't use X here")
+                    CpuRegister.X -> out("  inx |  lda  P8ESTACK_LO,x |  tax")
                     CpuRegister.Y -> out("  inx |  ldy  P8ESTACK_LO,x")
                 }
             }
@@ -722,7 +751,7 @@ internal class AsmGen(private val program: Program,
                 expressionsAsmGen.translateExpression(index)
                 when(register) {
                     CpuRegister.A -> out("  inx |  lda  P8ESTACK_LO,x")
-                    CpuRegister.X -> throw AssemblyError("can't use X here")
+                    CpuRegister.X -> out("  inx |  lda  P8ESTACK_LO,x |  tax")
                     CpuRegister.Y -> out("  inx |  ldy  P8ESTACK_LO,x")
                 }
             }
@@ -792,16 +821,38 @@ internal class AsmGen(private val program: Program,
             }
 
     private fun translate(stmt: IfStatement) {
-        expressionsAsmGen.translateExpression(stmt.condition)
-        translateTestStack(stmt.condition.inferType(program).typeOrElse(DataType.STRUCT))
-        val elseLabel = makeLabel("if_else")
-        val endLabel = makeLabel("if_end")
-        out("  beq  $elseLabel")
-        translate(stmt.truepart)
-        out("  jmp  $endLabel")
-        out(elseLabel)
-        translate(stmt.elsepart)
-        out(endLabel)
+        when {
+            stmt.elsepart.containsNoCodeNorVars() -> {
+                // empty else
+                expressionsAsmGen.translateExpression(stmt.condition)
+                translateTestStack(stmt.condition.inferType(program).typeOrElse(DataType.STRUCT))
+                val endLabel = makeLabel("if_end")
+                out("  beq  $endLabel")
+                translate(stmt.truepart)
+                out(endLabel)
+            }
+            stmt.truepart.containsNoCodeNorVars() -> {
+                // empty true part
+                expressionsAsmGen.translateExpression(stmt.condition)
+                translateTestStack(stmt.condition.inferType(program).typeOrElse(DataType.STRUCT))
+                val endLabel = makeLabel("if_end")
+                out("  bne  $endLabel")
+                translate(stmt.elsepart)
+                out(endLabel)
+            }
+            else -> {
+                expressionsAsmGen.translateExpression(stmt.condition)
+                translateTestStack(stmt.condition.inferType(program).typeOrElse(DataType.STRUCT))
+                val elseLabel = makeLabel("if_else")
+                val endLabel = makeLabel("if_end")
+                out("  beq  $elseLabel")
+                translate(stmt.truepart)
+                out("  jmp  $endLabel")
+                out(elseLabel)
+                translate(stmt.elsepart)
+                out(endLabel)
+            }
+        }
     }
 
     private fun translateTestStack(dataType: DataType) {

compiler/src/prog8/compiler/target/c64/codegen/ExpressionsAsmGen.kt

@@ -47,7 +47,7 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                             RegisterOrPair.AY -> asmgen.out("  sta  P8ESTACK_LO,x |  tya |  sta  P8ESTACK_HI,x |  dex")
                             RegisterOrPair.X -> {
                                 // return value in X register has been discarded, just push a zero
-                                if(CompilationTarget.machine.cpu==CpuType.CPU65c02)
+                                if(CompilationTarget.instance.machine.cpu==CpuType.CPU65c02)
                                     asmgen.out("  stz  P8ESTACK_LO,x")
                                 else
                                     asmgen.out("  lda  #0 |  sta  P8ESTACK_LO,x")
@@ -56,7 +56,7 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                             RegisterOrPair.AX -> {
                                 // return value in X register has been discarded, just push a zero in this place
                                 asmgen.out("  sta  P8ESTACK_LO,x")
-                                if(CompilationTarget.machine.cpu==CpuType.CPU65c02)
+                                if(CompilationTarget.instance.machine.cpu==CpuType.CPU65c02)
                                     asmgen.out("  stz  P8ESTACK_HI,x")
                                 else
                                     asmgen.out("  lda  #0 |  sta  P8ESTACK_HI,x")
@@ -64,7 +64,7 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                             }
                             RegisterOrPair.XY -> {
                                 // return value in X register has been discarded, just push a zero in this place
-                                if(CompilationTarget.machine.cpu==CpuType.CPU65c02)
+                                if(CompilationTarget.instance.machine.cpu==CpuType.CPU65c02)
                                     asmgen.out("  stz  P8ESTACK_LO,x")
                                 else
                                     asmgen.out("  lda  #0 |  sta  P8ESTACK_LO,x")
@@ -85,7 +85,7 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                 when(expr.type) {
                     DataType.UBYTE, DataType.BYTE -> {}
                     DataType.UWORD, DataType.WORD -> {
-                        if(CompilationTarget.machine.cpu==CpuType.CPU65c02)
+                        if(CompilationTarget.instance.machine.cpu==CpuType.CPU65c02)
                             asmgen.out("  stz  P8ESTACK_HI+1,x")
                         else
                             asmgen.out("  lda  #0  |  sta  P8ESTACK_HI+1,x")
@@ -103,14 +103,9 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                         asmgen.out(""" 
                             lda  P8ESTACK_LO+1,x
                             ora  #$7f
-                            bmi  +""")
-                        if(CompilationTarget.machine.cpu==CpuType.CPU65c02)
-                            asmgen.out("""
-+                               stz  P8ESTACK_HI+1,x""")
-                        else
-                            asmgen.out("""
-                                lda  #0
-+                               sta  P8ESTACK_HI+1,x""")
+                            bmi  +
+                            lda  #0
++                           sta  P8ESTACK_HI+1,x""")
                     }
                     DataType.FLOAT -> asmgen.out(" jsr  c64flt.stack_b2float")
                     in PassByReferenceDatatypes -> throw AssemblyError("cannot cast to a pass-by-reference datatype")
@@ -212,9 +207,6 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
         }
     }
 
-    private val optimizedByteMultiplications = setOf(3,5,6,7,9,10,11,12,13,14,15,20,25,40)
-    private val optimizedWordMultiplications = setOf(3,5,6,7,9,10,12,15,20,25,40)
-
     private fun translateExpression(expr: BinaryExpression) {
         val leftIDt = expr.left.inferType(program)
         val rightIDt = expr.right.inferType(program)
@@ -308,40 +300,40 @@ internal class ExpressionsAsmGen(private val program: Program, private val asmge
                         val amount = value.number.toInt()
                         when(rightDt) {
                             DataType.UBYTE -> {
-                                if(amount in optimizedByteMultiplications) {
+                                if(amount in asmgen.optimizedByteMultiplications) {
                                     translateExpression(expr.left)
-                                    asmgen.out(" jsr  math.mul_byte_$amount")
+                                    asmgen.out(" jsr  math.stack_mul_byte_$amount")
                                     return
                                 }
                             }
                             DataType.BYTE -> {
-                                if(amount in optimizedByteMultiplications) {
+                                if(amount in asmgen.optimizedByteMultiplications) {
                                     translateExpression(expr.left)
-                                    asmgen.out(" jsr  math.mul_byte_$amount")
+                                    asmgen.out(" jsr  math.stack_mul_byte_$amount")
                                     return
                                 }
-                                if(amount.absoluteValue in optimizedByteMultiplications) {
+                                if(amount.absoluteValue in asmgen.optimizedByteMultiplications) {
                                     translateExpression(expr.left)
-                                    asmgen.out(" jsr  prog8_lib.neg_b |  jsr  math.mul_byte_${amount.absoluteValue}")
+                                    asmgen.out(" jsr  prog8_lib.neg_b |  jsr  math.stack_mul_byte_${amount.absoluteValue}")
                                     return
                                 }
                             }
                             DataType.UWORD -> {
-                                if(amount in optimizedWordMultiplications) {
+                                if(amount in asmgen.optimizedWordMultiplications) {
                                     translateExpression(expr.left)
-                                    asmgen.out(" jsr  math.mul_word_$amount")
+                                    asmgen.out(" jsr  math.stack_mul_word_$amount")
                                     return
                                 }
                             }
                             DataType.WORD -> {
-                                if(amount in optimizedWordMultiplications) {
+                                if(amount in asmgen.optimizedWordMultiplications) {
                                     translateExpression(expr.left)
-                                    asmgen.out(" jsr  math.mul_word_$amount")
+                                    asmgen.out(" jsr  math.stack_mul_word_$amount")
                                     return
                                 }
-                                if(amount.absoluteValue in optimizedWordMultiplications) {
+                                if(amount.absoluteValue in asmgen.optimizedWordMultiplications) {
                                     translateExpression(expr.left)
-                                    asmgen.out(" jsr  prog8_lib.neg_w |  jsr  math.mul_word_${amount.absoluteValue}")
+                                    asmgen.out(" jsr  prog8_lib.neg_w |  jsr  math.stack_mul_word_${amount.absoluteValue}")
                                     return
                                 }
                             }

compiler/src/prog8/compiler/target/c64/codegen/FunctionCallAsmGen.kt

@@ -4,6 +4,7 @@ import prog8.ast.IFunctionCall
 import prog8.ast.Program
 import prog8.ast.base.*
 import prog8.ast.expressions.*
+import prog8.ast.statements.RegisterOrStatusflag
 import prog8.ast.statements.Subroutine
 import prog8.ast.statements.SubroutineParameter
 import prog8.compiler.AssemblyError
@@ -18,7 +19,7 @@ internal class FunctionCallAsmGen(private val program: Program, private val asmg
         val sub = stmt.target.targetSubroutine(program.namespace) ?: throw AssemblyError("undefined subroutine ${stmt.target}")
         val saveX = CpuRegister.X in sub.asmClobbers || sub.regXasResult() || sub.regXasParam()
         if(saveX)
-            asmgen.saveRegister(CpuRegister.X)  // we only save X for now (required! is the eval stack pointer), screw A and Y...
+            asmgen.saveRegister(CpuRegister.X)
 
         val subName = asmgen.asmSymbolName(stmt.target)
         if(stmt.args.isNotEmpty()) {
@@ -63,33 +64,79 @@ internal class FunctionCallAsmGen(private val program: Program, private val asmg
     private fun registerArgsViaStackEvaluation(stmt: IFunctionCall, sub: Subroutine) {
         // this is called when one or more of the arguments are 'complex' and
         // cannot be assigned to a register easily or risk clobbering other registers.
+
+        if(sub.parameters.isEmpty())
+            return
+
+        // 1. load all arguments reversed onto the stack: first arg goes last (is on top).
         for (arg in stmt.args.reversed())
             asmgen.translateExpression(arg)
-        for (regparam in sub.asmParameterRegisters) {
+
+        var argForCarry: IndexedValue<Pair<Expression, RegisterOrStatusflag>>? = null
+        var argForXregister: IndexedValue<Pair<Expression, RegisterOrStatusflag>>? = null
+        var argForAregister: IndexedValue<Pair<Expression, RegisterOrStatusflag>>? = null
+
+        asmgen.out("  inx")     // align estack pointer
+
+        for(argi in stmt.args.zip(sub.asmParameterRegisters).withIndex()) {
             when {
-                regparam.statusflag==Statusflag.Pc -> {
-                    asmgen.out("""
-                        inx
-                        pha
-                        lda  P8ESTACK_LO,x
-                        beq  +
-                        sec  
-                        bcs  ++
-+                       clc
-+                       pla""")
+                argi.value.second.stack -> TODO("asmsub @stack parameter")
+                argi.value.second.statusflag == Statusflag.Pc -> {
+                    require(argForCarry == null)
+                    argForCarry = argi
                 }
-                regparam.statusflag!=null -> {
-                    throw AssemblyError("can only use Carry as status flag parameter")
+                argi.value.second.statusflag != null -> throw AssemblyError("can only use Carry as status flag parameter")
+                argi.value.second.registerOrPair in setOf(RegisterOrPair.X, RegisterOrPair.AX, RegisterOrPair.XY) -> {
+                    require(argForXregister==null)
+                    argForXregister = argi
                 }
-                regparam.registerOrPair!=null -> {
-                    val tgt = AsmAssignTarget.fromRegisters(regparam.registerOrPair, program, asmgen)
-                    val source = AsmAssignSource(SourceStorageKind.STACK, program, tgt.datatype)
-                    val asgn = AsmAssignment(source, tgt, false, Position.DUMMY)
-                    asmgen.translateNormalAssignment(asgn)
+                argi.value.second.registerOrPair in setOf(RegisterOrPair.A, RegisterOrPair.AY) -> {
+                    require(argForAregister == null)
+                    argForAregister = argi
                 }
-                else -> {}
+                argi.value.second.registerOrPair == RegisterOrPair.Y -> {
+                    asmgen.out("  ldy  P8ESTACK_LO+${argi.index},x")
+                }
+                else -> throw AssemblyError("weird argument")
             }
         }
+
+        if(argForCarry!=null) {
+            asmgen.out("""
+                lda  P8ESTACK_LO+${argForCarry.index},x
+                beq  +
+                sec
+                bcs  ++
++               clc
++               php""")             // push the status flags
+        }
+
+        if(argForAregister!=null) {
+            when(argForAregister.value.second.registerOrPair) {
+                RegisterOrPair.A -> asmgen.out("  lda  P8ESTACK_LO+${argForAregister.index},x")
+                RegisterOrPair.AY -> asmgen.out("  lda  P8ESTACK_LO+${argForAregister.index},x |  ldy  P8ESTACK_HI+${argForAregister.index},x")
+                else -> throw AssemblyError("weird arg")
+            }
+        }
+
+        if(argForXregister!=null) {
+
+            if(argForAregister!=null)
+                asmgen.out("  pha")
+            when(argForXregister.value.second.registerOrPair) {
+                RegisterOrPair.X -> asmgen.out("  lda  P8ESTACK_LO+${argForXregister.index},x |  tax")
+                RegisterOrPair.AX -> asmgen.out("  ldy  P8ESTACK_LO+${argForXregister.index},x |  lda  P8ESTACK_HI+${argForXregister.index},x |  tax |  tya")
+                RegisterOrPair.XY -> asmgen.out("  ldy  P8ESTACK_HI+${argForXregister.index},x |  lda  P8ESTACK_LO+${argForXregister.index},x |  tax")
+                else -> throw AssemblyError("weird arg")
+            }
+            if(argForAregister!=null)
+                asmgen.out("  pla")
+        } else {
+            repeat(sub.parameters.size - 1) { asmgen.out("  inx") }       // unwind stack
+        }
+
+        if(argForCarry!=null)
+            asmgen.out("  plp")       // set the carry flag back to correct value
     }
 
     private fun argumentViaVariable(sub: Subroutine, parameter: IndexedValue<SubroutineParameter>, value: Expression) {

compiler/src/prog8/compiler/target/c64/codegen/assignment/AsmAssignment.kt

@@ -159,6 +159,7 @@ internal class AsmAssignment(val source: AsmAssignSource,
                              val position: Position) {
 
     init {
-        require(source.datatype==target.datatype) {"source and target datatype must be identical"}
+        if(target.register !in setOf(RegisterOrPair.XY, RegisterOrPair.AX, RegisterOrPair.AY))
+            require(source.datatype.memorySize() == target.datatype.memorySize()) { "source and target datatype must be same storage class" }
     }
 }

compiler/src/prog8/compiler/target/c64/codegen/assignment/AssignmentAsmGen.kt

@@ -120,17 +120,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                     is IdentifierReference -> throw AssemblyError("source kind should have been variable")
                     is ArrayIndexedExpression -> throw AssemblyError("source kind should have been array")
                     is DirectMemoryRead -> throw AssemblyError("source kind should have been memory")
-//                    is TypecastExpression -> {
-//                        if(assign.target.kind == TargetStorageKind.STACK) {
-//                            asmgen.translateExpression(value)
-//                            assignStackValue(assign.target)
-//                        } else {
-//                            println("!!!!TYPECAST to ${assign.target.kind} $value")
-//                            // TODO maybe we can do the typecast on the target directly instead of on the stack?
-//                            asmgen.translateExpression(value)
-//                            assignStackValue(assign.target)
-//                        }
-//                    }
+                    is TypecastExpression -> assignTypeCastedValue(assign.target, value.type, value.expression, assign)
 //                    is FunctionCall -> {
 //                        if (assign.target.kind == TargetStorageKind.STACK) {
 //                            asmgen.translateExpression(value)
@@ -161,6 +151,52 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
         }
     }
 
+    private fun assignTypeCastedValue(target: AsmAssignTarget, targetDt: DataType, value: Expression, origAssign: AsmAssignment) {
+        val valueDt = value.inferType(program).typeOrElse(DataType.STRUCT)
+        when(value) {
+            is IdentifierReference -> {
+                if (valueDt == DataType.UBYTE || valueDt == DataType.BYTE) {
+                    if(targetDt in WordDatatypes) {
+                        assignVariableByteIntoWord(target, value, valueDt)
+                        return
+                    }
+                }
+            }
+            is DirectMemoryRead -> {
+                if(targetDt in WordDatatypes) {
+                    if (value.addressExpression is NumericLiteralValue) {
+                        val address = (value.addressExpression as NumericLiteralValue).number.toInt()
+                        assignMemoryByteIntoWord(target, address, null)
+                        return
+                    }
+                    else if (value.addressExpression is IdentifierReference) {
+                        assignMemoryByteIntoWord(target, null, value.addressExpression as IdentifierReference)
+                        return
+                    }
+                }
+            }
+            is NumericLiteralValue -> throw AssemblyError("a cast of a literal value should have been const-folded away")
+            else -> {}
+        }
+
+        when(value) {
+            is PrefixExpression -> {}
+            is BinaryExpression -> {}
+            is ArrayIndexedExpression -> {}
+            is TypecastExpression -> {}
+            is RangeExpr -> {}
+            is FunctionCall -> {}
+            else -> {
+                // TODO optimize the others further?
+                println("warning: slow stack evaluation used for typecast: into $targetDt at ${value.position}")
+            }
+        }
+
+        // give up, do it via eval stack
+        asmgen.translateExpression(origAssign.source.expression!!)
+        assignStackValue(target)
+    }
+
     private fun assignStackValue(target: AsmAssignTarget) {
         when(target.kind) {
             TargetStorageKind.VARIABLE -> {
@@ -260,16 +296,18 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                     DataType.UBYTE, DataType.BYTE -> {
                         when(target.register!!) {
                             RegisterOrPair.A -> asmgen.out(" inx |  lda  P8ESTACK_LO,x")
-                            RegisterOrPair.X -> throw AssemblyError("can't use X here")
+                            RegisterOrPair.X -> throw AssemblyError("can't load X from stack here - use intermediary var? ${target.origAstTarget?.position}")
                             RegisterOrPair.Y -> asmgen.out(" inx |  ldy  P8ESTACK_LO,x")
-                            else -> throw AssemblyError("can't assign byte to register pair word")
+                            RegisterOrPair.AX -> asmgen.out(" inx |  lda  P8ESTACK_LO,x |  ldx  #0")
+                            RegisterOrPair.AY -> asmgen.out(" inx |  lda  P8ESTACK_LO,x |  ldy  #0")
+                            else -> throw AssemblyError("can't assign byte from stack to register pair XY")
                         }
                     }
                     DataType.UWORD, DataType.WORD, in PassByReferenceDatatypes -> {
                         when(target.register!!) {
-                            RegisterOrPair.AX -> throw AssemblyError("can't use X here")
+                            RegisterOrPair.AX -> throw AssemblyError("can't load X from stack here - use intermediary var? ${target.origAstTarget?.position}")
                             RegisterOrPair.AY-> asmgen.out(" inx |  lda  P8ESTACK_LO,x |  ldy  P8ESTACK_HI,x")
-                            RegisterOrPair.XY-> throw AssemblyError("can't use X here")
+                            RegisterOrPair.XY-> throw AssemblyError("can't load X from stack here - use intermediary var? ${target.origAstTarget?.position}")
                             else -> throw AssemblyError("can't assign word to single byte register")
                         }
                     }
@@ -518,6 +556,59 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
         }
     }
 
+    private fun assignVariableByteIntoWord(wordtarget: AsmAssignTarget, bytevar: IdentifierReference, valueDt: DataType) {
+        if(valueDt == DataType.BYTE)
+            TODO("sign extend byte to word")
+
+        val sourceName = asmgen.asmVariableName(bytevar)
+        when(wordtarget.kind) {
+            TargetStorageKind.VARIABLE -> {
+                asmgen.out("""
+                    lda  $sourceName
+                    sta  ${wordtarget.asmVarname}
+                    lda  #0
+                    sta  ${wordtarget.asmVarname}+1
+                    """)
+            }
+            TargetStorageKind.ARRAY -> {
+                val index = wordtarget.array!!.arrayspec.index
+                when {
+                    wordtarget.constArrayIndexValue!=null -> {
+                        val scaledIdx = wordtarget.constArrayIndexValue!! * 2
+                        asmgen.out(" lda  $sourceName  | sta  ${wordtarget.asmVarname}+$scaledIdx |  lda  #0  | sta  ${wordtarget.asmVarname}+$scaledIdx+1")
+                    }
+                    index is IdentifierReference -> {
+                        asmgen.loadScaledArrayIndexIntoRegister(wordtarget.array, wordtarget.datatype, CpuRegister.Y)
+                        asmgen.out(" lda  $sourceName |  sta  ${wordtarget.asmVarname},y |  lda  #0 |  iny |  sta  ${wordtarget.asmVarname},y")
+                    }
+                    else -> {
+                        asmgen.out("  lda  $sourceName |  sta  P8ESTACK_LO,x |  lda  #0 |  sta  P8ESTACK_HI,x |  dex")
+                        asmgen.translateExpression(index)
+                        asmgen.out("  inx |  lda  P8ESTACK_LO,x")
+                        popAndWriteArrayvalueWithUnscaledIndexA(wordtarget.datatype, wordtarget.asmVarname)
+                    }
+                }
+            }
+            TargetStorageKind.REGISTER -> {
+                when(wordtarget.register!!) {
+                    RegisterOrPair.AX -> asmgen.out("  lda  $sourceName |  ldx  #0")
+                    RegisterOrPair.AY -> asmgen.out("  lda  $sourceName |  ldy  #0")
+                    RegisterOrPair.XY -> asmgen.out("  ldx  $sourceName |  ldy  #0")
+                    else -> throw AssemblyError("only reg pairs are words")
+                }
+            }
+            TargetStorageKind.STACK -> {
+                asmgen.out("""
+                    lda  #$sourceName
+                    sta  P8ESTACK_LO,x
+                    lda  #0
+                    sta  P8ESTACK_HI,x
+                    dex""")
+            }
+            else -> throw AssemblyError("other types aren't word")
+        }
+    }
+
     private fun assignRegisterByte(target: AsmAssignTarget, register: CpuRegister) {
         require(target.datatype in ByteDatatypes)
         when(target.kind) {
@@ -571,19 +662,25 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                         RegisterOrPair.A -> {}
                         RegisterOrPair.X -> { asmgen.out("  tax") }
                         RegisterOrPair.Y -> { asmgen.out("  tay") }
-                        else -> throw AssemblyError("attempt to assign byte to register pair word")
+                        RegisterOrPair.AY -> { asmgen.out("  ldy  #0") }
+                        RegisterOrPair.AX -> { asmgen.out("  ldx  #0") }
+                        RegisterOrPair.XY -> { asmgen.out("  tax |  ldy  #0") }
                     }
                     CpuRegister.X -> when(target.register!!) {
                         RegisterOrPair.A -> { asmgen.out("  txa") }
                         RegisterOrPair.X -> {  }
                         RegisterOrPair.Y -> { asmgen.out("  txy") }
-                        else -> throw AssemblyError("attempt to assign byte to register pair word")
+                        RegisterOrPair.AY -> { asmgen.out("  txa |  ldy  #0") }
+                        RegisterOrPair.AX -> { asmgen.out("  txa |  ldx  #0") }
+                        RegisterOrPair.XY -> { asmgen.out("  ldy  #0") }
                     }
                     CpuRegister.Y -> when(target.register!!) {
                         RegisterOrPair.A -> { asmgen.out("  tya") }
                         RegisterOrPair.X -> { asmgen.out("  tyx") }
                         RegisterOrPair.Y -> { }
-                        else -> throw AssemblyError("attempt to assign byte to register pair word")
+                        RegisterOrPair.AY -> { asmgen.out("  tya |  ldy  #0") }
+                        RegisterOrPair.AX -> { asmgen.out("  tya |  ldx  #0") }
+                        RegisterOrPair.XY -> { asmgen.out("  tya |  tax |  ldy  #0") }
                     }
                 }
             }
@@ -692,7 +789,9 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                 RegisterOrPair.A -> asmgen.out("  lda  #${byte.toHex()}")
                 RegisterOrPair.X -> asmgen.out("  ldx  #${byte.toHex()}")
                 RegisterOrPair.Y -> asmgen.out("  ldy  #${byte.toHex()}")
-                else -> throw AssemblyError("can't assign byte to word register apir")
+                RegisterOrPair.AX -> asmgen.out("  lda  #${byte.toHex()} |  ldx  #0")
+                RegisterOrPair.AY -> asmgen.out("  lda  #${byte.toHex()} |  ldy  #0")
+                RegisterOrPair.XY -> asmgen.out("  ldx  #${byte.toHex()} |  ldy  #0")
             }
             TargetStorageKind.STACK -> {
                 asmgen.out("""
@@ -708,7 +807,7 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
             // optimized case for float zero
             when(target.kind) {
                 TargetStorageKind.VARIABLE -> {
-                    if(CompilationTarget.machine.cpu == CpuType.CPU65c02)
+                    if(CompilationTarget.instance.machine.cpu == CpuType.CPU65c02)
                         asmgen.out("""
                             stz  ${target.asmVarname}
                             stz  ${target.asmVarname}+1
@@ -847,7 +946,9 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                     RegisterOrPair.A -> asmgen.out("  lda  ${address.toHex()}")
                     RegisterOrPair.X -> asmgen.out("  ldx  ${address.toHex()}")
                     RegisterOrPair.Y -> asmgen.out("  ldy  ${address.toHex()}")
-                    else -> throw AssemblyError("can't assign byte to word register apir")
+                    RegisterOrPair.AX -> asmgen.out("  lda  ${address.toHex()} |  ldx  #0")
+                    RegisterOrPair.AY -> asmgen.out("  lda  ${address.toHex()} |  ldy  #0")
+                    RegisterOrPair.XY -> asmgen.out("  ldy  ${address.toHex()} |  ldy  #0")
                 }
                 TargetStorageKind.STACK -> {
                     asmgen.out("""
@@ -875,7 +976,9 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
                         RegisterOrPair.A -> {}
                         RegisterOrPair.X -> asmgen.out("  tax")
                         RegisterOrPair.Y -> asmgen.out("  tay")
-                        else -> throw AssemblyError("can't assign byte to word register apir")
+                        RegisterOrPair.AX -> asmgen.out("  ldx  #0")
+                        RegisterOrPair.AY -> asmgen.out("  ldy  #0")
+                        RegisterOrPair.XY -> asmgen.out("  tax |  ldy  #0")
                     }
                 }
                 TargetStorageKind.STACK -> {
@@ -886,6 +989,63 @@ internal class AssignmentAsmGen(private val program: Program, private val asmgen
         }
     }
 
+    private fun assignMemoryByteIntoWord(wordtarget: AsmAssignTarget, address: Int?, identifier: IdentifierReference?) {
+        if (address != null) {
+            when(wordtarget.kind) {
+                TargetStorageKind.VARIABLE -> {
+                    asmgen.out("""
+                        lda  ${address.toHex()}
+                        sta  ${wordtarget.asmVarname}
+                        lda  #0
+                        sta  ${wordtarget.asmVarname}+1
+                        """)
+                }
+                TargetStorageKind.ARRAY -> {
+                    throw AssemblyError("no asm gen for assign memory byte at $address to array ${wordtarget.asmVarname}")
+                }
+                TargetStorageKind.REGISTER -> when(wordtarget.register!!) {
+                    RegisterOrPair.AX -> asmgen.out("  lda  ${address.toHex()} |  ldx  #0")
+                    RegisterOrPair.AY -> asmgen.out("  lda  ${address.toHex()} |  ldy  #0")
+                    RegisterOrPair.XY -> asmgen.out("  ldy  ${address.toHex()} |  ldy  #0")
+                    else -> throw AssemblyError("word regs can only be pair")
+                }
+                TargetStorageKind.STACK -> {
+                    asmgen.out("""
+                        lda  ${address.toHex()}
+                        sta  P8ESTACK_LO,x
+                        lda  #0
+                        sta  P8ESTACK_HI,x
+                        dex""")
+                }
+                else -> throw AssemblyError("other types aren't word")
+            }
+        } else if (identifier != null) {
+            when(wordtarget.kind) {
+                TargetStorageKind.VARIABLE -> {
+                    asmgen.loadByteFromPointerIntoA(identifier)
+                    asmgen.out(" sta  ${wordtarget.asmVarname} |  lda  #0 |  sta ${wordtarget.asmVarname}+1")
+                }
+                TargetStorageKind.ARRAY -> {
+                    throw AssemblyError("no asm gen for assign memory byte $identifier to array ${wordtarget.asmVarname} ")
+                }
+                TargetStorageKind.REGISTER -> {
+                    asmgen.loadByteFromPointerIntoA(identifier)
+                    when(wordtarget.register!!) {
+                        RegisterOrPair.AX -> asmgen.out("  ldx  #0")
+                        RegisterOrPair.AY -> asmgen.out("  ldy  #0")
+                        RegisterOrPair.XY -> asmgen.out("  tax |  ldy  #0")
+                        else -> throw AssemblyError("word regs can only be pair")
+                    }
+                }
+                TargetStorageKind.STACK -> {
+                    asmgen.loadByteFromPointerIntoA(identifier)
+                    asmgen.out(" sta  P8ESTACK_LO,x |  lda  #0 |  sta  P8ESTACK_HI,x |  dex")
+                }
+                else -> throw AssemblyError("other types aren't word")
+            }
+        }
+    }
+
     private fun storeByteViaRegisterAInMemoryAddress(ldaInstructionArg: String, memoryAddress: DirectMemoryWrite) {
         val addressExpr = memoryAddress.addressExpression
         val addressLv = addressExpr as? NumericLiteralValue

compiler/src/prog8/compiler/target/c64/codegen/assignment/AugmentableAssignmentAsmGen.kt

@@ -8,6 +8,7 @@ import prog8.compiler.target.CompilationTarget
 import prog8.compiler.target.CpuType
 import prog8.compiler.target.c64.codegen.AsmGen
 import prog8.compiler.toHex
+import kotlin.math.absoluteValue
 
 internal class AugmentableAssignmentAsmGen(private val program: Program,
                                            private val assignmentAsmGen: AssignmentAsmGen,
@@ -22,8 +23,7 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                 // A = -A , A = +A, A = ~A, A = not A
                 val type = value.inferType(program).typeOrElse(DataType.STRUCT)
                 when (value.operator) {
-                    "+" -> {
-                    }
+                    "+" -> {}
                     "-" -> inplaceNegate(assign.target, type)
                     "~" -> inplaceInvert(assign.target, type)
                     "not" -> inplaceBooleanNot(assign.target, type)
@@ -193,7 +193,7 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                         asmgen.translateExpression(memory.addressExpression)
                         // TODO buggy?:
                         asmgen.out("  jsr  prog8_lib.read_byte_from_address_on_stack |  sta  P8ZP_SCRATCH_B1")
-                        val zp = CompilationTarget.machine.zeropage
+                        val zp = CompilationTarget.instance.machine.zeropage
                         when {
                             valueLv != null -> inplaceModification_byte_litval_to_variable(zp.SCRATCH_B1.toHex(), DataType.UBYTE, operator, valueLv.toInt())
                             ident != null -> inplaceModification_byte_variable_to_variable(zp.SCRATCH_B1.toHex(), DataType.UBYTE, operator, ident)
@@ -250,7 +250,9 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                 else
                     asmgen.out("  sta  (P8ZP_SCRATCH_W1),y")
             }
-            "*" -> TODO("mul mem byte")// asmgen.out("  jsr  prog8_lib.mul_byte")  //  the optimized routines should have been checked earlier
+            "*" -> {
+                TODO("mul mem byte")// asmgen.out("  jsr  prog8_lib.mul_byte")
+            }
             "/" -> TODO("div mem byte")// asmgen.out(if(types==DataType.UBYTE) "  jsr  prog8_lib.idiv_ub" else "  jsr  prog8_lib.idiv_b")
             "%" -> {
                 TODO("mem byte remainder")
@@ -309,7 +311,9 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                 else
                     asmgen.out("  sta  (P8ZP_SCRATCH_W1),y")
             }
-            "*" -> TODO("mem mul")// asmgen.out("  jsr  prog8_lib.mul_byte")  //  the optimized routines should have been checked earlier
+            "*" -> {
+                TODO("mem mul")// asmgen.out("  jsr  prog8_lib.mul_byte")
+            }
             "/" -> TODO("mem div")// asmgen.out(if(types==DataType.UBYTE) "  jsr  prog8_lib.idiv_ub" else "  jsr  prog8_lib.idiv_b")
             "%" -> {
                 TODO("mem byte remainder")
@@ -367,8 +371,12 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                     asmgen.out("  sta  (P8ZP_SCRATCH_W1),y")
             }
             "*" -> {
-                TODO("mem mul byte litval")
-                // asmgen.out("  jsr  prog8_lib.mul_byte")  //  the optimized routines should have been checked earlier
+                if(value in asmgen.optimizedByteMultiplications) {
+                    TODO("optimized mem mul ubyte litval $value")
+                } else {
+                    TODO("mem mul ubyte litval $value")
+                    // asmgen.out("  jsr  prog8_lib.mul_byte")
+                }
             }
             "/" -> {
                 if(value==0)
@@ -443,7 +451,7 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
             "-" -> asmgen.out(" lda  $name |  sec |  sbc  P8ESTACK_LO+1,x |  sta  $name")
             "*" -> {
                 TODO("var mul byte expr")
-                // asmgen.out("  jsr  prog8_lib.mul_byte")  //  the optimized routines should have been checked earlier
+                // asmgen.out("  jsr  prog8_lib.mul_byte")
             }
             "/" -> {
                 TODO("var div byte expr")// asmgen.out(if(types==DataType.UBYTE) "  jsr  prog8_lib.idiv_ub" else "  jsr  prog8_lib.idiv_b")
@@ -572,9 +580,21 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
             "+" -> asmgen.out(" lda  $name |  clc |  adc  #$value |  sta  $name")
             "-" -> asmgen.out(" lda  $name |  sec |  sbc  #$value |  sta  $name")
             "*" -> {
-                // TODO what about the optimized mul_5 etc routines?
-                TODO("var byte mul litval")
-                // asmgen.out("  jsr  prog8_lib.mul_byte")  //  the optimized routines should have been checked earlier
+                if(dt == DataType.UBYTE) {
+                    if(value in asmgen.optimizedByteMultiplications) {
+                        asmgen.out("  lda  $name |  jsr  math.mul_byte_$value |  sta  $name")
+                    } else {
+                        TODO("var mul ubyte litval $value")
+                        // asmgen.out("  jsr  prog8_lib.mul_byte")
+                    }
+                } else {
+                    if(value.absoluteValue in asmgen.optimizedByteMultiplications) {
+                        asmgen.out("  lda  $name |  jsr  math.mul_byte_$value |  sta  $name")
+                    } else {
+                        TODO("var mul sbyte litval $value")
+                        // asmgen.out("  jsr  prog8_lib.mul_byte")
+                    }
+                }
             }
             "/" -> {
                 if (dt == DataType.UBYTE) {
@@ -669,19 +689,32 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                 }
             }
             "*" -> {
-                // TODO what about the optimized mul_5 etc routines?
-                asmgen.out("""
-                    lda  $name
-                    sta  P8ZP_SCRATCH_W1
-                    lda  $name+1
-                    sta  P8ZP_SCRATCH_W1+1
-                    lda  #<$value
-                    ldy  #>$value
-                    jsr  math.multiply_words
-                    lda  math.multiply_words.result
-                    sta  $name
-                    lda  math.multiply_words.result+1
-                    sta  $name+1""")
+                if(dt == DataType.UWORD){
+                    if(value in asmgen.optimizedWordMultiplications) {
+                        asmgen.out("  lda  $name |  ldy  $name+1 |  jsr  math.mul_word_$value |  sta  $name |  sty  $name+1")
+                    } else {
+                        TODO("var uword mul litval $value")
+                    }
+                } else {
+                    if(value.absoluteValue in asmgen.optimizedWordMultiplications) {
+                        asmgen.out("  lda  $name |  ldy  $name+1 |  jsr  math.mul_word_$value |  sta  $name |  sty  $name+1")
+                    } else {
+                        // TODO don't use stack here
+                        // TODO does this work for signed words?
+                        asmgen.out("""
+                        lda  $name
+                        sta  P8ZP_SCRATCH_W1
+                        lda  $name+1
+                        sta  P8ZP_SCRATCH_W1+1
+                        lda  #<$value
+                        ldy  #>$value
+                        jsr  math.multiply_words
+                        lda  math.multiply_words.result
+                        sta  $name
+                        lda  math.multiply_words.result+1
+                        sta  $name+1""")
+                    }
+                }
             }
             "/" -> {
                 if(value==0)
@@ -747,13 +780,13 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
             "&" -> {
                 when {
                     value == 0 -> {
-                        if(CompilationTarget.machine.cpu == CpuType.CPU65c02)
+                        if(CompilationTarget.instance.machine.cpu == CpuType.CPU65c02)
                             asmgen.out(" stz  $name |  stz  $name+1")
                         else
                             asmgen.out(" lda  #0 |  sta  $name |  sta  $name+1")
                     }
                     value and 255 == 0 -> {
-                        if(CompilationTarget.machine.cpu == CpuType.CPU65c02)
+                        if(CompilationTarget.instance.machine.cpu == CpuType.CPU65c02)
                             asmgen.out(" stz  $name")
                         else
                             asmgen.out(" lda  #0 |  sta  $name")
@@ -761,7 +794,7 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                     }
                     value < 0x0100 -> {
                         asmgen.out(" lda  $name |  and  #$value |  sta  $name")
-                        if(CompilationTarget.machine.cpu == CpuType.CPU65c02)
+                        if(CompilationTarget.instance.machine.cpu == CpuType.CPU65c02)
                             asmgen.out("  stz  $name+1")
                         else
                             asmgen.out("  lda  #0 |  sta  $name+1")
@@ -1088,11 +1121,11 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
         // because the value is evaluated onto the eval stack (=slow).
         println("warning: slow stack evaluation used (2):  $name $operator= ${value::class.simpleName} at ${value.position}") // TODO
         asmgen.translateExpression(value)
+        asmgen.out("  jsr  c64flt.pop_float_fac1")
         asmgen.saveRegister(CpuRegister.X)
         when (operator) {
             "**" -> {
                 asmgen.out("""
-                    jsr  c64flt.pop_float_fac1
                     lda  #<$name
                     ldy  #>$name
                     jsr  c64flt.CONUPK
@@ -1101,7 +1134,6 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
             }
             "+" -> {
                 asmgen.out("""
-                    jsr  c64flt.pop_float_fac1
                     lda  #<$name
                     ldy  #>$name
                     jsr  c64flt.FADD
@@ -1109,7 +1141,6 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
             }
             "-" -> {
                 asmgen.out("""
-                    jsr  c64flt.pop_float_fac1
                     lda  #<$name
                     ldy  #>$name
                     jsr  c64flt.FSUB
@@ -1117,7 +1148,6 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
             }
             "*" -> {
                 asmgen.out("""
-                    jsr  c64flt.pop_float_fac1
                     lda  #<$name
                     ldy  #>$name
                     jsr  c64flt.FMULT
@@ -1125,7 +1155,6 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
             }
             "/" -> {
                 asmgen.out("""
-                    jsr  c64flt.pop_float_fac1
                     lda  #<$name
                     ldy  #>$name
                     jsr  c64flt.FDIV
@@ -1133,7 +1162,6 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
             }
             else -> throw AssemblyError("invalid operator for in-place float modification $operator")
         }
-        // store Fac1 back into memory
         asmgen.out("""
             ldx  #<$name
             ldy  #>$name
@@ -1250,6 +1278,7 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                 """)
             }
             "*" -> {
+                // assume that code optimization is already done on the AST level for special cases such as 0, 1, 2...
                 asmgen.out("""
                     lda  #<$name
                     ldy  #>$name
@@ -1285,7 +1314,6 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
     private fun inplaceCast(target: AsmAssignTarget, cast: TypecastExpression, position: Position) {
         val outerCastDt = cast.type
         val innerCastDt = (cast.expression as? TypecastExpression)?.type
-
         if (innerCastDt == null) {
             // simple typecast where the value is the target
             when (target.datatype) {
@@ -1295,20 +1323,20 @@ internal class AugmentableAssignmentAsmGen(private val program: Program,
                         DataType.UBYTE, DataType.BYTE -> {
                             when(target.kind) {
                                 TargetStorageKind.VARIABLE -> {
-                                    if(CompilationTarget.machine.cpu == CpuType.CPU65c02)
+                                    if(CompilationTarget.instance.machine.cpu == CpuType.CPU65c02)
                                         asmgen.out(" stz  ${target.asmVarname}+1")
                                     else
                                         asmgen.out(" lda  #0 |  sta  ${target.asmVarname}+1")
                                 }
                                 TargetStorageKind.ARRAY -> {
                                     asmgen.loadScaledArrayIndexIntoRegister(target.array!!, target.datatype, CpuRegister.Y, true)
-                                    if(CompilationTarget.machine.cpu == CpuType.CPU65c02)
+                                    if(CompilationTarget.instance.machine.cpu == CpuType.CPU65c02)
                                         asmgen.out("  stz  ${target.asmVarname},y")
                                     else
                                         asmgen.out("  lda  #0 |  sta  ${target.asmVarname},y")
                                 }
                                 TargetStorageKind.STACK -> {
-                                    if(CompilationTarget.machine.cpu == CpuType.CPU65c02)
+                                    if(CompilationTarget.instance.machine.cpu == CpuType.CPU65c02)
                                         asmgen.out(" stz  P8ESTACK_HI+1,x")
                                     else
                                         asmgen.out(" lda  #0 |  sta  P8ESTACK_HI+1,x")

compiler/src/prog8/compiler/target/cx16/CX16MachineDefinition.kt

@@ -26,7 +26,6 @@ internal object CX16MachineDefinition: IMachineDefinition {
     override val ESTACK_HI = 0x0500        //  $0500-$05ff inclusive
 
     override lateinit var zeropage: Zeropage
-    override val initSystemProcname = "cx16.init_system"
 
     override fun getFloat(num: Number) = C64MachineDefinition.Mflpt5.fromNumber(num)
 
@@ -75,7 +74,6 @@ internal object CX16MachineDefinition: IMachineDefinition {
 
         override val SCRATCH_B1 = 0x79      // temp storage for a single byte
         override val SCRATCH_REG = 0x7a     // temp storage for a register
-        override val SCRATCH_REG_X = 0x7b   // temp storage for register X (the evaluation stack pointer)
         override val SCRATCH_W1 = 0x7c      // temp storage 1 for a word  $7c+$7d
         override val SCRATCH_W2 = 0x7e      // temp storage 2 for a word  $7e+$7f
 
@@ -96,16 +94,16 @@ internal object CX16MachineDefinition: IMachineDefinition {
             when (options.zeropage) {
                 ZeropageType.FULL -> {
                     free.addAll(0x22..0xff)
-                    free.removeAll(listOf(SCRATCH_B1, SCRATCH_REG, SCRATCH_REG_X, SCRATCH_W1, SCRATCH_W1 + 1, SCRATCH_W2, SCRATCH_W2 + 1))
+                    free.removeAll(listOf(SCRATCH_B1, SCRATCH_REG, SCRATCH_W1, SCRATCH_W1 + 1, SCRATCH_W2, SCRATCH_W2 + 1))
                 }
                 ZeropageType.KERNALSAFE -> {
                     free.addAll(0x22..0x7f)
                     free.addAll(0xa9..0xff)
-                    free.removeAll(listOf(SCRATCH_B1, SCRATCH_REG, SCRATCH_REG_X, SCRATCH_W1, SCRATCH_W1 + 1, SCRATCH_W2, SCRATCH_W2 + 1))
+                    free.removeAll(listOf(SCRATCH_B1, SCRATCH_REG, SCRATCH_W1, SCRATCH_W1 + 1, SCRATCH_W2, SCRATCH_W2 + 1))
                 }
                 ZeropageType.BASICSAFE -> {
                     free.addAll(0x22..0x7f)
-                    free.removeAll(listOf(SCRATCH_B1, SCRATCH_REG, SCRATCH_REG_X, SCRATCH_W1, SCRATCH_W1 + 1, SCRATCH_W2, SCRATCH_W2 + 1))
+                    free.removeAll(listOf(SCRATCH_B1, SCRATCH_REG, SCRATCH_W1, SCRATCH_W1 + 1, SCRATCH_W2, SCRATCH_W2 + 1))
                 }
                 ZeropageType.DONTUSE -> {
                     free.clear() // don't use zeropage at all
@@ -115,7 +113,6 @@ internal object CX16MachineDefinition: IMachineDefinition {
 
             require(SCRATCH_B1 !in free)
             require(SCRATCH_REG !in free)
-            require(SCRATCH_REG_X !in free)
             require(SCRATCH_W1 !in free)
             require(SCRATCH_W2 !in free)
 

compiler/src/prog8/optimizer/ConstantFoldingOptimizer.kt

@@ -150,7 +150,7 @@ internal class ConstantIdentifierReplacer(private val program: Program, private
                     if(rangeExpr==null && litval!=null) {
                         // arraysize initializer is a single int, and we know the size.
                         val fillvalue = litval.number.toDouble()
-                        if (fillvalue < CompilationTarget.machine.FLOAT_MAX_NEGATIVE || fillvalue > CompilationTarget.machine.FLOAT_MAX_POSITIVE)
+                        if (fillvalue < CompilationTarget.instance.machine.FLOAT_MAX_NEGATIVE || fillvalue > CompilationTarget.instance.machine.FLOAT_MAX_POSITIVE)
                             errors.err("float value overflow", litval.position)
                         else {
                             // create the array itself, filled with the fillvalue.

compiler/src/prog8/optimizer/ExpressionSimplifier.kt

@@ -297,6 +297,49 @@ internal class ExpressionSimplifier(private val program: Program) : AstWalker()
         return noModifications
     }
 
+    override fun after(functionCall: FunctionCall, parent: Node): Iterable<IAstModification> {
+        if(functionCall.target.nameInSource == listOf("lsb")) {
+            val arg = functionCall.args[0]
+            if(arg is TypecastExpression) {
+                val valueDt = arg.expression.inferType(program)
+                if (valueDt.istype(DataType.BYTE) || valueDt.istype(DataType.UBYTE)) {
+                    // useless lsb() of byte value that was casted to word
+                    return listOf(IAstModification.ReplaceNode(functionCall, arg.expression, parent))
+                }
+            } else {
+                val argDt = arg.inferType(program)
+                if (argDt.istype(DataType.BYTE) || argDt.istype(DataType.UBYTE)) {
+                    // useless lsb() of byte value
+                    return listOf(IAstModification.ReplaceNode(functionCall, arg, parent))
+                }
+            }
+        }
+        else if(functionCall.target.nameInSource == listOf("msb")) {
+            val arg = functionCall.args[0]
+            if(arg is TypecastExpression) {
+                val valueDt = arg.expression.inferType(program)
+                if (valueDt.istype(DataType.BYTE) || valueDt.istype(DataType.UBYTE)) {
+                    // useless msb() of byte value that was casted to word, replace with 0
+                    return listOf(IAstModification.ReplaceNode(
+                            functionCall,
+                            NumericLiteralValue(valueDt.typeOrElse(DataType.UBYTE), 0, arg.expression.position),
+                            parent))
+                }
+            } else {
+                val argDt = arg.inferType(program)
+                if (argDt.istype(DataType.BYTE) || argDt.istype(DataType.UBYTE)) {
+                    // useless msb() of byte value, replace with 0
+                    return listOf(IAstModification.ReplaceNode(
+                            functionCall,
+                            NumericLiteralValue(argDt.typeOrElse(DataType.UBYTE), 0, arg.position),
+                            parent))
+                }
+            }
+        }
+
+        return noModifications
+    }
+
     private fun determineY(x: Expression, subBinExpr: BinaryExpression): Expression? {
         return when {
             subBinExpr.left isSameAs x -> subBinExpr.right

compiler/src/prog8/optimizer/StatementOptimizer.kt

@@ -104,7 +104,7 @@ internal class StatementOptimizer(private val program: Program,
                 if(string!=null) {
                     val pos = functionCallStatement.position
                     if (string.value.length == 1) {
-                        val firstCharEncoded = CompilationTarget.encodeString(string.value, string.altEncoding)[0]
+                        val firstCharEncoded = CompilationTarget.instance.encodeString(string.value, string.altEncoding)[0]
                         val chrout = FunctionCallStatement(
                                 IdentifierReference(listOf("c64", "CHROUT"), pos),
                                 mutableListOf(NumericLiteralValue(DataType.UBYTE, firstCharEncoded.toInt(), pos)),
@@ -112,7 +112,7 @@ internal class StatementOptimizer(private val program: Program,
                         )
                         return listOf(IAstModification.ReplaceNode(functionCallStatement, chrout, parent))
                     } else if (string.value.length == 2) {
-                        val firstTwoCharsEncoded = CompilationTarget.encodeString(string.value.take(2), string.altEncoding)
+                        val firstTwoCharsEncoded = CompilationTarget.instance.encodeString(string.value.take(2), string.altEncoding)
                         val chrout1 = FunctionCallStatement(
                                 IdentifierReference(listOf("c64", "CHROUT"), pos),
                                 mutableListOf(NumericLiteralValue(DataType.UBYTE, firstTwoCharsEncoded[0].toInt(), pos)),
@@ -220,7 +220,7 @@ internal class StatementOptimizer(private val program: Program,
                 val size = sv.value.length
                 if(size==1) {
                     // loop over string of length 1 -> just assign the single character
-                    val character = CompilationTarget.encodeString(sv.value, sv.altEncoding)[0]
+                    val character = CompilationTarget.instance.encodeString(sv.value, sv.altEncoding)[0]
                     val byte = NumericLiteralValue(DataType.UBYTE, character, iterable.position)
                     val scope = AnonymousScope(mutableListOf(), forLoop.position)
                     scope.statements.add(Assignment(AssignTarget(forLoop.loopVar, null, null, forLoop.position), byte, forLoop.position))

compiler/test/UnitTests.kt

@@ -129,7 +129,7 @@ class TestC64Zeropage {
 
     @Test
     fun testNames() {
-        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(), false))
+        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(), false, "c64"))
 
         zp.allocate("", DataType.UBYTE, null, errors)
         zp.allocate("", DataType.UBYTE, null, errors)
@@ -142,37 +142,37 @@ class TestC64Zeropage {
 
     @Test
     fun testZpFloatEnable() {
-        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), false))
+        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), false, "c64"))
         assertFailsWith<CompilerException> {
             zp.allocate("", DataType.FLOAT, null, errors)
         }
-        val zp2 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.DONTUSE, emptyList(), true))
+        val zp2 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.DONTUSE, emptyList(), true, "c64"))
         assertFailsWith<CompilerException> {
             zp2.allocate("", DataType.FLOAT, null, errors)
         }
-        val zp3 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FLOATSAFE, emptyList(), true))
+        val zp3 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FLOATSAFE, emptyList(), true, "c64"))
         zp3.allocate("", DataType.FLOAT, null, errors)
     }
 
     @Test
     fun testZpModesWithFloats() {
-        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), false))
-        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.KERNALSAFE, emptyList(), false))
-        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(), false))
-        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FLOATSAFE, emptyList(), false))
-        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(), true))
-        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FLOATSAFE, emptyList(), true))
+        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), false, "c64"))
+        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.KERNALSAFE, emptyList(), false, "c64"))
+        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(), false, "c64"))
+        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FLOATSAFE, emptyList(), false, "c64"))
+        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(), true, "c64"))
+        C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FLOATSAFE, emptyList(), true, "c64"))
         assertFailsWith<CompilerException> {
-            C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), true))
+            C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), true, "c64"))
         }
         assertFailsWith<CompilerException> {
-            C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.KERNALSAFE, emptyList(), true))
+            C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.KERNALSAFE, emptyList(), true, "c64"))
         }
     }
 
     @Test
     fun testZpDontuse() {
-        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.DONTUSE, emptyList(), false))
+        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.DONTUSE, emptyList(), false, "c64"))
         println(zp.free)
         assertEquals(0, zp.available())
         assertFailsWith<CompilerException> {
@@ -182,19 +182,19 @@ class TestC64Zeropage {
 
     @Test
     fun testFreeSpaces() {
-        val zp1 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(), true))
+        val zp1 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(), true, "c64"))
         assertEquals(16, zp1.available())
-        val zp2 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FLOATSAFE, emptyList(), false))
+        val zp2 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FLOATSAFE, emptyList(), false, "c64"))
         assertEquals(91, zp2.available())
-        val zp3 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.KERNALSAFE, emptyList(), false))
+        val zp3 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.KERNALSAFE, emptyList(), false, "c64"))
         assertEquals(125, zp3.available())
-        val zp4 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), false))
+        val zp4 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), false, "c64"))
         assertEquals(238, zp4.available())
     }
 
     @Test
     fun testReservedSpace() {
-        val zp1 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), false))
+        val zp1 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), false, "c64"))
         assertEquals(238, zp1.available())
         assertTrue(50 in zp1.free)
         assertTrue(100 in zp1.free)
@@ -203,7 +203,7 @@ class TestC64Zeropage {
         assertTrue(200 in zp1.free)
         assertTrue(255 in zp1.free)
         assertTrue(199 in zp1.free)
-        val zp2 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, listOf(50 .. 100, 200..255), false))
+        val zp2 = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, listOf(50 .. 100, 200..255), false, "c64"))
         assertEquals(139, zp2.available())
         assertFalse(50 in zp2.free)
         assertFalse(100 in zp2.free)
@@ -216,7 +216,7 @@ class TestC64Zeropage {
 
     @Test
     fun testBasicsafeAllocation() {
-        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(), true))
+        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(), true, "c64"))
         assertEquals(16, zp.available())
 
         assertFailsWith<ZeropageDepletedError> {
@@ -239,7 +239,7 @@ class TestC64Zeropage {
 
     @Test
     fun testFullAllocation() {
-        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), false))
+        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.FULL, emptyList(), false, "c64"))
         assertEquals(238, zp.available())
         val loc = zp.allocate("", DataType.UWORD, null, errors)
         assertTrue(loc > 3)
@@ -269,7 +269,7 @@ class TestC64Zeropage {
 
     @Test
     fun testEfficientAllocation() {
-        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(),  true))
+        val zp = C64Zeropage(CompilationOptions(OutputType.RAW, LauncherType.NONE, ZeropageType.BASICSAFE, emptyList(),  true, "c64"))
         assertEquals(16, zp.available())
         assertEquals(0x04, zp.allocate("", DataType.WORD, null, errors))
         assertEquals(0x06, zp.allocate("", DataType.UBYTE, null, errors))

docs/source/syntaxreference.rst

@@ -33,6 +33,14 @@ This makes it easier to understand and relate the generated code. Examples::
 Directives
 -----------
 
+.. data:: %target <target>
+
+    Level: module.
+    Global setting, selects a compilation target from within the source file.
+    The default compilation target is "c64" which targets the Commodore-64 machine.
+    You can also omit this and use the ``-target`` command line option.
+
+
 .. data:: %output <type>
 
 	Level: module.
@@ -514,18 +522,20 @@ and returning stuff in several registers as well. The ``clobbers`` clause is use
 what CPU registers are clobbered by the call instead of being unchanged or returning a meaningful result value.
 
 
-Subroutines that are implemented purely in assembly code and which have an assembly calling convention (i.e.
-the parameters are strictly passed via cpu registers), are defined like this::
+User subroutines in the program source code that are implemented purely in assembly and which have an assembly calling convention (i.e.
+the parameters are strictly passed via cpu registers), are defined with ``asmsub`` like this::
 
-    asmsub  FREADS32() clobbers(A,X,Y)  {
+    asmsub  clear_screenchars (ubyte char @ A) clobbers(Y)  {
         %asm {{
-            lda  $62
-            eor  #$ff
-            asl  a
-            lda  #0
-            ldx  #$a0
-            jmp  $bc4f
-        }}
+            ldy  #0
+    _loop   sta  c64.Screen,y
+            sta  c64.Screen+$0100,y
+            sta  c64.Screen+$0200,y
+            sta  c64.Screen+$02e8,y
+            iny
+            bne  _loop
+            rts
+            }}
     }
 
 the statement body of such a subroutine should consist of just an inline assembly block.

docs/source/todo.rst

@@ -2,12 +2,15 @@
 TODO
 ====
 
-- optimize assignment codegeneration
-- get rid of all TODO's ;-)
+- get rid of all other TODO's in the code ;-)
+- line-circle-gfx examples are now a few hundred bytes larger than before. Why is that, can it be fixed?
+- compiler errors and warnings in standard format so the IDE shows them as clickable links; ./test.asm:2578:3: blablabla
+- further optimize assignment codegeneration
+- auto select correct library to import based on target, instead of having c64- and  cx16-  prefix variants
+- implement @stack for asmsub parameters
 - make it possible to use cpu opcodes such as 'nop' as variable names by prefixing all asm vars with something such as '_'
 - option to load the built-in library files from a directory instead of the embedded ones (for easier library development/debugging)
 - aliases for imported symbols for example perhaps '%alias print = c64scr.print' ?
-- investigate support for 8bitguy's Commander X16 platform https://www.commanderx16.com  and https://github.com/commanderx16/x16-docs
 - see if we can group some errors together for instance the (now single) errors about unidentified symbols
 
 

examples/cube3d-float.p8

@@ -57,7 +57,7 @@ main {
         float Azy = cosb*sinc
         float Azz = cosb*cosc
 
-        ubyte i
+        ubyte @zp i
         for i in 0 to len(xcoor)-1 {
             rotatedx[i] = Axx*xcoor[i] + Axy*ycoor[i] + Axz*zcoor[i]
             rotatedy[i] = Ayx*xcoor[i] + Ayy*ycoor[i] + Ayz*zcoor[i]
@@ -69,7 +69,7 @@ main {
 
         ; plot the points of the 3d cube
         ; first the points on the back, then the points on the front (painter algorithm)
-        ubyte i
+        ubyte @zp i
         float rz
         float persp
         ubyte sx

examples/cube3d-gfx.p8

@@ -1,12 +1,8 @@
 %import c64lib
 %import c64graphics
 
-
 main {
 
-    const uword width = 255
-    const uword height = 200
-
     ; vertices
     word[] xcoor = [ -100, -100, -100, -100,  100,  100,  100, 100 ]
     word[] ycoor = [ -100, -100,  100,  100, -100, -100,  100, 100 ]
@@ -69,7 +65,7 @@ main {
         word Azy = wcosb*wsinc / 128
         word Azz = wcosb*wcosc / 128
 
-        ubyte i
+        ubyte @zp i
         for i in 0 to len(xcoor)-1 {
             ; don't normalize by dividing by 128, instead keep some precision for perspective calc later
             rotatedx[i] = (Axx*xcoor[i] + Axy*ycoor[i] + Axz*zcoor[i])
@@ -78,17 +74,21 @@ main {
         }
     }
 
+    const uword screen_width = 320
+    const ubyte screen_height = 200
+
+
     sub draw_lines() {
-        ubyte i
+        ubyte @zp i
         for i in len(edgesFrom) -1 downto 0 {
-            ubyte vFrom = edgesFrom[i]
-            ubyte vTo = edgesTo[i]
-            word persp1 = 256 + rotatedz[vFrom]/256
-            word persp2 = 256 + rotatedz[vTo]/256
-            graphics.line(rotatedx[vFrom] / persp1 + 160.w as uword,
-                          rotatedy[vFrom] / persp1 + 100 as ubyte,
-                          rotatedx[vTo] / persp2 + 160.w as uword,
-                          rotatedy[vTo] / persp2 + 100 as ubyte)
+            ubyte @zp vFrom = edgesFrom[i]
+            ubyte @zp vTo = edgesTo[i]              ; TODO need compiler error for double declaration if also declared outside the for loop!
+            word @zp persp1 = 256 + rotatedz[vFrom]/256
+            word @zp persp2 = 256 + rotatedz[vTo]/256
+            graphics.line(rotatedx[vFrom] / persp1 + screen_width/2 as uword,
+                          rotatedy[vFrom] / persp1 + screen_height/2 as ubyte,
+                          rotatedx[vTo] / persp2 + screen_width/2 as uword,
+                          rotatedy[vTo] / persp2 + screen_height/2 as ubyte)
         }
     }
 }

examples/cube3d-sprites.p8

@@ -122,7 +122,7 @@ main {
         word Azy = wcosb*wsinc / 128
         word Azz = wcosb*wcosc / 128
 
-        ubyte i
+        ubyte @zp i
         for i in 0 to len(xcoor)-1 {
             ; don't normalize by dividing by 128, instead keep some precision for perspective calc later
             rotatedx[i] = (Axx*xcoor[i] + Axy*ycoor[i] + Axz*zcoor[i])
@@ -138,8 +138,8 @@ main {
 
         ; first sort vertices to sprite order so the back/front order is correct as well
         ; (simple bubble sort as it's only 8 items to sort)
-        ubyte i
-        ubyte i1
+        ubyte @zp i
+        ubyte @zp i1
         for i in 6 downto 0 {
             for i1 in 0 to i {
                 ubyte i2 = i1+1
@@ -154,7 +154,7 @@ main {
         ubyte[] spritecolors = [1,1,7,15,12,11,9,9]
 
         for i in 0 to 7 {
-            word zc = rotatedz[i]
+            word @zp zc = rotatedz[i]
             word persp = 300+zc/256
             ubyte sx = rotatedx[i] / persp + width/2 as ubyte + 20
             ubyte sy = rotatedy[i] / persp + height/2 as ubyte + 40

examples/cube3d.p8

@@ -62,7 +62,7 @@ main {
         word Azy = wcosb*wsinc / 128
         word Azz = wcosb*wcosc / 128
 
-        ubyte i
+        ubyte @zp i
         for i in 0 to len(xcoor)-1 {
             ; don't normalize by dividing by 128, instead keep some precision for perspective calc later
             rotatedx[i] = Axx*xcoor[i] + Axy*ycoor[i] + Axz*zcoor[i]
@@ -76,9 +76,9 @@ main {
         ; plot the points of the 3d cube
         ; first the points on the back, then the points on the front (painter algorithm)
 
-        ubyte i
-        word rz
-        word persp
+        ubyte @zp i
+        word @zp rz
+        word @zp persp
         byte sx
         byte sy
 

examples/cx16/arithmetic/aggregates.p8

@@ -0,0 +1,109 @@
+%target cx16
+%import cx16flt
+%import cx16textio
+%zeropage basicsafe
+
+main {
+
+    sub start() {
+        ubyte[]  ubarr = [100, 0, 99, 199, 22]
+        byte[]  barr = [-100, 0, 99, -122, 22]
+        uword[] uwarr = [1000, 0, 222, 4444, 999]
+        word[]  warr = [-1000, 0, 999, -4444, 222]
+        float[]  farr = [-1000.1, 0, 999.9, -4444.4, 222.2]
+        str name = "irmen"
+        ubyte ub
+        byte bb
+        word ww
+        uword uw
+        float ff
+
+        ; LEN/STRLEN
+        ubyte length = len(name)
+        if length!=5 txt.print("error len1\n")
+        length = len(uwarr)
+        if length!=5 txt.print("error len2\n")
+        length=strlen(name)
+        if length!=5 txt.print("error strlen1\n")
+        name[3] = 0
+        length=strlen(name)
+        if length!=3 txt.print("error strlen2\n")
+
+        ; MAX
+        ub = max(ubarr)
+        if ub!=199 txt.print("error max1\n")
+        bb = max(barr)
+        if bb!=99 txt.print("error max2\n")
+        uw = max(uwarr)
+        if uw!=4444 txt.print("error max3\n")
+        ww = max(warr)
+        if ww!=999 txt.print("error max4\n")
+        ff = max(farr)
+        if ff!=999.9 txt.print("error max5\n")
+
+        ; MIN
+        ub = min(ubarr)
+        if ub!=0 txt.print("error min1\n")
+        bb = min(barr)
+        if bb!=-122 txt.print("error min2\n")
+        uw = min(uwarr)
+        if uw!=0 txt.print("error min3\n")
+        ww = min(warr)
+        if ww!=-4444 txt.print("error min4\n")
+        ff = min(farr)
+        if ff!=-4444.4 txt.print("error min5\n")
+
+        ; SUM
+        uw = sum(ubarr)
+        if uw!=420 txt.print("error sum1\n")
+        ww = sum(barr)
+        if ww!=-101 txt.print("error sum2\n")
+        uw = sum(uwarr)
+        if uw!=6665 txt.print("error sum3\n")
+        ww = sum(warr)
+        if ww!=-4223 txt.print("error sum4\n")
+        ff = sum(farr)
+        if ff!=-4222.4 txt.print("error sum5\n")
+
+        ; ANY
+        ub = any(ubarr)
+        if ub==0 txt.print("error any1\n")
+        ub = any(barr)
+        if ub==0 txt.print("error any2\n")
+        ub = any(uwarr)
+        if ub==0 txt.print("error any3\n")
+        ub = any(warr)
+        if ub==0 txt.print("error any4\n")
+        ub = any(farr)
+        if ub==0 txt.print("error any5\n")
+
+        ; ALL
+        ub = all(ubarr)
+        if ub==1 txt.print("error all1\n")
+        ub = all(barr)
+        if ub==1 txt.print("error all2\n")
+        ub = all(uwarr)
+        if ub==1 txt.print("error all3\n")
+        ub = all(warr)
+        if ub==1 txt.print("error all4\n")
+        ub = all(farr)
+        if ub==1 txt.print("error all5\n")
+        ubarr[1]=$40
+        barr[1]=$40
+        uwarr[1]=$4000
+        warr[1]=$4000
+        farr[1]=1.1
+        ub = all(ubarr)
+        if ub==0 txt.print("error all6\n")
+        ub = all(barr)
+        if ub==0 txt.print("error all7\n")
+        ub = all(uwarr)
+        if ub==0 txt.print("error all8\n")
+        ub = all(warr)
+        if ub==0 txt.print("error all9\n")
+        ub = all(farr)
+        if ub==0 txt.print("error all10\n")
+
+        txt.print("\nyou should see no errors printed above (only at first run).")
+    }
+}

examples/cx16/arithmetic/bitshift.p8

@@ -0,0 +1,924 @@
+%target cx16
+%import cx16textio
+%zeropage basicsafe
+
+
+main {
+
+    sub start() {
+        ubyte A
+
+        txt.print("ubyte shift left\n")
+        A = shiftlb0()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftlb1()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftlb2()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftlb3()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftlb4()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftlb5()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftlb6()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftlb7()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftlb8()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftlb9()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        txt.print("enter to continue:\n")
+        void c64.CHRIN()
+
+        txt.print("ubyte shift right\n")
+        A = shiftrb0()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftrb1()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftrb2()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftrb3()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftrb4()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftrb5()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftrb6()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftrb7()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftrb8()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        A = shiftrb9()
+        txt.print_ubbin(A, true)
+        c64.CHROUT('\n')
+        txt.print("enter to continue:\n")
+        void c64.CHRIN()
+
+
+
+
+        txt.print("signed byte shift left\n")
+        byte signedb
+        signedb = shiftlsb0()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftlsb1()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftlsb2()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftlsb3()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftlsb4()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftlsb5()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftlsb6()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftlsb7()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftlsb8()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftlsb9()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        txt.print("enter to continue:\n")
+        void c64.CHRIN()
+
+        txt.print("signed byte shift right\n")
+        signedb = shiftrsb0()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftrsb1()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftrsb2()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftrsb3()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftrsb4()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftrsb5()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftrsb6()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftrsb7()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftrsb8()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        signedb = shiftrsb9()
+        txt.print_ubbin(signedb as ubyte, true)
+        c64.CHROUT('\n')
+        txt.print("enter to continue:\n")
+        void c64.CHRIN()
+
+
+
+
+        txt.print("uword shift left\n")
+        uword uw
+        uw = shiftluw0()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw1()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw2()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw3()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw4()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw5()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw6()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw7()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw8()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw9()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw10()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw11()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw12()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw13()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw14()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw15()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw16()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftluw17()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        txt.print("enter to continue:\n")
+        void c64.CHRIN()
+
+        txt.print("uword shift right\n")
+        uw = shiftruw0()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw1()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw2()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw3()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw4()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw5()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw6()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw7()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw8()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw9()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw10()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw11()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw12()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw13()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw14()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw15()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw16()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        uw = shiftruw17()
+        txt.print_uwbin(uw, true)
+        c64.CHROUT('\n')
+        txt.print("enter to continue:\n")
+        void c64.CHRIN()
+
+        txt.print("signed word shift left\n")
+        word sw
+        sw = shiftlsw0()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw1()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw2()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw3()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw4()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw5()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw6()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw7()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw8()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw9()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw10()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw11()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw12()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw13()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw14()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw15()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw16()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftlsw17()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        txt.print("enter to continue:\n")
+        void c64.CHRIN()
+
+        txt.print("signed word shift right\n")
+        sw = shiftrsw0()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw1()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw2()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw3()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw4()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw5()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw6()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw7()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw8()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw9()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw10()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw11()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw12()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw13()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw14()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw15()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw16()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+        sw = shiftrsw17()
+        txt.print_uwbin(sw as uword, true)
+        c64.CHROUT('\n')
+
+    }
+
+    sub shiftruw0() -> uword {
+        uword q = $a49f
+        return q >> 0
+    }
+
+    sub shiftruw1() -> uword {
+        uword q = $a49f
+        return q >> 1
+    }
+
+    sub shiftruw2() -> uword {
+        uword q = $a49f
+        return q >> 2
+    }
+
+    sub shiftruw3() -> uword {
+        uword q = $a49f
+        return q >> 3
+    }
+
+    sub shiftruw4() -> uword {
+        uword q = $a49f
+        return q >> 4
+    }
+
+    sub shiftruw5() -> uword {
+        uword q = $a49f
+        return q >> 5
+    }
+
+    sub shiftruw6() -> uword {
+        uword q = $a49f
+        return q >> 6
+    }
+
+    sub shiftruw7() -> uword {
+        uword q = $a49f
+        return q >> 7
+    }
+
+    sub shiftruw8() -> uword {
+        uword q = $a49f
+        return (q >> 8)
+    }
+
+    sub shiftruw9() -> uword {
+        uword q = $a49f
+        return (q >> 9)
+    }
+
+    sub shiftruw10() -> uword {
+        uword q = $a49f
+        return (q >> 10)
+    }
+
+    sub shiftruw11() -> uword {
+        uword q = $a49f
+        return (q >> 11)
+    }
+
+    sub shiftruw12() -> uword {
+        uword q = $a49f
+        return (q >> 12)
+    }
+
+    sub shiftruw13() -> uword {
+        uword q = $a49f
+        return (q >> 13)
+    }
+
+    sub shiftruw14() -> uword {
+        uword q = $a49f
+        return (q >> 14)
+    }
+
+    sub shiftruw15() -> uword {
+        uword q = $a49f
+        return (q >> 15)
+    }
+
+    sub shiftruw16() -> uword {
+        uword q = $a49f
+        return (q >> 16)
+    }
+
+    sub shiftruw17() -> uword {
+        uword q = $a49f
+        return (q >> 17)
+    }
+
+
+
+
+
+    sub shiftrsw0() -> word {
+        word q = -12345
+        return q >> 0
+    }
+
+    sub shiftrsw1() -> word {
+        word q = -12345
+        return q >> 1
+    }
+
+    sub shiftrsw2() -> word {
+        word q = -12345
+        return q >> 2
+    }
+
+    sub shiftrsw3() -> word {
+        word q = -12345
+        return q >> 3
+    }
+
+    sub shiftrsw4() -> word {
+        word q = -12345
+        return q >> 4
+    }
+
+    sub shiftrsw5() -> word {
+        word q = -12345
+        return q >> 5
+    }
+
+    sub shiftrsw6() -> word {
+        word q = -12345
+        return q >> 6
+    }
+
+    sub shiftrsw7() -> word {
+        word q = -12345
+        return q >> 7
+    }
+
+    sub shiftrsw8() -> word {
+        word q = -12345
+        return (q >> 8)
+    }
+
+    sub shiftrsw9() -> word {
+        word q = -12345
+        return (q >> 9)
+    }
+
+    sub shiftrsw10() -> word {
+        word q = -12345
+        return (q >> 10)
+    }
+
+    sub shiftrsw11() -> word {
+        word q = -12345
+        return (q >> 11)
+    }
+
+    sub shiftrsw12() -> word {
+        word q = -12345
+        return (q >> 12)
+    }
+
+    sub shiftrsw13() -> word {
+        word q = -12345
+        return (q >> 13)
+    }
+
+    sub shiftrsw14() -> word {
+        word q = -12345
+        return (q >> 14)
+    }
+
+    sub shiftrsw15() -> word {
+        word q = -12345
+        return (q >> 15)
+    }
+
+    sub shiftrsw16() -> word {
+        word q = -12345
+        return (q >> 16)
+    }
+
+    sub shiftrsw17() -> word {
+        word q = -12345
+        return (q >> 17)
+    }
+
+
+
+
+    sub shiftluw0() -> uword {
+        uword q = $a49f
+        return q << 0
+    }
+
+    sub shiftluw1() -> uword {
+        uword q = $a49f
+        return q << 1
+    }
+
+    sub shiftluw2() -> uword {
+        uword q = $a49f
+        return q << 2
+    }
+
+    sub shiftluw3() -> uword {
+        uword q = $a49f
+        return q << 3
+    }
+
+    sub shiftluw4() -> uword {
+        uword q = $a49f
+        return q << 4
+    }
+
+    sub shiftluw5() -> uword {
+        uword q = $a49f
+        return q << 5
+    }
+
+    sub shiftluw6() -> uword {
+        uword q = $a49f
+        return q << 6
+    }
+
+    sub shiftluw7() -> uword {
+        uword q = $a49f
+        return q << 7
+    }
+
+    sub shiftluw8() -> uword {
+        uword q = $a49f
+        return q << 8
+    }
+
+    sub shiftluw9() -> uword {
+        uword q = $a49f
+        return q << 9
+    }
+
+    sub shiftluw10() -> uword {
+        uword q = $a49f
+        return q << 10
+    }
+
+    sub shiftluw11() -> uword {
+        uword q = $a49f
+        return q << 11
+    }
+
+    sub shiftluw12() -> uword {
+        uword q = $a49f
+        return q << 12
+    }
+
+    sub shiftluw13() -> uword {
+        uword q = $a49f
+        return q << 13
+    }
+
+    sub shiftluw14() -> uword {
+        uword q = $a49f
+        return q << 14
+    }
+
+    sub shiftluw15() -> uword {
+        uword q = $a49f
+        return q << 15
+    }
+
+    sub shiftluw16() -> uword {
+        uword q = $a49f
+        return q << 16
+    }
+
+    sub shiftluw17() -> uword {
+        uword q = $a49f
+        return q << 17
+    }
+
+
+
+    sub shiftlsw0() -> word {
+        word q = -12345
+        return q << 0
+    }
+
+    sub shiftlsw1() -> word {
+        word q = -12345
+        return q << 1
+    }
+
+    sub shiftlsw2() -> word {
+        word q = -12345
+        return q << 2
+    }
+
+    sub shiftlsw3() -> word {
+        word q = -12345
+        return q << 3
+    }
+
+    sub shiftlsw4() -> word {
+        word q = -12345
+        return q << 4
+    }
+
+    sub shiftlsw5() -> word {
+        word q = -12345
+        return q << 5
+    }
+
+    sub shiftlsw6() -> word {
+        word q = -12345
+        return q << 6
+    }
+
+    sub shiftlsw7() -> word {
+        word q = -12345
+        return q << 7
+    }
+
+    sub shiftlsw8() -> word {
+        word q = -12345
+        return q << 8
+    }
+
+    sub shiftlsw9() -> word {
+        word q = -12345
+        return q << 9
+    }
+
+    sub shiftlsw10() -> word {
+        word q = -12345
+        return q << 10
+    }
+
+    sub shiftlsw11() -> word {
+        word q = -12345
+        return q << 11
+    }
+
+    sub shiftlsw12() -> word {
+        word q = -12345
+        return q << 12
+    }
+
+    sub shiftlsw13() -> word {
+        word q = -12345
+        return q << 13
+    }
+
+    sub shiftlsw14() -> word {
+        word q = -12345
+        return q << 14
+    }
+
+    sub shiftlsw15() -> word {
+        word q = -12345
+        return q << 15
+    }
+
+    sub shiftlsw16() -> word {
+        word q = -12345
+        return q << 16
+    }
+
+    sub shiftlsw17() -> word {
+        word q = -12345
+        return q << 17
+    }
+
+
+
+    sub shiftlb0() -> ubyte {
+        ubyte yy=$ed
+        return yy << 0
+    }
+    sub shiftlb1() -> ubyte {
+        ubyte yy=$ed
+        return yy << 1
+    }
+    sub shiftlb2() -> ubyte {
+        ubyte yy=$ed
+        return yy << 2
+    }
+    sub shiftlb3() -> ubyte {
+        ubyte yy=$ed
+        return yy << 3
+    }
+    sub shiftlb4() -> ubyte {
+        ubyte yy=$ed
+        return yy << 4
+    }
+    sub shiftlb5() -> ubyte {
+        ubyte yy=$ed
+        return yy << 5
+    }
+    sub shiftlb6() -> ubyte {
+        ubyte yy=$ed
+        return yy << 6
+    }
+    sub shiftlb7() -> ubyte {
+        ubyte yy=$ed
+        return yy << 7
+    }
+    sub shiftlb8() -> ubyte {
+        ubyte yy=$ed
+        return yy << 8
+    }
+    sub shiftlb9() -> ubyte {
+        ubyte yy=$ed
+        return yy << 9
+    }
+
+    sub shiftrb0() -> ubyte {
+        ubyte yy=$ed
+        return yy >> 0
+    }
+    sub shiftrb1() -> ubyte {
+        ubyte yy=$ed
+        return yy >> 1
+    }
+    sub shiftrb2() -> ubyte {
+        ubyte yy=$ed
+        return yy >> 2
+    }
+    sub shiftrb3() -> ubyte {
+        ubyte yy=$ed
+        return yy >> 3
+    }
+    sub shiftrb4() -> ubyte {
+        ubyte yy=$ed
+        return yy >> 4
+    }
+    sub shiftrb5() -> ubyte {
+        ubyte yy=$ed
+        return yy >> 5
+    }
+    sub shiftrb6() -> ubyte {
+        ubyte yy=$ed
+        return yy >> 6
+    }
+    sub shiftrb7() -> ubyte {
+        ubyte yy=$ed
+        return yy >> 7
+    }
+    sub shiftrb8() -> ubyte {
+        ubyte yy=$ed
+        return yy >> 8
+    }
+    sub shiftrb9() -> ubyte {
+        ubyte yy=$ed
+        return yy >> 9
+    }
+
+
+
+    sub shiftlsb0() -> byte {
+        byte yy=-123
+        return yy << 0
+    }
+    sub shiftlsb1() -> byte {
+        byte yy=-123
+        return yy << 1
+    }
+    sub shiftlsb2() -> byte {
+        byte yy=-123
+        return yy << 2
+    }
+    sub shiftlsb3() -> byte {
+        byte yy=-123
+        return yy << 3
+    }
+    sub shiftlsb4() -> byte {
+        byte yy=-123
+        return yy << 4
+    }
+    sub shiftlsb5() -> byte {
+        byte yy=-123
+        return yy << 5
+    }
+    sub shiftlsb6() -> byte {
+        byte yy=-123
+        return yy << 6
+    }
+    sub shiftlsb7() -> byte {
+        byte yy=-123
+        return yy << 7
+    }
+    sub shiftlsb8() -> byte {
+        byte yy=-123
+        return yy << 8
+    }
+    sub shiftlsb9() -> byte {
+        byte yy=-123
+        return yy << 9
+    }
+
+    sub shiftrsb0() -> byte {
+        byte yy=-123
+        return yy >> 0
+    }
+    sub shiftrsb1() -> byte {
+        byte yy=-123
+        return yy >> 1
+    }
+    sub shiftrsb2() -> byte {
+        byte yy=-123
+        return yy >> 2
+    }
+    sub shiftrsb3() -> byte {
+        byte yy=-123
+        return yy >> 3
+    }
+    sub shiftrsb4() -> byte {
+        byte yy=-123
+        return yy >> 4
+    }
+    sub shiftrsb5() -> byte {
+        byte yy=-123
+        return yy >> 5
+    }
+    sub shiftrsb6() -> byte {
+        byte yy=-123
+        return yy >> 6
+    }
+    sub shiftrsb7() -> byte {
+        byte yy=-123
+        return yy >> 7
+    }
+    sub shiftrsb8() -> byte {
+        byte yy=-123
+        return yy >> 8
+    }
+    sub shiftrsb9() -> byte {
+        byte yy=-123
+        return yy >> 9
+    }
+}

examples/cx16/arithmetic/div.p8

@@ -0,0 +1,104 @@
+%target cx16
+%import cx16flt
+%import cx16textio
+%zeropage basicsafe
+
+main {
+
+    sub start() {
+        div_ubyte(0, 1, 0)
+        div_ubyte(100, 6, 16)
+        div_ubyte(255, 2, 127)
+
+        div_byte(0, 1, 0)
+        div_byte(100, -6, -16)
+        div_byte(127, -2, -63)
+
+        div_uword(0,1,0)
+        div_uword(40000,500,80)
+        div_uword(43211,2,21605)
+
+        div_word(0,1,0)
+        div_word(-20000,500,-40)
+        div_word(-2222,2,-1111)
+
+        div_float(0,1,0)
+        div_float(999.9,111.0,9.008108108108107)
+    }
+
+    sub div_ubyte(ubyte a1, ubyte a2, ubyte c) {
+        ubyte r = a1/a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("ubyte ")
+        txt.print_ub(a1)
+        txt.print(" / ")
+        txt.print_ub(a2)
+        txt.print(" = ")
+        txt.print_ub(r)
+        c64.CHROUT('\n')
+    }
+
+    sub div_byte(byte a1, byte a2, byte c) {
+        byte r = a1/a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("byte ")
+        txt.print_b(a1)
+        txt.print(" / ")
+        txt.print_b(a2)
+        txt.print(" = ")
+        txt.print_b(r)
+        c64.CHROUT('\n')
+    }
+
+    sub div_uword(uword a1, uword  a2, uword c) {
+        uword  r = a1/a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("uword ")
+        txt.print_uw(a1)
+        txt.print(" / ")
+        txt.print_uw(a2)
+        txt.print(" = ")
+        txt.print_uw(r)
+        c64.CHROUT('\n')
+    }
+
+    sub div_word(word a1, word a2, word c) {
+        word r = a1/a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("word ")
+        txt.print_w(a1)
+        txt.print(" / ")
+        txt.print_w(a2)
+        txt.print(" = ")
+        txt.print_w(r)
+        c64.CHROUT('\n')
+    }
+
+    sub div_float(float  a1, float a2, float  c) {
+        float r = a1/a2
+        if abs(r-c)<0.00001
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+
+        txt.print("float ")
+        c64flt.print_f(a1)
+        txt.print(" / ")
+        c64flt.print_f(a2)
+        txt.print(" = ")
+        c64flt.print_f(r)
+        c64.CHROUT('\n')
+    }
+}

examples/cx16/arithmetic/minus.p8

@@ -0,0 +1,112 @@
+%target cx16
+%import cx16flt
+%import cx16textio
+%zeropage basicsafe
+
+main {
+
+    sub start() {
+        minus_ubyte(0, 0, 0)
+        minus_ubyte(200, 0, 200)
+        minus_ubyte(200, 100, 100)
+        minus_ubyte(100, 200, 156)
+
+        minus_byte(0, 0, 0)
+        minus_byte(100, 100, 0)
+        minus_byte(50, -50, 100)
+        minus_byte(0, -30, 30)
+        minus_byte(-30, 0, -30)
+
+        minus_uword(0,0,0)
+        minus_uword(50000,0, 50000)
+        minus_uword(50000,20000,30000)
+        minus_uword(20000,50000,35536)
+
+        minus_word(0,0,0)
+        minus_word(1000,1000,0)
+        minus_word(-1000,1000,-2000)
+        minus_word(1000,500,500)
+        minus_word(0,-3333,3333)
+        minus_word(-3333,0,-3333)
+
+        minus_float(0,0,0)
+        minus_float(2.5,1.5,1.0)
+        minus_float(-1.5,3.5,-5.0)
+    }
+
+    sub minus_ubyte(ubyte a1, ubyte a2, ubyte c) {
+        ubyte r = a1-a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("ubyte ")
+        txt.print_ub(a1)
+        txt.print(" - ")
+        txt.print_ub(a2)
+        txt.print(" = ")
+        txt.print_ub(r)
+        c64.CHROUT('\n')
+    }
+
+    sub minus_byte(byte a1, byte a2, byte c) {
+        byte r = a1-a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("byte ")
+        txt.print_b(a1)
+        txt.print(" - ")
+        txt.print_b(a2)
+        txt.print(" = ")
+        txt.print_b(r)
+        c64.CHROUT('\n')
+    }
+
+    sub minus_uword(uword a1, uword  a2, uword c) {
+        uword  r = a1-a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("uword ")
+        txt.print_uw(a1)
+        txt.print(" - ")
+        txt.print_uw(a2)
+        txt.print(" = ")
+        txt.print_uw(r)
+        c64.CHROUT('\n')
+    }
+
+    sub minus_word(word a1, word a2, word c) {
+        word r = a1-a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("word ")
+        txt.print_w(a1)
+        txt.print(" - ")
+        txt.print_w(a2)
+        txt.print(" = ")
+        txt.print_w(r)
+        c64.CHROUT('\n')
+    }
+
+    sub minus_float(float  a1, float a2, float  c) {
+        float r = a1-a2
+        if abs(r-c)<0.00001
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+
+        txt.print("float ")
+        c64flt.print_f(a1)
+        txt.print(" - ")
+        c64flt.print_f(a2)
+        txt.print(" = ")
+        c64flt.print_f(r)
+        c64.CHROUT('\n')
+    }
+}

examples/cx16/arithmetic/mult.p8

@@ -0,0 +1,106 @@
+%target cx16
+%import cx16flt
+%import cx16textio
+%zeropage basicsafe
+
+main {
+
+    sub start() {
+        mul_ubyte(0, 0, 0)
+        mul_ubyte(20, 1, 20)
+        mul_ubyte(20, 10, 200)
+
+        mul_byte(0, 0, 0)
+        mul_byte(10, 10, 100)
+        mul_byte(5, -5, -25)
+        mul_byte(0, -30, 0)
+
+        mul_uword(0,0,0)
+        mul_uword(50000,1, 50000)
+        mul_uword(500,100,50000)
+
+        mul_word(0,0,0)
+        mul_word(-10,1000,-10000)
+        mul_word(1,-3333,-3333)
+
+        mul_float(0,0,0)
+        mul_float(2.5,10,25)
+        mul_float(-1.5,10,-15)
+    }
+
+    sub mul_ubyte(ubyte a1, ubyte a2, ubyte c) {
+        ubyte r = a1*a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("ubyte ")
+        txt.print_ub(a1)
+        txt.print(" * ")
+        txt.print_ub(a2)
+        txt.print(" = ")
+        txt.print_ub(r)
+        c64.CHROUT('\n')
+    }
+
+    sub mul_byte(byte a1, byte a2, byte c) {
+        byte r = a1*a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("byte ")
+        txt.print_b(a1)
+        txt.print(" * ")
+        txt.print_b(a2)
+        txt.print(" = ")
+        txt.print_b(r)
+        c64.CHROUT('\n')
+    }
+
+    sub mul_uword(uword a1, uword  a2, uword c) {
+        uword  r = a1*a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("uword ")
+        txt.print_uw(a1)
+        txt.print(" * ")
+        txt.print_uw(a2)
+        txt.print(" = ")
+        txt.print_uw(r)
+        c64.CHROUT('\n')
+    }
+
+    sub mul_word(word a1, word a2, word c) {
+        word r = a1*a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("word ")
+        txt.print_w(a1)
+        txt.print(" * ")
+        txt.print_w(a2)
+        txt.print(" = ")
+        txt.print_w(r)
+        c64.CHROUT('\n')
+    }
+
+    sub mul_float(float  a1, float a2, float  c) {
+        float r = a1*a2
+        if abs(r-c)<0.00001
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+
+        txt.print("float ")
+        c64flt.print_f(a1)
+        txt.print(" * ")
+        c64flt.print_f(a2)
+        txt.print(" = ")
+        c64flt.print_f(r)
+        c64.CHROUT('\n')
+    }
+}

examples/cx16/arithmetic/plus.p8

@@ -0,0 +1,110 @@
+%target cx16
+%import cx16flt
+%import cx16textio
+%zeropage basicsafe
+
+main {
+
+    sub start() {
+        plus_ubyte(0, 0, 0)
+        plus_ubyte(0, 200, 200)
+        plus_ubyte(100, 200, 44)
+
+        plus_byte(0, 0, 0)
+        plus_byte(-100, 100, 0)
+        plus_byte(-50, 100, 50)
+        plus_byte(0, -30, -30)
+        plus_byte(-30, 0, -30)
+
+        plus_uword(0,0,0)
+        plus_uword(0,50000,50000)
+        plus_uword(50000,20000,4464)
+
+        plus_word(0,0,0)
+        plus_word(-1000,1000,0)
+        plus_word(-500,1000,500)
+        plus_word(0,-3333,-3333)
+        plus_word(-3333,0,-3333)
+
+        plus_float(0,0,0)
+        plus_float(1.5,2.5,4.0)
+        plus_float(-1.5,3.5,2.0)
+        plus_float(-1.1,3.3,2.2)
+    }
+
+    sub plus_ubyte(ubyte a1, ubyte a2, ubyte c) {
+        ubyte r = a1+a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("ubyte ")
+        txt.print_ub(a1)
+        txt.print(" + ")
+        txt.print_ub(a2)
+        txt.print(" = ")
+        txt.print_ub(r)
+        c64.CHROUT('\n')
+    }
+
+    sub plus_byte(byte a1, byte a2, byte c) {
+        byte r = a1+a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("byte ")
+        txt.print_b(a1)
+        txt.print(" + ")
+        txt.print_b(a2)
+        txt.print(" = ")
+        txt.print_b(r)
+        c64.CHROUT('\n')
+    }
+
+    sub plus_uword(uword a1, uword  a2, uword c) {
+        uword  r = a1+a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("uword ")
+        txt.print_uw(a1)
+        txt.print(" + ")
+        txt.print_uw(a2)
+        txt.print(" = ")
+        txt.print_uw(r)
+        c64.CHROUT('\n')
+    }
+
+    sub plus_word(word a1, word a2, word c) {
+        word r = a1+a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("word ")
+        txt.print_w(a1)
+        txt.print(" + ")
+        txt.print_w(a2)
+        txt.print(" = ")
+        txt.print_w(r)
+        c64.CHROUT('\n')
+    }
+
+    sub plus_float(float  a1, float a2, float  c) {
+        float r = a1+a2
+        if abs(r-c)<0.00001
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+
+        txt.print("float ")
+        c64flt.print_f(a1)
+        txt.print(" + ")
+        c64flt.print_f(a2)
+        txt.print(" = ")
+        c64flt.print_f(r)
+        c64.CHROUT('\n')
+    }
+}

examples/cx16/arithmetic/postincrdecr.p8

@@ -0,0 +1,141 @@
+%target cx16
+%import cx16flt
+%import cx16textio
+%zeropage basicsafe
+
+main {
+
+    sub start()  {
+
+        txt.plot(0,24)
+
+        ubyte Y
+        ubyte ub=200
+        byte bb=-100
+        uword uw = 2000
+        word ww = -1000
+        float fl = 999.99
+        ubyte[3] ubarr = 200
+        byte[3] barr = -100
+        uword[3] uwarr = 2000
+        word[3] warr = -1000
+        float[3] flarr = 999.99
+
+        txt.print("++\n")
+        ub++
+        bb++
+        uw++
+        ww++
+        fl++
+        ubarr[1]++
+        barr[1]++
+        uwarr[1]++
+        warr[1]++
+        flarr[1] ++
+
+        check_ub(ub, 201)
+        Y=100
+        Y++
+        check_ub(Y, 101)
+        check_fl(fl, 1000.99)
+        check_b(bb, -99)
+        check_uw(uw, 2001)
+        check_w(ww, -999)
+        check_ub(ubarr[0], 200)
+        check_fl(flarr[0], 999.99)
+        check_b(barr[0], -100)
+        check_uw(uwarr[0], 2000)
+        check_w(warr[0], -1000)
+        check_ub(ubarr[1], 201)
+        check_fl(flarr[1], 1000.99)
+        check_b(barr[1], -99)
+        check_uw(uwarr[1], 2001)
+        check_w(warr[1], -999)
+
+        txt.print("--\n")
+        ub--
+        bb--
+        uw--
+        ww--
+        fl--
+        ubarr[1]--
+        barr[1]--
+        uwarr[1]--
+        warr[1]--
+        flarr[1] --
+        check_ub(ub, 200)
+
+        Y=100
+        Y--
+        check_ub(Y, 99)
+        check_fl(fl, 999.99)
+        check_b(bb, -100)
+        check_uw(uw, 2000)
+        check_w(ww, -1000)
+        check_ub(ubarr[1], 200)
+        check_fl(flarr[1], 999.99)
+        check_b(barr[1], -100)
+        check_uw(uwarr[1], 2000)
+        check_w(warr[1], -1000)
+    }
+
+    sub check_ub(ubyte value, ubyte expected) {
+        if value==expected
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print(" ubyte ")
+        txt.print_ub(value)
+        c64.CHROUT(',')
+        txt.print_ub(expected)
+        c64.CHROUT('\n')
+    }
+
+    sub check_b(byte value, byte expected) {
+        if value==expected
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print(" byte ")
+        txt.print_b(value)
+        c64.CHROUT(',')
+        txt.print_b(expected)
+        c64.CHROUT('\n')
+    }
+
+    sub check_uw(uword value, uword expected) {
+        if value==expected
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print(" uword ")
+        txt.print_uw(value)
+        c64.CHROUT(',')
+        txt.print_uw(expected)
+        c64.CHROUT('\n')
+    }
+
+    sub check_w(word value, word expected) {
+        if value==expected
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print(" word ")
+        txt.print_w(value)
+        c64.CHROUT(',')
+        txt.print_w(expected)
+        c64.CHROUT('\n')
+    }
+
+    sub check_fl(float value, float expected) {
+        if value==expected
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print(" float ")
+        c64flt.print_f(value)
+        c64.CHROUT(',')
+        c64flt.print_f(expected)
+        c64.CHROUT('\n')
+    }
+}

examples/cx16/arithmetic/remainder.p8

@@ -0,0 +1,48 @@
+%target cx16
+%import cx16textio
+%zeropage basicsafe
+
+main {
+
+    sub start() {
+        remainder_ubyte(0, 1, 0)
+        remainder_ubyte(100, 6, 4)
+        remainder_ubyte(255, 2, 1)
+        remainder_ubyte(255, 20, 15)
+
+        remainder_uword(0,1,0)
+        remainder_uword(40000,511,142)
+        remainder_uword(40000,500,0)
+        remainder_uword(43211,12,11)
+    }
+
+    sub remainder_ubyte(ubyte a1, ubyte a2, ubyte c) {
+        ubyte r = a1%a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("ubyte ")
+        txt.print_ub(a1)
+        txt.print(" % ")
+        txt.print_ub(a2)
+        txt.print(" = ")
+        txt.print_ub(r)
+        c64.CHROUT('\n')
+    }
+
+    sub remainder_uword(uword a1, uword  a2, uword c) {
+        uword  r = a1%a2
+        if r==c
+            txt.print(" ok  ")
+        else
+            txt.print("err! ")
+        txt.print("uword ")
+        txt.print_uw(a1)
+        txt.print(" % ")
+        txt.print_uw(a2)
+        txt.print(" = ")
+        txt.print_uw(r)
+        c64.CHROUT('\n')
+    }
+}

examples/cx16/arithmetic/sgn.p8

@@ -0,0 +1,132 @@
+%target cx16
+%import cx16flt
+%import cx16textio
+%zeropage basicsafe
+
+main {
+
+    sub start() {
+        byte b1
+        byte b2
+        ubyte ub1
+        ubyte ub2
+        word w1
+        word w2
+        uword uw1
+        uword uw2
+        float f1
+        float f2
+
+        b1 = 10
+        b2 = 10
+        if sgn(b2-b1) != 0
+            txt.print("sgn1 error1\n")
+
+        b1 = -100
+        b2 = -100
+        if sgn(b2-b1) != 0
+            txt.print("sgn1 error2\n")
+
+        ub1 = 200
+        ub2 = 200
+        if sgn(ub2-ub1) != 0
+            txt.print("sgn1 error3\n")
+
+        w1 = 100
+        w2 = 100
+        if sgn(w2-w1) != 0
+            txt.print("sgn1 error4\n")
+
+        w1 = -2000
+        w2 = -2000
+        if sgn(w2-w1) != 0
+            txt.print("sgn1 error5\n")
+
+        uw1 = 999
+        uw2 = 999
+        if sgn(uw2-uw1) != 0
+            txt.print("sgn1 error6\n")
+
+        f1 = 3.45
+        f2 = 3.45
+        if sgn(f2-f1) != 0
+            txt.print("sgn1 error7\n")
+
+
+        ; -1
+        b1 = 11
+        b2 = 10
+        if sgn(b2-b1) != -1
+            txt.print("sgn2 error1\n")
+
+        b1 = -10
+        b2 = -100
+        if sgn(b2-b1) != -1
+            txt.print("sgn2 error2\n")
+
+        ub1 = 202
+        ub2 = 200
+        if sgn(ub2 as byte - ub1 as byte) != -1
+            txt.print("sgn2 error3\n")
+
+        w1 = 101
+        w2 = 100
+        if sgn(w2-w1) != -1
+            txt.print("sgn2 error4\n")
+
+        w1 = -200
+        w2 = -2000
+        if sgn(w2-w1) != -1
+            txt.print("sgn2 error5\n")
+
+        uw1 = 2222
+        uw2 = 999
+        if sgn((uw2 as word) - (uw1 as word)) != -1
+            txt.print("sgn2 error6a\n")
+        if sgn(uw2 - uw1) != 1      ; always 0 or 1 if unsigned
+            txt.print("sgn2 error6b\n")
+
+        f1 = 3.45
+        f2 = 1.11
+        if sgn(f2-f1) != -1
+            txt.print("sgn2 error7\n")
+
+        ; +1
+        b1 = 11
+        b2 = 20
+        if sgn(b2-b1) != 1
+            txt.print("sgn3 error1\n")
+
+        b1 = -10
+        b2 = -1
+        if sgn(b2-b1) != 1
+            txt.print("sgn3 error2\n")
+
+        ub1 = 202
+        ub2 = 205
+        if sgn(ub2-ub1) != 1
+            txt.print("sgn3 error3\n")
+
+        w1 = 101
+        w2 = 200
+        if sgn(w2-w1) != 1
+            txt.print("sgn3 error4\n")
+
+        w1 = -200
+        w2 = -20
+        if sgn(w2-w1) != 1
+            txt.print("sgn3 error5\n")
+
+        uw1 = 2222
+        uw2 = 9999
+        if sgn(uw2-uw1) != 1
+            txt.print("sgn3 error6\n")
+
+        f1 = 3.45
+        f2 = 5.11
+        if sgn(f2-f1) != 1
+            txt.print("sgn3 error7\n")
+
+        txt.print("should see no sgn errors\n")
+    }
+ }

examples/cx16/cobramk3-gfx.p8

@@ -0,0 +1,211 @@
+%target cx16
+%import cx16lib
+%import conv
+
+main {
+
+    ; storage for rotated coordinates
+    word[shipdata.totalNumberOfPoints] rotatedx
+    word[shipdata.totalNumberOfPoints] rotatedy
+    word[shipdata.totalNumberOfPoints] rotatedz
+
+
+    ; TODO hidden surface removal - this needs surfaces data and surface normal calculations.
+    ; TODO add all other Elite's ships, show their name, advance to next ship on keypress
+
+    sub start()  {
+        uword anglex
+        uword angley
+        uword anglez
+
+        void cx16.screen_set_mode($80)
+        cx16.r0 = 0
+        cx16.GRAPH_init()
+        cx16.GRAPH_set_colors(13, 6, 6)
+        cx16.GRAPH_clear()
+        print_ship_name()
+
+
+        repeat {
+            rotate_vertices(msb(anglex), msb(angley), msb(anglez))
+
+            cx16.GRAPH_set_colors(0, 0, 0)
+            cx16.r0 = 32
+            cx16.r1 = 10
+            cx16.r2 = 256
+            cx16.r3 = 185
+            cx16.r4 = 0
+            cx16.GRAPH_draw_rect(true)
+
+            cx16.GRAPH_set_colors(1, 0, 0)
+            draw_lines()
+
+            anglex -= 505
+            angley += 217
+            anglez += 452
+        }
+    }
+
+    sub print_ship_name() {
+        cx16.r0 = 32
+        cx16.r1 = 8
+        ubyte c
+        for c in "ship: "
+            cx16.GRAPH_put_char(c)
+        for c in shipdata.shipName
+            cx16.GRAPH_put_char(c)
+
+        cx16.r0 += 16
+        print_number_gfx(shipdata.totalNumberOfPoints)
+        for c in " vertices, "
+            cx16.GRAPH_put_char(c)
+        print_number_gfx(shipdata.totalNumberOfEdges)
+        for c in " edges, "
+            cx16.GRAPH_put_char(c)
+        print_number_gfx(shipdata.totalNumberOfFaces)
+        for c in " faces"
+            cx16.GRAPH_put_char(c)
+    }
+
+    asmsub print_number_gfx(ubyte num @ A) clobbers(A,Y) {
+    	%asm {{
+    		phx
+    		jsr  conv.ubyte2decimal
+    		phx
+    		pha
+    		cpy  #'0'
+    		beq  +
+    		tya
+    		jsr  cx16.GRAPH_put_char
+    		pla
+    		jsr  cx16.GRAPH_put_char
+    		jmp  _ones
+    +       pla
+            cmp  #'0'
+            beq  _ones
+            jsr  cx16.GRAPH_put_char
+    _ones   pla
+    		jsr  cx16.GRAPH_put_char
+    		plx
+    		rts
+    	}}
+    }
+
+    sub rotate_vertices(ubyte ax, ubyte ay, ubyte az) {
+        ; rotate around origin (0,0,0)
+
+        ; set up the 3d rotation matrix values
+        word wcosa = cos8(ax)
+        word wsina = sin8(ax)
+        word wcosb = cos8(ay)
+        word wsinb = sin8(ay)
+        word wcosc = cos8(az)
+        word wsinc = sin8(az)
+
+        word wcosa_sinb = wcosa*wsinb / 128
+        word wsina_sinb = wsina*wsinb / 128
+
+        word Axx = wcosa*wcosb / 128
+        word Axy = (wcosa_sinb*wsinc - wsina*wcosc) / 128
+        word Axz = (wcosa_sinb*wcosc + wsina*wsinc) / 128
+        word Ayx = wsina*wcosb / 128
+        word Ayy = (wsina_sinb*wsinc + wcosa*wcosc) / 128
+        word Ayz = (wsina_sinb*wcosc - wcosa*wsinc) / 128
+        word Azx = -wsinb
+        word Azy = wcosb*wsinc / 128
+        word Azz = wcosb*wcosc / 128
+
+        ubyte @zp i
+        for i in 0 to shipdata.totalNumberOfPoints-1 {
+            ; don't normalize by dividing by 128, instead keep some precision for perspective calc later
+            rotatedx[i] = Axx*shipdata.xcoor[i] + Axy*shipdata.ycoor[i] + Axz*shipdata.zcoor[i]
+            rotatedy[i] = Ayx*shipdata.xcoor[i] + Ayy*shipdata.ycoor[i] + Ayz*shipdata.zcoor[i]
+            rotatedz[i] = Azx*shipdata.xcoor[i] + Azy*shipdata.ycoor[i] + Azz*shipdata.zcoor[i]
+        }
+    }
+
+    const uword screen_width = 320
+    const ubyte screen_height = 200
+
+    sub draw_lines() {
+        ubyte @zp i
+        for i in shipdata.totalNumberOfEdges -1 downto 0 {
+            ubyte @zp vFrom = shipdata.edgesFrom[i]
+            ubyte @zp vTo = shipdata.edgesTo[i]
+            word persp1 = 200 + rotatedz[vFrom]/256
+            word persp2 = 200 + rotatedz[vTo]/256
+            cx16.r0 = rotatedx[vFrom] / persp1 + screen_width/2 as uword
+            cx16.r1 = rotatedy[vFrom] / persp1 + screen_height/2 as uword
+            cx16.r2 = rotatedx[vTo] / persp2 + screen_width/2 as uword
+            cx16.r3 = rotatedy[vTo] / persp2 + screen_height/2 as uword
+            cx16.GRAPH_draw_line()
+        }
+    }
+}
+
+shipdata {
+
+    ; Ship model data converted from BBC Elite's Cobra MK 3
+    ; downloaded from http://www.elitehomepage.org/archive/index.htm
+
+const ubyte totalNumberOfEdges = 51
+const ubyte totalNumberOfFaces = 22
+const ubyte totalNumberOfPoints = 34
+str shipName = "cobra-mk3"
+; vertices
+word[] xcoor = [ 32,-32,0,-120,120,-88,88,128,-128,0,-32,32,-36,-8,8,36,36,8,-8,-36,-1,-1,-80,-80,-88,80,88,80,1,1,1,1,-1,-1 ]
+word[] ycoor = [ 0,0,26,-3,-3,16,16,-8,-8,26,-24,-24,8,12,12,8,-12,-16,-16,-12,-1,-1,-6,6,0,6,0,-6,-1,-1,1,1,1,1 ]
+word[] zcoor = [ 76,76,24,-8,-8,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,-40,76,90,-40,-40,-40,-40,-40,-40,76,90,76,90,76,90 ]
+; edges and faces
+ubyte[] edgesFrom = [ 0,1,0,10,1,0,2,0,4,0,4,7,2,1,1,3,8,3,2,5,6,5,6,16,15,14,14,18,13,12,12,26,25,25,22,23,22,20,28,21,20,28,29,30,31,30,32,20,21,20,20 ]
+ubyte[] edgesTo = [ 1,2,2,11,10,11,6,6,6,4,7,11,5,5,3,5,10,8,9,9,9,8,7,17,16,15,17,19,18,13,19,27,26,27,23,24,24,28,29,29,21,30,31,31,33,32,33,32,33,33,29 ]
+ubyte[] facesPoints = [
+     0,1,2 ,255,
+     11,10,1,0 ,255,
+     0,2,6 ,255,
+     6,4,0 ,255,
+     4,7,11,0 ,255,
+     5,2,1 ,255,
+     1,3,5 ,255,
+     10,8,3,1 ,255,
+     9,2,5 ,255,
+     9,6,2 ,255,
+     3,8,5 ,255,
+     4,6,7 ,255,
+     5,8,10,11,7,6,9 ,255,
+     17,16,15,14 ,255,
+     19,18,13,12 ,255,
+     27,26,25 ,255,
+     22,23,24 ,255,
+     20,28,29,21 ,255,
+     30,28,29,31 ,255,
+     33,31,30,32 ,255,
+     20,32,33,21 ,255,
+     29,31,33,20 ,255
+]
+ubyte[] facesEdges = [
+     0,1,2 ,255,
+     3,4,0,5 ,255,
+     2,6,7 ,255,
+     8,9,7 ,255,
+     10,11,5,9 ,255,
+     12,1,13 ,255,
+     14,15,13 ,255,
+     16,17,14,4 ,255,
+     18,12,19 ,255,
+     20,6,18 ,255,
+     17,21,15 ,255,
+     8,22,10 ,255,
+     21,16,3,11,22,20,19 ,255,
+     23,24,25,26 ,255,
+     27,28,29,30 ,255,
+     31,32,33 ,255,
+     34,35,36 ,255,
+     37,38,39,40 ,255,
+     41,38,42,43 ,255,
+     44,43,45,46 ,255,
+     47,46,48,40 ,255,
+     42,44,49,50 ,255
+]
+
+}

examples/cx16/cube3d-float.p8

@@ -0,0 +1,110 @@
+%target cx16
+%import cx16flt
+%import cx16textio
+%zeropage basicsafe
+
+main {
+
+    const uword width = 80
+    const uword height = 60
+
+    ; vertices
+    float[] xcoor = [ -1.0, -1.0, -1.0, -1.0,  1.0,  1.0,  1.0, 1.0 ]
+    float[] ycoor = [ -1.0, -1.0,  1.0,  1.0, -1.0, -1.0,  1.0, 1.0 ]
+    float[] zcoor = [ -1.0,  1.0, -1.0,  1.0, -1.0,  1.0, -1.0, 1.0 ]
+
+    ; storage for rotated coordinates
+    float[len(xcoor)] rotatedx=0.0
+    float[len(ycoor)] rotatedy=0.0
+    float[len(zcoor)] rotatedz=-1.0
+
+    sub start()  {
+        float time=0.0
+        ubyte timer_jiffies
+
+        repeat {
+            rotate_vertices(time)
+            txt.clear_screenchars(' ')
+            draw_edges()
+            time+=0.1
+
+            txt.plot(0,0)
+            txt.print("3d cube! (floating point calc) ")
+
+            %asm {{
+                phx
+                jsr  c64.RDTIM      ; A/X/Y
+                sta  timer_jiffies
+                lda  #0
+                jsr  c64.SETTIM
+                plx
+            }}
+            txt.print_ub(timer_jiffies)
+            txt.print(" jiffies/fr = ")
+            txt.print_ub(60/timer_jiffies)
+            txt.print(" fps")
+        }
+    }
+
+    sub rotate_vertices(float t) {
+        ; rotate around origin (0,0,0)
+
+        ; set up the 3d rotation matrix values
+        float cosa = cos(t)
+        float sina = sin(t)
+        float cosb = cos(t*0.33)
+        float sinb = sin(t*0.33)
+        float cosc = cos(t*0.78)
+        float sinc = sin(t*0.78)
+
+        float cosa_sinb = cosa*sinb
+        float sina_sinb = sina*sinb
+        float Axx = cosa*cosb
+        float Axy = cosa_sinb*sinc - sina*cosc
+        float Axz = cosa_sinb*cosc + sina*sinc
+        float Ayx = sina*cosb
+        float Ayy = sina_sinb*sinc + cosa*cosc
+        float Ayz = sina_sinb*cosc - cosa*sinc
+        float Azx = -sinb
+        float Azy = cosb*sinc
+        float Azz = cosb*cosc
+
+        ubyte @zp i
+        for i in 0 to len(xcoor)-1 {
+            rotatedx[i] = Axx*xcoor[i] + Axy*ycoor[i] + Axz*zcoor[i]
+            rotatedy[i] = Ayx*xcoor[i] + Ayy*ycoor[i] + Ayz*zcoor[i]
+            rotatedz[i] = Azx*xcoor[i] + Azy*ycoor[i] + Azz*zcoor[i]
+        }
+    }
+
+    sub draw_edges() {
+
+        ; plot the points of the 3d cube
+        ; first the points on the back, then the points on the front (painter algorithm)
+        ubyte @zp i
+        float rz
+        float persp
+        ubyte sx
+        ubyte sy
+
+        for i in 0 to len(xcoor)-1 {
+            rz = rotatedz[i]
+            if rz >= 0.1 {
+                persp = (5.0+rz)/(height as float)
+                sx = rotatedx[i] / persp + width/2.0 as ubyte
+                sy = rotatedy[i] / persp + height/2.0 as ubyte
+                txt.setcc(sx, sy, 46, 1)
+            }
+        }
+
+        for i in 0 to len(xcoor)-1 {
+            rz = rotatedz[i]
+            if rz < 0.1 {
+                persp = (5.0+rz)/(height as float)
+                sx = rotatedx[i] / persp + width/2.0 as ubyte
+                sy = rotatedy[i] / persp + height/2.0 as ubyte
+                txt.setcc(sx, sy, 81, 1)
+            }
+        }
+    }
+}

examples/cx16/cube3d.p8

@@ -0,0 +1,100 @@
+%target cx16
+%import cx16textio
+
+main {
+
+    const uword screen_width = 80
+    const uword screen_height = 60
+
+    ; vertices
+    word[] xcoor = [ -40, -40, -40, -40,  40,  40,  40, 40 ]
+    word[] ycoor = [ -40, -40,  40,  40, -40, -40,  40, 40 ]
+    word[] zcoor = [ -40,  40, -40,  40, -40,  40, -40, 40 ]
+
+    ; storage for rotated coordinates
+    word[len(xcoor)] rotatedx
+    word[len(ycoor)] rotatedy
+    word[len(zcoor)] rotatedz
+
+    sub start()  {
+
+        uword anglex
+        uword angley
+        uword anglez
+        ubyte timer_jiffies
+
+        repeat {
+            rotate_vertices(msb(anglex), msb(angley), msb(anglez))
+            txt.clear_screenchars(' ')
+            draw_edges()
+            anglex+=500
+            angley+=215
+            anglez+=453
+        }
+    }
+
+    sub rotate_vertices(ubyte ax, ubyte ay, ubyte az) {
+        ; rotate around origin (0,0,0)
+
+        ; set up the 3d rotation matrix values
+        word wcosa = cos8(ax)
+        word wsina = sin8(ax)
+        word wcosb = cos8(ay)
+        word wsinb = sin8(ay)
+        word wcosc = cos8(az)
+        word wsinc = sin8(az)
+
+        word wcosa_sinb = wcosa*wsinb / 128
+        word wsina_sinb = wsina*wsinb / 128
+
+        word Axx = wcosa*wcosb / 128
+        word Axy = (wcosa_sinb*wsinc - wsina*wcosc) / 128
+        word Axz = (wcosa_sinb*wcosc + wsina*wsinc) / 128
+        word Ayx = wsina*wcosb / 128
+        word Ayy = (wsina_sinb*wsinc + wcosa*wcosc) / 128
+        word Ayz = (wsina_sinb*wcosc - wcosa*wsinc) / 128
+        word Azx = -wsinb
+        word Azy = wcosb*wsinc / 128
+        word Azz = wcosb*wcosc / 128
+
+        ubyte @zp i
+        for i in 0 to len(xcoor)-1 {
+            ; don't normalize by dividing by 128, instead keep some precision for perspective calc later
+            rotatedx[i] = Axx*xcoor[i] + Axy*ycoor[i] + Axz*zcoor[i]
+            rotatedy[i] = Ayx*xcoor[i] + Ayy*ycoor[i] + Ayz*zcoor[i]
+            rotatedz[i] = Azx*xcoor[i] + Azy*ycoor[i] + Azz*zcoor[i]
+        }
+    }
+
+    sub draw_edges() {
+
+        ; plot the points of the 3d cube
+        ; first the points on the back, then the points on the front (painter algorithm)
+
+        ubyte @zp i
+        word @zp rz
+        word @zp persp
+        byte sx
+        byte sy
+
+        for i in 0 to len(xcoor)-1 {
+            rz = rotatedz[i]
+            if rz >= 10 {
+                persp = 500 + rz/64
+                sx = rotatedx[i] / persp as byte + screen_width/2
+                sy = rotatedy[i] / persp as byte + screen_height/2
+                txt.setcc(sx as ubyte, sy as ubyte, 46, 7)
+            }
+        }
+
+        for i in 0 to len(xcoor)-1 {
+            rz = rotatedz[i]
+            if rz < 10 {
+                persp = 500 + rz/64
+                sx = rotatedx[i] / persp as byte + screen_width/2
+                sy = rotatedy[i] / persp as byte + screen_height/2
+                txt.setcc(sx as ubyte, sy as ubyte, 81, 7)
+            }
+        }
+    }
+}

examples/cx16/datetime.p8

@@ -1,6 +1,7 @@
-; CommanderX16 text clock example!
+; CommanderX16 text datetime example!
 ; make sure to compile with the cx16 compiler target.
 
+%target cx16
 %import cx16textio
 %zeropage basicsafe
 
@@ -8,28 +9,6 @@ main {
 
     sub start() {
 
-;        %asm {{
-;            lda  #$80
-;            jsr  cx16.screen_set_mode
-;        }}
-;        cx16.r0=0
-;        cx16.GRAPH_init()
-;        %asm {{
-;            lda  #4
-;            ldy  #0
-;            ldx  #1
-;            jsr  cx16.GRAPH_set_colors
-;        }}
-;        cx16.GRAPH_clear()
-;        cx16.r0=10
-;        cx16.r1=10
-;        cx16.r2=100
-;        cx16.r3=150
-;        cx16.GRAPH_draw_line()
-;
-;        repeat {
-;        }
-
         cx16.r0 = mkword(8, 2020 - 1900)
         cx16.r1 = mkword(19, 27)
         cx16.r2 = mkword(0, 16)

examples/cx16/floats.p8

@@ -1,23 +0,0 @@
-; CommanderX16 floating point example!
-; make sure to compile with the cx16 compiler target.
-
-%import cx16textio
-%import cx16flt
-%zeropage basicsafe
-
-main {
-
-    sub start() {
-        float f1 = 5.55
-        float f2 = 33.3
-        float f3 = f1 * f2
-
-        c64flt.print_f(f1)
-        c64.CHROUT('*')
-        c64flt.print_f(f2)
-        c64.CHROUT('=')
-        c64flt.print_f(f3)
-        c64.CHROUT('\n')
-    }
-}
-

examples/cx16/mandelbrot-gfx.p8

@@ -1,3 +1,4 @@
+%target cx16
 %import cx16textio
 %import cx16flt
 %zeropage basicsafe
@@ -5,7 +6,7 @@
 main {
     const uword width = 256
     const uword height = 200
-    const ubyte max_iter = 16       ; 32 looks pretty nice
+    const ubyte max_iter = 16       ; 32 actually looks pretty nice but takes longer
 
     sub start()  {
         initialize()

examples/cx16/mandelbrot.p8

@@ -1,3 +1,4 @@
+%target cx16
 %import cx16textio
 %import cx16flt
 %zeropage basicsafe

examples/cx16/simplegraphics.p8

@@ -1,6 +1,7 @@
-; CommanderX16 text clock example!
+; CommanderX16 simple graphics example!
 ; make sure to compile with the cx16 compiler target.
 
+%target cx16
 %zeropage basicsafe
 
 main {

examples/cx16/swirl-float.p8

@@ -1,3 +1,4 @@
+%target cx16
 %import cx16textio
 %import cx16flt
 %zeropage basicsafe

examples/cx16/swirl.p8

@@ -1,3 +1,4 @@
+%target cx16
 %import cx16lib
 %import cx16textio
 %zeropage basicsafe

examples/mandelbrot-gfx.p8

@@ -38,9 +38,7 @@ main {
                 }
 
                 if iter & 1
-                    ; TODO fix plot() so we don't have to use separate internal variable
-                    graphics.internal_plotx = pixelx
-                    graphics.internal_plot(pixely)
+                    graphics.plot(pixelx, pixely)
             }
         }
 

examples/tehtriz.p8

@@ -385,9 +385,9 @@ waitkey:
     }
 
     sub drawBlock(ubyte x, ubyte y, ubyte character) {
-        ubyte i
+        ubyte @zp i
         for i in 15 downto 0 {
-            ubyte c=blocklogic.currentBlock[i]
+            ubyte @zp c=blocklogic.currentBlock[i]
             if c
                 txt.setcc((i&3)+x, (i/4)+y, character, c)
         }
@@ -534,7 +534,7 @@ blocklogic {
     }
 
     sub noCollision(ubyte xpos, ubyte ypos) -> ubyte {
-        ubyte i
+        ubyte @zp i
         for i in 15 downto 0 {
             if currentBlock[i] and txt.getchr(xpos + (i&3), ypos+i/4)!=32
                 return false

examples/test.p8

@@ -1,8 +1,12 @@
+%import c64lib
+%import c64graphics
 %import c64textio
+;%import c64flt
+;%option enable_floats
 %zeropage basicsafe
 
+
 main {
-    sub start() {
 
 ;asmsub  clear_screen (ubyte char @ A, ubyte color @ Y) clobbers(A)  { ...}
 ; TODO dont cause name conflict if we define sub or sub with param 'color' or even a var 'color' later.
@@ -10,6 +14,7 @@ main {
 ;   sub color(...) {}
 ;   sub other(ubyte color) {}    ; TODO don't cause name conflict
 
-        c64.CHROUT('\n')
+
+    sub start()  {
     }
 }

parser/antlr/prog8.g4

@@ -126,7 +126,7 @@ unconditionaljump :  'goto'  (integerliteral | scoped_identifier) ;
 
 directive :
 	directivename=('%output' | '%launcher' | '%zeropage' | '%zpreserved' | '%address' | '%import' |
-                       '%breakpoint' | '%asminclude' | '%asmbinary' | '%option')
+                       '%breakpoint' | '%asminclude' | '%asmbinary' | '%option' | '%target' )
         (directivearg? | directivearg (',' directivearg)*)
         ;
 
@@ -273,7 +273,7 @@ asmsub_decl : identifier '(' asmsub_params? ')' asmsub_clobbers? asmsub_returns?
 
 asmsub_params :  asmsub_param (',' EOL? asmsub_param)* ;
 
-asmsub_param :  vardecl '@' (identifier | stack='stack') ;      // A,X,Y,AX,AY,XY,Pc,Pz,Pn,Pv allowed
+asmsub_param :  vardecl '@' identifier ;      // A,X,Y,AX,AY,XY,Pc,Pz,Pn,Pv allowed.  TODO implement  stack='stack'
 
 asmsub_clobbers : 'clobbers' '(' clobber? ')' ;